Tools to Fight the Terrorist Threat

Chemical, biological, radiological, nuclear, and explosives (CBRNE) terrorist threats put law enforcement and soldiers at risk both at home and abroad. Law enforcement and soldiers must be provided with tools and knowledge to stay ahead of the capabilities of terrorists. Hexamethylene Triperoxide Diamine (HMTD) is a homemade explosive easily synthesized from hexamine, citric acid, and hydrogen peroxide. Although HMTD is very sensitive and prone to stability problems, it has a history of terrorists use, such as in the London bombing of 2005. Because law enforcement personnel must handle this material with no guarantee of purity nor indication of additives, for the sake of safety, knowledge of the stability and reactivity of HMTD was expanded in order to make handling safer. Differential scanning calorimetry was utilized to screen the compatibility of HMTD with various additives. It was found that water and weak acids, such as citric acid, destabilize HMTD. Gas chromatography / mass spectrometry (GC/MS) was employed to characterize both headspace gases (e.g. trimethylamine and dimethylformamide) and condensed phase decomposition products. Monitoring the decomposition of HMTD at room temperature and with gentle heating (60 0C) under various levels of humidity proved that humidity plays a major role in the kinetics of HMTD decomposition. Liquid chromatography / mass spectrometry was helpful for identification of condensed phase decomposition products and monitoring isotopic labeling studies. Through a labeling study with equimolar N and N hexamine during the synthesis of HMTD, it was found that hexamine dissociates before the formation of HMTD. There is currently a need for specialized pyrotechnic materials to combat the threat of biological weapons. Materials have been characterized and will be chosen based on their potential to produce heat and iodine to kill spore-forming bacteria (e.g. anthrax). One formulation, already proven to kill anthrax simulants, is diiodine pentoxide with aluminum; however, it suffers from poor stability and storage problems. The heat and iodine output from this mixture and candidate replacement mixtures were measured with bomb calorimetry and extraction and analysis of iodine by UV-Vis spectroscopy. Of the mixtures analyzed, calcium iodate and aluminum was found to be the highest producer of iodine gas. The heat output of this mixture and others can be increased by adding more fuel, with the cost of some iodine produced. Products of combustion were analyzed by thermal analysis, XPS, XRD, and LC/MS. Evidence was collected supporting the formation of metal iodides and metal oxides. One key reaction explaining the loss of iodine with increase in aluminum content is the reaction between aluminum and iodine, which forms aluminum triiodide. As seen in multiple cases, including the Boston Marathon bombing, improvised explosives may be as simple as a fuel/oxidizer (FOX) mixture initiated by a hot wire. The knowledge of which materials or compositions are explosive is incomplete, and tests for explosivity are currently conducted at specific scales. For example, ammonium nitrate is classified as an oxidizer because it does not explode at the pound scale, but can become explosive at a larger scale or with a fuel added. Herein, a bomb calorimeter with a pressure transducer has been studied for its use as a small scale metric (2 g) for predicting whether fuel/oxidizer mixtures will be explosive at larger scales. These results have been compared with calculated and measured detonation velocities, and measured air blast pressures. A positive correlation was observed between heat of burning and detonation velocity, and between heat of burning and air blast TNT equivalence.

There is currently a need for specialized pyrotechnic materials to combat the threat of biological weapons. Materials have been characterized and will be chosen based on their potential to produce heat and iodine to kill spore-forming bacteria (e.g. anthrax).
One formulation, already proven to kill anthrax simulants, is diiodine pentoxide with aluminum; however, it suffers from poor stability and storage problems. The heat and iodine output from this mixture and candidate replacement mixtures were measured with bomb calorimetry and extraction and analysis of iodine by UV-Vis spectroscopy.
Of the mixtures analyzed, calcium iodate and aluminum was found to be the highest producer of iodine gas. The heat output of this mixture and others can be increased by adding more fuel, with the cost of some iodine produced. Products of combustion were analyzed by thermal analysis, XPS, XRD, and LC/MS. Evidence was collected supporting the formation of metal iodides and metal oxides. One key reaction explaining the loss of iodine with increase in aluminum content is the reaction between aluminum and iodine, which forms aluminum triiodide.
As seen in multiple cases, including the Boston Marathon bombing, improvised explosives may be as simple as a fuel/oxidizer (FOX) mixture initiated by a hot wire.
The knowledge of which materials or compositions are explosive is incomplete, and tests for explosivity are currently conducted at specific scales. For example, ammonium nitrate is classified as an oxidizer because it does not explode at the pound scale, but can become explosive at a larger scale or with a fuel added. Herein, a bomb calorimeter with a pressure transducer has been studied for its use as a small scale metric (2 g) for predicting whether fuel/oxidizer mixtures will be explosive at larger scales. These results have been compared with calculated and measured detonation velocities, and seek God, and I would place my cause before God; Who does great and unsearchable things, wonders without number (Job 5:7-9). The fear of the Lord leads to life so that one may sleep satisfied, untouched by evil (Proverbs 19:23). In this you greatly rejoice, even though now for a little while, if necessary, you have been distressed by various trials so that the proof of your faith, being more precious than gold which is perishable, even though tested by fire, may be found to result in praise and glory and honor at the revelation of Jesus Christ (1 Peter 1:6-7).                 (Error bars in heat are too small to be seen; Table 1

Introduction
HMTD is synthesized from the reaction of hexamine with hydrogen peroxide.
The oxidation is catalyzed by acid, usually citric acid. It was discovered in 1885 by Legler using formaldehyde, ammonium sulfate, and hydrogen peroxide [1]. The structure was proposed in 1900 by Baeyer and Villiger [2]. Von Girsewald was the first to use hexamine, citric acid, and hydrogen peroxide [3]. X-ray diffraction showed exactly planar 3-fold coordination about the two bridgehead nitrogen atoms rather than pyramidal structure [4,5]. This ring strain in HMTD may account for its low thermal stability and high sensitivity to friction [6,7]

Synthesis of HMTD with other acids or no acid
Using same amounts of hexamine and hydrogen peroxide as above, but no acid added, precipitation of HMTD was not observed for 7 days at room temperature. After 9 days of stirring, 261 mg HMTD was recovered, ~7 % yield assuming 1:1 molar ratio hexamine:HMTD. Other diprotic and triprotic acids used, in place of citric acid, included sulfuric acid, phosphoric acid, and oxalic acid; like citric acid, they were added in 1.1 to 1 molar ratios hexamine:acid. Monoprotic acids gave poor yields (Table 5) if added in 1.1 to 1 molar ratios. If these (acetic acid, trifluoroacetic acid, formic acid, and nitric acid) were added in a 2.2 to 1 molar ratio hexamine:acid, yields were comparable to those achieved with citric acid.

Synthesis of HMTD with formaldehyde ( 13 C or 12 C)
Formaldehyde, up to 6 moles per mole hexamine, appeared to accelerate the reaction and increased the yield to over 100% based 1 to 1 hexamine:HMTD. For example, HMTD was synthesized by adding hexamine (0.4499 g, 3.22 mmol) to a solution of 13 C formaldehyde in water (2.0153 g of solution, 20 wt%, 13.43mmol) in an ice bath. Hydrogen peroxide was then slowly added (1.7871 g of solution, 50 wt%, 26.28 mmol) and later, anhydrous citric acid (0.6817 g, 3.55 mmol). HMTD started to precipitate within 2 hours, in contrast to the 5 to 6 hours required without formaldehyde.
The reaction was allowed to continue overnight as the ice bath warmed up. Aliquots of the reaction mix were taken every 0.5 hour for 4 hours after the addition of the acid, and the final aliquot was taken 27 hours later. The crude HMTD was vacuum-filtered, washed with distilled water (~200 mL) to remove acid and then HPLC grade methanol (~200 mL) to aid drying (dried several hours by vacuum filtration). yield was about 60 % (assuming 1:1 molar ratio hexamine:HMTD). adding acetic-formic anhydride to hexamethylenetetramine at room temperature, using the method of Gilbert [10]. Tetramethylene diperoxide diamine dialdehyde (TMDDD) was synthesized by the route of Wierzbicki [5]. N,N'-methylenebisformamide (m/z 102,  15 N into the condensed-phase decomposition products.

Decomposition of HMTD with deuterium oxide high humidity
HMTD (20mg) was heated at 60 o C in a small vial which was sealed in a 10mL headspace vial with 1mL of deuterium oxide (D2O) between the inner and outer vials so that HMTD did not directly make contact. HMTD was decomposed in a similar configuration with a saturated NaCl/D2O solution (analogous to 75 %RH conditions) between outer and inner vials for 5 days. The pH of the D2O and analogous experiments with water was found to be highly acidic (pH of 1). These samples were extracted with 30mL of acetonitrile, and run on GC/MS and on LC/MS to track the exchange of deuterium into the condensed phase decomposition products. Headspace analysis was also conducted according to the method described in section 2.13 using SPME. NMR

Differential Scanning Calorimeter (DSC)
DSC samples were prepared by measuring 150 to 200 mg of sample into a glass capillary tube, which was then flame sealed. For samples with an additive, 5 to 30 wt% additive was gently stirred into a 20 mg HMTD sample, and this mixture was placed in the capillary tube. If additives were liquid, 2 mL of the liquid was added to 150-200 mg of HMTD, and then sealed in capillary tubes. The sealed micro-ampules were weighed before and after DSC analysis to verify no leakage during testing. Samples were run on a TA Instruments Q100 DSC from 25 to 300°C with a ramp rate of 20 °C/min under nitrogen flow. Results were processed via TA's Universal Analysis software.

Monitoring Rate of HMTD Formation
Aliquots (100 μL) were removed and diluted with 5 mL of HPLC grade acetonitrile with sodium bicarbonate and magnesium sulfate added to neutralize acid and dry the solvent, respectively. This mixture was then diluted 1/10 v/v and analyzed by GC/MS.

Condensed Phase Analysis -GC/MS
Analysis of the acetonitrile samples, generated as described above, was

Headspace Analysis -GC/MS
Headspace of the HMTD was sampled via gas-tight syringe (5 mL or 1 mL) or Solid Phase Microextraction (SPME) fiber (SUPELCO fused silica coated with 65 μm of PDMS/DVB). The former was used for permanent gases; the latter for volatile amines. SPME fibers were flushed under helium 45 min at 250 o C prior to use. They the flow rate, at 2.5 mL/min. Two injection methods were used. A 5 μL injection with 5:1 split ratio was used to detect O2 and N2 signals; a 1 mL injection with a 1:1 split ratio was used for traces of other small molecules. The mass spectrometer scan parameters were from 10 to 100 m/z at12.89 scans/sec.

Condensed Phase Analysis -LC/MS
Liquid chromatography / mass spectrometry (LC/MS) analysis was conducted using modified procedures recently published [12]. HMTD samples were typically This method consisted of an initial mobile phase of 95 % solvent B (0.1 % acetic acid) and 5 % solvent C (acetonitrile). It was held for 2 minutes and then linearly ramped to 5 % B and 95 % C over 18 minutes. This was held for 2 minutes, returned to initial conditions over 1 minute and the re-equilibrated for 5 minutes. A second HPLC system was developed for optimum analysis of HMTD and hexamine; it employed an Advantage PFP column (100 x 2.1 mm, 5 μm) (Analytical Sales & Service, Pompton Plains). In order to gain some retention of hexamine, neutral pH conditions were preferable, but this caused broadening of the HMTD peak shape. To remedy this problem, three different mobile phase solvents were used to provide both pH and solvent strength gradients. Initially, 95 % solvent A (10 mM ammonium acetate, pH 6.8) and 5 % solvent C (acetonitrile) were held for 3 minutes following injection to retain hexamine. The system was then rapidly ramped to 85% solvent B (0.1 % acetic acid), 5 % solvent A and 10 % solvent C over the next 3 minutes. Organic levels increased slowly for 9 minutes to 35% C, 60% B and 5 % A, then rapidly for 3 minutes to 90 % C and 5 % of both A and B. This was held for 2 minutes before returning to initial conditions and re-equilibrated for 5 minutes prior to the next injection. Although this method revealed HMTD and most of the decomposition products, e.g. hexamine, a substantial number of species were still so polar that they were negligibly retained by this method. A third system employed an aqueous normal phase method using an Analytical Sales and Service Advantage 100 Silica column (150 mm x 2.1 mm, 5 μm).
Initial conditions of 95 % solvent C and 5 % solvent D (methanol) were held for 2 minutes before ramping to 5 % C and 95 % D over 6 minutes. Solvent C was then replaced with solvent B over 1 minute and then ramped to 60 % B to 40 % D over 10 minutes. After holding this for 2 minutes, it was ramped back to 95 % D and 5 % B 12 over 2 minutes then 95 % D and 5 % C over 1 minute. Initial conditions were returned over 2 minutes and held for 5 minutes before the next injection. This method required the use of electrospray ionization (ESI); however, this ionization mechanism is not optimal for HMTD detection.

Results and Discussion
Previously reported were thermal decomposition kinetics of HMTD determined by manometry [Ea 107 kJ/mol and A =4.21 x 10 10 s -1 ] and HMTD fragmentation by electron impact mass spectrometry [13][14][15]. Here, we examine factors which influence the stability of HMTD. It is the standard protocol of this lab that following synthesis a purification step is performed to promote stability. Unfortunately, HMTD had only limited solubility even in the most polar solvent requiring large volumes of ethyl acetate and acetonitrile for recrystallization which were almost impossible to remove completely from HMTD. For that reason, many of the studies were conducted with both crude and recrystallized HMTD to ensure the presence of trace solvent had not biased results.

HMTD Headspace
Since HMTD decomposition was readily observed at 60 o C, significant decomposition at ambient temperature was probable. In fact, when HMTD was  Figure 1 shows that these compounds were found in headspace of HMTD sample stored at room temperature for one year. In addition, while permanent gases, oxygen and nitrogen, were not found, carbon monoxide and carbon dioxide occurred in significant amounts. HMTD was not observed in the headspace by GC/MS under dry, moist, acidic, or basic conditions.
Since HMTD could be identified in ACN solutions, either HMTD content in headspace was below the detection limits of our GC/MS system or due to its reactivity, occurrence was not sustainable in the headspace.

Effect of Additives on HMTD Decomposition
The effect of additives on HMTD stability was screened by DSC. A general trend was readily observed: acids lower the temperature at which the exothermic maximum appeared (Table 1). We had previously demonstrated that concentrated mineral acid could be used to destroy HMTD [16]. We and others also observed that aqueous basic solutions rapidly decompose HMTD [17]. To determine the effect of select additives without water, HMTD was held at 60 oC for a week at 30 %RH, and of these additives, only citric acid markedly accelerated HMTD decomposition (Tables 2).
The fact that water and citric acid, both used in the synthesis of HMTD, lowered its  15 thermal stability markedly emphasizes the need to thoroughly rinse and dry HMTD.
Headspace monitoring revealed that water, citric acid or any acidity sped up the production of TMA and DMF in the gas phase.

Effect of Humidity on HMTD Decomposition
In 1924, it was reported: "That H.M.T.D. is stable at temperatures up to at least 60 o C; it is not affected by storage under water; but it is slowly affected when subjected to high humidity at maximum summer temperature….It is practically nonhygroscopic" [17]. Since DSC results did not support this statement, samples of crude

Mass Spectral Analysis of Condensed-Phase Synthesis and Decomposition
Products HMTD was heated at 60 o C under various conditions. Products were examined by GC/MS and LC/MS; and assignments are shown in Table 3 and Table 4, respectively.
Assignments are based on comparison with the authentic samples where compositions could be assigned to within 5 ppm of their calculated mass (Table   4). Examining the HMTD decomposition products, it is tempting to suggest HMTD thermolysis produces a number of small molecular fragments, e.g. CH2O, NH3, CH2NH or CH(O)NH2 which undergo further reaction, such as an aldehyde-amine condensation.
Thermal degradation of hexamine forms hexahydro-1,3,5-triazine, octahydro-1,3,5,7tetrazocine, and 1,3,5,7-tetrazabicyclo-[3.3.1]-nonane [20]. Bachman found that performing the nitration of hexamine in acetic anhydride with ammonium nitrate allowed two moles of RDX to be produced rather than one via direct nitration [21]. The question was whether the extra RDX came from fragments of hexamine or nitramines CH2NNO2 or directly from hexamine. On the basis of the observed by-products, Aristoff et al concluded that degradation of hexamine, itself, and not combination of smaller fragments, was the route by which RDX is formed [22]. Gilbert also confirmed this later by showing that RDX can be obtained by the direct nitrolysis of substituted triazine rings [10].
In the synthesis of HMTD from hexamine the question of stoichiometry arises.
Under the normal synthetic route as it is describe in equation 1; our yield, based on hexamine, was not more than 60%. However, if excess formaldehyde was added to the reaction mixture, yields of greater than 100% (based on 1 HMTD to 1 hexamine) were observed, and the reaction rate increases (precipitation of HMTD started to occur in 2 hrs compared to 5-6 hrs without formaldehyde). Equation 2 describes that reaction and 20 may also describe what occurs when no extra formaldehyde is added and the reaction must wait for the degradation of part of the hexamine to form formaldehyde ( Figure 4).
Indeed, hexamine is frequently used as a source of formaldehyde [18,23]. Although not shown in the above reactions, without citric acid formation of HMTD takes days. Furthermore, the reaction is sensitive to the type and amount of acid (   A mechanism for HMTD formation was proposed on data from isotopic ratio mass spectrometry [24]. Because it required the formation of a triperoxy tertiary amine and protonated methylene imine, we sought alternative proposals. Tentative proposals are illustrated in Figures 5 and 6. In Figure 5 hexamine is broken into small molecules, and from the formaldehyde/hydrogen peroxide reaction bis(hydroxymethyl) peroxide (BHMP) is formed, while from the imine/ hydrogen peroxide reaction bis(methylamine) 26 peroxide (BMAP) is formed. The latter reacts with 2 molecules of BHMP, forming tetramethylene diperoxide diamine (TMDD) as an intermediate, to create HMTD. The mechanism in Figure 6 also postulates the formation of BHMP but allows hexamine to remain moderately intact until fairly late in the reaction. Both mechanisms speculate that the reaction proceeds to HMTD faster in the presence of excess formaldehyde because formation does not require initial degradation of hexamine into formaldehyde.
The key to both mechanisms is the formation of BHMP, first synthesized in 1914 by Fenton from hydrogen peroxide and formaldehyde and later studied by Satterfield [25].
It is likely this species was generated in situ in the reported syntheses of several caged peroxides having planar bridgehead nitrogen atoms [26]. Once a methylene is lost from hexamine as formaldehyde the resulting octahydro-1,3,5,7-tetrazocine would be subject to rapid ring inversion and isomerization from which BHMP could bridge across two nitrogens.  To discriminate between the mechanisms proposed in Figures 5 and 6, synthesis of HMTD was done with a 1 to 1 mixture of 14

32
The following step in this pathway is the decomposition of INT3 into two new species or isomerization into a 7-member ring as shown in Figure 9. The formation of two radials, INT4 and INT5, is favorable according to the entropy changes; however, INT6 formation should be favorable due to a lower energy barrier to overcome. Next, we considered the decomposition of the HMTD molecule in an acidic environment. A proton can attach to either an oxygen atom or a nitrogen atom.
Protonated HMTD forms spontaneously without an appreciable energy barrier. When a proton is attached to one of the nitrogen atoms, the first step in decomposition of the cation will proceed via a C─N bond rupture. The energy barrier associated with this event is much higher than that obtained for the first step in the decomposition of a protonated oxygen atom in the HMTD molecule. Moreover, the barrier associated with the O-atom protonation is also smaller than the magnitude of the energy barrier associated with TS1 in Figure 8.
without or very low barrier   Table 7) and it leads to the formation of two 5-member ring radicals.
The structure and characteristics of the intermediate species in Path B of HMTD with protonated oxygen are presented in Table 8. The most important in this route is the possibility that INT1HOB is neutralized by an anion (several anions were tested, OH -, Cl -, SO4 2-, HSO4 -) to produce INT1 shown for neutral decomposition in Figure 8 and as entry 2 in Table 8. This pathway allows one to return to the neutral HMTD decomposition without the necessity to overcome a barrier 32.5 (30.5) kcal/mol.
Most neutral intermediates can be protonated without an appreciable energy barrier. The intermediates described in Table 8 suggest the possible intermediates with quite large molar mass similar to those presented in Tables 3 and 4. All these decomposition steps proceed without barriers or with small energy barriers; hence, most of these species are accessible. The highest barrier is related to the formation of formaldehyde (entry 6, Table 8). We also tested the fate of the relatively stable intermediate INT2 (entry 2, Table 6). The structure and properties of the intermediates observed during the decomposition of its protonated form are presented in Table 9. All the decomposition steps that lead to the formation of these intermediates proceed via barriers smaller than 30 kcal/mol. In most cases a much lower barrier or even no barrier is associated with the intermediate. Most of the species listed in Table   9 are rather small and resemble some of the species listed in Tables 3 and 4. Protonation of a nitrogen in the HMTD molecule as the initial step was also considered. The attachment of a proton to nitrogen is preferred by 2.2 kcal/mol over its addition to one of the oxygen atoms in the molecule; however, there are only two nitrogen atoms compared to six oxygen atoms in an HMTD molecule. The initial steps in the decomposition of a nitrogen protonated HMTD are shown in Figure 11.  Table   10. Thus, HMTD decomposition is also expected to occur in basic environment as was observed in the experimental part of this study.

Conclusion
Since HMTD is destabilized by water and citric acid, it is important to purify it after initial synthesis. It is recommended to rinse with water to remove acid, then with methanol to remove water. Ignoring the degrading effects of water and acid may lead was observed that hexamine, substituted triazines, and linear amines are formed in the condensed phase, and the observation of these products is humidity dependant. The mechanism of formation of HMTD was found to proceed through a complete breakdown of hexamine, involving formaldehyde exchange. Positive identification of synthesis intermediates remains as a future work.

Introduction
Previously we examined a series of oxidizers and fuels to determine their potential as explosive threats [1]. In the current work we examine, in detail, performance of oxides of iodine with the goal of determining their effectiveness as biocides. The biological threat of particular concern is spore production by Bacillus anthracis. While kill methods are diverse and not completely understood, it is known that a combination of heat and molecular iodine is effective [2,3]. A number of iodate and periodate salts were examined by formulating them with fuels and measuring heat evolution and molecular iodine release. Diiodine pentoxide has been used as a benchmark because it contains the highest weight percentage of iodine. Unfortunately, its long-term stability with a favored fuel, aluminum, is poor. Herein we examine the fuels aluminum and boron carbide.

Aging Studies
For aging studies, loose powder pyrotechnic mixtures were aged at 60°C and 75% RH (relative humidity). Time points were at 3 days and 14 days. Fresh samples and aged samples were analyzed by simultaneous differential scanning calorimetry/thermogravimetric analysis (TA Instruments, Q600 SDT, 20C/min, 50 to 1000 °C); infrared (IR) spectroscopy (Thermo Nicolet 6700 FR-IR with ATR cell, 32 scans, resolution 4 cm -1 , 650-4000 cm -1 ); and visual observation. IR was used specifically to detect oxygen-hydrogen bonds, indicating uptake of water. The burn characteristics of fresh and aged samples were also noted.

Simultaneous Differential Scanning Calorimetry Thermogravimetric Analysis (SDT)
A TA instruments Q600 SDT was used to characterize the original pyrotechnic

Titration for Oxide Content
In the case of 80/20 Ca(IO3)2/Al combustion products (pH 11 when mixed with water), an acid base titration was performed. Hydrochloric acid (30 mL of 0.100 M) was added to 50-150 mg of combustion products and allowed to stir for 20min. The solution was then back-titrated with 0.100 M sodium hydroxide, with bromothymol blue indicator.

X-Ray Photoelectron Spectroscopy (XPS)
A Thermo Scientific K-Alpha XPS (Aluminum source, 1486.7 eV) was used to help determine bomb calorimetry combustion products of NaIO3/Al, Bi(IO3)3/Al, KIO3/Al, Ca(IO3)2/Al, and I2O5/Al. The pass energy was 50 eV with a resolution of ±0.05eV. Samples and standards were prepared in a nitrogen glove box (from Genesis).
Charge effects were corrected based on the peak signal from the corresponding cation of an appropriate standard (i.e. KIO3/Al combustion products were corrected from K2p3/2 from KI).

Liquid Chromatography / Mass Spectrometry (LCMS)
Water and methanol extracts of bomb calorimetry combustion products of Ca(IO3)2/Al and I2O5/Al were prepared and infused into a Thermo Exactive Orbitrap Mass spectrometer with an electrospray ionization interface (ESI). This method was modified from a method used to analyze aluminum chloride in ESI negative mode with no additives in water [7]. The tune conditions (10 μl/min) were as follows: spray voltage

Friction Testing (BAM method)
Testing was conducted according to the UN method (on an FS-12A BAM machine from OZM research) where the threshold initiation level (TIL) of a sample (in N force) is reported where 1 out of 6 samples were a "go" with a snapping sound [8].
A sample size of 10 mm 3 was used.

Drop-weight impact (Modified BOE method)
This test was conducted with a BOE machine manufactured by SMS (10 mg sample, 3.63 kg weight) using the UN method [8]. Ca(IO3)2/Al was tested seven times at the highest height of the instrument (75 cm). A Dh50 number was obtained with an up/down method (14 samples, where 50 % of the samples were a "go") with RDX (class 1, Holston) for comparison. A test was considered a "go" when an explosion or flash occurred.

Electrostatic Sensitivity Testing (ARDEC method 1032)
This test was conducted with a machine manufactured by UTEC Corporation, LLC using ARDEC method 1032 [9]. Testing starting at the 0.25 J level, and the energy level was stepped down until a TIL energy value was reached with 0 out of 20 samples 53 were a "go". A test was considered a "go" when a flash considerably brighter than a blank occurred and the tape holding the sample down split open.

Results and Discussion
Choice of oxidizers was governed by availability as well as reported iodine production ( Table 1). (Iodoform was considered but not examined because it was neither an oxidizer nor a good fuel.) Because aluminum is often used to create heat-producing pyrotechnic mixtures, oxidizers were initially compared using it as the fuel (Figure 1). Boron carbide was also examined because recent studies reported when it was used in delay mixtures of periodate, iodine production was observed ( Figure 2)   with boron carbide. Diiodine pentoxide did not burn with boron carbide under argon.
As Figure 1 shows, diiodine pentoxide was most effective in both iodine and heat production. However, long term stability was poor. In the presence of moisture this oxide is reportedly converted to iodic acid, also a white solid [11].

Oxidizer/B 4 C (80/20) Powder in Argon
Iodine Production Heat of Reaction 55 powder ( Figure 3). It may be the reaction of aluminum with iodic acid which causes the rapid color change observable in Figure 3. Evidence of the presence of iodic acid can be found in the SDT of I2O5 aged under the same conditions ( Figure S2; water loss at 112 ⁰C and 219 ⁰C). At the same temperature and humidity, visual observations as well as infrared spectrometry (IR), thermal gravimetric analysis (TGA), and differential scanning calorimetry (DSC) suggested that calcium iodate, sodium iodate, and sodium periodate, (and mixtures with fuel) were stable ( Figure S1-S32). All oxidizers alone remained white solids through the aging study. When an original 75/25 calcium iodate/aluminum mixture was allowed to age two weeks under these conditions, no change is observed in its appearance, production of iodine or thermal trace, suggesting acceptable thermal stability ( Figure S22-S24). Even without considering the efficiency of I2 production, it would be difficult for other species to match diiodine pentoxide (I2O5) in terms of iodine formation because they do not contain as much iodine per mass of oxidizer ( The alkali iodates normally decompose to make the iodide salt (eq. 2) and oxygen with perhaps up to 30% forming the oxide instead (eq. 1) [12]. The addition of a fuel eliminates the free oxygen, but in the case of aluminum fuel, excess aluminum may promote the formation of Al2I6 [13]. Six oxidizers and I2O5 were examined with aluminum, boron carbide and a mixture of the two ( Table 2). The data reported was obtained in an argon atmosphere in a closed-bomb (Parr); iodine (I2) was collected after combustion and usually quantified by UV-Vis spectroscopy. The reported results are averages of at least three tests. Average heat released under argon (across all mixes) was 3975 J/g; similar to heat released from 80/20 I2O5/Al (4414 J/g). Iodine production was more sensitive to the fuel/oxidizer ratio than was heat output ( Table 3). Review of the data sorted in Table 3 indicated that as the oxidizer/fuel ratio moved from stoichiometric (roughly 80/20) to a more fuel rich formulation (60/40), I2 production decreased and heat generally increased. We attributed this to oxygen deficiency, which caused the fuel to combine with the iodine species (acting as oxidant) preventing the release of molecular iodine. Indeed, preliminary data suggested that both iodine production and heat release are improved by the presence of oxygen. and analyzed by X-ray photoelectron spectroscopy (XPS) and simultaneous thermal gravimetric/differential scanning calorimetry (DSC/SDT). XPS results in Table 4 show electron binding energies of the combustion products, which are consistent with oxidation state assignments of I -, O 2-, Al +3 , Ca +2 , N 3-, Na + , K + , Bi +3 . The resulting elemental analysis is shown in Table 5, noting that all results show more oxygen than anticipated. This is attributed to the presence of moisture or surface oxidation; oxidation of iodides is explained later from SDT experiments ( for Ca(IO3)2/Al from mix 60 (which is stoichiometric) there is not sufficient iodide (I -) found to support the required 1:2 ratio for CaI2. When the aluminum fuel content was raised from 20 wt% to 40 wt%, the ratio was consistent with CaI2 production, but this  Theoretical iodine (wt%) is iodine content of original mixture; I 2 recovered (wt%) is mass I 2 extracted from combustion products (quantification by UV-Vis) relative to original mix mass; I 2 yield /theory is mass of I 2 relative to theoretical amount.   The SDT allowed observation of heat released or absorbed concomitant with weight loss in the iodine-containing samples during heating as opposed to burning with fuel. Table 6 summarizes the observations when these fresh samples were heated in unsealed SDT pans. Table 7 analyzes the remaining solid products produced from the reactions outlined in Table 6 although the actual residue was collected from the bomb calorimetry experiments (Table 2). Neat I2O5 decomposed at ~438 °C and did not appear to react with aluminum ( Figure S4) While NaI and KI have been identified from the DSC melt and XPS examination of the combustion products, we know also from the basicity of the combustion products and presence of molecular iodine that equation 1 is also operative [12,14]. The sodium and potassium salts show an increase in iodine production when boron carbide, rather than aluminum, was used as the fuel (Figures 12 and 13).
Bismuth triiodate, upon heating, exhibited two modest endotherms at 550 °C and at 579 °C [4,15]. These are assigned as the stepwise oxidation of bismuth iodate to the oxide Bi2O3 with release of I2 (eq. 5, 6). Indeed there was also one small endotherm at 817 °C, the melting point of Bi2O3 [16].
5Bi(IO3)3  Bi5O7I + 7I2 + 19O2 (5) 2Bi5O7I + 1/2O2 5Bi2O3 + I2 (6) When aluminum was added the two endotherms were visible at slightly lower temperatures, 528 °C and 566 °C (accompanied by ~40% weight loss), and an exotherm near the melting point of aluminum (641°C) appeared. There is little heat released at this exotherm and almost no weight loss (Table 6). This cannot be explained by a direct reaction of Bi2O3 with Al. When reagent grade Bi2O3 and Al were examined under the same experimental conditions, no reaction was observed until the oxide melted (814 °C). The combustion of bismuth triiodate with aluminum in a sealed vessel under argon yielded a black product that exhibited only one endotherm at ~365 °C. This melt as well as its UV-Vis spectrum confirmed this product as BiI3 (m.p. 390 °C) [6,16].
Indeed, little molecular iodine was produced if the combustion was in an inert atmosphere. Unlike the alkali iodate salts, less, rather than more, molecular iodine was produced when the bismuth or calcium iodates were burned with boron carbide rather than aluminum (Table 2).
Calcium iodate, like the bismuth iodate, exhibited two modest endotherms at 656 °C and 736 °C. The first endotherm is ascribed to the decomposition of Ca (IO3)2 to Ca5(IO6)2, iodine and oxygen and the second endotherm to the complete oxidation of the calcium salt to calcium oxide with further generation of iodine and oxygen [12,17].
When aluminum is mixed with the calcium iodate, where the decomposition of Ca(IO3)2 and melt of aluminum coincide at 650 °C, an exothermic reaction occurs which forms both calcium and aluminum oxide as well as iodine ( Table 6). The formation of calcium oxide is claimed based on the basicity of the combustion product (from closed bomb calorimetry in argon) from the 80/20 Ca(IO3)2/Al mixture (pH 11), the ratio of elements in the XPS (Table 5, mix 60); and the fact that when the residue from the combustion was examined by SDT, neither endotherms nor exotherms were observed and weight loss was only 6%. These combustion products were shown by titration to form 11% CaO (assuming this is the product). Some XRD peaks characteristic of γ-Al2O3 were observed in the 80/20 Ca(IO3)2/Al combustion products, but no good matches for a particular iodide (although some peaks match for CaI2•6.5H2O). If aluminum was introduced into the calcium iodate in excess, e.g. 60/40 Ca(IO3)2/Al, then the DSC/SDT scan of the product mixture showed an endotherm at 652 °C, characteristic of the melt of excess aluminum. XRD peaks of these products match γ-Al2O3 and more closely with CaI2•6.5H2O than the products of the 80/20 Ca(IO3)2/Al mix ( Figure 5).
Furthermore, the SDT of the combustion products shows a mass loss of 31%, rather than 6%, and the pH was pH 5, instead of 11. These observations along with the great 66 reduction in produced I2 (42% with 20% Al down to 2% with 40% Al, see Tables 2 and   3) suggest some formation of Al2I6, a Lewis acid. The peak binding energies of the iodine signal from XPS suggests that the combustion products from Ca(IO3)2/Al (both 80/20 and 60/40) as well as other iodate/Al mixtures, contain iodine present as iodide (    formation of CaI2. LC/MS of the water extract of 60/40 I2O5/Al combustion products shows peaks consistent with Al2I6 ( Figure 8), but they were not observed in the 80/20 I2O5/Al combustion products. What is also interesting to note, is that the extract of a fresh mixture of 80/20 I2O5/Al produced LC/MS peaks consistent with known hydration products of I2O5 (IO3from HIO3, and I2O5•IO3from I2O5•HIO3) [11]. The methanol extract of fresh 80/20 Ca(IO3)2/Al did not contain any identifiable peaks (Figures 6 and   8).   is not sensitive to friction or impact, but does have a similar and sometimes more sensitive response than RDX to ESD (Table 8). Adding a binder did not change the impact or friction sensitivity, and seemed to improve the ESD sensitivity.    is approached [18]. We have studied 80/20 Ca(IO3)2/Al, Bi(IO3)3 /Al, and NaIO3/Al at both 515 kPa (60 psig) and 2515 kPa (350 psig) pressures (Table 2) to determine if the reaction can be driven to produce more molecular iodine. Interestingly, 80/20

80/20 Ca(IO 3 ) 2 /Al
Bi(IO3)3/Al produced very little free iodine (possibly further combination of Bi + I2); 80/20 Ca(IO3)2/Al produced slightly more iodine (45% vs. 42%); and 80/20 NaIO3/Al produced considerably more iodine (28% vs. 16%). The increase in iodine produced from 80/20 NaIO3/Al would likely be coming from further oxidation of NaI. Table 9 summarizes the reactions observed with the various iodate and periodate salts. The production of molecular iodine is opposed by both the potential for the original cation (Na + , K + , Ca 2+ , Bi 3+ ) as well as the aluminum to form the iodide salts.
Aluminum preferentially forms the oxide if there is sufficient oxygen available in the mixture, but the alkali ions preferentially form the iodide (MI), reducing molecular iodine formation. Calcium and bismuth form oxides, but bismuth oxide undergoes a metathesis reaction with aluminum to form, ultimately, bismuth iodide, which probably forms through elemental bismuth reacting with elemental iodine. In aluminum heavy mixtures, calcium iodate may form calcium iodide and aluminum iodide, although it is difficult to tell the difference between having calcium oxide and aluminum iodide in the products (with post reaction with moisture to form CaI2•6.5H2O), or having a mixture of calcium and aluminum iodides. In general, excess aluminum reduces I2 formation.
The fact that more molecular iodine is released when there is more oxygen available to the fuels indicates that most of the metals would rather be oxides than iodides. This is supported by the Gibbs free energy and enthalpy of oxidation of iodide salts to metal oxides (Table 10). The oxidation of the alkali iodide salts is endothermic, with a positive Gibbs free energy suggesting that they are less likely to produce iodine gas than the other iodide salts listed. The oxidation of the alkali earth iodides, aluminum iodide, and bismuth iodide is exothermic, with a negative Gibbs free energy suggesting release of iodine to be more favorable than that of the alkalis. All the iodide salts (KI, NaI, CaI2, BiI3, and Al2I6) were run on SDT under air as well as under nitrogen. Under air, calcium iodide and aluminum iodide produced traces with small exotherms and large mass losses. In contrast, under nitrogen, calcium iodide showed no decomposition as heat flow and mass loss below its melting point, and aluminum iodide produced an endotherm during its melt with some significant mass loss (moderate sublimation).
These differences suggest significant oxidation in air for these two salts. The sodium, potassium, and bismuth iodide salts showed little difference between air and nitrogen, with their melts accompanying almost total mass loss, which is presumed to be mostly sublimation (Table 7).  The potential for molecular iodine to be released may depend on the relative oxophilicity of aluminum relative to the cation accompanying the iodate (Table 10).
This would especially be important in oxygen deficient situations such as experiments performed under inert atmosphere. With insufficient oxygen the iodide may be formed instead. We believe this to be the case with bismuth iodate, due to the favorable reaction between bismuth oxide and aluminum, which frees up bismuth for a reaction with iodine. Because the reaction of some metal oxides (calcium and magnesium) with aluminum is not as favorable, it is likely that excess aluminum in this case would react with iodine directly in an oxygen deficient environment.
We have noted that use of a combination of boron carbide (B4C) and aluminum as fuels resulted in more iodine formation from the alkali iodates than the use of either fuel alone ( Table 2). The exact nature of the reactions have not been ascertained. Boron carbide has been examined by bomb calorimetry, and diboron trioxide and carbon dioxide were formed [19,20] Furthermore, the combustion products of boron with potassium nitrate and potassium perchlorate under argon were found to be KB5O8 . 4H2O and KB5O6(OH)4 . 2H2O, respectively [21]. The authors of that article speculate that reaction 8 occurs: 2KClO4 +2B  2KBO2 + Cl2 +2O2 (8) Using that model we suggest a similar reaction (eq 9). Indeed, over time a boron carbide mixture with sodium iodate evolved molecular iodine at room temperature.
Perhaps the reason the combination fuel Al/B4C results in higher amounts of evolved I2 can be attributed to the alkali metal being removed from the competition with aluminum for the freed oxygen. Thus, both the alkali metal and the aluminum are incorporated in a stable species allowing molecular iodine to be evolved.

Introduction
Hundreds of years ago the field of energetic materials began with the creation of a fuel-oxidizer mixture of charcoal, sulfur, and potassium nitrate, which became known as black powder [1]. Within the last century the fuel-oxidizer mixture of ammonium nitrate and fuel oil (ANFO) became popular as a commercial blasting agent [2] and later as a terrorist tool [2,3]. In the intervening period, the discovery of nitration resulted in a number of high-density organic molecules-nitrate esters, nitroarenes, nitramines. Because these molecules have become the basis of military weaponry much effort has been expended in modeling their detonation performance. Fuel/oxidizer (FOX) mixtures, when examined by the same protocols, have been termed "non-ideal" explosives because the models usually over-predict performance. Nevertheless, it has become imperative that we understand FOX mixtures since their ease of creation-simply mixing a fuel and oxidizer together-has made them a common choice in illicit bombing.
We have previously reported a series of 11 oxidizers and 13 fuels examined by differential scanning calorimetry (DSC), simultaneous DSC/TGA (SDT), and by open burn. DSC is usually the first step in evaluating the energy content of an energetic formulation because the technique can use less than a milligram of material. In preparing the fuel/oxidizer DSC samples, great care was taken to make the samples homogeneous.
Nevertheless, the DSC traces were difficult to interpret due to the small size of the prepared batches and the presence of multiple thermal events [4]. Herein we report a re-investigation of these and other FOX mixtures using isoperibol calorimetry-a Parr bomb-recording heat release and dynamic pressure rise of 2 gram samples during reaction under argon. Initiation of detonation of select formulations was attempted on the pound-scale (~10 lb FOX with 1 lb C4 Military Explosive), and data was recorded by high-speed photography and pressure transducer.

Simultaneous DSC/TGA (SDT)
A TA Q600 simultaneous DSC/TGA was used to run samples of 4-6 mg in open aluminum oxide pans, and scanned at 20 °C/min under 100 mL/min nitrogen flow. The temperature was calibrated by running Zinc with melting point of 419.5 °C. The temperature range was usually 50 °C to 1000 °C. Oxidizer / aluminum mixtures were analyzed with this technique due to exotherms appearing at higher temperatures than the DSC limits.

Bomb Calorimetry with Pressure Transducer
Heat output and pressure/time curves were determined using a Parr 6200 calorimeter and Parr 1108 bomb, fitted with a pressure transducer (Parr 6976 pressure recording system, including a 5108A Kistler piezoelectric coupler, and a 211B2 Kistler piezoelectric pressure transducer with a calibrated sensitivity of 1.096 mV/psi). The Parr bomb was calibrated (i.e. 10 trials) with benzoic acid ignited with fuse wire (9.6232 J/cm) and cotton string (167.36 J) in 2515 kPa oxygen (ΔHcomb = 26434 J/g). In an oxygen atmosphere, the string was in contact with the fuse wire and sample and was ignited by the fuse wire to aid ignition of the sample. The FOX samples (three to six 2 g samples under each set of conditions) were ignited with a fuse wire under argon (2859 kPa, 400 psig). This pressure represented the maximum initial pressure in which the regulator could handle. It appeared to be a good balance allowing rapid initiation of burn, and minimizing heat losses with the walls of the Parr bomb [5]. With some energetic materials, it has been observed that there is a critical pressure of ignition 85 associated with a specified input energy [6,7]. Igniting samples at a higher initial pressure is more likely to overcome the critical pressure of the sample. A National Instruments USB-6210 data acquisition card (maximum sample rate of 250 kS/s) and LabView software were used to collect the pressure/time data at a rate of 10 kS/s. This sample collection rate of 100 µs between pressure points was high enough resolution to result in pressure/time plots that appeared continuous on the ms time-scale (see Figures   9 and 10). Figure 1 outlines the protocol followed.
Where α ( Figure 6) is the incident angle from vertical, measured by taking the inverse tangent of two points on the side of the pipe (X1, Y1) and (X2, Y2 ): If two points are taken from the calibrated coordinate system (i.e. for 70:30

Predictive Tools
Cheetah 7.0 from Lawrence Livermore National Laboratory (product library: sandia, jczs revision 1923) was used to predict detonation velocity, detonation pressure, and total energy of reaction. Each mixture was run with Cheetah using the density that was measured for its large scale test [8].
The blast effects calculator (BEC V5.1) was used to obtain air blast TNT equivalence from the measured peak air blast pressures [9,10,11]. For each experiment, a goal seek method was used with the empirical fits for pressure (as a function of scaled distance, m/kg 1/3 ) to find the total amount of TNT needed to achieve the same peak pressure. However, the booster also has a contribution to the air blast pressure. This  (Table 4) of each test to find the TNT equivalence of the sample (TNT Equivalence = TNT equivalent mass of sample / sample weight).

Parr Bomb Calorimetry
A Parr bomb calorimeter was used primarily to estimate the energy available from FOX mixtures. Combustion was accomplished under argon gas instead of oxygen gas to determine heat of reaction without excess oxygen ( Table 1)   The change in internal energy of the formulations, as judged by the heat of decomposition measured at the sub-milligram-scale by DSC (far right column, Table 1) and heat of reaction observed in the 2 g Parr bomb samples (penultimate right column,   Table 3 shows FOX mixtures for which initiation of detonation was attempted.

Detonation Testing
Four of the mixtures failed to propagate detonation although the velocity of the burn front is recorded under the velocity of km/s. Figure 11 provides screen captures of the reactions observed. The detonation front was taken to be the bright line running ahead of the booster debris cloud (bottom). A detonation rather than a burn was judged by the rapid PVC wall expansion immediately behind the front. Figure 12 shows KNO3:sucrose as an example of a mixture which failed to support detonation. Figure 12 also shows KNO3:aluminum as an example of a mixture where the detonation failed and transited to a rapid burn. In this case the mixture is more flammable than detonable.      Cheetah. (Error bars in heat are too small to be seen; Table 1 shows relative standard deviation.) Observed detonation velocities tracked with the Cheetah predicted detonation velocities. Figure 15 shows the non-detonations (i.e. potassium nitrate formulations) with an X and separates the shots done with aluminum fuel from those done with sucrose and from those done with formulations including high explosives (in red, two TNT shots and one that was 50% RDX).  Table 1 shows relative standard deviation.) Figure 16 suggests there may be a minimum energy (~2.8 kJ/g) needed for detonation. However, the data as well as detonation theory dictates that energy alone does not guarantee detonation. The rate of energy release by the formulation must be fast enough to support detonation. If we make the rather speculative assumption that the rates of all the oxidizer/sucrose reactions are similar because the rate of reaction in these low density powders is diffusion controlled, then we might expect a linear relationship between energy of reaction and detonation velocity. No Detonation speculated that aluminum can provide enough additional energy during its oxidation to push a low-energy formulation to detonation; this was not the case in these studies.
Ammonium nitrate and perchlorate sucrose mixtures were detonable; substitution of aluminum for sucrose increased the heat released in the calorimeter, but detonation velocity decreased. We attribute this result to the lower density of the aluminum formulation due to the small aluminum particle size. Not surprisingly the air blast in terms of TNT equivalence increased with the addition of aluminum. It is well known that aluminum does not react rapidly enough to contribute all its energy to the detonation front; hence, the provision in Cheetah to make some of the aluminum content "inert." In fact, air blast in terms of TNT equivalence is proportional to the heat observed in the Parr calorimeter ( Figure 17). (Error bars in heat are too small to be seen; Table 1 shows relative standard deviation.)

Conclusions
Measurement or calculation (Cheetah) of heat of reaction is a useful first step in determining whether a formulation is potentially detonable. It appears there is some minimum energy which a formulation must possess to be detonable. However, examination of Table 4 clearly shows that some materials with high reaction energy (i.e. KNO3/Al) do not detonate, while others with low reaction energy (i.e. AN/sucrose) do.
Clearly any small-scale test or model must take into account the rate of reaction as well as energy. The potassium nitrate/sucrose mixture exhibited low heat release in the Parr bomb, and it did not detonate in the field-scale configuration. The substitution of aluminum for sucrose dramatically increased the energy released (as measured in the calorimeter), but the mixture (KNO3:Al) still did not detonate in field trials. The rate recorded in Table 3 is a burn, as judged by video record and discussed above ( Figure   12). The potassium nitrate/sucrose mixture was prodded into detonation by spiking it with 5wt% RDX or 7wt% potassium chlorate. Both these chemicals were capable of rapidly adding energy to the mixture. However, the total energy released by these potassium nitrate/sucrose mixture with these additives was only a little over half that of potassium nitrate / aluminum. This observation points to the importance of the rate at which the energy is provided. Figure 18 recasts the Parr data found in Figures 9 and 10 colorizing Parr pressure data to reflect the outcome in the large-scale tests. In general, FOX mixtures, which exhibited a rapid rise to peak pressure, detonated on the large 108 scale. Those FOX mixtures, which reached peak pressure more slowly, did not detonate at the large scale, with the exception of ammonium nitrate and sucrose. Ultimate outcome at the 5 to 6 kg scale is shown by colorred for FOX which detonated; blue for FOX which did not detonate. The ammonium nitrate:sugar mixture is so slow that it has its own time axis (above plot).
With aluminum mixtures at the large scale, it has already been mentioned that due to the slowness of the reaction only some fraction of the energy released in the aluminum oxidation can support the detonation front [13]. The rest is manifest in the Taylor wave expansion, i.e. air blast. The fuel/oxidizer mixture has as similar problem with reaction rate. Detonation velocity is strongly dependent on density [14]. FOX mixtures are far from dense, and a significant amount of time must be spent in diffusion 109 and compaction of the fuel and oxidizer. High explosives, such as PETN or RDX, have reaction zone lengths of approximately 1-2 mm, reacting rapidly enough so that much of their energy can support the detonation front [15]. This in contrast to a non-ideal explosive, such as ANFO, with a reaction zone length estimated as 8-12 mm [15]. With these FOX mixtures the fraction of energy released to the front must be significantly less. How much less and the role of compaction in these composite materials will be the subject of a number of future studies.