ENERGETIC SALTS: DEGRADATION AND TRANSFORMATION

The first part of this work mainly focuses on distinguishing and characterizing two white solids namely urea nitrate (UN) and nitrourea (NU) with similar melting point (~160 o C). Urea Nitrate is an inorganic salt. It has found use as an improvised explosive and was used in the first world trade center bombings. Nitrourea, a dehydrated product of UN, is an organic salt. In this study we reported and compared two routes to NU synthesis. We also proposed a decomposition route for UN. The second part of the work investigates perchlorate contamination in soils following fireworks displays. A total of two hundred and twenty two soil specimens were collect before and after ten individual July 4 th fireworks display events from 2007 to 2012. Soils were extracted in water and analyzed using ion chromatography by a modified EPA Method 314.0. Our study showed that soils are free of perchlorates before and four months after fireworks, but the soil samples collected within 24 hrs of fireworks showed perchlorate contamination from below the detection limit up to hundreds or thousands of nanograms per gram of soil. Through this study we also suggest that poor adsorption of perchlorates by soil matrices results in ground water contamination. The third part of the manuscript explores the methodologies used to prevent solid phase changes in ammonium nitrate (AN) for expanding its usage in applications. These solid phase changes occur because of absorption of moisture from the atmosphere. AN was thoroughly dried to stop these phase changes and then attempts were made to maintain dryness of AN by coating with polymers. Dried AN and polymer coated AN performance was tested using differential scanning calorimetry (DSC). We have succeeded in drying and encapsulating AN. Many questions are unanswered at this moment, such as the amount of polymerization, the uniformity of coating, the amount of absorption of moisture and if there are any more methodologies or other polymers which would improve the performance. This study is underway.


Infrared (IR), Raman and Nuclear Magnetic Resonance (NMR) Spectroscopy
Fourier Transform Infrared (FTIR) spectra were obtained with Nicolet 6700 FTIR with a 5 min hold followed by a 20°C/min ramp to 250°C and a post-run at 310°C for 1 min. The transfer line temperature was 250°C and the mass selective detector source and quadrupole temperatures were 230°C and 150°C, respectively. Chemical ionization (positive mode) was used with methane as the reagent gas.

Liquid Chromatography High-Resolution Mass Spectrometry (LCMS)
Both UN and NU were analyzed using liquid chromatography/high-resolution mass spectrometry (LCMS). The mass spectrometer was operated in both negative and positive ion mode under multiple ionization conditions coupled with the HPLC or direct injection into the atmospheric pressure chemical ionization (APCI) source via 7 syringe pump. The vaporizer temperature was set at 150°C, and capillary was set to 125°C. The discharge current ranged from 5-10 μA, and the sheath gas and auxiliary gas were operated at 25 and 10 arbitrary units, respectively, for negative mode, and 20 and 0 for positive mode. When directly injected into the ionization source via syringe pump, sample flow rate was ~10 μL/min, and the material was dissolved in a 50:50 methanol-water solution. If adduct formation was desired, 0.25wt% carbon tetrachloride or 0.23wt% aqueous ammonium acetate was added to the methanol-water mixture. MS resolution was set to high (50,000 at 2 Hz), and the maximum injection time was 250 ms.

Differential Scanning Calorimeter (DSC)
Analyses were performed using a TA Instruments model Q-100 DSC. Samples were run under nitrogen flow (50mL/min), and the system was calibrated against indium (m.p. 156.60ºC, H f 28.71 J/g). Sample amounts ranged from 0.2 to 0.5mg and were sealed in glass micro-ampoules (1.5mm O.D., 0.28mm wall thickness, and 8mm length) [4]. Samples were run in triplicate.

Raman Spectroscopy
Raman spectra of UN, NU, U and AN are given in Figure 3. Literature assignments are available for the prominent 1043 cm -1 and 1057 cm -1 lines in AN and UN, respectively. They are due to internal symmetric stretching of the NO 3 anion [7,8].
The 715 cm -1 in AN and the 537 cm -1 in UN may also be due to internal covalent stretches of NO 3 anion. A C-N symmetric stretching mode has been assigned to the 9 possibly the 989 cm -1 line in NU. The weak lines in urea at 1625 and 1649 cm -1 have been assigned to NH 2 deformations. A similar assignment is possible in UN and NU.
The 1540 cm -1 line in urea has been assigned to C-O stretching. Raman lines at 1583 cm -1 (NU) and 1574 cm -1 (UN) may also result from C-O stretches.

1 H and 13 C NMR Spectroscopy
Proton NMR spectra of NU in d 6 -acetone showed 2 broad peaks--one around 12 ppm, assigned to the proton adjacent to nitro group, and one around 7 ppm, assigned to the proton attached to the amino group. In contrast, UN showed a single peak at about 8 ppm; indicating that the hydronium ion exchanges with amino protons and cannot be observed by proton NMR. Proton decoupled 13 C NMR in d 6 acetone yielded a single peak-at 151 ppm for nitrourea and at 163 ppm for UN. The latter is higher due to proton-bonding on the oxygen site of the carbonyl group.

Gas Chromatography Mass Spectrometry (GC/MS)
No attempt was made to obtain GC/MS of UN. NU in acetonitrile solution was analyzed by chemical ionization GC/MS. NU, made by methods 1 and 2, as well as crude and recrystallized NU, were examined. The chromatographic peak of NU was broad and asymmetric with mass fragments: 44 (medium), 63 (large), 91 (small), 106 (small, NU+H + ).

Liquid Chromatography High-Resolution Mass Spectrometry (LCMS)
Mass spectra for UN and NU (HPLC and direct injection) are given in Figures 4 and 5.
The solvent system was equal portions water and methanol spiked with 0.25 wt% carbon tetrachloride (CCl 4 ) for adduct formation. Most of the major peaks remained the same for NU spectrum whether the sample was introduced via HPLC or direct injection. The negative ion spectrum with CCl 4 (Figure 4)  injections in methanol/water without adduct former for both NU and UN did not provide useful information.
Mass spectrum of nitrourea changed as the solution aged. In a freshly made methanol/water solution, NU -H + (104.010) was the largest peak in the negative mode of the MS fragmentation pattern. However, when a methanol/water solution of NU is allowed to sit one week under ambient conditions all evidence of the NU -H + fragment disappears and essentially only dicyanic acid -H + (85.044) becomes the major peak. This corresponds to the decomposition of NU previously reported [9].

Thermal Gravimetric Analyzer
A DSC trace of NU is shown is Figure 7. There was no discernable difference between the TGA traces of the crude and re-crystallized NU or between the NU made by the first or second synthesis methods. The TGA traces of UN (Figure 8,bottom) and NU (Figures 7) were distinctly different. Weight-loss of both UN and NU started at ~140 o C; NU exhibited a single weight-loss event, while UN experienced three thermal events. The TGA trace of urea ( Figure 8, top) also exhibited three thermal events before being completely consumed. FT-IR spectra of the gases from decomposition of NU, UN and urea are compared in a stack plot ( Figure 9) that also includes a mixture of CO 2 and N 2 O gas (second spectrum from top). Overlapping spectra of neat CO 2 and N 2 O gases and a mixture of the two, shown in Figure 10, distinguish peaks associated with these two gases. Comparison of the IR spectral region between 2100 cm -1 and 2400 cm -1 ( Figure 11) for UN, NU, U and CO 2 /N 2 O mixture suggest that UN decomposition produced significant amounts of CO 2 while NU and urea decomposition did not. Further, N 2 O gas appeared to be present in UN, NU and urea as demonstrated by prominent peak at 2240 cm -1 and an isolated doublet centered at 1288 cm -1 ( Figure 10). The 1355 cm -1 peak in Figure 10, adjacent to the doublet and only appearing in UN was not assigned. Closer inspection of all three materials ( Figure 11) reveals another peak centered between the N 2 O peak (2240 cm -1 ) and CO 2 peak (2360 cm -1 ). This peak (2283 cm -1 ) is assigned to gas phase isocyanic acid, literature value 2269 cm -1 [10]. Since NU was completely consumed, a suggested mechanism is as shown by eqs. 4 and 5. Isocyanic acid (HN=C=O) was also observed in the mass spectrum fragmentation pattern of NU and UN ( Table 3). The other product nitramide (H 2 NNO 2 ) is believed to decompose via eq. 5.

13
The formation and subsequent hydrolysis of isocyanic acid to ammonia and carbon dioxide has been demonstrated [9]. This hydrolysis had been observed by mass spectrometry in the aging of aqueous methanol solutions of NU via eq. 6.
The first weight-loss event for UN (~140 o C) only consumed about 40% of the sample ( Figure 8). The weight-loss events for NU and UN do not appear to proceed via the same mechanism as their decomposition products are notably different. The decomposition gases of UN exhibit peaks at 2400 and 1350 cm -1 not observed in the spectra of the NU off gases. Formation of urea from UN (eq. 7) and the subsequent reactions of urea (eqs. [8][9][10][11][12][13] to form melamine, cyanourea, biuret and cyanuric acid results in species which are likely to survive at high temperatures [10]. [ O=C(NH 2 ) 2  NH 3 + HCNO (8) 14 melamine cyanuric acid The IR spectrum corresponding with the weight loss in the NU thermogram is shown in Figure 9. The location of the single weight-loss event for NU and the IR spectrum of its decomposition gases match those observed for urea (also Figure 9). The IR spectrum of urea decomposition gases match those predicted in eq. 5. The remaining solid residue was likely one of the products in eqs. [8][9][10][11][12][13]. Not only is the TGA thermogram of UN more complex than that of NU, but the IR spectra of its decomposition gases also differed. Thus, urea nitrate decomposes via urea (eq. 7) and the decomposition route of urea (eqs. [8][9][10][11][12][13], while nitrourea decomposes via isocyanic acid and nitramide (eq. 4).

Conclusion
UN and NU can be distinguished by a number of their physical properties ( Table 1). As expected UN behaves like an ionic compound and decomposes to urea and nitric acid. A difficulty in differentiating between UN and NU has been that they are both white solids with similar melting points.    Table 3. Fragmentation patterns of NU and UN. Assignments with Cladduct were supported by additional mass peaks corresponding to the 37 Cladduct.    . Overlap magnified IR spectra of urea nitrate, nitrourea, urea, and a mixture of CO 2 and N 2 O.
This high enrichment was attributed to natural sources. It is now evident that trace amounts of perchlorate ions appear to be widespread in areas around the world not usually associated with anthropogenic activity. In Arctic snow, having no known anthropogenic sources, concentrations ranged from about 1 to18 pg per g of snow (Furdui.et.al., 2010). Arid areas, where evaporation of moisture (i.e. from rain or 33 snow) leaves dry deposition, tend to have elevated amounts of perchlorate in soil.
Studies in unsaturated zones of semi-arid and arid regions of New Mexico, Nevada, Texas, Utah and Arizona indicated amounts ranging from 1-13 ng of perchlorate ions per gram of soil (subsurface/sediment samples) (Rao.et.al., 2007). Current theories suggest atmospheric reactions of chlorine compounds with ozone and other reactants result in formation of perchlorates. (Simonaitis-Heicklen., 1975, Prasad-Lee., 1994, Dasgupta.et.al., 2005, Gu.et.al., 2006 The authors of the Arctic snow study related ice depth to seasonal variation in perchlorate ions and suggested perchlorate formation was maximal during Arctic summer months. They concluded that peak production of perchlorate probably resulted from multiple mechanisms but the major contributor likely involved sodium chloride aerosols in the presence of electrical discharges.
These authors point out that there are heightened baseline perchlorate levels at latitudes near the Gulf of Mexico where frequency of lightning flashes are highest in contrast with the Arctic where lightning activity is minimal and perchlorate concentrations are relatively low. They also concluded that perchlorate formation does not contribute significantly to chlorine removal from the atmosphere.
Proliferation of perchlorates is almost entirely the result of activities considered essential in the civilized world. Significant anthropogenic sources include rocket fuels, road flares, munitions, blasting agents and pyrotechnics (fireworks) ). Detection of perchlorate from specific sources is elusive because perchlorate salts are mobilized when in contact with water and ultimately move into the water table. Some monitoring studies of soils, ground water, tap water and bottled water from industrialized countries are available in the literature (Kosaka. et. al., 2007, Kannan. et. al., 2009, Wu. et. al., 2010, Her. et. al., 2011. Values range from tenths of ng/g to about 10 ng/g. There are a few studies monitoring soils and ground water located near fireworks manufacturing facilities and active fireworks displays. The occurrence of perchlorate ions at these sites are as high as 7700 ng/g in contaminated ground water surrounding a fireworks manufacturing facility in India (Isobe et.al., 2012), 15 ng/g from soil samples immediately following a fireworks display in Germany ( Scheytt. Et. al., 2011), 44 ng/g following a display in Oklahoma (Wilkin. et. al., 2007), 34000 ng/g (with extensive fireworks debris included) following a Massachusetts display (Mass. DEP, 2006). One study on Long Island, New York collected wet/dry atmospheric deposition in surrounding areas following a fireworks display (Munster. et. al., 2008). Average perchlorate concentrations were 0.21 ng/g with a maximum value of about 3 ng/g. In general for areas not involved in fireworks displays perchlorate concentrations were below the drinking water limit set by Massachusetts (i.e. 2.0 ng/g) (Daley, 2006). A study of large surface water reservoirs in the Great Lakes Basin and proximal to highly industrialized communities detected perchlorates near detection limits (0.2 ng/g) in only 8 out of 55 collection locations (Backus et. al., 2005).  (Kucharzyk. et. al., 2009).
In this study we investigated the amounts of perchlorate recovered from soils at four sites conducting July 4 th fireworks celebrations. A total of 222 soil samples collected from ten individual fireworks displays from 2007 to 2012 were analyzed.
The purpose of this study was to evaluate perchlorate concentrations before and after such events. In most cases selected sites had been used for fireworks displays in previous years. With one exception perchlorate contamination, prior to the specific event, was not observed. Perchlorate content in measured masses of soils was evaluated from water extracts by ion chromatography according to a modified EPA Method 314.

Sampling
Soil samples were collected before and after July 4

Sample Preparation
Collected soil specimens (i.e. 1-2 lb each) were homogeneously mixed and air dried overnight, then sieved (mesh # 500μm) to remove small stones and other organic debris (MacMillan. et. al., 2007). Triplicate samples of soil (about 5.0 grams) were extracted in 18.0MΩ cm Milli Q water (10.0mL). Samples were agitated overnight in a shaker bath at 250 rpm and ambient temperature. The resulting slurries were centrifuged to collect the supernatant layer (2500rpm, 30min, and 25˚C). The resulting supernatant layer was centrifuged (4000rpm, 15min, and 25˚C) to remove fine particles and filtered through a 0.45-μm PTFE syringe filter and analyzed for perchlorates via EPA Method 314.0.

Sample Analysis
A Dionex ICS-2100 Reagent Free Ion Chromatograph was used for analysis of the supernatant layer for perchlorates, using a 100 μL sample loop, an IonPac AG16 guard column (4 X 50 mm I.D.), an IonPac AS16 analytical column (4 X 250 mm I.D.); 50.0mM potassium hydroxide mobile phase, flow rate of 1.50 mL/min and ASRS (4 mm). The heated conductivity cell suppression current was 190 mA. Samples were run in triplicates. Detection limit for perchlorates was about 5 μg/L (0.50 ng).
Samples containing excess perchlorate were diluted accordingly. An external calibration method was used. Calibration standards were prepared from dilutions of a certified standard solutions (999mg/ml, VHGLabs, Cerilliant) with very low conductivity water (18.0MΩ·cm) obtained from a Milli Q Water System. Water blanks and quality assurance samples were introduced following every 6-9 authentic samples.

Recovery Studies
A protocol analogous to that for authentic samples (Section 2. spiked samples, with and without soil, were compared to assess recoveries. In all cases recoveries from soils was close to 100%.

Results
A total of 10 fireworks events were evaluated over a period of six years. The sites with geographical locations are summarized in Table 1 Table 3. Perchorate contamination ranged from BDL to about 2700 ng/g of soil. Results form Site 3 and 4 are given in Tables 4 and 5, respectively. Perchlorate contamination ranged from BDL to 25000 ng/g and 3000 ng/g, respectively.

Discussion
Strong oxidants, ammonium and postssium perchlorate, are major components of almost all pyrotechnic including fireworks. The high water solubility of perchlorate anions results in rapid transport to the ground water when rainfall is extensive. Under these conditions residence times for perchlorates within soils are likely to be short.  (Schilt, 1979). Cell free extracts of nitrate adapted Bacillus cereus is capable of reducing perchlorates, (Urbansky, E.T. 1988), Vibrio dechloraticans was patented by Korenkov et al. in 1976 as perchlorate reductase. Ethanol and acetate have been used as a source of nutrients to these bacteria (Rikken et al;. Wolinella succinogens, strain HAP-1 was first isolated from a municipal anaerobic digestor and has been shown to metabolize perchlorate and chlorate. It has been used at AFRL (U.S. air Force 1994, Wallace and Attaway 1994). They have demonstrated that perchlorate concentration of 3000 μg/mL can be reduced to less than 0.5 μg/mL. In the present study we monitored the occurrence of perchlorate anion following fireworks displays.
Prior to the July 4 th events, except for one isolated sample, perchlorates were not detected. To our knowledge the only reason for the perchlorate contamination at these sites were the fireworks displays. After the fireworks displays the amounts of perchlorates ranged BDL to a maximum of 24,585ng/g at Site 3. To further test our assumption that fireworks were the exclusive source of perchlorates, soil samples were collected and analyzed about four months following the event. Perchlorates were not 40 observed at the detection limits of this study. Our results are consistent with a MassDEP study of soils collected following a fireworks display in 2004 (Mass DEP, 2006). Prior to the event perchlorates were not observed. Following the event they reported samples with as much as 560 ng/g in soil.

Conclusion
Fireworks are a source of perchlorate contamination. Increase in the concentration of perchlorates in soils has been noted after fireworks displays. Figure 1 summarizes the results of this study as average values for all sampled locations at a given site. Site 3 which was monitored for two years showed the greatest overall contamination following fireworks displays (4065 and 1894 ng/g soil). Least contamination was at Site 1 where overall average values ranged from 118 to 472 ng/g soil. These amounts are higher than the set EPA interim limits in drinking water (15ng/g water) (USEPA, 2008). The fact that perchlorates were not detected at sites repeatedly used for fireworks displays over the years suggested they are highly mobile in soil due to their poor adhesion to the soil matrix facilitating seepage into ground water.

Abstract:
The goal was to make ammonium nitrate more useable in applications where volume changes cannot be tolerated, e.g. solid rocket propellant. To accomplish this, a way was sought to eliminate solid phase changes which resulted in volumetric changes. In particular, elimination of the IV -III AN phase change was targeted. Elimination of this was thought to be possible through extreme drying; vacuum drying with heat was used. To maintain dry AN coating methodologies were explored. Polystyrene or polyacrylonitrile were applied using precipitation polymerization method. The performance of the dried and the coated ammonium nitrate was analyzed by differential scanning calorimetry (DSC).

Introduction:
Ammonium nitrate (AN) was discovered by Glauber in 1659 and used as an ingredient in explosive formulations as early as the 1860. 1  The phase change near room temperature (32 to 55 o C) III → IV results in a volume change slightly more than 3.5% (Fig. 1)  To reduce the III/IV phase change various "phase stabilizers" are added-potassium, magnesium, or metal oxides like copper, zinc or nickel. [5][6][7][8][9][10][11][12][13][14][15][16][17][18][19][20][21] These can be added while the AN is aqueous or when it is a hot melt, or they can be introduced by chemical reaction. 9,10,[19][20][21] The additives generally serve one of two functions; they fit into the AN crystal structure or they act as internal desiccants. Potassium nitrate is an example of the former; and magnesium nitrate, an example of the latter. 5-10 Although a number of potential solutions to the III/IV problem have been fielded, 10-21 none have been fully successful. The transition III/IV temperature changes as a function of water content. In fully dry AN, state IV changes directly to state II around 51 o C (Table 1).
In a comprehensive study of the properties of AN, Kiiski of Yara, notes that phase III AN can be avoided; the change can be II to IV: "if there is no water, phase III will never exist…." 22 As water content is increased from 0.01 to 0.35% it is 54 reported that the temperature of the IV  III transition temperature decreased, 23 and an increase in water content also increases the rate of the IV  III phase change. One way to reduce water content is temperature cycling across the III/IV transition. 24,25 The above discussion suggests that producing AN, which does not have a phase change at room temperature, is possible, if it is dry. Our goal, therefore, is to "encase" the AN in such a manner that it remains dry for its lifetime. We believe the way to do that is with a coating or encapsulating species.
Generally, immediately after prilling, AN prills are coated with a thin layer (1 to 2 microns) of chemicals to reduce hygroscopicity, caking and cracking and dust formation. Coatings may be clay or diatomaceous earth or anionic alkylarylsulfonates or cationic long chain fatty amines which are applied as melts or oils on the order of a few kilograms per ton. 2,2'-Azobis(2-methyl propionitrile), 98% (AIBN)was purchased from Sigma Aldrich.

Drying AN
Ammonium nitrate (AN) was ground in a motor and pestle. Drying was either by holding the AN overnight at 110˚C in a vacuum oven or at 100˚C in a conventional oven. Dried AN was stored in a desiccator until use.

Coating AN Method 1
To a 250 mL round-bottomed flask, equipped with a football stirrer, and reflux condenser, chloroform (5.0 mL), PS (1.00 g, 9.52 x 10 -5 mol) and BPO (100 mg,9.83 x 10 -3 mol) were added. The mixture was heated to 60ºC in silicon oil bath and stirred until the entire polymer is dissolved. AN (4.00g, 1.19 mol) was added to the mixture and stirred four hours to allow the reaction to go to completion. Precipitation of the 58 coated AN was facilitated by addition of cyclohexane (5.0 mL). The coated AN was collected on filter paper by vacuum filtration. The product was rinsed with three 2 mL aliquots of cyclohexane, and dried in a desiccator.

Coating AN Method 2
Cyclohexane (5.0 mL) was poured into a 250 mL round-bottomed flask equipped with a football stirrer and reflux condenser. AcrN (1.00 mL, 0.322 mol) and AIBN (100 mg, 10.32 mol) were added and heated to 60ºC in silicon oil bath. AN (4.00g, 0.847 mol]) was added to the mixture and stirred ten hours to allow the reaction to go to completion. Hexane (5.0 mL) was added and the mixture was stirred for 10 minutes to facilitate precipitation. The AcrN-coated AN with was collected on filter paper by vacuum filtration, rinsed with three 2 mL aliquots of hexane, and dried in a desiccator.

Differential Scanning Calorimeter (DSC)
Thermal analyses were performed using a TA Instruments model Q-100 DSC.

Field Emission Scanning Electron Microscope (FESEM)
The surface morphology of the coated AN was examined using a ZEISS Sigma-VP Field Emission Scanning Electron Microscope (FESEM).

Infrared Spectroscopy (IR)
Attenuated total reflectance (ATR) spectra of coated and uncoated AN were obtained with Nicolet 6700 FTIR Spectrometer using 32 scans, 650 to 4000 cm -1 spectral range and at resolution of 4.0 cm -1 .

Thermal Gravimetric Analyzers (TGA)
TGA thermograms were obtained using a TA Q5000 TGA using nitrogen purge gas to constantly sweep the balance (10 mL/min) and furnace (25 mL/min).
Coated and uncoated AN (10-15 mg) were held in open platinum pans (100 µL). Runs were performed in triplicates, ramping the samples at 5 °C/min from 40 °C to 325°C.

Differential Scanning Calorimeter (DSC)
The solid-state phase changes of ammonium nitrate are often written as simply as the diagram at the top of Table 2, but the lower half of that These solid state phase changes involve large structural changes in the AN crystal (Fig. 1); thus they can be observed by thermal analysis [43][44][45][46][47][48][49][50][51] For that reason 60 DSC was used to assess the effectiveness of the AN drying protocol as well as the state of the AN. Figure 2 shows the entire DSC thermogram of AN as received.
Generally, endotherms were visible before the AN decomposition exotherm at 316 o C (which released 1000 J/g on average). An additional endotherm was observed immediately after the exotherm. This has been shown to be the phase change of water formed during the ammonium nitrate decomposition. 41 Occasionally (4 out of 27 samples), five endotherms were visible before the exotherm. Figure 3 expands this region for two of the anomalous "as received" AN samples. In one two endotherms were observe-one at 39 o C and one at 52 o C-and the other shows only one endotherm but it is at 41 o C. There was a fair amount of variation among samples; therefore Table 3 tabulates the position the results for 28 "as received" samples. The first endotherm, the IV to III phase transition, was not observed at 32 o C where it is often cited in the literature; 40, 52 and is pictured in the top of Table 2. Instead it was observed as low as 41 o C and as high as 56 o C ( Table 3). The variability in the temperature at which the phase changes between 32 o C and 84 o C occur has been previously reported. It has been suggested that this endotherm around 50 o C may actually be two endotherms: the IV to III phase transition of dry AN and the IV to II transition. 22,23,43 This may also explain the two endotherms near 50 o C visible in Figure 3. Figure 4 shows the DSC trace of AN after drying, and Table 4  Since the DSC traces suggested AN was successfully dried, encapsulation techniques were sought to act as a long-term moisture barrier. Polystyrene 29 and acrylonitrile. 30 were used since the methods previously applied match the resources we had at hand. DSC scans of the coated AN were collected. Figure 5 and Table 5 record the DSC data from samples of AN coated with polystyrene; 138 out of 187 samples evidenced no endotherm at ~90 o C. Figure 6 and Table 6 record the same data for AN coated with acrylonitrile. In this case 12 out of 18 samples lacked the ~90 o C endotherm. These results show the level of moisture present in the uncoated AN ( Fig.   3) can be maintained throughout the encapsulation protocol. It is also noticeable that more heat is released during the exotherm ~1500 J/g versus 1000 J/g due to the fact fuel is present. Figures 7 and 8 show the FESEM images of the uncoated (Fig. 7) and polystyrene (PS) coated AN (Fig. 8).The samples were observed at a same scale. The uncoated AN presents a smooth, almost amorphous surface compared to a AN coated with PS. The image of the coated AN is rather difficult to read, but it appears that the PS coating may be around 2 mm thick.

Water absorption Data
To determine how effectively the "encapsulated" AN was protected from moisture sorption, samples (~0.5 g each) of the polystryene (PS) and polyacrylonitrile (ACrN) coated AN as well as uncoated AN were sealed in a container along with an open dish of water. The weights of the AN samples were monitored every half hour for about 5 hrs for 24 hours. In the first 5 hours of observation there was no visible difference in the coated AN samples, whereas uncoated AN became visibly "wet." After 15 hours in the humidity chamber, the coated AN samples had water droplets, appeared as sweat, on the surface. Even though there was a visual difference, the coated AN gained as much weight as the uncoated AN within the first 5 hours in the chamber. This suggested the AN was not sufficiently or uniformly coated. The results are shown in Table 7. Table 8 shows the weight gain of the AN and the PS-coated AN during initial exposure to a moist atmosphere. The coated sample showed significant resistance to water absorption to the first five hours.

Infrared (IR)
The infrared (IR) spectra of AN, neat polystyrene (PS) and PS-coated AN are shown together in Figure 9. AN exhibits NH 4 + stretching, symmetric and asymmetric modes, near 3234 cm, -1 3058 cm, -1 and 3025cm. -1 Total stretching and in-plane and out-of-plane deformation modes of NO 3 were observed around 2500 cm. -1 In the region 1300-1500 NH 4 + and NO 3 asymmetric stretching and deformations have been reported by Chan. 52 Polystyrene shows aromatic C−H stretching at 2900 cm -1 and aromatic C=C stretching around 1600 cm. -1 The latter was observed in PS-coated AN although the majority of resonances are due to AN. Acrylonitrile has a charecteristic band around 2200cm -1 corresponding to C≡N stretching. 53 This was also observed in AN coated with acrylonitrile as shown in Figure 10.

Thermal Gravimetric Analyzers (TGA)
In an attempt to determine the amount of polymer coating present on the ammonium nitrate, TGA was performed. TGA of uncoated AN is shown in Figure 11. Each thermogram showed only a single-step weight loss. While we had expected to see the polymer-coated AN evidence a two-step weight loss-first loss of polymer and then loss of AN, apparently AN is thermolyzed more readily than the polymer. Thus, the polymer coating delays the release of AN. Apparently, the polymer coating must be breached before AN is lost.