FLORIDA STATE UNIVERSITY COLLEGE OF ARTS AND SCIENCES

STABILITY AND DEGRADATION PROCESSES OF ENERGETIC MATERIALS

By Melissa Mileham

A Dissertation submitted to the Department of Chemistry and Biochemistry in partial fulfillment of the requirements for the degree Doctor of Philosophy

Degree Awarded: Summer Semester, 2008

The members of the Committee approve the dissertation of Melissa Mileham defended on July 1, 2008.

______________________________ Albert E. Stiegman Professor Directing Dissertation

______________________________ Vincent Salters Outside Committee Member

______________________________ Kenneth Goldsby Committee Member

______________________________ John Dorsey Committee Member

Approved: _______________________________________________________ Joseph Schlenoff, Chair, Department of Chemistry and Biochemistry

The Office of Graduate Studies has verified and approved the above named committee

ii

ACKNOWLEDGMENTS I would like to thank Dr. Albert E. Stiegman for his invaluable guidance and support throughout my graduate career. Thank you for answering all of my questions and not laughing (too hard) at all of my mistakes. My experiences in the lab as well as the people I have met have truly been priceless. I would also like to thank Dr. Lambertus J. van de Burgt for sharing even just a fraction of your knowledge about lasers with me. I really appreciate all of your help and guidance over the last couple of years. Dr. Michael P. Kramer, thank you for not only providing funding for my research, but always having an active role in it as well. Thank you for giving me this opportunity, and I can honestly say that there was never a dull moment. Surprising yes, but never dull. Also, many thanks to all of the staff members, especially the machine and glass shop, without whom some of my work would not have even been possible.

iii

TABLE OF CONTENTS Acknowledgments..iii List of Tables ..v List of Figures vi List of Abbreviations .....ix Abstract ...x INTRODUCTION...1 GENERAL EXPERIMENTAL ..6 1. SURFACE STABILITY AND DEGRADATION STUDIES OF TNT ON METAL OXIDES.13 2. SURFACE STABILITY AND DEGRADATION STUDIES OF PETN ON METAL OXIDES.28 3. LASER INITIATION PROCESSES IN THERMITE ENERGETIC MATERIALS STUDIED BY A LASER DESORPTION IONIZATION (LDI) TECHNIQUE..41 4. PHOTO-THERMAL INIATION PROCESSES OF ORGANIC-INORGANIC HYBRID METASTABLE INTERSTITIAL COMPOSITE (MIC) MATERIALS..53 SUMMARY...64 REFERENCES..66 BIOGRAPHICAL SKETCH.70

iv

LIST OF TABLES Table 1. TNT coverage levels on metal oxides surveyed..17 Table 2. Surface coverage of PETN on each metal oxide.32 Table 3. Effects of PETN on heat and energy given off by MIC composites...54

LIST OF FIGURES Figure 1. Molecular structures of common high energy materials..2 Figure 2. Schematic of an isocratic pumping system in HPLC...7 Figure 3. FT-IR spectrometer schematic.9 Figure 4. Principle of the MALDI process10 Figure 5. Scheme of a time-of-flight mass spectrometer...11 Figure 6. The four-level pumping system of the Nd:YAG laser...12 Figure 7. Percent decomposition over time of TNT deposited on various metal oxide substrates17 Figure 8. Differential scanning calorimetry trace of pure TNT.19 Figure 9. DSC of a) 1.4 and b) 4 monolayers of TNT deposited on MnO2...20 Figure 10. DSC of a) 1 monolayer and b) 4 monolayers of TNT deposited on CuO20 Figure 11. DSC traces of TNT deposited on KBr, LiF, and SiO2..21 Figure 12. HPLC chromatograms of the products of 1.4 monolayer of TNT deposited on MnO2 (a) initially and at 50 C for (b) 16 (c) 24 and (d) 42 days.23 Figure 13. HPLC chromatograms of the products of 1 monolayer of TNT deposited on CuO (a) initially and at 50 C for (b) 31 (c) 39 and (d) 58 days23 Figure 14. (a) Gas chromatogram showing two principle components at 11.83 and 11.9 minutes whose mass spectra identify them as (b) trinitrobenzene and (c) unreacted TNT25 Figure 15. Plot of the decomposition data of TNT on MnO2 over time fit to equation 2......................................................................................................................................26 Figure 16. The evolution of a brown gas from PETN on the surface of MoO3 after being stored at 50 C33 Figure 17. Differential Scanning Calorimetry (DSC) of pure PETN with a vented sample pan (note exothermic processes are above the baseline)...34

vi

Figure 18. DSC of (a) <2, (b) 2, and (c) 6 monolayer coverage of PETN deposited on nano-scale MoO3................................................................................................................36 Figure 19. DSC of (a) <2, (b) 2, and (c) 6 monolayer coverage of PETN deposited on micron-scale MoO337 Figure 20. DSC traces of PETN deposited on (a) SiO2 and (b) KBr.37 Figure 21. Percent PETN remaining on MoO3 vs. time38 Figure 22. FT-IR spectroscopy of PETN on the surface of MoO3 at 100 C after a) 1, b) 5, c) 24, d) 27.5, and e) 46 hours (=CO2, =N2O, =N2O4, =NO2)39 Figure 23. SEM images of (a) 50 nm, (b) 100 nm, and (c) micron-scale aluminum particles..43 Figure 24. LDI-TOF mass spectra of 50 nm aluminum at a laser energy density of (a) 1.74 and (b) 1.90 J/cm2..45 Figure 25. Aluminum ions formed at a laser energy of 1.90 J/cm2 for each aluminum particle size47 Figure 26. Major aluminum ions formed ([Al]+ + [Al2]+ + [Al2O]+) for each particle size at increasing energy densities48 Figure 27. LDI-TOF mass spectra of iron(III) oxide at a laser density of (a) 1.58 and (b) 1.74 J/cm2..49 Figure 28. LDI-TOF mass spectra of thermite mixtures at a laser energy density of (a) 1.42, (b) 1.58, (c) 1.74, and (d) 1.90 J/cm251 Figure 29. Relative amounts of pure iron cluster species formed during laser desorption of the Fe2O3 control and thermite samples at an energy density of 2.054 J/cm2...51 Figure 30. Mixed Al/Fe oxide species formed at an energy density of 2.054 J/cm2 for 50 and 100 nm and micron aluminum thermite mixtures...52 Figure 31. Photo-thermal initiation of Fe2O3/100 nm Al samples with increasing amounts of PETN at 1064 nm (arrow indicates the position of the laser plume emission).57 Figure 32. Initiation time (ms) as a function of PETN coverage in mg at 1064 nm..57 Figure 33. Deflagration duration time (ms) as a function of the amount of PETN at 1064 nm..58

vii

Figure 34. Initiation time of Al/Fe2O3 MIC materials with various concentrations of PETN initiated with a single pulse of 532 nm radiation61 Figure 35. Energy density (J/cm2) required in order to initiate the thermite/PETN mixture at 1064 nm..62

viii

LIST OF ABBREVIATIONS TNT TNB RDX HMX PETN MIC LDI/TOF MS BET YAG FE-SEM DSC TGA HPLC GC-MS FID FT-IR 2,4,6-trinitrotoluene trinitrobenzene hexahydro-1,3,5-trinitro-1,3,5-triazine (Royal Demolition eXplosive) octahydro-1,3,5,7-tetranitro-1,3,5,7-tetrazocine (High Molecular weight rdX) pentaerythritol tetranitrate metastable interstitial composite laser desorption ionization time-of-fight mass spectrometry Brunauer-Emmett-Teller (mathematical formulation for surface area anaylsis) yttrium aluminum garnet (common garnet crystal used in lasers) field emission scanning electron microscopy differential scanning calorimetry thermogravimetric analysis high-performance liquid chromatography gas chromatography mass spectrometry flame ionization detector fourier-transform infrared

ix

ABSTRACT The use of binary inorganic solid-state fuel/oxidant redox processes typified by the classic aluminum/iron(III) oxide thermite reaction in combination with traditional energetic materials such as 2,4,6-trinitrotoluene (TNT), hexahydro-1,3,5-trinitro-1,3,5triazine (RDX), and octahydro-1,3,5,7-tetranitro-1,3,5,7-tetrazocine (HMX) has been of interest in order to produce higher output explosives. This dissertation focuses on the stability and degradation processes that occur with the combination of binary inorganic fuel/oxidant systems with high energy materials. First is a discussion on the stability of 2,4,6-trinitrotoluene (TNT) as well as pentaerythritol tetranitrate (PETN) deposited onto the surface of metal oxides through a wet impregnation technique, which is followed by a discussion of the study of the initiation processes of the aluminum/iron(III) oxide thermite reaction using laser induced desorption-ionization time-of-flight mass spectrometry. Finally, the photo-thermal initiation of an aluminum/iron(III) oxide thermite with PETN deposited on the surface of the iron(III) oxide in increasing increments was studied using a single pulse of a Nd:YAG laser at differing wavelengths in order to understand the effects of the presence of PETN on the time to initiation, as well as the deflagration duration of the thermite reaction.

INTRODUCTION

Stability and Degradation Processes of TNT and PETN on Metal Oxides Recently there has been a new approach to the fabrication of high output explosives and pyrotechnics in which binary inorganic solid-state reactive materials are used in combination with traditional organic high energy explosives such as 2,4,6trinitrotoluene (TNT); hexahydro-1,3,5-trinitro-1,3,5-triazine (RDX); octahydro-1,3,5,7tetranitro-1,3,5,7-tetrazocine (HMX); and pentaerythritol tetranitrate (PETN) as shown in figure 1. The binary inorganic materials are typified by the classic thermite reaction (equation 1), where a stoichiometric amount of a metal (fuel) and a metal oxide (oxidant) are mixed and a highly exothermic redox reaction occurs. Fe2O3 + 2Al Al2O3 + 2Fe H = 3.97 kJ/g (1) These binary inorganic materials have the advantage of having a high energy density, in fact several times higher than that of a traditional high energy material. However, the rate of energy release is much slower than that of conventional explosives making it difficult to exploit these materials for explosive purposes. Therefore, in order to overcome this problem, new composite materials have been developed in which the binary inorganic systems are combined with conventional explosives such as TNT, PETN, RDX, and HMX. Thus the energy release in these combined systems is driven by the conventional explosive, resulting in a rapid release of energy in a controlled fashion. Previous studies have indicated that the combination of these hybrid organic/inorganic materials provides great promise in the exploitation of the high energy density provided by the binary inorganic mixtures. TNT, for example, has been combined with various oxidant phases and fuels such as MnO2 and Al as well as Al-Zr and Mg-Al alloys for use as high energy density explosives for high penetration applications and as components for boosters and primers.1-3 PETN, on the other hand, has been studied with binary inorganic materials where Al, Al alloys, and Zr have been used as fuels and Pb3O4, Fe2O3, and MnO3 have been used as the oxidant for their use as boosters and primers as well as their use in a laser initiated system.2, 4

Figure 1. Molecular structures of common high energy materials There is one problem that exists in the incorporation of conventional organic high energy explosives with fuel/oxidant binary inorganic mixtures, however, in that there is a possibility that an inherent incompatibility between the inorganic and organic phases may occur. This incompatibility arises from the deleterious surface chemistry between the organic material and the metal or metal oxide phase in the composite, which may arise from an interfacial chemical process that may occur either as a reduction at the surface of the metal or an oxidation at the surface of the oxide. The magnitude of these processes may be affected by the degree of contact that occurs between the organic materials and the metal or metal oxide surfaces, as well as the temperature, humidity, and other ambient conditions of the storage and handling of these systems. The chemical stability of both TNT and PETN when deposited onto the surface of metal oxides such as MnO2, CuO, WO3, MoO3, Bi2O3, SnO2, and Fe2O3 has been investigated and is reported hereafter. In order to imitate storage conditions of such materials, the samples were placed in an oven set at 50 C, periodically removing aliquots to be analyzed chromatographically in order to determine if a decomposition of the energetic materials had occurred at the surface of the oxide. In a case where 2

degradation of the energetic material occurred, subsequent studies were completed to understand the degradation process as well as any products that are produced at the surface of the oxide.

Laser Initiation Processes of Thermite Energetic Materials Using Laser Desorption Ionization (LDI) The traditional thermite reaction (equation 1) in which stoichiometric amounts of aluminum and iron oxide react to form aluminum oxide and iron metal has become a great interest in the production of high energy materials due to their high energy density. However, these materials have not been broadly useful due to slow mass-transport processes of the reaction, which leads to a slower energy release and subsequently a lower obtained power. It has recently been determined though that the energetic properties of these binary inorganic materials may be enhanced by having at least one of the components in the mixture of nanometer size, usually the fuel. The nano-scale fuel causes a more intimate contact between the fuel and the oxidizer, which enhances the mass-transport and therefore leads to an increase in power in these materials known as metastable interstitial composites (MIC).5, 6 The thermite reaction is highly exothermic and thermal initiation of the reaction occurs at high temperatures once the melting point of one of the components, typically the metal has been achieved.7 This initiation begins the reactive processes which liberate heat causing the reaction to accelerate. Ignition of the reaction occurs once the reaction becomes self-sustaining and propagates. The incident laser energy put into a sample for laser initiation causes high localized heating to occur, which then leads to ignition of the bulk sample. The time to ignition is defined as the time at which the energy released by the reaction becomes greater than or equal to the energy put into the composite by the laser. Both the laser excitation event and the subsequent ignition and propagation are extremely high temperature events, which generate both liquid and gas phase (plasma) species. The chemical reactions occurring within and between these phases as well as with the solid material all contribute to the net combustion process. This complexity and the fact that the reaction occurs so quickly has made it difficult to observe the species

formed during the reaction other than the final products and therefore thoroughly understand the reaction. Recently, however, this has been addressed by Dlott et al. by using time-resolved spectroscopy, where a reaction between aluminum and nitrocellulose was initiated by flash-heating the mixture with a 100 ps laser pulse in the near infrared.8, 9 Short-pulse photo-thermal initiation allows for the use of time-resolved spectroscopic techniques in order to monitor reaction dynamics, interpret some of the intermediate species of the reaction, and to observe specific structural changes in the reactive components. Time-resolved studies of the combustion of Al/MoO3 MIC composites have been studied and have identified neutral species such as AlO that form in the process.10 Previous studies of standard binary fuel/oxidant systems have involved thermochemical measurements such as temperature profiles, burn rates, and nonisothermal calorimetry (DSC, DTA, and TGA).11-13 Recent studies of MIC materials have used laser-induced photo-thermal initiation of binary inorganic systems in order to determine properties such as ignition time and burn rate as well as study the effect of particle size on combustion properties. Pantoya et al. studied the combustion velocities and ignition times of an Al/MoO3 thermite system with varying aluminum particle diameter by observing the reaction over time using high-speed cameras.14, 15 A similar study on the ignition process of a magnesium and barium oxide pyrotechnic mixture were completed by stmark et al. where the mixture was initiated using a CO2 laser and the subsequent ignition process was observed using high-speed photography which was synchronized with the laser pulse.16 Each study found that the combustion properties of binary inorganic mixtures is dependant upon the particle size of the fuel. The use of photo-thermal laser initiation coupled with time-of-flight mass spectrometry (LDI/TOF MS) in order to directly observe the ionic species formed in the plasma phase of both conventional Al/Fe2O3 thermite and the corresponding MIC materials is discussed. Though this does not show the reaction to completion, the species formed in the plasma allows for the study of the reactive processes of various aluminum sizes in the reaction. This may be the first use of this technique to characterize reactive species in the initiation phase of binary inorganic reactions.

Photo-thermal Initiation Processes of Hybrid Organic/Inorganic Metastable Interstitial Composite (MIC) Materials

Two approaches used separately or in tandem have been devised in order to overcome the slow energy release found in binary inorganic materials compared to that of traditional energetic materials. One is the development of metastable interstitial composite (MIC) materials in which one of the components (typically the fuel) is of nanoscale dimensions. Improvements in the rate of energy release are generally attributed to better mixing of the components and a more intimate fuel/oxidant contact.5,6 Another approach, applied to conventional (i.e. micron scale) thermite-type compositions as well at to MIC materials, is to mix traditional organic high-energy materials such as 2,4,6trinitrotoluene (TNT); hexahydro-1,3,5-trinitro-1,3,5-triazine (RDX); and pentaerythritol tetranitrate (PETN) into the inorganic system to form an organic/inorganic composite. In these systems the energy release of the binary fuel/oxide system is driven by the conventional explosive, thereby releasing energy much more rapidly and in a controlled fashion. Studies to date have indicated that this approach affords great promise in exploiting the high energy density provided by binary inorganic energetic compositions.14

It was determined from thermal analysis studies that depositions of at least 6 monolayers resulted in thermal properties for the decomposition of the organic species that were the same as the bulk. This suggests that even in small amounts incorporation of the organic phase may result in observable changes in the energy release properties of the composite, which are amenable to laboratory study. We report here a study of the energy release dynamics of the Al/Fe2O3 MIC materials with depositions of 16-127.9 mg of PETN per gram of thermite. The study is carried out using short pulse (ns) photo-thermal initiation, with the dynamics of the process studied by time-resolved spectroscopic techniques.

GENERAL EXPERIMENTAL

Liquid Chromatography Liquid Chromatography (LC) is an analytical technique where the components of a mixture are separated in solution at room temperature. It is one of the most commonly used separation techniques due to its versatility in that there are two interactive phases (stationary and mobile phases) and this interaction may be altered in order to change the selectivity of the system. Each component in a mixture should interact differently with each of the two phases causing a separation to occur. The time it takes each component to travel through the column to the detector is the retention time and is what allows for the characterization of the species. High Performance Liquid Chromatography (HPLC) describes a liquid chromatography technique where the mobile phase is mechanically pumped through a column containing the stationary phase; therefore, the instrument consists of an injector, pump, column, and detector (figure 2). Reversed-phase chromatography is the most widely used technique and is used to separate neutral molecules by their degree of hydrophobicity. The typical stationary phase has an organic functional group such as CH3, -C4H9, -C8H17 and C18H37 chemically attached to silica. The functional group affects the retention time, which increases exponentially with chain length, as well as the column selectivity and efficiency. The typical mobile phases used in reversed-phase chromatography include a polar solvent, usually water, which is mixed with a slightly less polar solvent such as methanol or acetonitrile. The polarity of the solvent is inversely proportional to its eluting strength and the mixture used as the mobile phase is chosen to give the desired separation.17, 18

Figure 2. Schematic of an isocratic pumping system in HPLC

Gas Chromatography - Mass Spectroscopy Gas chromatography is a method used for separating and analyzing mixtures of volatile compounds. Similar to liquid chromatography, gas chromatography takes advantage of varying interactions occurring between each component and the two phases (stationary and mobile). The main difference, however, is that gas chromatography uses a gas as a mobile phase, and therefore in order to use this technique all components must be volatile and thermally stable making it somewhat more limited than liquid chromatography. Due to the fact that there is little interaction between the molecules in gas phase, the mobile phase primarily acts as a way to move the components through the system and thus the distribution equilibria of the components is determined by their vapor pressure and sorption by the stationary phase. Since vapor pressure plays a key role in the separation, the column must be heated to a temperature high enough to provide an appropriate analysis time. The most commonly used carrier gases are helium and nitrogen; however, argon, hydrogen, and carbon dioxide may also be used. The carrier gas is determined by its compatibility with the detector of the system, which in this study a flame ionization detector (FID) was used due to its ability to detect organic molecules. The column can either be a packed column or a capillary open tubular column, each with its advantages and several options for stationary phases which will depend on the types of solutes being separated. The gas chromatograph can also be coupled to a mass spectrometer, which ionizes the components and subsequently separates and detects them

based on their mass-to-charge (m/z) ratio. This feature allows the confirmation of the identity of each analyte in a mixture as well as any unknowns.19, 20

Differential Scanning Calorimetry (DSC) Differential Scanning Calorimetry (DSC) is a thermal analysis technique that measures the heat flow associated with thermally driven endothermic or exothermic transitions in a material. The DSC is capable of giving thermodynamic information such as phase changes, melting, crystallization, product stability, and oxidative stability. Two identical pans, one reference and one containing the sample, are placed on platforms that are connected to individual furnaces. In a general experiment, the two pans are simultaneously heated at a specific rate, typically not more than 20 C/min, measuring the difference in heat flow between the sample pan and the reference pan. The result is a plot of heat flow versus temperature on which the features correspond either to an absorption or release of energy in the sample upon heating due to either phase transitions or a decomposition of the material.21

Gas Physisorption Gas adsorption/desorption techniques are used in order to find characteristics of porous materials such as surface area, pore size, and pore volume. Nitrogen gas is used at incrementally higher pressures to dose a sample at liquid nitrogen temperatures. This known dosing pressure is compared to a measured actual pressure, the difference between which yields the amount of gas adsorbed by the sample. An isotherm plot is given by plotting the measured volumes of adsorbed gas versus the relative pressure at which these adsorptions took place. There are six classifications of physisorption isotherms, each of which indicate the different adsorbent-adsorbate interactions. The first few points of the isotherm are used to calculate the surface area, typically using a mathematical equation developed by Brunauer, Emmett, and Teller (BET), while other equations are used to calculate the pore size and volume.22

Fourier-Transform Infrared Spectroscopy (FT-IR) Fourier-transform infrared (FT-IR) spectroscopy is a technique used for observing the vibrations of the atoms in a molecule. It has become a very versatile technique, in that almost any type of sample can be used, e.g. liquids, powders, films, gases, and surfaces. The spectrum is obtained by using an interferometer, where radiation from the source passes through before hitting the sample, in which interference of radiation between two beams produces the transmitted beam to the sample at 90 from the input beam. The fraction of incident radiation that is absorbed by the sample at a particular energy is then determined and is converted to an absorbance spectrum using a fourier transformation process. A schematic of the instrument is shown in figure 3.23

Figure 3. FT-IR spectrometer schematic

Laser Desorption Ionization Time-of-Flight Mass Spectrometry (LDI-TOF MS) Matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF MS) is a technique that is most often used for, but not limited to, the analysis of biomolecules because significant decomposition of the fragile molecules does not occur and large mass ranges can be detected. MALDI-TOF mass spectrometry allows species to be converted from the solid phase to gas-phase ions with little fragmentation due to the simultaneous vaporization and ionization processes. Since these processes occur in a single step, it is referred to as a desorption/ionization technique. This singlestep process is accomplished because the samples are placed under a vacuum of at least 10-6 torr, which allows the sample to sublime directly into the ion source with little or no heating.24

The sample is usually mixed in a large excess of a matrix compound, which is typically a weak acid that absorbs light readily at the wavelength of the laser. The sample is then deposited onto the solid surface of the target, which is made out of a conducting metal, most often stainless steel. The mixtures used in this case act as a matrix themselves, and thus a typical matrix is not needed. Therefore, small holes were drilled into the sample plate into which the mixture was tightly packed in order to ensure that the sample would stay in place, especially under a high vacuum and the process itself is referred to as laser desporption ionization (LDI) mass spectrometry. Once the proper vacuum is reached, the sample may be hit with a brief laser pulse. The irradiated spot is rapidly heated by the laser energy and becomes vibrationally excited, which results in a portion of the deposited samples surface to be released from the irradiated spot, allowing the matrix to carry a portion of the analyte into the vapor phase with little or no heating as illustrated by figure 4.24, 25 Since the LDI is equipped with a time-of-flight mass spectrometer each of the accelerated ions from the source has a different velocity, which is based off of the mass and charge of the ion. The accelerated ions then move from the source into the drift region of the analyzer where the different velocities separate the ions in order to keep them from hitting the detector at the same time. The now mass separated ions hit the detector region, where they are counted, giving rise to a specific signal for each m/z value proportional to the number of ions present, producing a mass spectrum.(figure 5).24

Figure 4. Principle of the MALDI process (Wilkins et al., 2006).

10

Figure 5. Scheme of a time-of-flight mass spectrometer (Wilkins et al., 2006). Nd:YAG Laser The Nd:YAG laser uses a neodynium yttrium-aluminum-garnet (Y3Al5O12) crystal, a common lasing medium for solid-state lasers, where a small amount (~1 %) of yttrium ions are replaced by neodynium ions in the crystal structure. A flashlamp is used to excite the neodynium electrons from the ground state and upon relaxation a slow decay back to the ground state results in a dominant fluorescence around 1064 nm with a slow decay time (ms), which results in the laser action. The neodynium ion actually has a fourlevel pumping system with several different laser transitions that result in fast nonradiative decay to the upper laser level (E2). A simplified version of this system is shown in figure 6 in which level 4 (E3) represents the combination of all the levels above the upper laser level in the real atomic system. Level 3 (E2) is the upper laser level and usually is a long-lived level, while level 2 (E1) is the lower laser level and level 1 (Eo) represents the ground level The laser may be operated both in a pulsed or continuous mode. If the laser is pulsed, it is most often operated in the Q-switching mode, which is an optical switch that opens once the maximum population inversion of the neodynium ions is acquired, allowing the light wave to run through the cavity, depopulating the excited laser medium and resulting in a laser pulse. This pulse is less than ten nanoseconds and gives an output power of 20 mW.26

11

Figure 6. The four-level pumping system of the Nd:YAG laser

12

CHAPTER 1

SURFACE STABILITY AND DEGRADATION STUDIES OF TNT ON METAL OXIDES

Introduction An important issue related to the incorporation of such conventional organic highenergy materials with metal/metal oxide binary compositions is one of inherent incompatibility between the organic and inorganic phases. This incompatibility arises from deleterious surface chemistry between the organic and metal or metal oxide phases in the composite. These deleterious processes arise from interfacial chemical processes including, but not limited to, reduction at the metals (fuel) surface and oxidation at the oxide surface. The magnitude of these processes will likely be exacerbated by the degree of contact between the organic materials and the metal or metal oxide surfaces and will be affected by temperature, humidity, and other ambient conditions of storage and handling. The chemical stability of 2,4,6-trinitritoluene (TNT) when placed in physical contact with metal oxide surfaces was investigated. The TNT was deposited at 1-3 monolayers of coverage on the surface of microcrystalline MnO2, CuO, WO3, MoO3, Bi2O3, SnO2, and Fe2O3 by wet impregnation techniques. The samples were placed in a 50 C oven and allowed to react over a period of ten months. Periodically, a small portion of the sample was removed and analyzed chromatographically to determine if products were being formed at the surface of the oxide. TNT proved to be inert to most oxides; however, both CuO and MnO2 effected a clean decomposition of the TNT molecule to trinitrobenzene (TNB).

13

Experimental

Materials MnO2 (Aldrich, reagent grade, >90% purity, ~10mm particle size), CuO (Aldrich, nanopowder, ~33 nm), Fe2O3 (Fisher, anhydrous), MoO3 (nano- and micron-scale, Climax Molybdenum), SnO2 (Keeling & Walker Ltd.), WO3 (Atlantic Equipment Engineering) and Bi2O3 (Aldrich) were used as received from the manufacturer. 2,4,6Trinitrotoluene (TNT) was obtained from Chemservice and sublimed prior to use. Acetonitrile (Acros, reagent grade) and methanol (EMD Chemicals, HPLC grade) was also used as received.

Long-term surface reactivity studies Samples for long-term reactivity studies were prepared using a wet impregnation technique. Sixty milligrams of TNT, dissolved in acetonitrile, was slurried with one gram of the metal oxide, after which the acetonitrile was removed under vacuum (10-3 torr) in order to leave a dry powder of the oxide with TNT deposited onto its surface. The approximate coverage area of TNT is 986 m2/g, which is based on the estimated area of a TNT molecule obtained from treating the molecule as a disk with the outer circumference defined by the oxygen atoms of the nitro groups (obtained from the crystal structure of 1,3-dinitrotoluene) with a radius extending from the center of the aromatic ring.27 Sealed containers containing the samples were placed in an oven held at 50 C. Periodically, a small portion of the powder (~0.10 g) was withdrawn and stirred into one milliliter of acetonitrile. The solids were allowed to settle and the supernatant liquid containing the organic species was removed and analyzed by HPLC using a 60:40 H2O/CH3OH solution as the mobile phase. The presence of new compounds formed from interfacial chemistry was evident in the chromatogram by comparison to the control samples using a UV detector set to 254 nm. The percent degradation was determined from the ratio of the integrated peak area of the TNT in the HPLC with the sum of all the peak areas. For samples that showed only minimal degradation, no attempt was made to determine the products. For oxides that induced a significant amount of decomposition the product was characterized by GC-MS.

14

BET surface area analysis The surface area of each metal oxide used in the survey was found by BET methods using a Micromeretics ASAP 2020 physisorption surface area and porosity analyzer. Approximately 0.25 g of each metal oxide was placed in a sample tube and was degassed under a vacuum of 20 mHg at 90 C for sixty minutes in order to remove any water from the sample. A subsequent heating at 340 C for 240 minutes was performed in order to completely degas the material. The analysis was performed at liquid nitrogen temperatures in order to obtain the surface area for each metal oxide.

HPLC Liquid chromatography experiments were performed on a Beckman Coulter System Gold HPLC equipped with a 125 Solvent Module, 166 Detector and 508 Autosampler using a Beckman C18 column that is 250 x 4.6 mm. The detector was set at 254 nm in order to detect any organic compounds that may be present. A flow rate of 0.75 mL/min was used for the mobile phase, an isocratic mixture of 60:40 H2O/methanol. The chromatogram was taken over a period of 30 minutes, allowing all compounds time to travel through the column. TNT appears at approximately 18 minutes at this flow rate, assuming only degradation products of the TNT molecule occur, they should elute faster through the column due to their smaller size.

GC-MS The GC-MS data were collected on an Agilent 6890+ GC coupled with an HP 5973 MSD (mass selective detector). The GC was equipped with a DB-5 capillary column which had a length of 30 meters and an i.d. of 0.25 m using helium as the carrier gas with a flow rate of 1.5 mL/min. The oven temperature ramp was set as isothermal at 140 C and the injector and detector temperatures were set to 200 and 250 C, respectively. The detector used was a flame ionization detector (FID), which is used primarily for organic compounds due to its ability to easily detect hydrocarbons. The sample was also ionized and analyzed by mass (m/z) in the mass selective detector. The source of the MSD was set at 100 C. This allows the mass of any of the degradation

15

products of TNT corresponding to elution time through the gas chromatogram to be determined as well as the products themselves to be identified.

Thermal Analysis Differential scanning calorimetry (DSC) was performed on a TA Instruments Q1000 DSC under both O2 and an inert N2 environment. Samples were prepared using the wet impregnation technique as described above. A two milliliter aliquot of TNT (0.13 M in acetonitrile) was added to one gram of MnO2 or CuO and the solvent was removed under vacuum. The dry powder containing TNT deposited onto the surface of the metal oxide was then analyzed using a scan rate of 10 C/min with a pin prick in the pan in order to vent the sample pan. When the sample is analyzed in a sealed container, the decomposition becomes a deflagration event with an abrupt release of energy. Samples of pure TNT and TNT deposited on the surface of inert materials such as silica gel and KBr as well as LiF were prepared and analyzed under the same conditions in order to differentiate surface chemistry specific to the metal oxides.

Results and Discussion

Degradation studies The long-term degradation studies were designed to simulate long-term ambient storage of multicomponent high-energy materials where TNT would be in intimate contact with metal oxide surfaces. Relatively low concentrations (~1-3.5 monolayers) of TNT were deposited onto the oxide surfaces to isolate specific interfacial chemistry and to allow sensitive detection of reaction products relative to the TNT (i.e. without the chromatogram being overwhelmed by a large amount of TNT). Table 1 shows the oxides surveyed in the study and their TNT surface coverage in the prepared samples.

16

Table 1. TNT coverage levels on metal oxides surveyed Metal Oxide MnO2 CuO WO3 MoO3 (large) MoO3 (nano) Bi2O3 SnO2 Fe2O3 BET Surface Area (m2/g) 40 32 17 25 55 26 21 33 TNT coverage (equivalent monolayers) 1.5 1.9 3.5 2.4 1.1 2.3 2.8 1.8

A chromatogram of the samples was taken immediately upon preparation to establish the initial purity. They were then placed in a 50 C oven under the ambient atmosphere and allowed to react over the course of approximately eleven months with samples withdrawn periodically over that time and analyzed by HPLC. The amount of decomposition of the TNT over time for each oxide surface where a measurable decomposition is observed is shown in figure 7.

Percent Decomposition of TNT on Metal Oxides

100 90 80 70 60 Percent Decomposition 50 40 30 20 10 0 309 253

190

168

98

85

Figure 7. Percent decomposition over time of TNT deposited on various metal oxide substrates

17

Bi2O3

SnO2

Fe2O3

32

24

CuO

42

time (days)

MnO2

59

TNT proved to be inert or slightly reactive on the majority of the oxide surfaces. In particular, no detectable decomposition was observed for MoO3 and WO3. As indicated in figure 7, small amounts of decomposition (<10 %) were observed for Bi2O3, Fe2O3, and SnO2 over the duration of the study. For two of the oxides, CuO and MnO2, more extensive decomposition was observed with 21 % of the TNT consumed over CuO and essentially complete (100 %) consumption observed for MnO2. For the case of CuO, the decomposition reached its maximum after about one month (figure 7), but did not react further after that. This result suggests that this reaction is stoichiometric with a specific surface site, which, once consumed, ceases to react further. Conversely, for MnO2, the reaction proceeds to completion suggesting either a large amount of surface sites on the MnO2 or that the reaction is catalytic. An important aspect of TNT reactivity on both of these surfaces is that the HPLC data indicates that the reaction is quite clean, going primarily to a discrete product of trinitrobenzene as opposed to decomposition into multiple species with MnO2 being somewhat cleaner than the CuO. Moreover, based on a comparison of the retention times, the product is the same for both oxides.

Reactivity of TNT on MnO2 Thermal analysis. The thermal analysis of pure TNT using differential scanning

calorimetry (DSC) is shown in figure 8. The thermal scan shows a sharp endotherm at 80 C, which corresponds to the melting and an exotherm, whose observed peak at this scan rate (10 C/min) is at 306 C, that corresponds to the decomposition of the molecule. This thermal decomposition process has been studied previously and was found to initially involve oxidation of the methyl group but, even in the early stages of the reaction, it is quite complex and ultimately yields telomeric or polymeric materials.28-31 When considering composites of TNT with inorganic oxides, the effect of the organic-inorganic interface on these well-known thermal processes is of interest. The DSC plot of TNT deposited at approximately 1.4 monolayer loadings on MnO2 is shown in figure 9. At

18

Figure 8. Differential scanning calorimetry trace of pure TNT (note that the exothermic processes are above the baseline) these coverage levels, the melting endotherm is not observed. This result is expected since no crystalline bulk phase is present. More importantly, however, the thermal decomposition process occurs at 229 C, which is significantly lower than the bulk. When the amount of TNT is increased to >4 monolayers, two exothermic processes are observed, one whose peak is close to that resolved in the bulk (321 C) and a low temperature exothermic process at 278 C. Our preliminary interpretation of this complex thermal behavior is that it represents the superposition of interfacial and bulk processes observable at the higher loadings of TNT though it may reflect a more complex decomposition process than is initiated at the surface. A similar, but less dramatic, interfacial effect is observed on CuO with monolayer coverage yielding decomposition exotherms at 211 C and 294 C due to the interfacial effect of the CuO, while excess loadings showing interfacial and bulk decomposition at 294 and 318 C respectively (figure 10).

19

Figure 9. DSC of a) 1.4 and b) 4 monolayers of TNT deposited on MnO2

Figure 10. DSC of a) 1 monolayer and b) 4 monolayers of TNT deposited on CuO 20

It is important to determine whether the interfacial mediation of the decomposition temperature is specific to those oxides or constituted a general surface effect. Thermal analysis of monolayers of TNT deposited on inert oxides and salts indicates that in almost all cases the decomposition temperature is modified (figure 11). On fumed silica the decomposition temperature is very close to that of the bulk at 310.8 C indicating that the silica surface is inert. Interestingly, on simple salts the interfacial effect is pronounced. Specifically, LiF shows a decomposition temperature of 293.0 C while KBr shows an even stronger surface effect with a decomposition temperature of 270.9 C. These results do not provide a chemical rationale for the interfacial effect, nor do they provide any specifics on the decomposition pathway(s) on the surface. However, it is clear from the thermal analysis data that there is a pronounced and surprisingly general surface effect between TNT and solid surfaces. This effect might well mitigate energy release processes or affect long-term stability of heterogeneous composite mixtures.

Figure 11. DSC traces of TNT deposited on KBr, LiF, and SiO2 (note exotherms are above the line).

21

Kinetics and product analysis. Reaction processes and decomposition pathways of trinitrotoluene deposited on the surface of microcrystalline MnO2 were monitored by HPLC. Initial chromatographic data shows only a sharp peak at a retention time (RT) of 18.57 minutes, which is assigned to TNT itself by comparison to a standard solution of TNT in acetonitrile (figure 12). No other significant products were observed. The samples were kept at a constant temperature of 50 C and reanalyzed after 16 days. The chromatogram shows two peaks, the TNT peak and a new species eluting earlier with a retention time of 10.26 minutes. After sitting at 50 C for another 8 days, the chromatogram indicates that this species was still being produced and was present in approximately equal amounts to the starting material. Also present in the chromatogram was small amounts of additional species that elute significantly faster. The observed pattern of more quickly eluting species appearing over time is suggestive of a general decomposition of the TNT into lower molecular weight species. Notably, however, these faster eluting species represent a very small amount of the total dissolved mass and, in fact, the chromatographic data shows that TNT is actually being converted relatively cleanly into a single product in a solid-state reaction on the surface of the MnO2. Finally, chromatograms collected after 42 days show that the new species is now predominant, suggesting that most of the TNT has been converted. The faster eluting species, while still present, has not increased significantly in concentration. Similar observations were made from the surface reaction of TNT with CuO. Chromatographic data indicates the formation of the same species eluting at 18.57 minutes after 31 days of reacting at 50 C. While the species that forms is the same, the extent of reaction is much more limited and, after 39 days, there is little additional formation of this product. As with MnO2, faster eluting species are also present which, while still minor, make up a larger fraction of the total dissolved material than they do in the MnO2 (figure 13).

22

Figure 12. HPLC chromatograms of the products of 1.4 monolayer of TNT deposited on MnO2 (a) initially and at 50 C for (b) 16 (c) 24 and (d) 42 days.

Figure 13. HPLC chromatograms of the products of 1 monolayer of TNT deposited on CuO (a) initially and at 50 C for (b) 31 (c) 39 and (d) 58 days.

23

Characterization of the main product of the solid-state reaction was carried out by GC-MS (figure 14). Injection of the product mixture, after reaction for 31 days on MnO2, resolved two relatively intense peaks at 11.83 and 11.9 minutes retention time on the GC. The mass spectrum of the 11.9 minute peak shows the classic fragmentation pattern for TNT with a weak parent-ion peak at 227.0 m/z and a more intense peak at 210.0 m/z corresponding to [TNT-OH]+ due to the loss of O from an o-nitro group and a H from the methyl group with ion bombardment to produce [C7H4(NO2)2(NO)]+.32 The mass spectrum of the species that elutes at 11.83 minutes shows a strong peak at 213.0 m/z/ and a series of peaks at 167.0, 120.0, and 75.0 m/z. The 213.0 m/z peak can be assigned as the parent peak of trinitrobenzene (TNB), C6H3(NO2)3, and the progression comes from sequential loss of NO2 (or NO2 and H) groups. This assignment is confirmed unambiguously by comparison to the published fragmentation pattern of a trinitrobenzene. The mass spectrum was also collected on some of the minor species resolved by the gas chromatogram. They have not been fully characterized but, as would be expected, the slower eluting species appear to be higher molecular weight biphenyl species and the faster eluting species are lower molecular weight decomposition fragments. Similar data was collected for the CuO catalysis. The primary product of the solid-state reaction is TNB, though, as indicated by the HPLC data, conversion to TNB is not as extensive as it is in the case of the MnO2. The gas chromatogram indicates that there are generally more species, both slower and faster eluting, than form with MnO2, suggesting that conversion to TNB is less clean and that there is more net decomposition of the TNT. While the mechanistic details of the reaction are not completely known, water has been observed forming in bulk studies. As such, this suggests that the net reaction is:

24

Figure 14. (a) Gas chromatogram showing two principle components at 11.83 and 11.9 minutes whose mass spectra identify them as (b) trinitrobenzene and (c) unreacted TNT. A likely pathway is the initial oxidation of the methyl group followed by a rapid decarbonylation process. Kinetically, the reaction is extremely slow and the plots of the relative concentration of TNT as a function of time are qualitatively consistent with a reaction that is pseudo first order in TNT. Further analysis of the data using the generalized approximate rate equation (equation 2) derived by Wilkinson, where p is the fraction reacted, t is the time, n is the reaction order, and K is the apparent rate constant is shown in figure 15.33 Evaluation of the slope (n/2) for the reaction, which is only linear when p 0.4, p values greater than 0.4 give a reaction order to the nearest half order.

t nt 1 = + p 2 K
25

(2)

Evaluation of the slope in this instance yields a reaction order of 1.4, consistent with a first order dependence on TNT and an apparent rate constant of 0.02/day for MnO2. The half-life for the reaction is 35 days. In the absence of more detailed data about the reaction products and mechanism, the kinetic analysis cannot be taken too far and only serves to indicate the relative efficiencies of the reaction. It also should be noted that it is not known whether the reaction is catalytic, using O2 as the oxidant, or whether it is stoichiometric with oxygen taken from specific sites on the metal oxide surface.

Figure 15. Plot of the decomposition data of TNT on MnO2 over time fit to equation 2. The net reaction is of considerable interest from a synthetic standpoint, since there are, to the best of our knowledge, no previous reports of the direct demethylation of aromatics. Moreover, there is no cost effective synthesis for TNB even though, from an energetic standpoint, it is a superior high-energy material. The most direct synthesis reported to date involves the conversion of phloroglucinol into trioximes with hydroxylamine followed by oxidation to trinitrobenzene with nitric acid.34 As such, a single-step heterogeneous synthesis may be of commercial utility in producing TNB. Unfortunately, attempts to scale up as a solution-solid reaction over MnO2 or as a solidstate reaction involving larger amounts of TNT were not successful. From all the data obtained so far, the reaction appears to occur as a very slow interfacial reaction. 26

Conclusions From the standpoint of compatibility, it is clear that the interfacial chemistry between TNT and inorganic surfaces can be pronounced. As indicated by the thermal analysis work, this involves at one level, changes in the normal thermal decomposition pathways of the molecule which may or may not have a net effect on stability or energy release in bulk composite materials. For redox active metal oxide phases the interfacial processes can result in a net chemical reaction that will slowly convert the TNT to another species, and possibly to degradation products. For the specific oxides reported here, the interfacial reaction converts TNT to TNB relatively cleanly and quantitatively for MnO2 and with more overall decomposition than for the case of CuO. Clearly, this result indicates that in composite energetic materials containing these components the TNT will be changing over time, which suggests that the stability and energy release may also be variable. Moreover, once conversion to TNB is accomplished, further decomposition may occur and adversely affect properties. Both MnO2 and CuO react with TNT under relatively mild conditions. For both oxides a major product of this solidstate reaction is the demethylation of TNT to form TNB. Over CuO this reaction does not appear to be exclusive as many other species are formed nor does it proceed to completion.

27

CHAPER 2

SURFACE STABILITY AND DEGRADATION STUDIES OF PETN ON METAL OXIDES

Introduction As previously stated the use of binary inorganic solid-state reactive materials in combination with traditional organic high explosives such as 2,4,6-trinitrotoluene (TNT), hexahydro-1,3,5-trinitro-1,3,5-triazine (RDX), octahydro-1,3,5,7-tetranitro-1,3,5,7tetrazocine (HMX) and pentaerythritol tetranitrate (PETN) has recently become a new approach to the fabrication of high output explosives. The binary inorganic materials are stoichiometric mixtures of a metal (fuel) and a metal oxide (oxidant) that react through a highly exothermic redox reaction such as the well-known thermite reaction. These binary inorganic materials have the advantage of having energy densities several times higher than that of conventional explosives, though the rate of energy release is significantly lower. In order to take advantage of the energy density of these binary materials, new composites involving the mixture of conventional explosives with these binary fuel/oxidant systems have been developed. These new composites allow the energy release of the system to be driven by the conventional explosive, thus driving the kinetics of the binary inorganic materials and thereby releasing energy much more rapidly and in a controlled fashion. However, it is unknown if an incompatibility between conventional explosives and metal/metal oxide binary compositions exists. This incompatibility could occur due to interfacial chemical processes occurring at the surface between the organic and inorganic phase of the composite. These chemical processes may include a reduction at the surface of the metal (fuel) and oxidation at the surface of the metal oxide. The magnitude of these processes will likely be increased by the degree of contact between the organic materials and the metal or metal oxide surfaces, and may be affected by temperature and humidity along with other ambient conditions of storage and handling.

28

The chemical stability of pentaerythritol tetranitrate (PETN) when placed in physical contact with metal oxides is investigated. In this study PETN was placed on the surfaces of a range of microcrystalline metal oxides including MnO2, CuO, MoO3, WO3, Bi2O3, SnO2, and Fe2O3 in coverages of 1-3.5 monolayers by a wet impregnation technique. Samples were then placed in a controlled temperature environment at 50 C and checked periodically for the presence of decomposition products using liquid chromatography. PETN proved to be inert over all of the oxides except MoO3 which showed the relatively rapid evolution of a brown gas over a period of 48 hours. Analysis of the evolved gas indicated that it is primarily NO2 along with N2O4, N2O, and CO2.

Experimental

Materials MnO2 (Aldrich, reagent grade, >90% purity, ~10mm particle size), CuO (Aldrich, nanopowder, ~33 nm), Fe2O3 (Fisher, anhydrous), MoO3 (nano- and micron-scale, Climax Molybdenum), SnO2 (Keeling & Walker Ltd.), WO3 (Atlantic Equipment Engineering) and Bi2O3 (Aldrich) were used as received from the manufacturer. Pentaerythritol tetranitrate (PETN) was prepared according to the literature and was stored in acetone (Aldrich, HPLC grade) for safety.35 Acetonitrile (Fisher, HPLC grade) and toluene (Aldrich, HPLC grade) were used as received.

Long-term surface reactivity studies A wet impregnation technique was used in preparing the samples for the longterm reactivity studies. A PETN standard solution (208 L, ~0.50 M) was diluted with one milliliter of acetone and subsequently added to one gram of the metal oxide. The acetone was removed under a vacuum (10-3 torr) leaving a dry powder containing the metal oxide and the PETN on its surface. The approximate coverage area of the PETN is 1543 m2/g, which is based on the estimated area of a PETN molecule, assuming the molecule is a disk with the outer circumference defined by the oxygen atoms of the nitro groups (obtained from the crystal structure of PETN) with the radius extending from the central carbon atom.36

29

The samples were placed in sealed containers and held at 50 C in an oven over a period of several months. Periodically, a small portion of the sample was removed (~0.10 g) and stirred into one milliliter of acetone. The solids were allowed to settle, and the supernatant was collected containing any organic species and analyzed by HPLC using a 65:35 CH3CN/H2O solution as the mobile phase and toluene as an internal standard. There was no evidence of new compounds forming in the chromatograph; however, a comparison between the control sample and each subsequent sample removed showed a decomposition of the PETN was occurring at the surface of the MoO3. The percent degradation was determined from the ratio of the integrated peak area of the PETN in the HPLC to that of the sum of all of the peak areas. For oxides that induced a significant amount of degradation, the products were analyzed using FT-IR spectroscopy.

BET surface area analysis The surface area of each metal oxide used in the survey was found by BET methods using a Micromeretics ASAP 2020 physisorption surface area and porosity analyzer. Approximately 0.25 g of each metal oxide was placed in a sample tube and was degassed under a vacuum of 20 mHg at 90 C for sixty minutes in order to remove any water from the sample. A subsequent heating at 340 C for 240 minutes was performed in order to completely degas the material. The analysis was performed at liquid nitrogen temperatures in order to obtain the surface area for each metal oxide.

HPLC Liquid chromatography experiments were performed on a Beckman Coulter System Gold HPLC equipped with a 125 Solvent Module, 166 Detector and 508 Autosampler using a Beckman C18 column that is 250 x 4.6 mm. The detector was set at 254 nm in order to detect any organic compounds that may be present. A flow rate of 0.75 mL/min was used for the mobile phase, an isocratic mixture of 65:35 acetonitrile/H2O and toluene as an internal standard. The chromatogram was taken over a period of 15 minutes, allowing all compounds time to travel through the column. PETN appears at approximately 3.6 minutes at this flow rate with another peak appearing at 7.5

30

minutes. No other products were seen in the chromatogram due to the fact that the degradation products appear to be solely gases.

FT-IR The FT-IR spectra were collected on a Thermo Nicolet Avatar 360 FT-IR in transmission mode. The sample, prepared as described above, was placed in a quartz tube and attached to a gas phase IR cell with sodium chloride windows, which was under vacuum (10-3 torr) allowing only gases evolved from a reaction between MoO3 and PETN to be seen. A background was taken, and the sample was then opened to the gascell allowing any evolved gas to enter the cell. The sample was heated to 100 C allowing a faster evolution of any reaction products and subsequent spectra were taken in the spectral range of 750 to 4000 cm-1 with a resolution of 4 cm-1 at first every 15 minutes and then at longer intervals until no further product evolution occurred.

Thermal analysis Differential Scanning Calorimetry (DSC) was performed on a TA Instruments Q1000 Series DSC under O2 at a ramp rate of 10 C/min from 40 to 450 C. Samples were prepared using the wet impregnation method as described above. An aliquot of 84 L of a PETN standard solution (~0.50 M) was diluted with a small amount of acetone and added to 0.25 grams of MoO3 (both nano and micron sizes). The solvent was removed under a vacuum, and the dry powder containing the PETN on the surface of the metal oxide was then analyzed. Samples of pure PETN and PETN on the surface of silica gel and KBr were also analyzed to differentiate surface chemistry specific to the metal oxides.

Results and Discussion

Degradation studies The long-term degradation studies were designed to simulate long-term ambient storage conditions of multicomponent high-energy materials in direct contact with metal oxide surfaces. Relatively low concentrations (1-3.5 monolayers) of PETN were

31

deposited onto the oxide surfaces in order to isolate specific interfacial chemistry as well as to allow the sensitive detection of reaction products. Table 2 shows the oxides surveyed in the study and their PETN surface coverage in the prepared samples. Table 2. Surface coverage of PETN on each metal oxide Metal Oxide MnO2 CuO WO3 MoO3 (large) MoO3 (nano) Bi2O3 SnO2 Fe2O3 BET Surface Area (m2/g) 40 32 17 25 55 26 21 33 PETN Coverage (equivalent monolayers) 2.0 2.5 2.9 2.0 0.91 3.1 3.8 2.4

A chromatogram of the samples was taken immediately following preparation to establish the initial purity. The samples were then placed in a 50 C oven under ambient atmosphere and allowed to react over the course of a couple of months, with aliquots periodically being withdrawn for analysis using HPLC. The amount of decomposition over time of PETN for most oxide surfaces was found to be very small. PETN proved to be inert or only slightly reactive on many of the metal oxide surfaces. The amount of PETN on the surface of Bi2O3, CuO, WO3, Fe2O3, MnO2, and SnO2 remains fairly constant over the period analyzed (~60 days) with only small changes in the concentration of PETN, which were within the experimental error of the measurement. However, PETN did show some decomposition on the surface of MoO3 as evidenced by a decline in the amount of PETN and the concomitant production of a brown gas, which can be observed visually in the sample container (figure 16). Increasing the ambient temperature at which the sample is stored results in a more rapid production of the gas.

32

Figure 16. The evolution of a brown gas from PETN on the surface of MoO3 after being stored at 50 C Reactivity of PETN with MoO3 Thermal analysis. Thermal analysis of pure pentaerythritol tetranitrate (PETN) using differential scanning calorimetry (DSC) is shown is figure 17. The thermal scan shows a sharp endotherm at 141 C, which corresponds to the melting and an exotherm, whose observed peak at this scan rate (10 C/min) is at 205 C, corresponds to the decomposition of the molecule. When the sample is analyzed in a sealed container, the decomposition becomes a deflagration event with an abrupt release of energy. When considering composites of PETN with inorganic oxides, the effect of the inorganicorganic interface on these well-understood thermal processes is of interest. The DSC plot of PETN at ~2 monolayer coverage on nano-scale MoO3 is shown in figure 18. At slightly less than 2 monolayer coverage of PETN on MoO3, the decomposition occurs at a significantly lower temperature of 142 C, with no melting endotherm being observed. The lack of a melting endotherm is expected since there is no crystalline bulk phase present. When the coverage of PETN is increased slightly to approximately 2 monolayers a melting endotherm begins to be observed at 138 C. The melting endotherm is

33

superimposed on the middle of the broad decomposition exotherm thereby splitting it into two apparent peaks at 112 and 144 C. There is also an exothermic process located at 192 C, which is lower than for the pure PETN presumably due to interfacial mediation at the surface of the oxide. An endotherm is also present at 292 C and 287 C for the <2 and 2

Figure 17. Differential Scanning Calorimetry (DSC) of pure PETN with a vented sample pan (note exothermic processes are above the baseline) monolayer samples respectively, which likely represents a desorption of some of the products from the initial decomposition. A sample with excess PETN on the surface of MoO3 shows an endotherm at 140.6 C and two exothermic processes, one whose peak is close to that resolved in the bulk (193 C), and a low temperature exothermic process at 164 C. This may be interpreted as coming from the superposition of the interfacial and the bulk processes observable at higher loadings of PETN, wherein the melting endotherm at 140.6 C, which is more pronounced at the higher loading, now dominates and obliterates most of the 141 C interfacial exotherm leaving only a residual spike at 164 C. The size of the 164 C exotherm, however, is quite large, therefore a more complex surface mediated decomposition process that results from the interfacial 34

reactions with the melt cannot be ruled out. An endothermic process just below 300 C is not seen in the bulk sample as it is in the monolayer samples due to the fact that it may be attributed to a decomposition that occurs at the surface, and is therefore solely an interfacial process. The DSC data of PETN deposited on nano-scale MoO3 is compared to that of PETN deposited on the larger micron size MoO3, which has a smaller surface area (figure 19). At a surface coverage of less than two monolayers, an endothermic process at 141 C indicating the melt of PETN is observed, as well as an exotherm at 195.4 C for the decomposition process. This exothermic process occurs at a lower temperature than that of the bulk PETN (205 C) showing that there is some interfacial mediation occurring, however, the fact that there is a melt present and only one decomposition peak indicates that there is mostly bulk crystalline phase present. In fact, similar processes are observed for the samples with a higher loading of PETN, showing that the same interfacial processes occurring on nano-scale MoO3 do not also occur on the micron scale MoO3. In short, when the surface area gets small enough, interfacial effects become less significant and are not observed in the thermal analysis or, alternatively, the large surface area provided by nano-scale materials will give rise to greater surface degradation effects. Comparable studies were performed on the remaining metal oxide surfaces, where a surface mediation effect was only observed at the surfaces of MnO2 and CuO, which showed the decomposition peak occurring at a lower temperature than that of pure PETN (149 and 175 C, respectively). It is important to determine whether the interfacial mediation of the decomposition temperature is specific to certain oxides or constituted a general surface effect. Thermal analysis of monolayers of PETN deposited on inert oxides and salts indicates that the decomposition temperature is not modified on all surfaces (figure 20). On fumed silica the decomposition temperature is very similar to that of the bulk PETN at 201 C. On simple salts such as KBr, there is also no detectable change in the decomposition temperature from that of the bulk. Thus the interfacial effect is not a general effect, but rather a specific interaction that occurs between PETN and the surface of nano-scale MoO3.

35

Figure 18. DSC of (a) <2, (b) 2, and (c) 6 monolayer coverage of PETN deposited on nano-scale MoO3

These results do not provide a chemical rationale for the interfacial effect, nor do they provide any specifics on the decomposition pathway(s) on the surface. It is clear from the thermal analysis that there is a significant surface reaction with MoO3. This effect might well mitigate energy release processes or affect long-term stability of heterogeneous composite mixtures.

36

Figure 19. DSC of (a) <2, (b) 2, and (c) 6 monolayer coverage of PETN deposited on micron-scale MoO3

Figure 20. DSC traces of PETN deposited on (a) SiO2 and (b) KBr Product analysis. The reaction processes of pentaerythritol tetranitrate (PETN)

deposited on the surface of molybdenum (VI) oxide were studied using liquid 37

chromatography and fourier-transform infrared spectroscopic techniques. The HPLC study was conducted over a period of three months wherein small aliquots of the sample were removed periodically for analysis. The chromatogram shows a modest (1.3 %) decrease in the amount of PETN present after 49 days (figure 21); however, no additional peaks in the chromatogram that might represent the decomposition products were observed, suggesting that the primary decomposition products are likely gaseous species that diffuse out of the solid. Only a small amount of net decomposition is observed chromatographically, which suggests that the reaction is localized at the interface and the catalytic decomposition of the bulk does not occur, at least at the temperatures studied.

100

99.5

99

98.5

Percent Remaining

98

97.5

97

96.5

96

95.5

95 0 1 3 7 14 time (days) 21 35 49 70

Figure 21. Percent PETN remaining on MoO3 after vs. time The sample for the gas phase FT-IR study was prepared as described previously. The container holding the sample was connected to a gas phase IR cell that had previously been evacuated. The sample was set to heat to 100 C, and spectra were periodically recorded over a period of two days (figure 22). Peaks associated with the decomposition of PETN began to appear during the first hour of heating. The products 38

observed by FT-IR were NO2 (1629, 1597 cm-1), N2O4 (1743 and 1270 cm-1), N2O (2236, 2212 cm-1), and CO2 (2359, 2342 and 3727-3600 cm-1).37 The band that occurs at 2918 cm-1 is due to the grease from the stopcocks of the gas cell heating slightly during the reaction and releasing a small amount of the grease into the cell. The band at 1270 cm-1 has been assigned to N2O4, which is in accordance with the reported values found by Mlen et al.38 However, this band seems to overlap with other fundamental vibrational bands of NO2 as well as N2O, which are at 1320 and 1285 cm-1 respectively. A control sample of the MoO3 without PETN prepared for FT-IR analysis showed that none of the products found were from the surface of the MoO3 alone.

e 350

300

d % Transmittance (a.u.) 250

200 c

150 b

100 a

50 3.75E+03 3.25E+03 2.75E+03 2.25E+03 1.75E+03 1.25E+03 7.50E+02 Wavenumbers (cm-1)

Figure 22. FT-IR spectroscopy of PETN on the surface of MoO3 at 100 C after a) 1, b) 5, c) 24, d) 27.5, and e) 46 hours (=CO2, =N2O, =N2O4, =NO2)

39

Conclusions From the standpoint of compatibility, it is clear that the interfacial chemistry between PETN and inorganic surfaces can be pronounced. As indicated by the thermal analysis studies, this involves at one level, changes in the normal thermal decomposition pathways of the molecule which may or may not have a net effect on the stability or energy release in bulk composite materials. For redox active metal oxide phases the interfacial processes can result in a net chemical reaction that will slowly convert PETN into another species, and possibly into degradation products. For the specific oxides reported here, the surface interactions of PETN with MoO3 produces a gas containing NO2, N2O, N2O4, and CO2 species. Clearly this result indicates that in composite energetic materials containing forms of MoO3, the PETN will be changing over time, which suggests that stability and energy release may also be variable.

40

CHAPTER 3

LASER INITIATION PROCESSES IN THERMITE ENERGETIC MATERIALS STUDIED BY A LASER DESORPTION IONIZATION (LDI) TECHNIQUE

Introduction The use of binary solid-state inorganic fuel/oxidant redox processes such as the classic aluminum/iron oxide thermite reaction in the production of energetic materials is of great interest due to their high energy densities. It has recently been discovered that the energetic properties of these materials may be enhanced when at least one of the components in the mixture is on the nanometer size scale. These materials, which are referred to as metastable interstitial composites (MIC) have a rapid release of energy that can be attributed to a more intimate fuel/oxidant contact that occurs in the nano-scale, which enhances mass transport and provides more power.5,6 Thermal initiation of thermite materials typically involves heating above the melting point of one of the components, typically the metal. This starts the reaction processes that liberate heat, causing the reaction to accelerate. Ignition occurs when the reaction becomes self-sustaining and propagates. For laser initiation, the incident energy put into the sample by the laser causes high localized heating, which leads to ignition of the bulk. The time to ignition is defined as the time at which the energy released by the reaction becomes greater than or equal to the energy put into the composite by the laser. Both the laser excitation event and the subsequent ignition and propagation of the mixture are extremely high-temperature events which generate both liquid and gas phase (plasma) species. Chemical reactions between species occurring within and between these phases and with the solid material all contribute to the net combustion process. This complexity coupled with the high temperatures makes these reactions difficult to study on a microscopic level. This has been addressed recently by Dlott et al. by using timeresolved spectroscopy and initiating the reaction between aluminum and nitrocellulose by flash-heating with a 100 ps laser pulse in the near infrared.8,9 Short pulse photo-thermal initiation allows the use of time-resolved spectroscopic techniques to monitor reaction

41

dynamics, elucidate some of the intermediate species produced during the reaction, and to observe specific structural changes in the reactive components. The reactive processes that occur during the laser initiation of aluminum/iron (III) oxide metastable intermolecular composites (MIC) have been studied by laser-induced desorption ionization time-of-flight mass spectrometry. The ions observed in the plume from the aluminum show fragments from the ablation of the oxide coating and from the metal core. Ablation of the iron oxide component consists primarily of pure iron species such as [Fe]+ and [Fe2]+ and small oxides such as [FeO]+ and [Fe2O]+ ions. In smaller quantities, metal oxide clusters that are either oxygen deficient, [Fe(FeO)x]+, or oxygen equivalent [(FeO)x]+, are observed. When the thermite composite is initiated, mixed metal species are observed in the plume, which correspond to the aluminum substitution analogues of the iron oxide clusters, specifically, [FeAl2O3]+, [AlFe2O3]+, and [AlFe2O2]+. Notably, the amounts of these mixed metal products that form are inversely proportional to the size of the aluminum particles. This suggests that the decrease in ignition time observed in MIC materials is due to the more facile liberation of reactive metallic aluminum when the particle size is small.

Experimental

Materials The 50 and 100 nm aluminum samples were obtained from Argonide, while the micron aluminum was purchased from Alfa Aesar. All aluminum samples were used as received. The iron (III) oxide powder, <0.25 microns, was purchased from Aldrich and was used as received. The size and morphology of the aluminum samples were characterized by field emission scanning electron microscopy (figure 23). The nano-scale materials are somewhat polydispersed, with the average particle size being around 50 and 100 nm for the two sizes, although both showed some particles above 200 and below 50 nm. Both samples showed some agglomeration of the particles, which was more pronounced in the 50 nm material. The aluminum, which we designate as micron scale, was a very polydispersed material composed largely of particles between 0.5 and 2.5 m in size with

42

only a few large particles around 5 m observable. The average size of the micron sample, averaged from the particles observable in the SEM, was 1.75 m. The active aluminum content and the estimated oxide thickness for the three samples were determined from weight gain measurements using thermogravimetric analysis on a TA Instruments Q500 TGA. The active aluminum content for the 50 and 100 nm and the 2.5 m materials was 71.5, 75.5, and 98.0 %, respectively, while the oxide thickness was determined to be 2.6, 4.5, and 6.0 nm, respectively.

Figure 23. SEM images of (a) 50 nm, (b) 100 nm, and (c) micron-scale aluminum particles Preparation of samples The thermite samples were prepared by mixing Fe2O3 with each of the aluminum particles at a 1:1 molar aluminum to iron ratio. The samples were thoroughly mixed by grinding the components in a mortar and pestle to ensure a homogenous distribution. Analysis of the samples after grinding using powder X-ray diffraction indicated that no new phases were formed as a result of sample preparation. The samples were then packed into a sample holder for laser desorption ionization time-of-flight mass spectrometry (LDI/TOF MS) along with the control samples of Fe2O3 and the 50 and 100 nm and the micron aluminum. LDI/TOF mass spectrometry The LDI/TOF mass spectra were collected using a commercial 1999 Bruker Biflex III matrix-assisted laser desorption and ionization time-of-flight mass spectrometer (MALDI-TOF) (Bruker Daltronics, Inc.) fitted with a solid sample holder for the introduction of inorganic material. The instrument was equipped with a pulsed nitrogen

43

laser at 337 nm with a peak width of 4 ns. All spectra were collected in linear mode with positive ion detection averaging 200 shots from the laser. In order to thermally initiate thermite, temperatures at or above the melting point of aluminum must be obtained. For laser pulses in the nanosecond regime, there is enough time for thermal transport into the material. The temperature attained in each component of the thermite can be estimated from the absorbed energy per unit volume, Ev, using equation 3, where J is the fluence at the center of the beam, R is the reflection coefficient, D is the thermal diffusion length, which is the depth of the sample that is heated with a single pulse from the laser, C(T) is the heat capacity, and is the density. The thermal diffusion length is calculated from the thermal diffusivity (D) and the pulse length of the laser (equation 4).8,9,39 Ev = J (1 - R) / D = D = (2Dt)1/2

Tf Ti

C (T )dT

(3) (4)

It was found that all of the aluminum was melted when the laser reached an energy density of 0.158 J/cm2. The volume of aluminum that was heated per pulse was approximately 7.42 x 10-9 cm3. Although the volume of the sample that was hit by the laser did not change, the active aluminum content would vary based on the size of the aluminum particle. The Fe2O3 phase composite also absorbed at 337 nm and was found to reach its melting point of 1565 C at an energy density of 0.474 J/cm2. Temperatures required to generate a plume from which species could be detected in TOF-MS were found to be above these threshold values of melting.

UV-Vis diffuse reflectance The reflectivity of the independent Al and Fe2O3 components and the compounded thermite were measured as total reflectance against a calibrated Spectralon diffuse scattering reference in a Perkin-Elmer Lambda 900 spectrophotometer equipped with a 160 mm integrating sphere.

44

Results and Discussion Aluminum samples in the three particle sizes, 3-5 m and 50 and 100 nm, were placed as tightly pressed powders in the laser desorption ionization instrument. Time-offlight mass spectra observed for the 50 nm Al as a function of incident laser power are shown in figure 24. No gas-phase ionic products were observed below a threshold energy density of 0.948 J/cm2. Above this threshold, the major products included [Al]+, [Al2O]+, and [Al2]+ at m/z of 26.9, 69.9, and 53.9, respectively. At power densities between 1.74 and 1.90 J/cm2, small amounts of more complex AlxOy products were also observed including [Al2O2]+, [Al3O]+, and [Al3O2]+ at m/z 85.9, 96.9, and 113.0, respectively (figure 24b). Since the system was evacuated, the oxide species must have originated either directly or indirectly from the native oxide layer on the aluminum powder. Laser desorption of Al2O3 produced a large number of neutral species, with the dominant ones

Figure 24. LDI-TOF mass spectra of 50 nm aluminum at a laser energy density of (a) 1.74 and (b) 1.90 J/cm2 being AlO, Al2O, and Al2O2.40,41 In fact, AlO has been observed in the emission spectrum of combusting Al/MoO3 thermite materials.10 Since the observed aluminum oxide ion clusters are oxygen deficient, they are likely formed from ion-molecule reactions in the plasma between Al+ and neutrals such as those given in equations 5-7. Clearly, other pathways may also be operating. The species AlO+ and Al2O2+ are only observed at a higher laser power and may represent direct ion yields from ablation of the Al2O3 layer.

45

Al+ + AlO Al+ + Al2O Al+ + Al2O2

Al2O+ Al3O+ Al3O2+

(5) (6) (7)

Studies by Granier and Pantoya on the fuel particle size dependence of laser ignition of Al/MoO3 MIC materials showed a general trend of faster ignition times as the Al particle size dropped to nanometer dimensions.15 This effect was attributed to melting point depression that accompanies the size reduction in the aluminum. Also important in the laser ignition process is the presence of a native oxide layer on the aluminum. The oxide layer is thickest on micron-size aluminum and decreases to approximately 2-4.5 nm as the aluminum particle size reaches the nanometer dimension. Notwithstanding the thinner oxide layer, the percent of active aluminum metal available will also decrease (i.e, Al2O3 represents a progressively higher percent of the mass of the particle). A direct effect of this is a reduction of the burn rate at very small particle sizes. The process of laser initiation will involve melting of the aluminum, the density change of which is sufficient to critically stress and break the oxide layer.42 As indicated above, the ions that are observed in the plasma can be associated with both the oxide and the aluminum phase of the nanoparticles. Since the oxide has an extremely high melting point and a weak optical absorption, most of the heating during the pulse will be of the metal itself. Above the energy density threshold, the three dominant species are [Al]+, [Al2]+, and [Al2O]+, which reflect contributions from the volatilization and ionization of the aluminum metal and the breakdown of the aluminum oxide shell. The amounts of these species observed for all three Al particle sizes at an energy density of 1.90 J/cm2 are shown in figure 25. The 100 nm particles show a large yield of [Al]+ and a small yield of [Al2O]+, while the converse is true for the 50 nm particles, which show a larger yield of [Al2O]+ than [Al]+. This difference likely arises from the higher percent of oxide present relative to the active aluminum content in the 50 nm material, the neutrals of which will deplete more of the available [Al]+ in the plume, according to equation 5. The 100 nm sample has a higher active aluminum content relative to the amount of oxide present, which gives a higher [Al]+ ion yield. Interestingly, the amount of each Al species observed for the micron-size sample is between that of the 50 and 100 nm samples. It yields more [Al2O]+ than the 100

46

30000

micron Al 50 nm Al 100 nm Al

25000

20000

Intensity 15000

10000

5000

0 [Al]+ [Al2]+ [Al2O]+

Figure 25. Aluminum ions formed at a laser energy of 1.90 J/cm2 for each aluminum particle size nm Al, consistent with its much thicker oxide layer, but significantly less [Al]+, which seems inconsistent with its high available Al content. The origin of this may lie in its higher melting point, which means that less of the heat deposited by the laser pulse is used to vaporize the sample. The power dependence of the ion yields is consistent with this suggestion. The total yield of the most prevalent ions ([Al]+, [Al2]+, and [Al2O]+) as a function of energy density indicates that at the threshold, ion yields are low and relatively independent of particle size (figure 26). As the power increases, the nanometer-scale particles produce the highest ion yields, which are consistent with the shorter ignition times observed in the bulk materials. At the very highest energy density, however, the micron-size sample begins to exceed that of the nano-scale aluminum. The trends observed are explainable in the context of the free energy of the active aluminum and the oxide layer as a function of particle size. The free energy of the active aluminum in the volume becomes more positive (i.e. less stable) as the size decreases, which gives rise to the melting point depression. Conversely, the free energy of the surface oxide, which becomes the dominant contribution to the total free energy as the particles become smaller, becomes more negative, making the oxide layer more stable and, hence, less easily ablated into the plume.43 Notably, the results are also extremely consistent with the 47

recent melt dispersion model of Levitas et al., which suggests that the stronger oxide layer on the nanoparticles results in large pressure changes during rapid melting, causing pressure-induced spallation of the oxide layer followed by a rapid dispersion of aluminum clusters.44 It is important to note that the laser-induced ignition of a thermite reaction will necessarily involve many reactions taking place in different phases of the irradiated materials. In this study, while we are observing a subset of those reactions, the species we are observing do afford support at a molecular level of the proposed advantages of nanoscale fuels on factors such as ignition time and burn rate.15

50000 45000 40000 35000 30000 Intensity 25000 20000 15000 10000 5000 0

micron 50 nm Al 100 nm Al

1.738

1.896 Energy Density (J/cm )

2

2.054

Figure 26. Major aluminum ions formed ([Al]+ + [Al2]+ + [Al2O]+) for each particle size at increasing energy densities Irradiation of the Fe2O3 sample results in ionic products at a threshold power density of 1.58 J/cm2. At the threshold, two primary species are observed, [Fe]+ and [FeO]+, along with K and Na impurities (figure 27a). As the laser power is increased, more complex species are observed, in particular, [Fe2O]+ and [Fe2O2]+, with minor products including [Fe2]+, [Fe3O2]+, and [Fe3O3]+ (figure 27b). Maunit et al. studied the laser ablation/ionization of iron oxides in considerable detail, and our spectra parallel the ones obtained by these authors under nonresonant excitation conditions and below the threshold of Fe-O bond dissociation.45 As suggested in this previous study, neutral FeO is an important constituent of the plume and contributes to the cluster formation through 48

ion-molecule reactions. The formation of [Fe2]+ was found to be through an ion-molecule reaction between Fe+ and neutral FeO that generates oxygen (equation 8), and FeO+ and Fe2O+ are subsequently formed through an ion-molecule reaction with the generated oxygen (equations 9 and 10). The larger clusters that are formed in the plume are either Fe+ + FeO Fe+ + O2 Fe2+ + O2 Fe2+ + O2 FeO+ Fe2O+ (8) (9) (10)

oxygen deficient, Fe(FeO)x, or oxygen equivalent, (FeO)x, and are also formed from ionmolecule reactions between neutrals such as FeO and precursor ions.

Figure 27. LDI-TOF mass spectra of iron(III) oxide at a laser density of (a) 1.58 and (b) 1.74 J/cm2 The TOF mass spectrometry characterization of the species forming in the plumes of Al/Fe2O3 thermite mixtures was carried out over a range of laser powers. Of interest are species that form between the two components, which represent gas-phase reactions that are part of the laser ignition process. No products were observed until an energy density of 1.11 J/cm2 was reached. At this threshold, however, the only products observed were those produced from the independent components. It was not until an energy density of 1.58 J/cm2 was reached that products consisting of Al/Fe mixed components were observed. At this threshold, small amounts of [AlOFe]+ began to appear in the spectrum (figure 28a,b). As discussed above, in the pure iron oxide, [Fe2O]+ 49

was formed from the reaction of [Fe2]+ with O2 (equation 10). While this reaction is possible for the formation of [AlOFe]+, only trace amounts of the precursor ion, [AlFe]+, were observed in the spectrum. This suggests that the [AlOFe]+ likely results from an ion-molecule reaction between neutral FeO and aluminum ions (equation 11). Obviously, FeO + Al+ FeOAl+ (11)

since species such as AlO and Fe+ are also present in the plume, their reaction may also contribute to the mixed metal product. As photo-thermal heating of the thermite mixture increases, more complex Al/Fe oxide species appear (figure 28d). All of the observed clusters are aluminum-substituted analogues of the larger ion clusters; specifically, [Al2FeO3]+ and [AlFe2O3]+ are related to [Fe3O3]+, while [AlFe2O2]+ is an analogue of [Fe3O2]+. The relative amounts of the pure iron clusters produced from the Fe2O3 control and from the thermite are shown in figure 29. It can be seen that [FeO]+ and [Fe2O]+ are actually produced in higher quantities in the thermite, while the larger pure iron clusters are only produced in very small amounts. This is consistent with the formation of the mixed oxide clusters at the expense of the pure iron clusters due to competition in the plume between Al+ and Fe+ for reaction with molecular FexOy species. The fact that [Fe2O]+ appears to be independent of the formation of [AlOFe]+ is consistent with their respective formation through different mechanistic pathways. The production of mixed metal products correlates strongly with the size of the aluminum. The energy density threshold at which mixed metal products are observed is lower for the nano-scale particles than that for the micron-scale particles (1.58 vs. 1.73 J/cm2). More significantly, the quantities of all of the mixed metal components increase with decreasing aluminum particle size (figure 30). This provides direct evidence for the suggestion that the shorter ignition times that accompany smaller particle sizes are due to the lower melting point of the nano-scale materials and the active aluminum content available in the particle that, for a given laser power, produces more reactive species.

50

Figure 28. LDI-TOF mass spectra of thermite mixtures at a laser energy density of (a) 1.42, (b) 1.58, (c) 1.74, and (d) 1.90 J/cm2

16000

Fe2O3 standard Thermite Reaction

14000

12000

10000

Intensity

8000

6000

4000

2000

0 [FeO]+ [Fe2O]+ [Fe2O2]+ [Fe3O3]+ [Fe3O2]+ [Fe4O4]+

Figure 29. Relative amounts of pure iron cluster species formed during laser desorption of the Fe2O3 control and thermite samples at an energy density of 2.054 J/cm2 51

16000

50 nm Al 100 nm Al micron Al

14000

12000

10000

Intensity

8000

6000

4000

2000

0 [AlOFe]+ [AlO2Fe]+ [Al2O3Fe]+ [AlO2Fe2]+ [AlO3Fe2]+

Figure 30. Mixed Al/Fe oxide species formed at an energy density of 2.054 J/cm2 for 50 and 100 nm and micron aluminum thermite mixtures Conclusions This study has provided insight into some of the reactions that occur between aluminum and iron oxide species in the plasma phase of the binary energetic system during laser initiation. In particular, ionic species generated directly from the laser desorption and through various ion molecules are observed for both the aluminum and iron oxide components. When the binary thermite is laser-initiated, mixed metal ionic species are produced in the plume. These species are mixed metal analogues of the iron clusters and are believed to form from the competition between Al+ and Fe+. The amount of mixed metal clusters that form in the plasma increases as the size of the aluminum particles decreases. This provides a direct verification that the shorter ignition times observed with decreasing fuel particle size are due to a lower melting point that liberates more reactive aluminum during laser incidence.

52

CHAPTER 4

PHOTO-THERMAL INIATION PROCESSES OF ORGANIC-INORGANIC HYBRID METASTABLE INTERSTITIAL COMPOSITE (MIC) MATERIALS

Introduction The organic high-energy material pentaerythritol tetranitrate (PETN) was incorporated at low concentrations into Al (100 nm)/Fe2O3 metastable interstitial composites (MIC) to form a hybrid organic/inorganic high-energy material. Studies of the dynamics of energy release were carried out by initiating the reaction photo-thermally with a single 8 ns pulse of the 1064 nm fundamental of a Nd:YAG laser. The reaction dynamics were measured using time-resolved spectroscopy of the light emitted from the deflagrating material. Two parameters were measured: the time to initiation and the duration of the deflagration. The presence of small amounts of PETN (16 mg/g MIC) results in a dramatic decrease in the initiation time. This is attributed to a contribution to the temperature of the reacting system from the combustion of the PETN that, at lower loadings, appears to follow an Arrhenius dependence. The presence of PETN was also found to reduce the energy density required for single-pulse photo-thermal initiation by an order of magnitude, suggesting that hybrid materials such as this may be engineered to optimize their use as an efficient photodetonation medium.

Experimental

Materials The iron (III) oxide powder, <0.25 m, was purchased from Aldrich, while the 100 nm aluminum was obtained from Argonide. All materials were used as received. Pentaerythritol tetranitrate (PETN) was prepared according to the literature and stored in acetone (Aldrich, HPLC grade) for safety purposes.35

53

Preparation of samples Pentaerythritol tetranitrate (PETN) was deposited onto the iron (III) oxide by a wet impregnation technique. A small amount, up to 1.053 mL, of PETN standard solution (~ 0.5 M) was added to 0.5 grams Fe2O3 in order to obtain varying loading of PETN in the final composite (Table 3). The acetone of the stock solution of PETN was subsequently removed under a vacuum (10-3 torr) leaving a dry powder. The thermite samples were prepared by mixing Fe2O3/PETN samples with 100 nm aluminum particles in a 1:1 molar aluminum to iron ratio. The samples were thoroughly mixed by grinding the components together using a mortar and pestle in order to ensure a homogenous mixture. Approximately 0.25 grams of this mixture was compressed in a die at an applied pressure of 46,000 psi. The resulting pellets were 6.39 mm in diameter by 3.28 mm giving an average density of 2.38 g/cm3. Table 3. Effects of PETN on heat and energy given off by MIC composites mg PETN/g MIC composite 16.0 32.0 64.0 95.9 128 Fraction of heat given off by PETN 0.0343 0.0615 0.116 0.164 0.208 Energy PETN/ g composite (kJ/g) 0.130 0.260 0.521 0.781 1.04

Nd:YAG laser A Spectra-Physics DCR-3G Nd:YAG laser using single pulses of both 1064 and 532 nm light was used in order to combust the thermite/PETN samples. At 1064 nm an 8 ns 900 mJ pulse was focused through a fused silica 75 mm focal length lens onto the pellet, while at 532 nm the laser pulse was 6 ns and 360 mJ. The focused beam diameter was calculated to be 11 m giving an energy density of 8.8 x 105 J/cm2 at 1064 nm and 5.7 m with an energy density of 1.4 x 106 J/cm2 at 532 nm. The laser beam was focused through a hole in a protective steel plate, with the hole measuring 2.54 cm in diameter,

54

which was covered by a fused silica window measuring 24x23x2 mm in order to catch any debris from the combustion of the pellet. The kinetics of the combustion flash was recorded using three photodiode detectors (Thorlabs DET210, 0.8 mm2 Si PIN, 1 ns rise time) and recorded by digital storage oscilloscopes (LeCroy LC 564A, 1 GHz bandwidth, 10-50 s sampling rate for slow signals). The photodiodes were placed 45 degrees from the laser beam axis, looking down at the sample through a fused silica window and filtered by either a heat absorbing filter (Hoya HA30) to reduce the 1064 nm scattered light (fast signal) or a 520 nm bandpass filter (Thorlabs FB520-10, 10 nm FWHM) for both fast and slow signals. Spectra were collected through a 3 mm diameter liquid light guide (Oriel 77554, NA/0.47, >50% transmittance from 270 to 720 nm) placed above the sample at 45 degrees from the laser beam axis coupled into a 300 mm spectrograph (Acton Research Corporation Spectra Pro 308i, 150 gr/mm grating blazed at 300 nm) using a biconvex fused silica lens (Spex, 1 inch diameter, f = 28.6 mm) and the light was dispersed onto a back-thinned CCD (1340x400 pixels, 20 m square, Princeton Instruments LN/CCD-400EB-G1 operated at -90 C). The timing for single shot operation was coordinated by triggering the CCD shutter, laser, and oscilloscope using a digital delay generator (EG&G PAR 9650) to trigger the CCD shutter to open for 30 ms and delaying the laser and oscilloscope triggers for 10 ms. The CCD exposure time, set in the software to 10 ms, was determined to be 30 ms by adjusting the laser trigger delay while observing the second harmonic 532 nm laser light.

Results and Discussion The effect of PETN incorporation on the dynamics of Al/Fe2O3 MIC initiation and deflagration was investigated using time resolved spectroscopic techniques. In general, thermal initiation of the thermite reaction occurs when the reaction temperature reaches or exceeds the melting temperature of the fuel, in this case aluminum. At this point there is a breakdown of the native oxide layer on the aluminum and the mass transport between the fuel and the oxide becomes large enough to propagate the reaction.42 The combustion of the samples in this instance was achieved photo-thermally

55

using a single 8 ns pulse of the 1064 nm fundamental from a Spectra-Physics DCR-3G Nd:YAG laser, which allows the use of time-resolved spectroscopic techniques. As such, the dynamics were monitored over the total time regime of the process by light emitted from the reacting material, collected through a band-pass filter with a max = 520 nm. The time-resolved data for the Fe2O3/100 nm Al MIC materials containing various amounts of PETN is shown in figure 31. The emissive plume generated by the laser incidence event on the surface of the sample can be seen as a sharp spike, indicated by an arrow in the figure, while the irregular intensity that follows in time is the plasma emission from the deflagration of the sample. An important quantitative parameter of the reaction dynamics is the ignition time. For photo-thermally initiated binary fuel-oxidant materials this is defined as the time at which the energy released by the reaction becomes greater than or equal to the energy put into the composite by the laser. For data collected photonically, the ignition time is generally taken as the time between the laser excitation and when emitted light can be detected. In the time-resolved system employed here, the sensitivity and dynamic ranges of the photodiodes is high, allowing the very early stages of combustion to be observed. In analyzing the data, the ignition time is defined as the period between the sharp plume emission from the initial laser incidence and the point at which the intensity of the light emitted from the combusting sample is twice the intensity of the noise, as measured in the pre-trigger region of the data. A plot of the initiation time against the amount of PETN in each sample at 1064 nm is shown in figure 32. The ignition time at 1064 nm for the pure thermite sample was found to be 6.05 0.28 ms, however, with the incorporation of PETN (16 mg/g thermite) the ignition time drops precipitously to 3.35 0.28 ms and continues to decrease until a minimum value of 1.60 0.354 ms is observed with 64.0 mg PETN/g thermite. Above this amount of PETN there is no further decrease and, in fact, it appears to plateau or even increases slightly. For safety reasons it was not possible to obtain data for higher concentrations of PETN.

56

Figure 31. Photo-thermal initiation of Fe2O3/100 nm Al samples with increasing amounts of PETN at 1064 nm (arrow indicates the position of the laser plume emission)

7 Initiation Time Exponential Decay Fit 6

Initiation Time (ms)

0 0 20 40 60 80 100 120 140 PETN (mg/g thermite)

Figure 32. Initiation time (ms) as a function of PETN coverage in mg at 1064 nm

57

The intense, irregular peak that follows ignition (figure 31) is the time-evolution of the plasma emission from the deflagration of the sample. The apparent oscillations in the emission intensity are attributed to modulation of the light by the deflagration process with the duration of this emission being directly related to the total time it takes for the sample to deflagrate. From the data, the deflagration duration is measured from the previously determined ignition point to the time where the emission intensity has returned to a value of twice the noise level. A plot of the deflagration duration as a function of PETN loading is shown in figure 33. As can be seen in figures 31 and 33, the incorporation of a small amount of PETN into the MIC material has a pronounced effect on the deflagration duration. The deflagration of a control sample of Al/Fe2O3 MIC materials with no organic phase present shows a weak broad emission with a long (282 ms) duration. The addition of a small amount of PETN (16.0 mg/g thermite) results in a significant decrease in the deflagration duration to 88 ms. The deflagration duration continues to decrease as more PETN is added, albeit more gradually, until it plateaus at 31.5 ms for 64.0 mg of PETN per gram of thermite. Scrutiny of the spectral data in figure 31 shows that, in general, the decrease in deflagration duration is accompanied by an increase in the emission intensity, which suggests that, as expected, more energy is being released in a short time period.
300 1064 nm 532 nm

250

Deflagration Duration Time (ms)

200

150

100

50

0 0 20 40 60 80 100 120 140 PETN (mg/g thermite)

Figure 33. Deflagration duration time (ms) as a function of the amount of PETN at 1064 nm

58

The changes observed in both the ignition time and the duration of deflagration as a function of PETN addition appear to follow the same basic trend: increasing amounts of PETN result in a decrease in the time to ignition and the deflagration duration, the effect of which ultimately plateaus at about 64.0 mg per gram of thermite. In general these effects can be thought of as originating from the contribution of the heat of combustion of the organic phase to that of the fuel/oxidant inorganic matrix, which acts to accelerate the binary reaction. PETN is a high-energy molecule with a heat of combustion of 2,572 kJ/mol.46 The amount of energy provided by the PETN and its contribution to the total energy output of the composite material at various levels of incorporation is given in Table 3. At the highest loadings studied the PETN accounts for about 20% of the total energy liberated by the composite during combustion. The effect of the PETN on the initiation time is likely due to its combustion during or immediately after the laser excitation event. Specifically, the laser excitation rapidly heats a volume of the composite defined by the diameter of the beam and the thermal diffusion length of the material. During the course of the 8 ns pulse both the aluminum and the iron oxide are heated with the relative temperature attained by each phase being dependant primarily on its optical absorbance at 1064 nm and its thermal diffusivity. The PETN is not absorbing at this wavelength so its combustion occurs from heat flow from the other components. This will happen relatively efficiently since the combustion temperature of the PETN is 205 C while the temperature reached by the inorganic components will be at or above their melting point. The PETN will contribute heat back into the inorganic components when combusted, thereby raising the temperature further and accelerating the initiation processes. Since even at the lowest incorporation level the combustion of PETN provides sufficient heat to melt all of the aluminum in the composite, its contribution to the initiation process will be substantial; this is directly reflected in the data. The decrease in the initiation time as a function of PETN concentration (figure 32), particularly in the region before the plateau in the time is reached, is not linear but instead appears to be approximately exponential. This suggests possible Arrhenius behavior in the rate (lifetime) of the initiation. Earlier studies in the rates of binary fueloxidant systems have suggested that behavior of the rate is Arrhenius in the early stages

59

before auto-acceleration of the reaction occurs due to the extreme exothermicity.11-13 This suggests that Arrhenius-type behavior may also be valid during the initiation process prior to deflagration. The rate (lifetime) of the initiation in the composite will depend on the temperature of the material, which is set to an initial value by the laser pulse and is, subsequently, augmented by the combustion of the PETN. The contribution of the PETN to the final temperature will be approximately proportional to the product of the mass of PETN and the heat of combustion: T H* m. An Arrhenius expression may be written exploiting this and shown in equation 12, where the initial temperature, To, attained from the laser pulse and the contribution from the PETN is a function of the mass, m, with the constant, C, representing the heat of combustion of the PETN and the thermal conductivity and heat capacity of the inorganic components that are being heated by it.

tinit

= Ae

Ea R (T0 +Cm )

(12)

This form of the Arrhenius equation fits well to the initiation time data, particularly at low PETN loadings, which is consistent with the basic model of how the PETN affects the dynamics as shown here. The deviation from Arrhenius behavior as initiation time plateaus at PETN loadings greater than 64 mg/g thermite mixture is somewhat unclear but may simply be that the maximum rate has been reached and that the additional heat placed into the melting and vaporization of the inorganic components will have only a minimal effect. As discussed, the deflagration duration (figure 33) of the reaction decreases rapidly with the addition of PETN and, like the initiation time, it reaches a relatively constant value at 64 mg/g. The lifetime of the deflagration is not described by the Arrhenius temperature dependence; however, this is expected due to the high exothermicity and concomitant auto-acceleration of the reaction. The leveling off of the deflagration time may represent a limiting value determined by the PETN combustion time, initiated in this fashion, as it begins to dominate the energetics of the composites.

60

Studies of the ignition time and deflagration duration were carried out with visible excitation at 532 nm. Since this wavelength is achieved through frequency doubling of the 1064 nm fundamental the highest energy density that could be realized was 1.4 x 106 J/cm2. At this energy density it was not possible to initiate the combustion of the pure Al/Fe2O3 MIC materials; however the presence of a small amount of PETN (10.0 mg/g thermite) resulted in reproducible single-pulse initiation. This effect is explainable in terms of prior discussion. While the energy provided at 532 nm is insufficient to initiate the pure thermite it does provide sufficient heat to combust the PETN, which then provides enough heat to drive the combustion. Notably, PETN does not absorb at 532 nm so there is not optical contribution to the effect. Consistent with this interpretation, the initiation time for 532 nm excitation (figure 34) is always longer than what is observed for 1064 nm excitation which is expected from less total heat going into the system and the fact that initiation is achieved secondarily through the PETN. The deflagration duration is essentially the same as that observed with 1064 nm initiation (figure 33) which is consistent with the idea that the duration of deflagration is dictated by the energy release of the composite system and is independent of the initiation conditions.

4 Inititation Time (ms)

0 0 20 40 60 80 100 120 140 PETN (mg/g thermite)

Figure 34. Initiation time of Al/Fe2O3 MIC materials with various concentrations of PETN initiated with a single pulse of 532 nm radiation

61

The results at 532 nm excitation suggest that one of the propitious effects of PETN incorporation is to lower the threshold energy required for single pulse initiation. This could have an impact on strategies for developing efficient photodetonation systems. To quantify this effect, the initation threshold energy was determined for the range of PETN loadings using 1064 nm initiation. The threshold energy density required to initiate pure Al/Fe2O3 MIC samples at 1064 nm is quite high at 4.7 x 105 J/cm2. The addition of PETN results in a relatively linear decrease in threshold energy required for initiation at 1064 nm up to 64.0 mg PETN/g thermite with an energy reaching a minimum value of 5.03 x 104 J/cm2 at 95.9 mg PETN/g thermite (figure 35) which represents almost an order of magnitude decrease in the necessary energy density.

6.00E+05

5.00E+05

Energy Density (J/cm2)

4.00E+05

3.00E+05

2.00E+05

1.00E+05

0.00E+00 0 20 40 60 80 100 120 140 PETN (mg/g thermite)

Figure 35. Energy density (J/cm2) required in order to initiate the thermite/PETN mixture at 1064 nm

62

Conclusions In conclusion, the addition of high energy materials, in this case pentaerythritol tetranitrate (PETN), to metastable interstitial composites (MIC) such as thermite-tyoe Al/Fe2O3 compositions changes the energy release dynamics compared to that of traditional MIC compositions. This is observed as a decrease in the initiation time and the duration of deflagration of the sample. This effect is attributed to the contribution to the overall heating of the sample due to the combustion of the PETN. Since the PETN does not absorb light, the process occurs through the indirect heating of the PETN from the inorganic phases to the point where combustion occurs. The initiation time and deflagration duration decreases with increasing PETN loadings but reaches a minimum at about 64.0 mg PETN/g thermite. Compositions containing PETN were found to require a much lower threshold energy for photo-thermal initiation. These results indicate that the incorporation of an organic phase, even at low concentrations, can have a profound effect on the reaction dynamics of MIC materials. This affords the possibility of tailoring the organic phase to better optimize desired properties of the materials.

63

SUMMARY This study has provided some insight into the stability and degradation processes of traditional energetic materials, such as 2,4,6-trinitrotoluene (TNT) and pentaerythritol tetranitrate (PETN) when placed in physical contact with metals or metal oxides. It was found that in the presence of the metal oxides MnO2 and CuO, a clean demethylation of TNT occurred, leaving the product trinitrobenzene (TNB). A rapid decomposition of PETN, on the other hand, to gaseous species primarily made up of NO2 along with N2O4, N2O, and CO2 occurred when placed on the surface of MoO3. These results indicate that an interfacial surface effect occurs, which was observed as a change in the normal thermal decomposition pathways of the molecule and may or may not have a net effect on the stability or energy release in bulk composite materials. Also of interest was the reaction processes occurring during the initiation of binary inorganic metastable interstitial composite (MIC) materials of aluminum and iron(III) oxide as well as the energy release dynamics of hybrid organic/inorganic MIC materials through laser initiation techniques. A laser desorption-ionization (LDI) technique coupled with a time-of-flight mass spectrometer was used in order to show any species occurring during the initiation process of MIC materials as ions reacting in the plasma phase. Mixed metal oxides, such as [AlOFe]+, form from analogues of the iron clusters the competition between Al+ and Fe+ in the thermite reaction. The energy release dynamics of the classic thermite reaction in combination with traditional high-energy materials, such as pentaerythritol tetranitrate (PETN), were studied by time-resolved spectroscopy in order to monitor the light emitted from the plasma of the reacting material using a Nd:YAG laser. A study of the initiation time, measured as the time between the plume of the reacting materials brought on by the laser pulse and the point where the intensity of the light becomes twice that of the noise, was performed. This study shows a decrease in the initiation time as a function of the amount of PETN present. Also, the total deflagration duration, measured as the total light emitted from the reacting plasma, decreases with increasing amounts of PETN. The threshold energy required in order to initiate the thermite reaction was also found to decrease significantly as the amount of PETN present increases. Since PETN does not absorb light

64

at the wavelengths used, but enough energy is provided to melt the inorganic material, heat must be transferred from the Al/Fe2O3 mixture to the PETN, allowing combustion to occur. The heat of combustion of the PETN then provides enough heat back into the inorganic mixture to affect the energy release dynamics of the system. This effect suggests that possible tailoring of the organic phase may be possible in order better optimize the properties of these materials.

65

REFERENCES 1. Cross, C.S. Explosive Compositions, Chem. Abs., 1969:526727, U.S. Dept. of the Army: Great Britain, (1969). 2. DeVilliers, R. Molecular Explosives Containing Metal-Metal Oxide Mixtures as Booster or Primer Compositions, Chem. Abs., 2004:675703, Metlite Alloys: South Africa, (2004). 3. Jones, J.W. High-Energy-Density Composite Explosive Containing Thermite Compositions Dispersed in Primary High Explosives, Chem. Abs., 2002:832736, U.S.P. Office, Lockheed Martin: USA, (2002). 4. Kloeber, M. And W. Schwartz. Laser-Sensitive Igniters, Chem. Abs., 1987:499288, Diehl G.m.b.H. and Co.: Germany, (1987). 5. Armstrong, A.W., Baschung, B., Booth, B.W., and Samirant, M. Nano Lett. 2003, 3, 253. 6. Pantoya, M.L., Granier, J.J. Propellants, Explosives, and Pyrotechnics. 2005, 30, 53. 7. Wang, L.L., Munir, Z.A., and Maximov, Y.M. Journal of Materials Science. 1993, 28, 3693. 8. Wang, S.F., Yang, Y.Q., Sun, Z.Y., and Dlott, D.D. Chemical Physics Letters. 2003, 368, 189. 9. Yang, Y.Q., Sun, Z.Y., Wang, S.F., and Dlott, D.D. Journal of Physical Chemistry B. 2003, 107, 4485. 10. Moore, D.S., Son, S.E., and Asay, B.W. Propellants, Explosives, and Pyrotechnics. 2004, 29, 106. 11. Boddington, T., Cottrell, A., Laye, P.G., and Singh, M. Thermochimica Acta. 1986, 106, 253. 12. Boddington, T., Laye, P.G., Tipping, J., Whalley, D. Combustion and Flame. 1986, 63, 359. 13. Rugunanan, R.A., Brown, M.E. Combustion Science and Technology. 1994, 95, 117. 14. Bockmon, B.S., Pantoya, M.L., Son, S.F., Asay, B.W., and Mang, J.T. Journal of Applied Physics. 2005, 98, 064903.

66

15. Granier, J.J. and Pantoya, M.L. Combustion and Flame. 2004, 138, 373. 16. stmark, H. and Roman, N. Journal of Applied Physics. 1993, 73(4), 1993. 17. Weston, A. and Brown, P.A. HPLC and CE: Principles and Practice. Academic Press: San Diego, 1997. 18. Gilbert, M.T. High Performance Liquid Chromatography. Wright: Bristol, 1987. 19. Niessen, W.M.A., ed. Current Practice of Gas Chromatography Mass Spectrometry. Marcel Dekker, Inc.: New York, 2001. 20. Willett, J.E. Gas Chromatography. John Wiley & Sons: Chichester, 1987. 21. Wendlandt, W.W.M. Thermal Analysis, 3rd ed. John Wiley & Sons: New York, 1986. 22. Rouquerol, F., Rouquerol, J., and Sing, K. Adsorption by Powders and Porous Solids. Academic Press: San Diego, 1999. 23. Stuart B., George, B., and McIntyre, P. Modern Infrared Spectroscopy. John Wiley & Sons: Chichester, 1996. 24. Wilkins, Charles L. and Jackson O. Lay. Identification of Microorganisms by Mass Spectrometry. Hoboken, N.J.: John Wiley & Sons, Inc., 2006. 25. Pasch, H. and Schrepp, W. MALDI-TOF Mass Spectrometry of Synthetic Polymers. New York: Springer, 2003. 26. Siegman, A.E. Lasers. Mill Valley, CA: University Science Books, 1986. 27. Graham, D., Kennedy, A.R., McHugh, C.J., Smith, W.E., David, W.I.F., Shankland, K., and Shankland, N. New Journal of Chemistry, 2004, 28, 161. 28. Brill, T.B. and James, K.J. Thermal-Decompostion of Energetic Materials. 62. Reconciliation of The Kinetics and Mechanisms of TNT on the Time-Scale from Microseconds to Hours. Journal of Physical Chemistry, 1993, 97(34), 8759. 29. Dacons, J.C., Adolph, H.G., and Kamlet, M.J. Some Novel Observations Concerning Thermal Decomposition of 2,4,6-Trinitrotoluene. Journal of Physical Chemistry, 1970, 74(16), 3035. 30. Long, G.T., Brems, B.A., and Wight, C.A. Autocatalytic Thermal Decomposition Kinetics of TNT. Thermochemica Acta, 2002, 388(1-2), 175.

67

31. Rogers, R.N. Combined Pyrolysis and Thin Layer Chromatography A Method for Study of Decomposition Mechanisms. Analytical Chemistry, 1967, 39(7), 730. 32. Yinon, J., Fraisse, D., and Dagley, I.J. Electron Impact, Chemical Ionization and Negative Chemical Ionization Mass Spectra, and Mass-Analysed Ion Kinetic Energy Spectrometry-Collision-Induced Dissociation Fragmentation Pathways of Some Dueterated 2,4,6-Trinitrotoluene Derivatives. Organic Mass Spectrometry, 1991, 26(10), 867. 33. Moore, J.W. and Pearson, R.G. Kinetics and Mechanisms. New York: John Wiley, 1981, p.21. 34. Bottaro, J.C., Malhorta, R., and Dodge, A. A facile two-step synthesis of 1,3,5trinitrobenzene. Synthesis, 2004, 4, 499. 35. Desseigne, G. Preparation and Nitration of Pure Pentaerythritol. Memorial des Poudres, 1951, 33, 169. 36. Brand, H.V. Ab Initio All-Electron Periodic Hartree-Fock Study of Hydrostatic Compression of Pentaerythritol Tetranitrate. Journal of Physical Chemistry B, 2005, 109(28), 13668. 37. Mlen, F. and Herman, M. Vibrational Bands of HxNyOz Molecules. J. Phys. Chem. Ref. Data, 1992, 21(4), 831. 38. Mlen, F., Pokorni, F., and Herman, M. Vibrational Band Analysis of N2O4. Chemical Physics Letters, 1992, 194(3), 181. 39. Amoruso, S., Bruzzese, R., Spinelli, N., and Velotta R. J. of Phys. B: At. Mol. Opt. Phys., 1999, 32, R131. 40. Amoruso, S., Amodeo, A., Berardi, V., Bruzzese, R., Spinelli, N., and Velotta, R. Applied Surface Science, 1996, 96-98, 175. 41. King, F.L., Dunlap, B.I., Parent, D.C. Journal of Chemical Physics, 1991, 94, 2578. 42. Rosenband, V. Combustion and Flame, 2004, 137, 366. 43. Kline, K.L.M.S. Thesis, University of Florida, Gainsville, Fl, 2003. 44. Levitas, V.I., Asay, B.W., Son, S.F., Pantoya, M. Applied Physics Letters. 2006, 89, 071909. 45. Maunit, B., Hachimi, A., Mannuelli, P., Calba, P.J., and Muller, J.F. Int. J. Mass Spectrom. Ion Process. 1996, 156, 173.

68

46. Ornallas, D.L., Carpenter, J.H., Gunn, S.R. Detonation Calorimeter and Results Obtained with Pentaerythritol Tetranitrate (PETN). Rev. Sci. Instrum. 1966, 37, 907.

69

BIOGRAPHICAL SKETCH

The author was born in Springfield, Illinois on August 2, 1981. She attended Butler University in Indianapolis, Indiana as an undergraduate where she received her Bachelor of Science degree in Chemistry in May 2003. She began her graduate studies in Chemistry at Florida State University in the fall of 2003 in the Department of Chemistry and Biochemistry under Dr. Albert E. Stiegman.

70