Combustion and Flame 162 (2015) 1350–1357

Contents lists available at ScienceDirect

Combustion and Flame

journal homepage: www.elsevier .com/locate /combustflame

Altering combustion of silicon/polytetrafluoroethylene with two-stepmechanical activation

http://dx.doi.org/10.1016/j.combustflame.2014.11.0050010-2180/� 2014 The Combustion Institute. Published by Elsevier Inc. All rights reserved.

⇑ Corresponding author. Fax: +1 (765) 494 0530.E-mail address: terry13@purdue.edu (B.C. Terry).

Brandon C. Terry a,⇑, Steven F. Son b, Lori J. Groven b,c

a School of Aeronautics and Astronautics, Purdue University, Zucrow Laboratories, 500 Allison Rd, West Lafayette, IN 47907, United Statesb School of Mechanical Engineering, Purdue University, Zucrow Laboratories, West Lafayette, IN 47907, United Statesc Chemical and Biological Engineering Department, South Dakota School of Mines and Technology, Rapid City, SD 57701, United States

a r t i c l e i n f o a b s t r a c t

Article history:Received 20 June 2014Received in revised form 3 November 2014Accepted 3 November 2014Available online 24 November 2014

Keywords:SiliconPolytetrafluoroethyleneMechanical activationHigh energy ball milling

Though silicon (Si) has been widely considered as a reactive fuel, the ability to tune its ignition and com-bustion characteristics remains challenging. One means to accomplish this may be mechanical activation(MA), which has been shown to be effective with aluminum fuels. In this work, a two-step MA processwas developed to prepare fuel-rich Si/polytetrafluoroethylene (PTFE) composite reactive powders. Theprocess consisted of cryogenic milling, followed by high intensity milling at room temperature. Thisresulted in particle refinement of the hard, brittle silicon particles and also dispersion within the moreductile PTFE matrix. Surface area of the as-milled powder was found to be moderate, ranging from1.75 ± 0.06 m2 g�1 (44/51 Si/PTFE) to 5.37 ± 0.10 m2 g�1 (90/10 Si/PTFE) and was driven by the fractionof refined Si powder not dispersed in a PTFE matrix. Combustion enthalpies ranged from 15.6 ± 0.4 kJ g�1

(44/51 Si/PTFE) to 21.8 ± 2.2 kJ g�1 (90/10 Si/PTFE) and were higher than a physical mixture of the pre-cursors. Combustion experiments showed that burning rates ranged from 1.6 to 2.1 mm s�1 and combus-tion temperatures (as measured from gray body emission) ranged from 1708 to 1889 K. The combustionperformance of MA Si/PTFE was comparable to mixtures prepared with nanoscale and nanoporous siliconpowders, indicating that the reactivity of silicon fuel had been successfully altered inexpensively withoutmany of the major drawbacks associated with using high specific surface area powders.

� 2014 The Combustion Institute. Published by Elsevier Inc. All rights reserved.

1. Introduction

Silicon (Si) has a long history as a reactive fuel in pyrotechnicapplications, such as delay and primer compositions, near infraredilluminants, and smoke formulations [1]. While there has histori-cally been significant efforts to study micron sized silicon powdersin energetic compositions [2–7], recent work has primarily focusedon the performance enhancement that nanoscale and nanoporoussilicon powders and wafers may provide [8–16]. Increasing thesurface area of reactive constituents by either particle size reduc-tion or by chemical processing may increase reactivity, lower igni-tion temperatures, and increase the intimacy of mixing [17–19].However, major drawbacks to using nanoscale and nanoporous sil-icon powders include: high cost; difficulty of synthesis (e.g., use ofstrong acids); processing issues due to the high specific surfacearea (SSA); and rapid oxidation (aging) of high SSA materials[17]. Recent studies have indicated that mechanical activation

(MA) provides an efficient, yet cost-effective, means for alteringreactivity of metal-based reactive composites, while still maintain-ing a micron scale morphology [18,20–25]. However, MA has notbeen fully explored for silicon-based reactive composites due tothe high hardness and low ductility of silicon (ultimate tensilestrength is ca. 7 GPa [26]).

Mechanical activation is the process of refining materials into areactive composite material [18] using low or high intensity mill-ing. While this process is similar to mechanical alloying, MA differsin that the process is interrupted prior to the constituents reacting.This process can yield micron-sized, nanostructured compositeparticles that are fully dense, resulting in altered reactivity dueto decreased diffusion length scales and increased interfacial con-tact area between constituents [18,21]. While MA often occurs atroom temperature, it has been reported by Umbrajkar et al. [27]that altering the milling temperature can have a substantial effecton material processing. By this same premise, milling at cryogenictemperatures (cryomilling) has been investigated as a means toproduce microstructural refinement [28–32], form nanopowders[33,34], and produce nanocomposite materials (both structural[35–37] and energetic [38,39]).

B.C. Terry et al. / Combustion and Flame 162 (2015) 1350–1357 1351

In the case of very hard, brittle materials, such as silicon, stan-dard high energy ball milling (HEBM) processes cannot readilyinduce plastic flow (yielding desired lamellar microstructures),but instead refines the brittle material through fracturing [40,41].When milling a mixed system with very different mechanicalproperties (i.e., a brittle material with a ductile material), the resultis generally a homogeneous dispersion of refined brittle materialwithin the ductile matrix [41]. However, in some cases the ductilematerial may act as a process control agent (PCA), hindering thefracturing of the brittle material. Milling at cryogenic temperaturesmay alleviate this problem by altering the relative ductility of thematerials [33]. Chung et al. [29] found that the inclusion of hardparticles during cryomilling can cause significant particle and grainsize reductions to both the brittle material and the ductile mate-rial. Therefore, a two-step process consisting of cryomilling fol-lowed by HEBM at ambient temperature may result in largemicron scaled composite particles with nanostructured intraparti-cle features comprised of refined constituents that are subse-quently cold-welded together by the more ductile matrix. Whilea two-step milling process (cryogenic and room temperature) hasbeen used to study the grain size evolution of single materials[34,42,43], it has not been investigated as a means to mechanicallyactivate reactive or non-reactive binary brittle/ductile compositesystems.

The objectives of this study were twofold: (i) to develop andcharacterize a two-step cryomilling and HEBM mechanical activa-tion process for preparation of a reactive, fuel-rich Si/polytetrafluo-roethylene (PTFE or Teflon™) composite material; and (ii) todetermine if mechanically activated Si/PTFE can achieve compara-ble reactivity to nanoscale Si and PTFE.

2. Experimental section

2.1. Mechanical activation and powder characterization

A summary of all mixtures evaluated in this study is presentedas Table 1. The materials used for mechanical activation of Si/PTFEincluded: Si (Sigma–Aldrich, �325 mesh, 99% trace metals) andPTFE (Sigma–Aldrich, 12 lm). The reactive mixtures preparedand studied were: 44/51 (46.3/53.7 wt.% Si/PTFE), 60/40, 70/30,80/20 and 90/10 wt.% Si/PTFE. One gram of reactive mixture wasplaced into a 25 mL polycarbonate tube (SPEX CertiPrep 6751C4vial) with hermetically sealed stainless steel end caps (SPEX Certi-Prep 6751E plug) and a single stainless steel impactor (SPEX Certi-Prep 6751P impactor). The container was then purged, filled withargon (99.997%), and placed into a cryomill (SPEX CertiPrep 6850Freezer/mill) using liquid nitrogen as the cryogenic medium. Afterpre-cooling the sample for 10 min, the material was milled for 6cycles (1 min ON and 1 min OFF). The milling container was thenremoved and allowed to warm up to room temperature beforethe reactive mixture was extracted in an argon-filled glove box.

Table 1A summary of all mixtures evaluated in this study.

Material Silicon(wt.%)

PTFE(wt.%)

FC-2175(wt.%)

MA 44/51 Si/PTFE 46.32 53.68 0MA 60/40 Si/PTFE 60 40 0MA 70/30 Si/PTFE 70 30 0MA 80/20 Si/PTFE 80 20 0MA 90/10 Si/PTFE 90 10 0MA 44/51 Si/PTFE (with Binder) 44 51 5SiTV (Ref. [11]) 44 51 544/5 nSi/FC-2175 89.80 0 10.2044/5 (Cryomilled Si)/FC-2175 89.80 0 10.20

Constituents were then placed into a 30 mL high density poly-ethylene (HDPE) container (Cole Palmer EW-62201-01) and milledfollowing the HEBM procedures outlined by Sippel et al. [20]. Spe-cifically, a charge ratio of 24 was used (5 large media, stainlesssteel, 9.5 mm diameter; and 15 small media, stainless steel,4.8 mm diameter). The container was then purged, filled withargon (99.997%), and placed into a high energy mill (SPEX,8000 M shaker/mill). The material was milled at room temperaturefor 20 cycles (1 min ON and 1 min OFF), during which the containerwas cooled using a fan. Following the HEBM procedure, materialwas extracted in an argon filled glove box and stored for furthercharacterization.

Powder characterization was performed on all MA Si/PTFE mix-tures using standard techniques. The particle size distribution ofthe milled materials was determined by laser diffraction (MalvernMastersizer Hydro 2000lP) using de-ionized water as the disper-sant medium. The specific surface area was determined byBrunauer, Emmett, and Teller (BET) analysis (Micromeritics Tristar3000) with nitrogen as the adsorbent gas. All BET samples weredegassed at 50 �C for 18 h. Imaging of the milled material was per-formed by optical microscopy (Hirox KH-8700), scanning electronmicroscopy (SEM), and energy dispersive spectroscopy (EDS) (FEIQuanta 3D-FEG). Phase composition of the milled materials wasaccomplished by X-ray diffraction (XRD) analysis (BrukerD8-Focus) using a scan rate of 2.0 deg. min�1.

Thermal and reaction characterization was also performed.Simple powder flame experiments were performed on a ventedburn plate using a butane torch ignition source in order to makequalitative observations of the reaction. The thermal behaviorwas determined by simultaneous differential scanning calorimetry(DSC) and thermo-gravimetric analysis (TGA) (TA InstrumentsQ600) at 10 K min�1 in ultra-high purity argon (99.999%). The heatof combustion of the milled material was determined by oxygencalorimetry (Parr 1281) using a custom-made alumina-silicate cru-cible [20]. All heats of combustion were compared to theoreticalvalues as calculated by the Cheetah (version 6.0) thermal equilib-rium code [44].

2.2. SiTV combustion experiments

Two fuel-rich reactive materials were prepared and pressed intopellets: 44/51 weight ratio Si/PTFE (as prepared from MA) and 44/51/5 wt.% Si/PTFE/FC-2175 (SiTV mixture [11]). The SiTV mixturewas comprised of 44/51 Si/PTFE (as prepared from MA) and FC-2175 (3 M Fluorel™ Fluoroelastomer [11,16], chemical equivalentto Viton A). SiTV mixtures were prepared and pressed into6.35 mm diameter pellets according to the SiTV procedures out-lined by Terry et al. [11]. Pellets were either pressed to a stop at78% theoretical maximum density (TMD) or pressed without a stopto the maximum attainable percent TMD. Vacuum grease (DowCorning

�High Vacuum Grease) was used as an inhibitor to avoid

flame spreading down the sides of the pellet.Spectroscopy and burning rate experiments were performed

using spectrometers and videography following the setup and pro-cedures outlined by Terry et al. [11]. Both a visual (VIS) spectrom-eter (Spectraline VS100 Imaging Spectrometer, 0.55 lm to1.02 lm, 0.5 mm � 5.0 mm slit, 12 kHz sampling rate) and aninfrared (IR) spectrometer (Spectraline ES100 Imaging Spectrome-ter, 1.11 lm to 4.75 lm, 0.5 mm � 6.0 mm slit, 1.32 kHz samplingrate) were used to investigate spectral emissions. All spectrome-ters were aligned and calibrated in accordance with Ref. [11].Videography (Cannon XL2 3CCD Digital Video Camcorder at30 fps, Cannon EF 100 mm f/2.8 Macro USM Lens) was used toanalyze pellet deflagration and measure burning rates.

Assumed gray body emission from the SiTV combustionexperiments were analyzed via coefficient of determination (R2)

1352 B.C. Terry et al. / Combustion and Flame 162 (2015) 1350–1357

optimization with Planck’s Law [11], which converted spectralintensity traces into time dependent temperature and effectiveemissivity profiles. A minimum R2 value (between measured spec-tral intensity and fit gray body profile) of 0.95 was selected in orderto avoid unacceptable uncertainty in the temperature and emissiv-ity measurements.

3. Results and discussion

3.1. Characterization of powder evolution and morphology

Initially, a single step mechanical activation of the Si/PTFE sys-tem was considered following the milling procedure outlined bySippel et al. [20]. This room temperature HEBM process resultedin a composite material without observed altered reactivity (i.e.,would not ignite via butane flame). It was hypothesized that thehigh disparity in relative ductility between silicon and PTFE causedPTFE to act as a PCA, hindering silicon refinement. A single stepcryomilling process was then attempted, which was found togreatly refine the Si and PTFE constituents into a fine powder,though again significant altered reactivity was not observed. How-ever, subsequent HEBM of the cryomilled mixture at room temper-ature yielded large, composite particles that did exhibit significantaltered reactivity. Specifically, the materials could be easily ignitedvia butane flame. While this two-step MA procedure was able toincrease the relative reactivity of the Si/PTFE precursors, furtherefforts are needed to optimize the milling conditions for each stage(e.g., milling time, charge ratio, etc.) as well as to elucidate theeffect of those milling conditions on the thermal behavior, ignitioncriteria, and reaction kinetics.

A proposed mechanism showing the morphology transforma-tion from a two-step cryomilling/HEBM of the Si/PTFE system isillustrated in Fig. 1. The reduced PTFE ductility at cryogenic tem-peratures should inhibit substantial lamellar formation, insteadpromoting grain and particle size refinement of both the ductileand brittle materials [29]. Once re-heated to room temperature,the HEBM process will plastically deform the ductile PTFE and pro-mote silicon particle encasement within the ductile matrix. Thisprocess results in large, nearly fully-dense composite particles.

Particle sizing of neat, cryomilled Si has shown that as millingtime increases (from 0 to 6 min) the frequency of 0.5–7 lm Si par-ticles steadily increases, which has been verified by microscopy tobe representative of the refined Si particle size in post-cryomilledSi/PTFE. Once the refined Si is HEBM into a PTFE matrix, particlesizing (Table 2) shows that the mean composite particles sizes(arithmetic, volume weighted, and surface weighted) all decreasewith decreasing PTFE content. Particle size distributions indicatedthat all MA Si/PTFE mixture ratios had a dominant peak at roughly80 lm. At higher PTFE concentrations (e.g., 44/51 Si/PTFE), a sec-ondary peak at roughly 480 lm becomes apparent, which appearsto be caused by the merging of�80 lm Si/PTFE composite particles(see Fig. 3). Excess PTFE allows for loose cold-welding to occur

Fig. 1. Proposed mechanism for the two-step cryomilling/HEBM process of the Si/PTFE system.

between particles, forming the larger �480 lm particles. PTFE leanmixtures (i.e., 70/30 and 90/10 Si/PTFE) do not contain enoughPTFE to encase all of the Si particles, leaving a loose Si fraction(0.5–7 lm) that is not incorporated into the PTFE matrix. Thisnon-encased fraction is also in agreement with BET SSA measure-ments (Table 2), which indicated that SSA increases with decreas-ing PTFE content.

The reported particle size distributions were supported by opti-cal and SEM microscopy of the as-milled powders, which indicateda bimodal distribution of the 44/51 Si/PTFE composite powder andloose, non-encased Si particles in the 70/30 and 90/10 Si/PTFEpowders. Both large particle flakes (on the order of 500 lm) andsmaller particles (on the order of 100 lm) were observed for the44/51 Si/PTFE material. The particle surfaces were rough andappeared to consist of consolidated PTFE fibers and approximately5 lm silicon particles. It was also observed that the edges of thecomposite particles had thin PTFE fibers decorated with refined sil-icon particles (Fig. 2).

These microstructures indicate that during the HEBM of thepost-cryomilled Si/PTFE powder, the PTFE is plastically deformedinto lamellar microstructures, which are then decorated by loose,refined Si particles. Those decorated lamellar microstructures aresubsequently rolled, cold-welded, and consolidated with otherPTFE microstructures, forming larger Si/PTFE composite particles(Fig. 3). While neat PTFE powder does initially have some lamellarmicrostructure, the intimate dispersion of refined Si decorating theMA PTFE lamellar microstructures in Fig. 2 suggests that they are aresult of the MA process.

The composite particles eventually reach a point where the highductility of the PTFE matrix promotes particle splitting and refine-ment of the larger composite particles. If sufficient PTFE is avail-able (e.g., 44/51 Si/PTFE), merging of composite particles mayoccur. This interfacial bonding between merged particles appearsto be weak, as evident by the distinct interfacial boundariesdepicted in Fig. 3 (dotted lines). Figure 3 also shows the Si particlerefinement and dispersion within the PTFE matrix, and the com-posite particles appear to have a high compact density (i.e., lowparticle porosity).

From XRD of the as-milled Si/PTFE composite particles, it wasobserved that only crystalline Si and PTFE are detected for any ofthe mixture ratios analyzed; no crystalline oxides, intermediates,or contaminants were observed. It was determined from EDS anal-ysis that a sectioned Si/PTFE composite particle contains 2.02 wt.%oxygen and 0.39 wt.% Fe, suggesting that low level contaminants/oxides are either below the XRD detection threshold or amorphous.These levels are consistent with EDS analysis of neat Si (2.23 wt.%oxygen and 0.25 wt.% Fe) indicating that they are inherent to the Siprecursor and were not introduced through the two-step MA pro-cess. The PTFE matrix in Si/PTFE composites may serve as a protec-tive layer during aging, inhibiting further oxidation of the siliconparticles beyond these initial values, but has not been assessed.

3.2. Thermal characterization

Oxygen bomb calorimetry was used to assess the relative com-bustion enthalpies of the as-milled Si/PTFE composite material incomparison to calculated theoretical values (see Table 3). All mate-rials tested had combustion enthalpies below calculated values(average reduction of 18.4 ± 2.5%), consistent with Wise et al.[45] which reported that oxygen bomb calorimetry of Si generallydoes not result in complete combustion. Neat Si powder yielded acombustion enthalpy of 24.1 ± 4.1 kJ g�1, which was a reduction of25.6%. The minimum combustion enthalpy deficit was 44/51Si/PTFE, which had a reduction of only 16.1% (15.6 ± 0.4 kJ g�1). Itwas also observed that 90/10 Si/PTFE produced the highest com-bustion enthalpy of the MA mixtures (21.8 ± 2.2 kJ g�1).

Table 2The mean particle sizes and BET specific surface areas for the 44/51, 70/30, and 90/10Si/PTFE composite materials.

Si/PTFE D50 (lm) D4,3 (lm) D3,2 (lm) BET specificMixture Arithmetic Volume

weightedSurfaceweighted

Surface area(m2 g�1)

44/51 78.2 142.4 31.7 1.75 ± 0.0670/30 73.2 92.2 25.6 3.83 ± 0.0390/10 48.2 84.7 13.0 5.37 ± 0.10

Fig. 3. Backscattered SEM image of a large sectioned and polished 44/51 Si/PTFEcomposite particle. Light gray corresponds to silicon, dark gray corresponds to PTFE.The dotted white lines indicate interfacial contact between merged particles.

Fig. 2. SEM image showing the plastic flow (lamellar microstructures) of the PTFEmatrix decorated with refined Si particles.

Table 3Calculated and measured oxygen bomb calorimetry of MA Si/PTFE.

Si/PTFE ratio Calculated (kJ g�1) Measured (kJ g�1)

Neat Si 32.4 24.1 ± 4.144/51 18.6 15.6 ± 0.460/40 22.2 18.5 ± 0.570/30 24.7 19.7 ± 1.180/20 27.3 21.6 ± 2.790/10 29.8 21.8 ± 2.2

Fig. 4. Calculated and measured oxygen bomb calorimetry of 44/51 Si/PTFEmixtures.

B.C. Terry et al. / Combustion and Flame 162 (2015) 1350–1357 1353

Though less than calculated values, the MA Si/PTFE compositesdid have higher combustion enthalpies than both physical mix-tures of the micron-sized precursors as well as physical mixturesusing nanoscale Si (CS#1, salt assisted combustion synthesized sil-icon powder [11]). Figure 4 shows this relative comparison for a44/51 Si/PTFE mixture ratio. The higher combustion enthalpy forMA Si/PTFE over the physical mixture indicates that the two-stepmechanical activation process improves the Si/PTFE combustionperformance of the system.

The thermal behavior of all neat materials used in this study isshown in Fig. 5 and is to be used as a baseline comparison for allcomposites and mixtures discussed below. As shown from DSC/TGA of MA Si/PTFE (Fig. 6), a PTFE melt endotherm is presentaround 327 �C for all mixtures. Decomposition of PTFE begins atroughly 470 �C and is complete around 600 �C, which is evidentby the onset and end of sample weight loss respectively. Theweight loss observed during PTFE decomposition is in agreementwith the PTFE content of the reactive mixtures: 54.2% (44/51),31.0% (70/30), and 11.5% (90/10). With 44/51 Si/PTFE powder, PTFEdecomposition is accompanied by a large endotherm, onset atroughly 530 �C and peaked at 571 �C. An overlapped endotherm/exotherm may exist within this temperature range, but becauseof the strong endothermic response due to PTFE decomposition(see Fig. 5) an exotherm is not apparent. In contrast, PTFE decom-position is accompanied by exothermic Si/PTFE reactions for the70/30 and 90/10 Si/PTFE mixtures, which are actually further fromstoichiometric conditions (22/78 Si/PTFE). Specifically, the exo-thermic fluorination of Si onset at 525 �C and had two peaks(543 and 580 �C) for 70/30 Si/PTFE and onset at 540 �C (peak at559 �C) for 90/10 Si/PTFE. Consistent with the work by Ksia _zczaket al. [46], it is clear that the Si/PTFE reaction is driven by PTFEdecomposition and that any exothermic reactions occur withinthe temperature range for which PTFE mass loss occurs. Furtherefforts are necessary to elucidate the detailed species interactions,which could be accomplished via DSC/TGA coupled with gas chro-matography and mass spectrometry.

The thermal analysis of silicon with a fluoroelastomer binder(FC-2175, SiTV composition) is also presented. Figure 7 showsDSC/TGA traces for both (cryomilled Si)/FC-2175 and nSi/FC-2175at the SiTV weight ratio of 44/5 Si/FC-2175. Decomposition of FC-2175 begins at roughly 430 �C and finishes around 500 �C, whichis evident by the onset and end of sample weight loss respectively(see Fig. 5). Decomposition of FC-2175 is also accompanied by anendotherm that onsets at 430 �C and peaks at 466 �C. This is con-gruent with the work by Burnham et al. [47] and Cuccuru et al.[48], which reported that Viton A (FC-2175 is an equivalent) begins

Fig. 5. DSC/TGA traces of the neat materials used in this study at 10 K min�1 inargon. Heat flows and sample weight percentages are offset for presentation.

Fig. 6. DSC/TGA traces of the as-milled Si/PTFE material at 10 K min�1 in argon.Heat flows and sample weight percentages are offset for presentation.

Fig. 7. DSC/TGA traces at 10 K min�1 in argon for nSi/FC-2175 and (cryomilled Si)/FC-2175 at the SiTV weight ratio of 44/5 Si/FC-2175. Heat flows and sample weightpercentages are offset for presentation.

1354 B.C. Terry et al. / Combustion and Flame 162 (2015) 1350–1357

to volatilize at 466 �C and 485 �C respectively when heated at10 �C min�1. The exothermic fluorination of Si onsets at 440 �Cand peaks at 476 �C for nSi/FC-2175 and onsets at 470 �C and peaksat 498 �C for cryomilled Si/FC-2175. Again, further efforts are nec-essary to elucidate the detailed Si/FC-2175 species interactions.

3.3. SiTV combustion experiments

The combined cryomilling/HEBM process was found to yieldmaterial that had similar reactivity to mixtures made with nano-scale constituents. Flame experiments of the as-milled powdersshowed that the cryomilled/HEBM 44/51, 60/40, and 70/30 Si/PTFEmixtures were easily ignited via butane igniter, which is notobserved either with a physical mixture of the initial precursors(micron scale) or from single step ambient milling. Observed reac-tivity was similar to nanoscale 44/51 Si/PTFE mixtures. Simpleflame experiments of the 80/20 and 90/10 Si/PTFE mixtures werecharacterized by a slowly propagating and quenching reaction

front, accompanied by PTFE decomposition gasses (no visible flamewas apparent).

The ductile PTFE matrix within MA Si/PTFE permits the com-posite particles to freely deform during compaction, producing ahigh compact density without requiring an additional binder.The following maximum attainable %TMD values were measuredfor MA 44/51 SiTV: 98.6 ± 0.2% (with FC-2175 binder) and98.5 ± 0.3% (without binder). In comparison to equivalent nano-scale Si/PTFE mixtures, the following maximum attainable%TMD values were measured when subject to the same pressingconditions (with binder): 87.8% (Vesta Ceramics PSi_E3, 5 nmpore diameter), 93.7% (US Research Nanomaterials, Inc.,nominally 30–50 nm), and 94.2% (CS#1, salt assisted combustionsynthesized silicon powder [11]).

Burning rates for all compacts (both 78% TMD and maximumattainable %TMD) yielded comparable results (Fig. 8), with burningrates ranging from 1.62 ± 0.15 mm s�1 (78%, no binder) to2.07 ± 0.50 mm s�1 (max TMD, with binder). While all deflagrationrates were of statistical variance of each other, two observationscan be inferred from these results. First, the variance and averagemagnitude of the burning rates are less without FC-2175 binder.Binders are generally assumed to increase repeatability (betterinterfacial contact and reduced void space). Second, the averageburning rates and variance do not appear to change with TMD.Increased pellet density often leads to lower burning rates andreduced variance [17]. If the mass burning rate of these particleswere assumed to remain constant for both densities, then theburning rate of the maximum attainable percent TMD pellets shouldhave decreased by roughly 21%. Instead, the mass burningrates of the pellets increased by 47% for the mixtures withbinder (0.308 ± 0.077 g cm�2-s�1 ? 0.454 ± 0.110 g cm�2-s�1) andincreased by 38% for the mixtures without binder (0.285 ±0.026 g cm�2-s�1 ? 0.392 ± 0.074 g cm�2-s�1). Increase mass burn-ing rate suggests that at higher densities, there may be increasedcontact area between particles, promoting an increased thermal dif-fusivity that leads to a somewhat higher average burning rate.

The burning rate of MA SiTV (1.77 ± 0.45 mm s�1, 78% TMD withbinder) was comparable to most nanoporous SiTV mixturesreported in the literature and statistically comparable to othernanoscale SiTV compositions (see Fig. 9). Terry et al. [11] have indi-cated that the silicon core particle size (D4,3) is the driving parame-ter in SiTV burning rate. As discussed previously, neat siliconcryomilled for 6 min is representative of the refined Si dispersedwithin MA Si/PTFE composite particles. Particle sizing of neat siliconcryomilled for 6 min indicates a Si core particle size of 5.9 lm. Using

Fig. 8. Burning Rates of MA 44/51 SiTV. All data points represent 5 measurementsand error bars represent one standard deviation from the mean. Fig. 10. Combustion temperatures of MA 44/51 SiTV. All data points represent 5

measurements and error bars represent one standard deviation from the mean.

Fig. 11. Combustion temperature of 44/51 SiTV (78%, with binder) in comparison toequivalent nanoscale and nanoporous silicon based SiTV mixtures [11]. All datapoints represent 5 measurements and error bars represent one standard deviationfrom the mean.

B.C. Terry et al. / Combustion and Flame 162 (2015) 1350–1357 1355

this measured core particle size, MA SiTV appears to follow the sameburning rate trend as the nanoscale SiTV reactive mixtures.

Thermal equilibrium calculations (Cheetah 6.0 [44]) predict that44/51 SiTV has theoretical combustion temperatures of 2837 K(with binder) and 2867 K (without binder). Measured averagecombustion temperatures, as determined by assumed gray bodyemission, were all within statistical variance of each other, rangingfrom 1708 ± 80 K (max TMD, no binder) to 1889 ± 91 K (78%, withbinder) (see Fig. 10). However, there was an average drop of 126 Kin combustion temperature for compositions without binder, sug-gesting that the fluoroelastomer binder may play a significant rolein SiTV combustion.

Terry et al. [11] showed that the active content in nanoscale Siwas the driving powder characteristic for measured SiTV combus-tion temperature (see Fig. 11). Using 98.8 ± 0.2% as the active Sicontent (measured by EDS analysis of neat silicon, cryomilled6 min), the measured combustion temperature of MA SiTV(1889 ± 91 K, 78% TMD with binder) was significantly lower thannanoscale SiTV mixtures with similar active Si content (>90%).

One variation between nanoscale SiTV and MA SiTV is the inti-macy in which the fluoroelastomer binder is incorporated into theformulation. Nanoscale SiTV mixtures had an intimate dispersionof the FC-2175 binder throughout the mixture, as it was evenlydispersed among the nanoscale Si and PTFE particles during thewetted mixing process. In contrast, with MA Si/PTFE the fluoroelas-tomer binder essentially thinly coated the as-milled Si/PTFE com-posite particles since the binder was introduced following the

Fig. 9. Burning rate of 44/51 SiTV (78%, with binder) in comparison to equivalentnanoscale and nanoporous silicon based SiTV mixtures [11]. All data pointsrepresent 5 measurements and error bars represent one standard deviation fromthe mean.

two-step milling process. Many investigations [49–52] into thecombustion behavior of various metal/PTFE systems have indi-cated that the incorporation of a fluoroelastomer binder does nothave a considerable effect on the combustion characteristics. How-ever, the present work suggests that an intimate incorporation offluoroelastomer binder within the SiTV mixture (intimate Si/FC-2175 contact) may increase combustion efficiency due to the highreactivity between Si and FC-2175 decomposition products.

To explore this, thermal analysis of a fuel-rich Si mixture withthe fluoroelastomer binder instead of with PTFE was conducted(see Fig. 7). Because decomposition of FC-2175 occurs at lowertemperatures than PTFE, the reactive compounds from FC-2175decomposition (HF, C2F2, and CHF2 [53]) would then be availableto react with Si before the PTFE decomposition products (:CF2

and CF4 [50,52]). This results in the fluorination of Si from FC-2175 occurring at a lower temperature than with PTFE. Therefore,a more intimate dispersion of FC-2175 binder within MA Si/PTFEcomposite particles may further increase MA SiTV combustionefficiency.

4. Conclusion

A two-step mechanical activation process was developed toproduce fuel-rich Si/PTFE reactive materials. This process involvedcryomilling the constituents at liquid nitrogen temperatures tochange the relative ductility of Si and PTFE, allowing for fracturingand refinement of the brittle silicon particles. High energy milling

1356 B.C. Terry et al. / Combustion and Flame 162 (2015) 1350–1357

of the material at room temperature then consolidated and cold-welded the refined silicon particles into a ductile PTFE matrix. Itwas shown that this process produces large (roughly 80 lm),high-purity Si/PTFE composite particles that are able to freelydeform and be pressed to near full density (98.5% TMD) withoutadditional binder. XRD analysis did not show the presence of crys-talline oxides, intermediates, or contaminants within the Si/PTFEcomposites. These results indicate that a two-step mechanical acti-vation process using both cryomilling and high energy milling maybe a suitable process for the MA of brittle/ductile systems with ahigh relative disparity in constituent ductility and should be con-sidered in other brittle/ductile systems, particularly for reactives.

Oxygen bomb calorimetry revealed that the measured combus-tion enthalpies ranged from 15.6 ± 0.4 kJ g�1 (44/51 Si/PTFE, 16.1%less than theoretical) to 21.8 ± 2.2 kJ g�1 (90/10 Si/PTFE, 27.1% lessthan theoretical). However, MA 44/51 Si/PTFE produced a combus-tion enthalpy that was 9.5% greater than a physical mixture of Siand PTFE, indicating that the two-step MA Si/PTFE composites havea higher combustion performance than a physical mixture of theirprecursors.

It was observed from combustion experiments that averageburning rates ranged from 1.6 to 2.1 mm s�1 and that average com-bustion temperatures ranged from 1708 to 1889 K. It was alsoshown that the fluoroelastomer binder may play a more significantrole in SiTV combustion than anticipated and should be furtherinvestigated in future work. All combustion results appeared tofollow the same trends and were comparable in magnitude toequivalent nanoscale and nanoporous SiTV mixtures.

MA Si/PTFE was shown to be a cost effective approach for pro-ducing high performance fuel-rich silicon powder without themajor drawbacks associated with nanoscale powders. Fuel-richMA Si/PTFE may be advantageous as a silicon fuel replacement inexisting Si/oxidizer formulations.

Acknowledgments

This research was conducted with Government support underand awarded by DoD, Air Force Office of Scientific Research,National Defense Science and Engineering Graduate (NDSEG)Fellowship, 32 CFR 168a. The authors would also like to acknowl-edge funding provided by AFOSR MURI Contract (#FA9550-13-1-0004) with Mitat Birkan as Program Manager. The authors wouldalso like to acknowledge Dr. Travis Sippel for technical supportwith mechanical activation and graduate student Matthew Beasonfor SEM/EDS analysis.

Appendix A. Supplementary material

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.combustflame.2014.11.005.

References

[1] E.-C. Koch, D. Clément, Propellants, Explos., Pyrotech. 32 (3) (2007) 205–212,http://dx.doi.org/10.1002/prep.200700021.

[2] H. Ellern, Military and Civilian Pyrotechnics, Chemical Pub. Co., New York,1968.

[3] R.A. Rugunanan, M.E. Brown, J. Therm. Anal. Calorim. 37 (6) (1991) 1193–1211,http://dx.doi.org/10.1007/BF01913854.

[4] R.A. Rugunanan, M.E. Brown, Combust. Sci. Technol. 95 (1–6) (1994) 85–99.[5] R.A. Rugunanan, M.E. Brown, Combust. Sci. Technol. 95 (1–6) (1994) 101–115.[6] R.A. Rugunanan, M.E. Brown, Combust. Sci. Technol. 95 (1–6) (1994) 61–83.[7] R.A. Rugunanan, M.E. Brown, Combust. Sci. Technol. 95 (1–6) (1994) 117–138.[8] B.A. Mason, L.J. Groven, R.A. Yetter, S.F. Son, J. Propul. Power 29 (6) (2013)

1435–1444.

[9] D. Clément, J. Diener, E. Gross, N. Künzner, V.Y. Timoshenko, D. Kovalev, Phys.Status Solidi 202 (8) (2005) 1357–1364, http://dx.doi.org/10.1002/pssa.200461102.

[10] S. Subramanian, T. Tiegs, S. Limaye, D. Kapoor, P. Redner, in: The 26th ArmyScience Conference: Transformational Army Science and Technology (S&T) –Harnessing Disruptive S&T for the Soldier, Orlando, FL, 2008.

[11] B.C. Terry, Y.-C. Lin, K.V. Manukyan, A.S. Mukasyan, S.F. Son, L.J. Groven,Propellants, Explos., Pyrotech. 39 (3) (2014) 337–347, http://dx.doi.org/10.1002/prep.201300058.

[12] W. Churaman, L. Currano, C. Becker, J. Phys. Chem. Solids 71 (2) (2010) 69–74,http://dx.doi.org/10.1016/j.jpcs.2009.07.022.

[13] D. Kovalev, V. Timoshenko, N. Künzner, E. Gross, F. Koch, Phys. Rev. Lett. 87 (6)(2001) 068301, http://dx.doi.org/10.1103/PhysRevLett.87.068301.

[14] B.A. Mason, S.F. Son, K.Y. Cho, C.D. Yarrington, J. Gesner, R.A. Yetter, B.W. Asay,in: 6th US National Combustion Meeting, Ann Arbor, MI, 2009.

[15] R. Thiruvengadathan, G.M. Belarde, A. Bezmelnitsyn, M. Shub, W. Balas-Hummers, K. Gangopadhyay, S. Gangopadhyay, Propellants, Explos., Pyrotech.37 (3) (2012) 359–372, http://dx.doi.org/10.1002/prep.201100129.

[16] C.D. Yarrington, S.F. Son, T.J. Foley, J. Propul. Power 26 (4) (2010) 734–743,http://dx.doi.org/10.2514/1.46182.

[17] R.A. Yetter, G.A. Risha, S.F. Son, Proc. Combust. Inst. 32 (2) (2009) 1819–1838,http://dx.doi.org/10.1016/j.proci.2008.08.013.

[18] E.L. Dreizin, Prog. Energy Combust. 35 (2) (2009) 141–167, http://dx.doi.org/10.1016/j.pecs.2008.09.001.

[19] D.G. Wei, R. Dave, R. Pfeffer, J. Nanopart. Res. 4 (1–2) (2002) 21–41.[20] T.R. Sippel, S.F. Son, L.J. Groven, Propellants, Explos., Pyrotech. 38 (2) (2013)

286–295, http://dx.doi.org/10.1002/prep.201200102.[21] M. Schoenitz, T.S. Ward, E.L. Dreizin, Proc. Combust. Inst. 30 (2) (2005) 2071–

2078, http://dx.doi.org/10.1016/j.proci.2004.08.134.[22] M. Schoenitz, T.S. Ward, E.L. Dreizin, Mat. Res. Soc. Symp. Proc. 800 (2004) 1–6.[23] T.R. Sippel, S.F. Son, L.J. Groven, Propellants, Explos., Pyrotech. 38 (3) (2013)

321–326, http://dx.doi.org/10.1002/prep.201200202.[24] K.V. Manukyan, B.A. Mason, L.J. Groven, Y.-C. Lin, M. Cherukara, S.F. Son, A.

Strachan, A.S. Mukasyan, J. Phys. Chem. C 116 (39) (2012) 21027–21038,http://dx.doi.org/10.1021/jp303407e.

[25] C. Gras, D. Vrel, E. Gaffet, F. Bernard, J. Alloys Compd. 314 (1–2) (2001) 240–250, http://dx.doi.org/10.1016/S0925-8388(00)01221-4.

[26] R. Hull, Properties of Crystalline Silicon, INSPEC, London, UK, 1999.[27] S.M. Umbrajkar, M. Schoenitz, S.R. Jones, E.L. Dreizin, J. Alloys Compd. 402 (1–

2) (2005) 70–77, http://dx.doi.org/10.1016/j.jallcom.2005.03.088.[28] D.B. Witkin, E.J. Lavernia, Prog. Mater. Sci. 51 (1) (2006) 1–60, http://

dx.doi.org/10.1016/j.pmatsci.2005.04.004.[29] K.H. Chung, J. He, D.H. Shin, J.M. Schoenung, Mater. Sci. Eng. A 356 (1–2) (2003)

23–31, http://dx.doi.org/10.1016/S0921-5093(02)00833-X.[30] F. Zhou, S.R. Nutt, C.C. Bampton, E.J. Lavernia, Metall. Mat. Trans., A 34 (9)

(2003) 1985–1992, http://dx.doi.org/10.1007/s11661-003-0163-4.[31] F. Zhou, D. Witkin, S.R. Nutt, E.J. Lavernia, Mater. Sci. Eng. A 375–377 (2004)

917–921, http://dx.doi.org/10.1016/j.msea.2003.10.235.[32] X. Zhang, H. Wang, J. Narayan, C.C. Koch, Acta Mater. 49 (8) (2001) 1319–1326,

http://dx.doi.org/10.1016/S1359-6454(01)00051-9.[33] D.M. King, Y. Zhou, L.F. Hakim, X. Liang, P. Li, A.W. Weimer, Ind. Eng. Chem.

Res. 48 (1) (2008) 352–360, http://dx.doi.org/10.1021/ie800196h.[34] C. Tiwary, A. Verma, S. Kashyp, K. Biswas, K. Chattopadhyay, Metall. Mat.

Trans., A 44 (4) (2013) 1917–1924, http://dx.doi.org/10.1007/s11661-012-1508-7.

[35] J.-H. Lee, S.J. Hwang, Rev. Adv. Mater. Sci. 18 (3) (2008) 289–292.[36] M.J. Luton, C.S. Jayanth, M.M. Disko, S. Matras, J. Vallone, in: L.E. McCandlsih,

D.E. Polk, R.W. Siegel, B.H. Kear (Eds.) MRS Fall Meeting, Pittsburgh, PA, vol.132, 1988.

[37] B.D. Sneed, Synthesis and Characterization of Al-Nanodiamond CompositePowders by High-energy Ball Milling. Naval Postgraduate School, 2011.

[38] S. Zhang, M. Schoenitz, E. Dreizin, in: 50th AIAA Aerospace Sciences Meetingincluding the New Horizons Forum and Aerospace Exposition, AmericanInstitute of Aeronautics and Astronautics, 2012.

[39] C. Badiola, M. Schoenitz, X. Zhu, E.L. Dreizin, J. Alloys Compd. 488 (1) (2009)386–391, http://dx.doi.org/10.1016/j.jallcom.2009.08.146.

[40] U. Anselmi-Tamburini, F. Maglia, S. Doppiu, M. Monagheddu, G. Cocco, Z.A.Munir, J. Mater. Res. 19 (05) (2011) 1558–1566, http://dx.doi.org/10.1557/jmr.2004.0209.

[41] R.M. Davis, B. Mcdermott, C.C. Koch, Metall. Trans. A 19 (12) (1988) 2867–2874.

[42] C.S. Tiwary, A. Verma, K. Biswas, A.K. Mondal, K. Chattopadhyay, Ceram. Int. 37(8) (2011) 3677–3686, http://dx.doi.org/10.1016/j.ceramint.2011.06.029.

[43] A. Verma, K. Biswas, C. Tiwary, A. Mondal, K. Chattopadhyay, Metall. Mat.Trans. A 42 (4) (2011) 1127–1137, http://dx.doi.org/10.1007/s11661-010-0490-1.

[44] S. Bastea, L.E. Fried, K.R. Glaesemann, W.M. Howard, I.-F. Kuo, P.C. Souers, in:6.0 ed.; Lawrence Livermore National Laboratory, 2010; vol. LLNL-SM-416166.

[45] S.S. Wise, J.L. Margrave, H.M. Feder, W.N. Hubbard, J. Phys. Chem. 67 (4) (1963)815–821, http://dx.doi.org/10.1021/j100798a024.

[46] A. Ksia _zczak, H. Boniuk, S. Cudziło, J. Therm. Anal. Calorim. 74 (2) (2003) 569–574, http://dx.doi.org/10.1023/B:JTAN.0000005195.46390.48.

[47] A.K. Burnham, R.K. Weese, Thermochim. Acta 426 (1–2) (2005) 85–92, http://dx.doi.org/10.1016/j.tca.2004.07.009.

B.C. Terry et al. / Combustion and Flame 162 (2015) 1350–1357 1357

[48] A. Cuccuru, P. Banditelli, S. Manfredini, L. Argiero, La Chimica e L’Indus 55(1973).

[49] N. Kubota, C. Serizawa, J. Propul. Power 3 (4) (1987) 303–307.[50] E.C. Koch, Propellants, Explos., Pyrotech. 27 (6) (2002) 340–351.[51] G. Boškovic, D. Negoicic, Sci.-Tech. Rev. 59 (1) (2009) 70–77.

[52] F.C. Christo, in: Aeronautical and Maritime Research Laboratory: Victoria,Australia, 1999.

[53] D.M. Chen, W.H. Hsieh, T.S. Snyder, V. Yang, T.A. Litzinger, K.K. Kuo, J. Propul.Power 7 (2) (1991) 250–257.