DECOMPOSITION OF AMMONIUM PERCHLORATE ENCAPSULATED NANOSCALE AND MICRON-SCALE CATALYST PARTICLES

Iron oxide is the most common catalyst in solid rocket propellant. We have previously demonstrated increased performance of propellant by encapsulating iron oxide particles within ammonium perchlorate (AP), but only nanoscale particles were used, and encapsulation was only accomplished in fine AP (~20 microns in diameter). In this study, we extended the size of particle inclusions to micron-scale within the AP particles as well the particle sizes of the AP-encapsulated catalyst particles (100s of microns) using fractional crystallization techniques with the AP-encapsulated particles as nucleation sites for precipitation. Here we report catalyst particle inclusions of micron-scale, as well as nanoscale, within AP and present characterization of this encapsulation. Encapsulating micron-sized particles and growing these composite particles could pave the way for numerous possible applications. A study of the thermal degradation of these AP-encapsulated particles compared against a standard mixture of iron oxide and AP showed that AP-encapsulated micron-scale catalyst particles exhibited similar behavior to AP-encapsulated nanoscale particles. Using computed tomography, we found that catalyst particles were dispersed throughout the interior of coarse AP-encapsulated micron-scale catalyst particles and decomposition was induced within these particles around catalyst-rich regions. particles and were compared against neat AP and a physical mixture of AP and iron oxide exposed to the same experimental conditions. This work identifies what impact the encapsulation procedure has on the then


ACKNOWLEDGMENTS
I would like to thank my advisor, Professor Steven Son, for helping to make me a better student, researcher and hopefully a better future employee. I would also like to thank my committee members, Dr. Chris Goldenstein and Dr. Terry Meyer for their help and support. My fellow researchers at Maurice J. Zucrow Laboratories and the Purdue Energetics Research Center also deserve a special thanks for helping me more than I could have ever asked them to.
I also would like to thank the many friends and family members who encouraged me to keep going and helped me to stay optimistic even when disappointments came. I don't know if I could have done it without you all.

ABSTRACT
Iron oxide is the most common catalyst in solid rocket propellant. We have previously demonstrated increased performance of propellant by encapsulating iron oxide particles within ammonium perchlorate (AP), but only nanoscale particles were used, and encapsulation was only accomplished in fine AP (~20 microns in diameter). In this study, we extended the size of particle inclusions to micron-scale within the AP particles as well the particle sizes of the AP-encapsulated catalyst particles (100s of microns) using fractional crystallization techniques with the APencapsulated particles as nucleation sites for precipitation. Here we report catalyst particle inclusions of micron-scale, as well as nanoscale, within AP and present characterization of this encapsulation. Encapsulating micron-sized particles and growing these composite particles could pave the way for numerous possible applications. A study of the thermal degradation of these APencapsulated particles compared against a standard mixture of iron oxide and AP showed that APencapsulated micron-scale catalyst particles exhibited similar behavior to AP-encapsulated nanoscale particles. Using computed tomography, we found that catalyst particles were dispersed throughout the interior of coarse AP-encapsulated micron-scale catalyst particles and decomposition was induced within these particles around catalyst-rich regions.

Background
The most common oxidizer used in solid rocket propellants is ammonium perchlorate. Often, a bimodal distribution of AP particle sizes is used for these composite propellants [1]. The bimodal distribution typically contains a mixture of coarse (>100 µm diameter) and fine (<20 µm diameter) particles [2][3][4]. As the coarse to fine ratio is changed, the flame structure changes, which results in different burning rates [5]. Further, adding a catalyst, such as iron (III) oxide (Fe2O3), reduces the surface activation energy for the decomposition reaction of AP and can also be used to tailor burning rates [6]. The highest propellant burning rate is seen when using nanoscale catalyst particles due to increased surface contact by the smaller particles [7][8][9][10][11]. However, increasing the catalyst amount beyond a certain point negatively impacts the propellant performance because it is an inert material in typical propellant reactions. Additionally, large amounts of fine particles increase the viscosity of the propellant to a point that makes it difficult to mix [12][13].
Encapsulating the catalyst particles averts this problem, and the ability to encapsulate nanoscale particles with AP has been demonstrated by multiple groups [14][15]. Additionally, the ability to encapsulate micron-sized particles, and to also grow those particles, may allow several other applications to be considered in future work with other materials as particle inclusions within the AP.
The encapsulation of micron-scale catalyst particles or the production of coarse AP particles with catalyst inclusions, to our knowledge, has not been performed prior to this study. In this study, we have encapsulated the nanoscale and micron-scale iron oxide particles within AP of varying particle sizes while averting the difficulty in propellant mixing that these catalyst particles create [5]. Moreover, the thermal decomposition of all encapsulated particles was also studied.
Characterization of these particles was focused on the decomposition of these particles under elevated temperatures from ambient to full decomposition temperatures. Digital microscopy, X-ray micro computed tomography (micro-CT), and differential scanning calorimetry (DSC) analyses were performed to fully characterize the encapsulated particles and were compared against neat AP and a physical mixture of AP and iron oxide exposed to the same experimental conditions. This work identifies what impact the encapsulation procedure has on the decomposition of AP, observes the locations of the iron oxide within or around the AP, and observes the impact of catalyst position on the decomposition of the AP.

Crystal Preparation Methods
Encapsulation of nanoscale particles was accomplished with a fast-crash solvent-antisolvent method, which is described in greater detail elsewhere [14]. In short, it can be summarized as overhead stirrer mixer to allow for complete equilibration before the particles were separated from the mother liquor. A No. 140 (106 µm) stainless steel sieve (Cole-Parmer, Vernon Hills, IL) was used to separate the coarse particles from fine particles as well as to prevent agglomeration, and the sieved particles were rinsed with hexanes three times and dried in a vacuum. We modified the procedures described in the literature from industrial scale to 1-gram batches. For these reasons, the process was carried out over a shorter time period than in these systems with a smaller proportional yield of coarse AP. A size distribution analysis of micron-scale iron oxide (1-5 µm, Sigma Aldrich, St. Louis, MO) is given elsewhere [16].
All tests were performed with neat and physically mixed samples with identical AP and iron oxide particle sizes. Fine AP was produced by performing the encapsulation procedure without adding catalyst. Coarse AP (RCS Rocket Motor Components, Cedar City, UT) used was 106 µm in diameter and larger. Physical mixing was carried out by mixing the AP with 1% by weight catalyst in hexane with an ultrasonic bath for five minutes followed by evaporation of the hexanes.

Characterization Methods
In other works utilizing the encapsulation procedure [5,14], encapsulation was quantitively DSC analyses were performed with heating from room temperature to 475 °C using a TA Instruments (New Castle, DE) Q600 with a heating rate of 10 °C/min under ultra-high purity argon (99.999%) at 100 mL/min with closed aluminum pans. A typical DSC profile of AP has three major events as associated with decomposition, in order of increasing temperature: first, an endothermic region associated with transition from orthorhombic to cubic phase [17] second, an exothermic energy release associated with low-temperature decomposition; and third, an endothermic or exothermic energy release associated with high-temperature decomposition. Cubic AP exhibits purely exothermic behavior in a closed pan, whereas endothermic behavior is observed in open pans [18]. While our previous work utilized open pan analysis [14], we have used closed pan analysis in this study to characterize the impact of encapsulation on all decomposition events.
Hot stage microscopy was performed using a Hirox (Hackensack, NJ) KH-8700 digital microscope coupled with a Linkam (Surrey, UK) TS1000 hot stage. Samples were heated from room temperature to 475 °C to replicate the heating rate used in the DSC analyses. A quartz pan was placed on top of a sapphire window to allow for closed pan testing in the hot stage.
Micro-CT scans were performed to identify the distribution and concentration of iron oxide particles inside the AP crystals. A Bruker (Kontich, Belgium) SkyScan 1272 with an operating voltage of 50 kV was employed with an aluminum 0.25 mm filter at a rotation step of 0.3°. The scanned images were reconstructed using NRecon software (Bruker, Kontich, Belgium) and analyzed using Avizo software (Thermo Fisher Scientific, Hillsboro, OR). The steps for analyzing each image in Avizo were the same, though settings were adjusted slightly to accommodate small variations in the images due to anomalies in scans, variations in scanner settings, and different reconstruction settings. A median filter was followed by interactive thresholding to distinguish AP and iron oxide particles from the surrounding free space and materials used to hold the sample in place during scanning. The crystal was then separated using the separate objects function.
Segmentation was used to identify free space (pores) within the particles.    Although these are qualitative measurements, the clearly show that encapsulation has occurred. Figure 1.2 shows that when encapsulated in AP, very few iron oxide particles pass through the filter paper. The AP particles are much larger than the iron oxide particles, and therefore settle in solution much more rapidly than the iron oxide particles as shown in Figure 1.3.

Encapsulation Results
By inspecting the bottom of the vials, one can observe that a similar amount of material has accumulated, but the overall transparency is vastly different between the physical mixtures and the encapsulated particles.

Thermal Decomposition Results
DSC curves of fine and coarse particles in Figures 1.4-1.5 with a)  The second exotherm has a similar effect, though its reduction is 1.9% of the physical mixture temperature and the largest variance of the second exotherm is 1.4%. These results show that confinement has a clear impact on the decomposition of these particles, though this impact seems to be dependent on the size of catalyst inclusions in the AP. Thermogravimetric analysis was used to determine average final mass content (as a percent of initial). When comparing encapsulated catalyst particles and grown AP-encapsulated catalyst particles with their corresponding physical mixtures, we found a difference of less than 1.0% which was determined to be insignificant.
Coarse AP-encapsulated catalyst particles did not exhibit consistent decomposition behavior in the exothermic region, but this may be due to multiple effects. Catalyst particles were visibly lost from the crystals during the growth procedure, and variations in the catalyst retention amounts likely has a strong impact on decomposition. Because crystallization is a purification process, and the encapsulated particles are essentially contaminants in the AP crystalline structure, it is expected that some iron oxide will not be maintained in the crystalline structure. A characterization of how much iron oxide is lost during this process was not performed because the purpose of this work was to establish the capability to create coarse AP particles with catalyst inclusions but may be of value in future work. Physical mixtures of these particles contained the same 1% weight as the fine particles. Further, as the particles grow, AP is deposited on the outside of the AP-encapsulated catalyst particles. This may create a "core-shell" effect and separates the catalyst particles from increased amounts of AP. We intentionally selected single particles for this analysis, but we did not resolve their structure to ensure that they were not polycrystalline. A more complex structure created by polycrystallinity may result in different decomposition behavior and could also contribute to the decomposition behavior observed in Fig. 5 and may be a valuable study for future work.

Hot-Stage and Micro-CT Analyses
A hot stage is essentially a small oven with viewports. In this study, this device was used with digital backlit microscopy to generate images of the decomposition as it occurred. In order to simulate a closed pan, a quartz pan was placed on top of a sapphire window. The thermal degradation of these samples is shown in Figs. 6-8. Figure 1.6 shows AP-encapsulated nanoparticles while heating. In the case of the encapsulated particles, no changes in the particles were seen from room temperature until about 245 °C. At this point, warping and movement of the crystals associated with a volume expansion indicated transition from orthorhombic to cubic crystalline structure occurred [17]. Opaque regions were seen propagating from iron oxide rich regions concentration near 270 °C, indicative of the onset of low-temperature decomposition.
Following this, the crystal structures begin to rapidly deteriorate as the particles transition into high-temperature decomposition, and the crystals eventually disintegrate and leave only catalyst remnants around 400 °C as previously reported [18]. A neat AP decomposition follows a similar process but transition from clear to opaque appears to spread throughout the crystals in a fairly even manner. Eventually the particles begin to decompose at the edges and begins to gradually retreat. it takes longer to spread throughout each particle. Neat AP gradually begins to darken (indicative of degradation) but does not appear to have a specific location where this process is concentrated.
In the grown encapsulated particles, however, dark spots are concentrated in specific locations. In the Figure 1.8, this process begins in a region where no surface defects are seen indicating that the decomposition process begins inside the particle. This dark spot spreads while other such regions begin forming and continues to grow until rapid decomposition begins occurring around 330 °C.
Large amounts of gas generated around 340 °C prevented further observations from being made, but it was apparent at this point that the particle was collapsing on itself and complete disintegration of the particle was underway.  Micro-CT analysis was attempted on coarse and fine AP-encapsulated particles, but only coarse particles were able to be satisfactorily processed. As the size of object being scanned decreases, the sensitivity to various scanning parameters become more significant, making scans of fine particles much more difficult. Future work will be devoted to exploring these parameters in order to analyze fine particles. For the purposes of this work, we focused on the analysis of coarse AP particles. Figs. 9-10 show reconstructed micro-CT images of grown AP-encapsulated micron-scale catalyst particles. An initial scan at room temperature was followed by heating the particles to multiple temperatures, with a CT scan after each heating event. Particles were heated with the hot stage with a 10 °C/min and held at each temperature for 5 minutes before cooling to room temperature without forced cooling. The areas of highest density are brightest colorthese indicate catalyst-rich regions. Figure 1.9 shows how the AP warps during heating above 245 °C due to phase transformation. Cracks form in the AP as a result of this process and are visible on the surface of the crystal. decomposition. Note that the scale bars represent a length of 500 µm, which is far larger than the mean iron oxide particle size of 1.0 µm. The most visible concentrations of iron oxide particles, therefore, are agglomerates of these particles. Though there are apparent voids in the particle prior to heating, the number of voids increases steadily with temperature. At 305 °C, after lowtemperature decomposition has begun, voids begin to appear around the iron oxide concentrations.
This continues at higher temperatures, and porosity increase throughout the particle is observed visually. Smaller iron oxide concentrations are suspected to be present at these locations. Further, the relative size of the particle has not changed significantly. These results agree closely with those observed in the hot stage analysis performed. Note that the catalyst particles appear to change slightly in sizethis is an image artifact resulting from the decomposition of AP near these particles.  After heating to 305 °C, the particle has passed through the first exothermic region of decomposition. This is evidenced by the surface of the crystal becomes slightly more porous, and the inside becomes significantly more porous. Iron oxide particles are also closer to the surface. The particle decreased slightly in size. After heating to 340 °C, the surface and the inner sections of the particle have become significantly more porous and more of the iron oxide is near the surface as the particle has decreased further in size. Of greatest significance is the observation that porosity is increasing in areas of iron oxide concentration which are located within the AP.
Closed porosity calculations are also included in Figure 1.11. Though the initial calculated porosity appears to be quite high, this is mostly a product of the irregular size and surface of the particle which make closed porosity calculations difficult. These results are best interpreted by observing the large increase in porosity as the particle begins to experience thermal decomposition that clearly favors concentrations of iron oxide. Avizo was also used to estimate the iron oxide concentration by volume, which was then used to calculate a mass concentration of 0.82%.
Because the iron oxide particles have a mean diameter on the order of the resolution of the scanner, the overall concentration may be higher. When an open pan is used rather than a closed pan, gaseous products are more readily removed from the reaction area following the first exothermic reaction. Condensed phase reactions dominate, and an endothermic region is observed. Conversely, gas phase reactions dominate during a closed pan test, resulting in a second exothermic region, and iron oxide is less able to catalyze these reactions [18]. In the first exothermic region, which has the same reaction behavior in both open and closed pan experiments, the encapsulated micron-scale particles were observed to have a significant impact on the reaction when compared to the physically mixed samples. The reason that a similar effect was not observed with AP-encapsulated nanoscale catalyst particles may be due to the relative size of the catalyst particles. Their proximity to the surface of the particles may be greater, inhibiting their ability to contribute to the gas-phase reaction. The crystal structure of the catalyst particles may also have an impact on these results. The manufacturer states that the nanoscale catalyst particles used have an amorphous crystal structure. This crystal structure has only been observed in particles less than 5 nm in diameter, whereas the micron-scale particles have a rhombohedral centered hexagonal structure [19].
The nature of the condensed-phase reaction, combined with significant losses of iron oxide during the growing procedure of coarse AP crystals, may be the reason similar trends were not observed in the DSC curves of coarse particles. Improvements in the procedure may result in higher retention of iron oxide, which is expected to result in greater catalyzation. Additionally, polycrystalline effects resulting from the synthesis of coarse AP-encapsulated catalyst particles may be another contributing reason that these trends were not observed in the DSC curves of coarse particles.
Digital microscope images of particles while being heated in a hot-stage offered additional insights. Visible transition from orthorhombic to cubic phase was observed between 245 °C and 27 250 °C. The degradation process was observed to begin faster with samples containing encapsulated iron oxide. It appeared to begin around areas of high concentration in fine AP crystals and from within coarse AP crystals. Observations are limited after the particles transitioned from low-temperature decomposition to high temperature decomposition, but iron oxide remnants were still visible following the full decomposition process. Micro-CT analysis of coarse APencapsulated micron-scale catalyst particles confirmed that iron oxide was dispersed throughout the inside of the particles, not concentrated on the surface, and that decomposition occurs at these particles. This work not only gives insight to how decomposition proceeds with encapsulated catalysts. It also shows that micron catalyst particles can be encapsulated. This is significant because other applications for larger encapsulated particles, such as the inclusion of thermographic phosphors or microwave absorbing particles that will likely need to be micron-scale. In addition, we have demonstrated that larger AP particles with encapsulated catalysts can be produced.