Investigating the oxidation mechanism of tantalum nanoparticles at high heating rates

Metal fuels possess higher energy densities than standard CHNO based energetics but traditionally suffer from slow reaction kinetics.^1–3 However, reduced diffusion length scales and increased specific surface areas of nanosized components have recently demonstrated increased reaction rates for these systems.^1,4 Most metal nanoparticles are passivated by a thin (typically 3–5 nm) nascent oxide shell, creating a barrier between the metal fuel and any oxidizer.⁵ As the particle size decreases, the mass fraction of metal oxide makes up a significant percentage of the particle. Understanding the interaction between the metal core and the metal oxide shell is crucial in understanding the oxidation mechanism of the system.^6,7

Aluminum is the most commonly used metal fuel due to its favorable energy density, low cost, and ignition characteristics.² Its oxidation mechanism has been extensively studied, and two prominent mechanisms for the ignition of Al nanoparticles (nAl) have been proposed: a diffusion based mechanism where the oxidizer and aluminum diffuse through the oxide shell boundary^8–10 and a melt dispersion based mechanism¹¹ where there is a violent rupturing of the oxide shell followed by spallation of the aluminum core. Previous studies have shown that decreasing the particle size of the fuel can dramatically decrease the ignition temperatures of Al based nanocomposites with the minimum recorded ignition temperatures being near the aluminum melting point.^10,12 High heating rate experiments have also shown that sintering of Al can occur on similar timescales as chemical reaction.^13,14

A relatively recent development is the commercial availability of nanopowders of a wide variety of metals. In this study, tantalum (Ta) nanopowder was chosen due to its high melting point in comparison to Al, 3017 °C vs 660 °C. It is known that the crystallization of the aluminum oxide shell (Al₂O₃) is an important part of the Al ignition mechanism.^7,15,16 Unfortunately, this is a difficult process to thoroughly investigate because oxide crystallization occurs at a temperature very close to the melting point of Al.⁷ Analogously, it has been observed that amorphous to crystalline transition in Ta₂O₅ thin films begins at ∼500 °C with crystallization occurring rapidly at higher temperatures;¹⁷ however unlike Al, the effect of the oxide shell crystallization on ignition of tantalum nanopowders (nTa) can be probed without interference from melting.

Ta/tungsten oxide (WO₃) thermite composites prepared by sol-gel synthesis and spark plasma sintering have been previously studied, but these composites employed micron-sized (5 μm average particle size) tantalum.^18–20 In our work, the oxidation of nTa (<50 nm average particle size) in an aerobic environment at both low and high heating rates was studied in addition to the reaction mechanisms of several tantalum based nanothermites at high heating rates. These experiments were used to examine the oxidation of nTa strictly by a gas phase oxidizer (O_2(g)) in addition to a condensed phase oxidizer (metal oxide) in a nanothermite mixture. Thermogravimetric analysis and differential scanning calorimetry (TGA/DSC) were employed to investigate the oxidation of Ta and Al nanopowders at slow heating rates. Temperature jump (T-Jump) ignition and high heating rate transmission electron microscopy (TEM) experiments were performed to examine the ignition behavior of Ta nanopowders (nTa) in the presence of oxygen, in addition to Ta based nanothermites in vacuum. Ta ignition was observed well below the melting points of Ta and tantalum oxide (Ta₂O₅), 3017 and 1872 °C, respectively. The experimental results allude to an oxygen diffusion dominated oxidation mechanism as the metal fuel is immobile prior to measured ignition temperatures. High heating rate TEM experiments also show two distinct oxidation mechanisms occurring depending on the maximum temperature achieved in the respective experiment.

nAl (∼50 nm) was purchased from Novacentrix with an active Al content of 81% by mass, determined by thermogravimetric analysis (TGA) as stated by the manufacturer. The nTa (<50 nm) used in this study was purchased from Global Advanced Metals. All metal oxide nanopowders (<50 nm) used in Ta based thermite mixtures for T-Jump and TEM analysis were purchased from Sigma Aldrich. Stoichiometric thermite suspensions were prepared by physically mixing Ta with the corresponding metal oxide in hexane (T-Jump) or ethanol (TEM) and sonication for 30 min. Because nTa disperses poorly in hexane, and it is actually an aid in the coating of highly concentrated nTa particles on the T-Jump filament. nTa is much more easily dispersed in ethanol and was used as the carrier solvent of low concentrations of particles needed for TEM sample deposition.

Thermogravimetric analysis and differential scanning calorimetry (TGA/DSC) were performed using a TA Instruments SDT Q600. For both nTa and nAl oxidation analysis runs, a heating rate of 20 °C/min and an O₂ flow rate of 100 mL/min were employed.

A home built temperature jump (T-Jump) thin platinum (Pt) wire ignition source was used with a heating rate of ∼10⁵ K/s.^21,22 The 76 μm diameter Pt wire was rapidly joule heated with a 3 ms current pulse, while a high-speed video was taken at 67 000 frames per second using a Phantom v12.0 digital camera running Phantom 692 software. Ignition time and temperature were then determined by mapping against the Callendar-Van Dusen temperature extrapolation from the temporal resistance of the Pt wire. T-Jump ignition experiments were performed in a pure O₂ environment at pressures ranging from 20.3 kPa to 1.89 MPa and under vacuum (∼0.13 mPa).

High heating rate studies in a TEM were conducted with Protochips Aduro heating chips with a Protochips TEM holder in a JEM 2100 field emission gun (FEG) TEM/scanning transmission electron microscope (STEM), equipped with Oxford energy dispersive spectrometry (EDS) and Gatan Tridiem electron energy loss (EELS) systems, with an accelerating voltage of 200 kV. Ta nanopowders were deposited onto an Aduro chip and heated in air, whereas the Ta based nanothermites were heated under vacuum within the TEM. All samples were heated at 10⁵ K/s and held for a period between 1 and 100 ms.

Ta/CuO nanothermite was evaluated using a previously described combustion cell.²³ 25 mg of sample was placed in the ∼13 cm³ cell and ignited by a resistively heated nichrome wire. An oscilloscope was used to record simultaneous pressurization and optical emission from a pressure transducer and a photomultiplier tube, respectively.

The oxidation of nTa was first studied using standard slow heating rate TGA/DSC under an O₂ flow (Fig. 1).

In the case of nTa, oxidation is a 1 step process beginning at ∼300 °C and ending at ∼500 °C. The exotherm occurring slightly above 600 °C possibly corresponds to the crystallization of Ta₂O₅ as the mass stays constant during this event. Complete oxidation of nTa would result in a final weight percent of ∼115%, indicating the nTa contains ∼70% unoxidized Ta before heating, as shown in Fig. 1(a). The onset temperature for oxidation is much lower for nTa (∼300 °C) than for nAl (∼600 °C), but the initial exotherm corresponding to oxidation occurs much more rapidly in the nAl case. This can be attributed to the outward diffusion of Al from the core when near the melting point of Al (660 °C). Trunov et al. have described the oxidation of nAl at slow heating rates as being a 4-step process with the predominant mass increases occurring in steps 3 and 4 corresponding to the growth of γ and α phase Al₂O₃, respectively.⁷ nTa does not appear to follow this multi-step oxidation mechanism as there is only a single stage of mass gain during the course of heating.

In the TGA/DSC analysis of nTa oxidation, both the fuel and metal oxide product are immobile in the temperature range of the experiment. Al begins to rapidly oxidize near the Al melting point (660 °C) due to the increased diffusion rates of liquid Al allowing for enhanced outward diffusion of Al through the nascent oxide shell. With Ta and Ta₂O₅ being immobile during the course of heating, oxidation must occur strictly by diffusion of O₂ through the oxide shell at slow heating rates. However, nTa undergoes a rapid self-sustaining oxidation when heated quickly which is not well captured using slow heating rate analytics such as TGA/DSC.

The ignition behavior of pure nTa at various pressures of O₂ was analyzed using the T-jump heating technique coupled with high speed videography. The measured ignition temperatures as a function of O₂ pressure are shown in Fig. 2.

The measured ignition temperatures are much lower than those of nAl at the same pressures using an identical technique.¹² With Ta having such a high melting temperature, the reaction must be heavily dependent on the diffusion of oxygen through the oxide shell to the Ta core. The nascent Ta₂O₅ shell on nTa has a melting point of 1872 °C, which is more than 1000 °C higher than the measured ignition temperatures. This suggests that something must happen to this shell to allow for enhanced diffusion rates of gaseous oxygen and the ignition event, leading to self-sustained oxidation of the Ta core. As stated above, Kim and Stebbins have shown that amorphous Ta₂O₅ films will crystallize above 500 °C;¹⁷ thus, it is reasonable to conjecture that the exothermic crystallization of Ta₂O₅ allows for enhanced diffusion of oxygen to Ta across the Ta₂O₅ boundary. This can be thought of as being similar to the hot spot ignition theory for CHNO energetics where localized points can begin to react exothermically, leading to a self-propagating reaction of the bulk material.²⁴ Potentially, only one point may rapidly crystallize, exposing the Ta core aiding in the rapid oxidation of Ta. The heat generated during the rapid oxidation will contribute to the further crystallization, and at a potential a temperature spike to enable melting of Ta₂O₅, exposing more of the Ta core. This theory was further explored using high heating rate TEM.

In order to probe the oxidation mechanism of nTa, the growth of the Ta₂O₅ oxide shell was investigated by depositing nTa onto Protochips Aduro heating chips. First, the chip was inserted into the Protochips TEM holder, multiple particles were imaged, and positions were recorded. In order to oxidize nTa, the holder was removed from the TEM and rapidly heated in air to 500 °C at a heating rate of 10⁵ K/s and held at this temperature for 10 ms. This process was repeated 4 times with re-insertion of the holder into the TEM to image the particles after each stage of heating. During the last heating run, the chip was held at 500 °C for 100 ms. The results in Fig. 3 show oxide shell thickness growth for 7 different locations on a single Ta aggregate. It is apparent that there is a difference in the growth rate depending on the location on the aggregate.

No crystallization of Ta₂O₅ is observed after any of the consecutive heating ramps, which agrees with the results seen in the TGA/DSC experiment above. In the TGA/DSC, the Ta₂O₅ crystallization exotherm occurs above 600 °C.

Slow heating rate studies on the oxidation of bulk tantalum in air indicate that oxidation follows a linear rate law in the 500 °C–800 °C temperature range.²⁵ The experimental results for the average oxide shell growth over time in Fig. 3 appear to follow a linear-parabolic trend similar to the Deal-Grove (D-G) model for the oxidation of silicon.²⁶ The Deal-Grove model originally developed for oxidation of silicon shows growth following a linear growth rate at short timescales (thin oxide layer) and a parabolic growth rate for long oxidation times (thick oxide layer). This implies that oxidation is kinetically limited in the initial stages, but once the shell reaches a certain thickness, oxidation is diffusion limited as shown below

where $Z$ is the oxide shell thickness, $Cs$ is the molar concentration of oxygen at the surface of the material, $ṽ$ is the molar volume of the formed metal oxide, $t$ is the time, and $k$ and $D$ represent the rate constant in the linear region and diffusion coefficient of O₂ in the formed metal oxide, respectively. Since both Si and Ta are immobile during the temperature range of oxidation and presuming no other factors come into play, D-G should be applicable to the Ta system. For the rapid heating of nTa in air to 500 °C shown in Fig. 3, the linear to parabolic transition occurs after ∼1.62 nm of shell growth (20 ms of total heating). The rate constant for the linear region, $k$⁠, is determined to be 160.4 nm/ms. A diffusion coefficient, $D$⁠, of 52 nm²/ms (5.2 × 10⁻¹⁴ m²/s) was extrapolated from the parabolic portion of the plot.

A previous study on the diffusion of oxygen in amorphous Ta₂O₅ thin films (100 nm thickness) reported a diffusion coefficient of 2.5 × 10⁻¹⁹ m²/s at ∼525 °C.²⁷ Our extrapolated value for oxygen diffusion through the oxide shell of nTa is more than 5 orders of magnitude larger. This is most likely due to the previous study being done on compact Ta₂O₅ films, while our oxygen diffusion coefficient was determined during oxidation of nTa and therefore more accurately represents oxygen diffusion during combustion.

Drastic surface changes of the oxide shell are clearly seen in Fig. 4, when Ta nanoparticles are heated in air to 850 °C and held at temperature for 1 ms. A temperature of 850 °C was chosen as the final temperature to ensure that all analyzed particles were sufficiently heated above 500 °C with only 1 ms of hold time, as a result of heat losses from the chip when heated in air. During oxidation, the particle's outer diameter increased from 32.83 to 38.16 nm, while the Ta core shrank by nearly 4 nm. This particle diameter growth is due to the density of Ta₂O₅ being approximately half that of the Ta metal core (8.2 g/cm³ vs. 16.7 g/cm³, respectively). After rapid heating to 850 °C, the oxide shell grew ∼4.6 nm over 1 ms. After 130 ms of total heating to 500 °C, the average shell growth was only ∼3.7 nm.

A crack observed in the post-heating TEM image in Fig. 4(d) is further evidence that changes in the oxide shell may be enhancing oxidation rates. The rapid crystallization and the stress on the particle may cause the observed cracking of the shell, which can enhance oxygen diffusion into the core. Elemental compositions from EDS point scans in Fig. 4 show that no oxygen exists in the center of the particle (a) prior to heating in air. After heating, a significant increase in the percentage of oxygen is observed in both the edge (b) and center (c) of the particle. As oxidation occurs, it is possible that local temperatures may reach the melting point of Ta₂O₅, explaining the lack of multiple cracks in the shell in the post-heating TEM images. Localized melting of Ta₂O₅ would seal the crack, but as this newly formed amorphous oxide crystallizes, new cracks may form.

The above D-G model shows that oxidation is kinetically limited when the oxide shell is sufficiently thin. If we assume that cracking of the shell prevents the reaction from proceeding to the diffusion limited regime, we can calculate the effective activation energy for oxidation using the calculated rate constant for the linear region and the Arrhenius equation. The relative time required to heat the particles to 850 °C at 10⁵ K/s with a 1 ms hold is significant in contrast to rapid heating to 500 °C and holding for 10 ms. Theoretical shell growth, assuming a kinetically limited oxidation mechanism, during the course of the entire heating to the 850 °C experiment is shown in Fig. S1 (supplementary material). Iterative calculations determined that $Ea$ required to match the experimentally measured shell growth of 4.6 nm is ∼65 kJ/mol (0.674 eV).

Selected area diffraction (SAD) patterns before and after rapid heating are shown in Fig. 5. Before heating, crystalline Ta from the core of the particle is observed, but no crystalline Ta₂O₅ is apparent. After heating, fringe lines corresponding to both crystalline Ta and crystalline Ta₂O₅ are visible. The clustered Ta₂O₅ fringe lines imply that the formed Ta₂O₅ has a preferred orientation with crystalline oxide growth occurring on specific planes.

High resolution TEM (HRTEM) images in Fig. 6(a) reveal little crystallization in the oxide shell after the series of heating ramps to 500 °C (30 ms in total), but crystallization is clearly observed after being heated to 850 °C for 1 ms in Fig. 6(b) (also confirmed by SAD in Fig. 5). These results show that rapid heating to 850 °C results in a faster diffusion based oxidation mechanism that may be attributed to cracks forming in the oxide shell.

Various nTa based nanothermites were investigated using the same T-jump ignition technique as above. The ignition temperatures were measured in a vacuum environment (∼10⁻⁶torr) to eliminate the impact of ambient gas phase oxygen on ignition. The oxygen release temperature of the pure metal oxides was plotted vs. ignition temperature of nTa and nAl based thermites as shown in Fig. 7.

Ignition temperatures for the Ta based nanothermites appear to be independent of oxidizer type. In some cases, the ignition temperatures vary greatly from the corresponding Al based nanothermites, with the greatest disparity being between the Fe₂O₃ nanothermites. Post-combustion products from the T-Jump ignition in air of Ta/CuO nanothermite were imaged (Fig. S2 in the supplementary material) using a previously published method, and results are discussed in the supplementary material.²⁸ One important result not shown in Fig. 7 is that the Ta/WO₃ nanothermite did not ignite under these heating conditions (∼24 °C to ∼1200 °C in 3 ms), but Al/WO₃ nanothermite has been previously shown to ignite at 757 °C in an identical setup.¹⁰ 

Table I shows the theoretical enthalpy of reaction per mole of metal fuel and adiabatic flame temperature for each nanothermite in Fig. 7. The theoretical adiabatic flame temperatures were calculated using the Cheetah 6.0 equilibrium code at a constant pressure.

The adiabatic flame temperatures are lower for all Ta based nanothermites, with the exception of the CuO containing nanothermite where the temperatures are approximately equivalent. There does not appear to be a trend between the enthalpy of reaction, ΔH_rxn, and the ignition temperatures measured in Fig. 7. The Ta/Fe₂O₃ system has a lower ΔH_rxn than Al/Fe₂O₃ yet ignites at a lower temperature.

One possible explanation for the discrepancies in ignition temperatures of Ta and Al based nanothermites is that alloying reactions unique to the Ta system may decrease the onset temperature of reaction. A phase diagram for Ta₂O₅ and Fe₂O₃ generated using density functional theory (DFT) calculations shows that the TaFeO₄ alloy can be formed with an enthalpy of formation of −2.5 kJ/mol.^29–31 No stable alloy exists for the Al₂O₃/Fe₂O₃ system. Unfortunately, this pattern does not hold true for the Co₃O₄ containing thermites, which also show a large ignition temperature difference. Neither system forms a stable alloy with Co₃O₄; therefore, it is more likely that something unique to nTa and independent of the metal oxide is the cause for the trend in Fig. 7. As stated above, the Ta/WO₃ nanothermite did not ignite under these conditions, contradictory to previous studies on sol-gel derived Ta/WO₃ thermites that ignited between 465 and 670 °C.¹⁸ This could be due to differences in the microstructure of the composite, heating environment, and heating rate. In addition, the sol-gel samples were pre-heated to 300–400 °C before ignition tests, and they also employed amorphous WO₃ where we used crystalline WO₃.

For each of the Ta based nanothermites, ignition occurs at or below the oxygen release temperature for the corresponding metal oxide. We propose that the metal oxide must wet the surface of the nTa for ignition to occur due to Ta being an immobile fuel under these conditions. Each of the Ta based nanothermites in Fig. 7 ignites fairly close to 700 °C, whereas the Al based nanothermites have a much wider range of ignition temperatures. The TGA/DSC results in Fig. 1(a) show an exotherm resulting from the crystallization of Ta₂O₅ with an onset temperature of 600 °C and peak temperature of 630 °C at low heating rates. The analysis of the crystallization of amorphous Ta₂O₅ thin films by Kim and Stebbins showed that crystallization rates rapidly increased at higher temperatures.¹⁷ We propose that at high heating rates, the crystallization of the amorphous shell releases enough energy in a short amount of time at the nTa/metal oxide interface to initiate reaction between the fuel and the oxidizer. In the case of Al based thermite reactions, Al goes through an endothermic phase change (melting) near the crystallization temperature of its oxide shell. With Al₂O₃ making up only 19 wt. % of nAl, the net process in this temperature regime would be endothermic as energy released from crystallization would go into the melting of Al. For the Ta case, the metal core does not go through any phase changes near the crystallization temperature of the oxide shell; therefore, the energy released from crystallization will locally heat the particle. Similar conclusions on crystallization induced ignition were drawn by Cervantes et al. for the sol-gel derived Ta/WO₃ thermite system where they saw decreased ignition temperatures when amorphous WO₃ was employed.²⁰ These types of surface interactions were further probed using high-high heating rate TEM for Ta/Fe₂O₃ and Ta/CuO nanothermites below (Fig. 8).

The Ta/Fe₂O₃ nanothermite system showed the largest disparity between nAl and nTa based systems, warranting a closer look at the reaction mechanism. Ta/Fe₂O₃ nanothermite was heated in the TEM at 10⁵ K/s to 1200 °C and held for 1 ms. Sullivan et al. preformed a similar study using the same in-situ high heating rate TEM technique with an Al based thermite system, and a reactive sintering reaction mechanism was supported.⁸ Figures 8(a) and 8(b) show before and after heating TEM images where Fe₂O₃ appears to wet the surface of Ta to allow for oxygen to diffuse through the Ta₂O₅ shell to oxidize Ta. Unlike Al, Ta does not sinter at areas that are not in contact with Fe₂O₃ under these heating conditions as shown by the maintained nanostructure of parts of the Ta aggregate shown in Fig. 8(b), which again proves that wetting the Ta fuel surface by the oxidizer is crucial for its ignition. The added heat at the interface may be promoting the reactive sintering between nTa and Fe₂O₃.

Ta/CuO nanothermite was also rapidly heated in the TEM to 1200 °C at a heating rate of 10⁵ K/s and held for 10 ms. TEM images shown in Fig. S3 (supplementary material) show a similar “wetting” phenomenon as seen in the Ta/Fe₂O₃ system. EDS results show that that CuO is fully reduced to Cu and even at a location away from the interface, Ta is oxidized, showing the ability for oxygen to be transported throughout the Ta aggregate, even in a vacuum environment.

Using our experimentally determined activation energy for oxidation (65 kJ/mol) described above, a kinetically limited shrinking core model was employed to determine a theoretical burn time for a stoichiometric Ta/CuO nanothermite. We primarily chose a shrinking core based on our prior studies on size resolved measurements on the burn rate of nanoscale titanium and zirconium particles.³² The model derivation is outlined in Ref. 33, and the final equation can be stated as

where ρ_Ta is the molar density of tantalum, r_c is the radius of the unreacted core, k is the reaction rate constant, C_O2 is the concentration of the gaseous oxidizer, and b is the stoichiometric coefficient of Ta in the oxidation reaction normalized per mole of oxidizer (0.8Ta + O₂ -> 0.4Ta₂O₅). The oxygen concentration used in the model was determined from the full decomposition of the metal oxide (CuO in this case). The model assumes that oxidation of Ta is a kinetically limited process and the timescale of oxygen release from CuO is negligible. The experimental burn time was obtained from evaluating the burn time within a fixed volume combustion cell. The amount of CuO used defines the totally available oxygen initially.

For our calculations, we assumed a 33 nm nTa particle as depicted in Fig. 5. We also ignored heat losses from convection and radiation but limited the maximum achieved temperature to the adiabatic flame temperature for Ta/CuO (2842 K) as shown in Table I. The rate constant at each time step in the model was calculated using the Arrhenius equation below with the experimentally determined activation energy of 65 kJ/mol

where $k$ is the reaction rate coefficient at a given temperature, $Ea$ is the activation energy for oxidation, $kB$ is the Boltzmann constant, and $T$ is the temperature. The maximal reaction rate coefficient, $ko$⁠, of 3.96 × 10⁶ nm/ms was determined from the value extrapolated from successive heating to 500 °C and single heating ramp to 850 °C. The oxygen concentration used in the model was determined by using a stoichiometric amount of CuO in the combustion cell with a total nanothermite mass of 25 mg and a volume of 13 cm³. The oxygen content in the model was finite; therefore, oxygen concentration decreased as it was consumed by nTa. The model results and combustion cell experimental results are shown in Fig. 9.

The kinetically limited shrinking core model shows that the core diameter has shrunk by ∼90% in 0.9 ms [Fig. 9(a)]. The FWHM of the optical emission from the combustion cell experiment, commonly used as the burn time, shows a duration of ∼0.8 ms [Fig. 9(b)], which suggests that a shrinking core model reasonably captures the reaction time for Ta high temperature oxidation based on parameters extracted at low temperature Deal-Grove oxidation.

Two distinct nTa oxidation mechanisms are apparent for nTa reacting with gas phase oxygen: a slow heating rate, low temperature oxidation resulting in the formation of amorphous Ta₂O₅, and a rapid oxidation caused by the crystallization of amorphous Ta₂O₅ creating cracks in the shell.

The Ta nanothermite reaction mechanism is controlled by reactive sintering with metal oxide. The metal oxide must “wet” Ta because Ta is immobile in the temperature range of ignition. Rapid crystallization of the oxide shell leads to heat release at the interface promoting sintering with the metal oxide. nAl also undergoes crystallization of its oxide shell, but the net energy at the interface is endothermic due to the melting of the Al core, which makes up 81% of the mass of a nAl particle. The exothermic crystallization of the Ta₂O₅ shell (Fig. 1) may be attributing to the initiation of localized reactive sintering with Fe₂O₃. This could also explain the trend in ignition temperatures seen in Fig. 7 where the ignition temperature for nTa based thermites seems to be independent of the metal oxide employed.

This study probed the importance of oxide shell crystallization in the oxidation mechanism of Ta nanoparticles and nanothermites. High heating rate TEM analysis showed that oxide shell growth occurred slowly when heated in air below 500 °C and was amorphous. When heated in air to 850 °C, oxide shell growth happened rapidly and crystallization of the shell led to cracks that may attribute to the enhanced oxygen diffusion rates. When investigating the nTa oxidation mechanism with a condensed phase oxidizer, ignition temperatures were independent of the employed metal oxide, unlike corresponding Al nanothermites. We propose that the rapid oxide shell crystallization of nTa contributes enough energy to the shell/oxidizer interface that reactive sintering is induced and kinetically limited oxidation occurs. An activation energy for oxidation was experimentally determined to be 65 kJ/mol from the D-G model and was used to model the burn time of the Ta/CuO nanothermite system.

See supplementary material for a figure (Fig. S1) showing the theoretical shell growth of nTa using the kinetically limited portion of the Deal-Grove model using an activation energy of 65 kJ/mol. There is also a discussion on T-Jump product collection and SEM images of Ta/CuO nanothermite combustion products (Fig. S2). Figure S3 shows TEM images before and after heating of Ta/CuO nanothermite in addition to EDS results.

This work was supported by the Army Research Office, the Defense Threat Reduction Agency, and the AIM lab (MRSEC Shared Facility, NSF DMR 05-20741) at the University of Maryland. Special thanks to Dr. Sz-Chian Liou and Dr. Wen-An Chiou for technical assistance.