Elsevier

Combustion and Flame

Volume 161, Issue 6, June 2014, Pages 1668-1677
Combustion and Flame

Melt dispersion mechanism for fast reaction of aluminum nano- and micron-scale particles: Flame propagation and SEM studies

https://doi.org/10.1016/j.combustflame.2013.11.021Get rights and content

Abstract

Flame propagation studies for Al nanoparticles (80 nm) and micron particles (3–4.5 μm) mixed with MoO3 in both an open and confined burn setup were examined. A scanning electron microscopy (SEM) analysis of the reactants and products reveals quantitative size data that contributes toward an understanding of the governing reaction mechanisms. For the confined burn tube experiments, nanoscaled reactants exhibited a flame speed of 960 m/s, the same as has been reported in previous experiments. Micron scale particles exhibited a flame speed of 402 m/s, much higher than the 244 m/s obtained previously for 1–3 μm particles. These flame speeds are in quantitative agreement with predictions based on the recently developed melt-dispersion mechanism (MDM) describing the reaction of Al particles. It also demonstrates that some micron particles can reach flame speeds just 58% lower than the fastest nanoparticles, while micron scale particles are less expensive and do not have the pre-combustion safety and environmental issues typical of nanoparticles. The SEM analysis reveals a significant (at least by factor of 3.7 for nanoparticles) reduction in Al particle size post combustion, which is in agreement with the MDM and in contrast to the predictions based on diffusion mechanisms. Open burn experiments with nanoscale reactants have flame speeds of 12 m/s and product particle sizes almost as small as those in the burn tube experiments. However, the presence of some large particles, which may grow based on the diffusion mechanism, exclude evaporation and the homogenous nucleation mechanism. For open burn experiments with micron reactants, with flames speeds of 9 m/s, SEM analysis shows a molten-resolidified product with no distinguishable particles and cavities containing numerous nanoparticles with a measured diameter of 36 nm.

Introduction

Aluminum particles have been integrated into various energetic formulations and technologies, ranging from propellant rate modifiers to ordnance applications and energy sources. Thermites composed of homogeneously mixed nanoscale aluminum and metal oxide particles have been shown to exhibit a three order of magnitude increase in flame propagation speed (reaching 1 km/s) and three order of magnitude reduction in ignition delay time in comparison to traditional 10–100 μm size thermite mixtures [1], [2], [3]. This has led to an increasing amount of research focused on nanoparticle reaction mechanisms [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19]. In contrast to the reaction of micron scale Al particles, which occurs in the gas phase, nanoparticles react in the condensed phase [5], [6], [7], [8], [13], [14], [15], [16], [17]. Since Al particles are covered by an oxide shell a few nm thick, the reaction’s limiting factor was considered to be the diffusion of Al and oxygen through the growing oxide shell. Various diffusion based models have been developed to describe this behavior [5], [11], [12], [17]. However, strong indications were reported that the diffusion mechanism was too slow to produce a flame propagation speeds of 1 km/s [9], [10], which led to the development of a new mechanochemical mechanism named the melt-dispersion mechanism (MDM) [9], [10].

According to the MDM the alumina shell does not fracture until the Al core melts, at least partially. The melting of Al is accompanied by a 6% volume increase, which leads to a compressive pressure of several GPa within the Al core as tensile hoop stresses on the order of the ultimate strength of alumina (∼11 GPa) develop in the alumina shell. This high tensile stress causes the shell’s dynamic fracture and spallation. After spallation of the shell, the pressure within the Al liquid core remains unchanged (on the order of 1–3 GPa) while at the bare surface the pressure is on the order of only 10 MPa due to the surrounding gas pressure and surface tension [10]. Due to this unbalanced pressure, an unloading wave propagates to the center of the Al core and creates a tensile pressure of 3–8 GPa. This pressure exceeds the cavitation limit of liquid Al, which causes molten Al to disperse into small, bare clusters (fragments) that fly at a high velocity of 100–250 m/s [10]. Oxidation of these clusters is not limited by diffusion through an initial oxide shell. Consequently, the MDM transforms a single Al particle covered by an alumina shell into hundreds or thousands of smaller bare particles. This explains the dramatic increase in Al particle reactivity at the nanoscale.

Due to the small time and space scales involved with nanoscaled Al, there are no in situ experimental observations of any proposed reaction mechanisms. However, for the MDM, various experimental results for the flame rates and burn times of nanoparticles are in good qualitative and quantitative agreement with theoretical predictions [9], [10], [20], [21], [22], [23], [24], [25], [26], [27], [28], while they qualitatively contradict the proposed diffusion mechanism. In particular, there is a quantitative description of the nontrivial experimental relationship between the flame rate and the reduced particle radius [9], [10], [20], including the independence of the flame rate from the particle size below some critical reduced particle radius. The MDM also quantitatively explains why a reduction (or increase) in the oxide shell thickness reduces (or increases) the flame rate (or vice versa), in direction opposition to the behavior proposed by the diffusion mechanism [20], [21]. The MDM explains the effect of heating rate on the reaction time [22] and the effect of density of the reactive mixture on the flame rate [23]. The MDM was also expanded to micron scale particles (i.e., 1–3 μm diameter particles) [24], [25], with flame rates close to the predicted values when Teflon was used as an oxidizer. When MoO3 oxidizer was used, the measured flame rate was only half the predicted value. This was explained in terms of non-optimal oxidation conditions. Also, a method for increasing the reactivity and flame rate for nano- and micron scale particles by producing pre-stressed core–shell structures, suggested based on the MDM, was quantitatively confirmed experimentally [27]. Other research groups have also discussed the MDM (e.g., in [16], [19], [30], [31]), their comments will be analyzed in Section 4.

The main goal of this paper is to study the particle size distribution before and after flame propagation experiments and to measure flame velocity in order to directly compare the MDM and diffusion mechanisms. Aluminum nano- and micron-scale particles with nano-scale MoO3 oxidizer were studied in a confined burn tube [1], [2] and an open burn configuration. Both configurations produced flame speeds in excess of 10 m/s, which is at the lower limit for the activation of the MDM [22]. According to the MDM, each particle disperses into multiple particles; thus, the particle size of post combustion products should be drastically reduced. According to the diffusion mechanism, a complete reaction of Al to Al2O3 leads to a factor of 1.25 volume increase; the particle diameter should then be increased by a factor of 1.251/3 = 1.08. In addition, particles may coalesce (through reactive sintering [14]), which would lead to an additional increase in particle size. Due to these opposing trends, it should be simple to distinguish between the smaller products produced by the MDM and larger products produced by diffusion mechanisms in these experiments.

Section snippets

Materials

Experiments were conducted using Al and MoO3 thermites. Two Al particle sizes (DAl, 3–4.5 μm and 80 nm) and one MoO3 particle size (Dox, 370 nm, supplied by Technanogy) were used. Loose powder mixtures were prepared with a slightly fuel rich stoichiometry (equivalence ratio of 1.2). The supplier of the micron scale Al particles (Sigma Aldrich) characterized the alumina shell thickness (δ) as 2 nm, but as the shell can slowly grow in air over time a shell thickness of 4 nm was also considered. This

Experimental results

Results from flame propagation experiments are presented in Table 2. The flame speed for composites made with nanoscaled Al in the confined tube configuration was 960 m/s, which is consistent with previous studies [1], [20], [24], [25] and was rationalized in terms of the MDM [9], [10], [20], [24], [25]. For composites made with micron scale Al, a record high flame speed of 402 m/s was measured. This is much higher than the previously reported flame speed of 205 m/s for a composite that used the

Inert atmosphere

Aluminum particles with a diameter of 80 nm have been studied in an Ar atmosphere in a shock tube with the goal of testing the existence of the MDM [16]. However, Al vapor was not detected at temperatures less than 2275 K, which is close to the melting temperature of the alumina shell. This led to the conclusion that the oxide shell was intact below this temperature, which contradicts the MDM. However, these experiments contradict any existing nano-aluminum ignition model that is based on the

Concluding remarks

Several new confirmations of the theoretical predictions based on the MDM for oxidation of Al particles mixed with MoO3 are presented. The flame speed for 3–4.5 μm Al particles in the burn tube reached 402 m/s, which is in quantitative agreement with predictions based on MDM. Such a flame speed is a record high, much higher than 205 m/s for the same Al particles previously reported [27] and higher than 244 m/s obtained previously for 1–3 μm Al particles [24], [25]. Since the only difference between

Acknowledgments

Financial support of the Office of Naval Research (Grant N00014-12-1-0525) with Dr. Cliff Bedford as Program Officer is gratefully acknowledged. Discussions with Dr. Nick Glumac are very much appreciated.

References (34)

  • K.B. Plantier et al.

    Combust. Flame

    (2005)
  • J. Granier et al.

    Combust. Flame

    (2004)
  • R.A. Yetter et al.

    Proc. Combust. Inst.

    (2009)
  • T. Bazyn et al.

    Combust. Flame

    (2006)
  • T. Bazyn et al.

    Proc. Combust. Inst.

    (2007)
  • V. Rozenband et al.

    Combust. Flame

    (1992)
  • Y. Ohkura et al.

    Combust. Flame

    (2011)
  • V.I. Levitas

    Combust. Flame

    (2009)
  • K.W. Watson et al.

    Combust. Flame

    (2008)
  • V.I. Levitas et al.

    Combust. Flame

    (2011)
  • B. Bockmon et al.

    J. Appl. Phys.

    (2005)
  • A. Rai et al.

    Combust. Theor. Model.

    (2006)
  • K. Park et al.

    J. Phys. Chem. B

    (2005)
  • V.I. Levitas et al.

    Appl. Phys. Lett.

    (2006)
  • V.I. Levitas et al.

    J. Appl. Phys.

    (2007)
  • V. Rozenband

    Thermo-mechanical aspects of the heterogeneous ignition of metals

    Combust. Flame

    (2004)
  • A. Rai et al.

    J. Phys. Chem. B

    (2004)
  • Cited by (0)

    View full text