Deflagration-to-detonation transition via the distributed photo ignition of carbon nanotubes suspended in fuel/oxidizer mixtures

Combustion and Flame 159 (2012) 1314–1320

Contents lists available at SciVerse ScienceDirect

Combustion and Flame

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

Deflagration-to-detonation transition via the distributed photo ignitionof carbon nanotubes suspended in fuel/oxidizer mixtures

Daniel J. Finigan, Brian D. Dohm, Jeffrey A. Mockelman, Matthew A. Oehlschlaeger ⇑Department of Mechanical, Aerospace, and Nuclear Engineering, Rensselaer Polytechnic Institute, Troy, NY, USA

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

Article history:Received 9 June 2011Received in revised form 18 August 2011Accepted 27 September 2011Available online 20 October 2011

Keywords:Deflagration-to-detonation transitionDetonation tubeIgnitionPhoto ignitionNanoparticles

0010-2180/$ - see front matter � 2011 The Combustdoi:10.1016/j.combustflame.2011.09.017

⇑ Corresponding author. Address: 110 8th St., JEC 2E-mail address: [email protected] (M.A. Oehlschlaeg

Here the promotion of flame acceleration and deflagration-to-detonation transition (DDT) using the dis-tributed photo ignition of photo-sensitive nanomaterials suspended in fuel/oxidizer mixtures is demon-strated for the first time. Distributed photo ignition was carried out by suspending single-walled carbonnanotubes (SWCNTs) with Fe impurity in quiescent C2H4/O2/N2 mixtures and flashing them with anordinary Xe camera flash. Following the flash, the distributed SWCNTs photo ignite and subsequentlyprovide a quasi-distributed ignition of the C2H4/O2/N2 mixture. In a closed detonation tube the quasi-dis-tributed photo ignition at one end of the tube leads to the promotion of flame acceleration and DDT and,for sensitive C2H4/O2 mixtures, appears to lead to direct detonation initiation or multiple combustionfronts. The DDT run-up distance, the distance required for the transition to detonation, was measuredusing ionization sensors and was found to be approximately a factor of 1.5� to 2� shorter for the distrib-uted photo ignition process than for traditional single-point spark ignition. It is hypothesized that theincreased volumetric energy release rate resulting from distributed photo-ignition enhances DDT dueto the decreased ignition delay and greater early-time flame area and turbulence levels, which in turnresult in accelerated formation and amplification of the leading shock and accelerated DDT.

� 2011 The Combustion Institute. Published by Elsevier Inc. All rights reserved.

1. Introduction

Flame acceleration and deflagration-to-detonation transition(DDT) have been the subject of numerous studies, see pertinentreviews [1–4], motivated by the extreme danger undesired detona-tion can pose to industrial processes involving combustible gasesand the potential for utilizing the rapid energy release and high over-pressures resulting from detonation for high-efficiency high-speedpropulsion cycles [5–7] (e.g., pulse detonation engines – PDEs).Due to the unsteady nature of propulsion cycles reliant on detona-tion, the time and length scales associated with the ignition andformation of a detonation wave in confined geometries is criticalto the performance of detonation engines [8].

A detonation wave can be directly initiated using a high-energysource or by flame acceleration resulting in deflagration-to-detona-tion transition (DDT), where the flame is initiated using a traditionallow-energy ignition source (e.g., spark). Because of the very high en-ergy requirements for the direct initiation of a detonation in gaseousfuel/air mixtures (order of kilojoules) [9], DDT is the most practicalmeans by which to generate a detonation in a propulsion engine.Following a localized deposition of energy (e.g., spark), DDT occursthrough several flame acceleration steps. First a laminar flame forms

ion Institute. Published by Elsevier

049, Troy, NY 12065, USA.er).

from the ignition kernel and quickly becomes wrinkled due to theLandau–Darrieus instability, intrinsic to freely expanding flames[3]. The wrinkled flame develops into a fully turbulent flame brushwhich accelerates with increasing levels of turbulence and corre-sponding growth in flame surface area [3]. As the turbulent flamebrush accelerates, compression waves are generated ahead of theflame, which coalesce into a leading shock wave [3]. Finally theaccelerating flame transitions into a detonation wave. The final tran-sition from a high-speed turbulent flame/shock front to a detonationis thought to involve a localized explosion somewhere in or ahead ofthe turbulent flame brush or in the boundary layer, due to the attain-ment of autoignition conditions, and the establishment of an induc-tion-time gradient enabling the SWACER (shock wave amplificationby coherent energy release) mechanism originally proposed by Zel’-dovich [3,10,11]. Many studies suggest that the localized explosionoccurs within a quenched volume of reactants within the turbulentflame brush or in the boundary layer [2].

The performance of pulse detonation engines (PDEs), where DDTis the means of detonation initiation, is dependent on the requisitetime and length scales for DDT run-up, due to the requirement ofsufficient engine length for DDT run-up and the limitations the timerequired for DDT run-up place on PDE cycle frequency. Many effortshave been made to promote flame acceleration and DDT through theuse of obstacles to induce turbulent fluctuations in the unburnedgas ahead of the accelerating flame and thereby increase the flame

Inc. All rights reserved.

D.J. Finigan et al. / Combustion and Flame 159 (2012) 1314–1320 1315

area and increase the rate of flame acceleration. Studies using a widerange of obstacle geometries have been reported, including helicalspirals, orifice plates, dimples, baffles, and swept ramps [12–15].Of course the use of obstacles to promote DDT results in a drag pen-alty to engine thrust.

Other non-fluid mechanic DDT promotion efforts include theaddition of sensitive fuel/O2 mixtures at the location of ignition tostart a detonation that then propagates into a fuel/air mixture[16]. Multipoint or distributed ignition sources that promote great-er levels of initial volumetric heat release and/or shock wave ampli-fication have also been investigated for DDT promotion. Frolov et al.[17] proposed a concept for promoting DDT based on triggering 11electric sparks, spaced down the length of a detonation tube,sequentially to amplify the leading shock wave developed fromthe coalescence of compression waves emanating from the acceler-ating flame. Their system, although effective, required a high volt-age source (2500 kV) and a total spark energy deposition of1.68 MJ/m2. Ciccarelli et al. [18] used four circumferentially-spacedspark igniters to initiate multiple ignition kernels at the head end ofa detonation tube and demonstrated 30% reductions in DDT run-updistance using the four-point ignition method in conjunction withorifice plate obstacles. Wang et al. [19] and Zhukov and Starikovskii[20] have demonstrated ignition and DDT promotion using a high-voltage nanosecond transient plasma ignition source (coronadischarge) that creates several high-energy ignition kernels aroundthe plasma zone. Wang et al. [19] demonstrated reductions in igni-tion delay and rise times of a factor of two to three for the distrib-uted transient plasma ignition when compared to single sparks ina flowing PDE. The literature contains various mechanistic explana-tions for the influence of ignition distribution on flame accelerationand DDT [17,18,20], it can be generalized that the increased rate ofvolumetric energy release due to the distribution of ignition sitesresults in greater early-time flame surface area, which increasesthe rate of turbulence generation and flame acceleration leadingto amplification of the compression from the accelerating turbulent

Fig. 1. Images of the distributed photo ignition of a stoichiometric C2H4/air mixture at 1weight), from Berkowitz and Oehlschlaeger [21].

flame brush, and, hence, results in accelerated leading shock forma-tion and DDT.

Recently, Berkowitz and Oehlschlaeger [21] investigated a dis-tributed ignition method for the quasi-homogenous ignition of com-bustible gaseous mixtures using the photo ignition of single-walledcarbon nanotubes (SWCNTs) containing Fe impurity suspended infuel/air mixtures. The photo-ignition of SWCNTs containing metalimpurities was first discovered by Ajayan et al. [22]. In subsequentstudies it was determined that the photo ignition phenomenon isdependent on the presence of metal nanoparticle impurities in theSWCNTs and the SWCNTs simply act to stabilize the naturally pyro-phoric metal nanoparticles [23–25]. It was also shown that thephoto ignition of these Fe-containing nanomaterials results in peaktemperatures in excess of 1500 �C based on nanoscale characteriza-tion of the products [24,25]. The exposure of the photo-sensitivenanomaterials to a low-energy Xe camera flash, when suspendedin combustible fuel/oxidizer mixtures, results in the rapid heatingand oxidation of the nanomaterials followed by the ignition of thefuel/oxidizer mixture. Because the nanomaterials can be distributedthroughout any given volume and exposed to a spatially-diffusivelight source, the ignition of the fuel/oxidizer mixture can be highlydistributed. Berkowitz and Oehlschlaeger demonstrated the quasi-homogenous ignition of ethylene/air mixtures through high-speedcamera images (see Fig. 1), which show the luminosity from thephoto-igniting nanomaterials, in lm sized clumps, and from the vol-umetric combustion of the ethylene/air. Experiments in a closedcombustion chamber also demonstrated reductions in ignitiondelay and rise times by up to a factor of two when compared to sin-gle-point spark ignition [21]. In other studies, the photo ignition ofSWCNTs has been demonstrated for combustion applications byChehroudi and Danczyk for the ignition of single fuel droplets inair [26] and Manaa et al. [27] for the ignition of solid explosives. Che-hroudi and Danczyk also patented the concept of using the photoignition of carbon nanotubes in distributed ignition applications[28,29]. Distributed photo ignition has potential in combustion

bar containing suspended single-walled carbon nanotubes with Fe impurity (70% by

1316 D.J. Finigan et al. / Combustion and Flame 159 (2012) 1314–1320

applications where a degree of distribution and/or control over thelocation and timing of ignition are desired. Here we demonstratethat the distributed photo ignition of SWCNTs with Fe impurity sus-pended in gaseous fuel/oxidizer mixtures can be used to promoteflame acceleration and DDT in a confined geometry.

Fig. 3. Example raw ion sensor signals for the photo ignition of a 30.3 kPastoichiometric C2H4/O2 mixture. S1–S7 labels indicate the ion sensor signals forsensors located sequentially from the ignition end of the detonation tube. In thiscase DDT occurred prior to sensor 3.

2. Experimental method

Experiments were performed in a closed detonation tube to com-paratively study DDT resulting from the distributed photo ignitionof carbon nanotubes with Fe impurity and a traditional single-pointspark ignition. A schematic of the detonation tube is shown in Fig. 2.The carbon steel tube is 1 m long with a 7.62 cm (3 in.) inner diam-eter and closed on both ends. The tube inner diameter is constantwith average tube bore roughness of Ra = 0.4–1.6 lm. For the mea-surement of combustion wave trajectories to determine DDT run-updistances, 12 ion sensors were axially spaced every tube diameter(7.62 cm, 3 in.) along the inside of the detonation tube side wall.The ion sensors were miniature spark plugs (Rimfire Mini ViperZ2, 0.5 mm spark gap, 5 mm outer thread diameter) supplied with9 V from an alkaline battery and connected in series to a computer-ized data acquisition system (1 MHz National Instruments system,automated LabView software interface). Upon passage of the com-bustion wave at each ion sensor location, the voltage monitoredby the data acquisition system sharply drops from the open circuitvalue (�9 V) due to the completion of the circuit at the ion sensorspark gap by the partially-ionized gases present in the combustionproducts; typical experimental ion sensor signals are shown inFig. 3. Although the detonation tube was outfitted with 12 sensors,only 7 were connected during a given experiment, due to the chan-nel limitation of the data acquisition system. Sensors were strategi-cally selected for each experiment to capture the DDT run-updistance.

Measurements of DDT run-up distance were carried out forboth the spark and photo ignition of quiescent stoichiometricC2H4/O2/N2 mixtures with three levels of N2 dilution (0%, 20%,and 40%), at initial pressures ranging from 25 to 170 kPa, and aninitial temperature of 297 ± 2 K. Mixture compositions and initialpressures were chosen such that DDT would occur within the1 m tube length. The C2H4/O2/N2 mixtures were made in a holdingtank via partial pressures and allowed to diffusively mix for 24 hbefore use. Prior to experiments the detonation tube was evacu-ated to 5 � 10�3 Torr and filled with the C2H4/O2/N2 mixture tothe desired pressure specified with a 1000 Torr Baratron MKSmanometer.

Combustion was initiated at the head end of the detonation tubeby either an automotive spark plug located in the tube end wall(modified MSD 6A capacitive discharge ignition controller, Cham-pion model QL82YC spark plug, 1 mm gap, single 105 mJ sparks)or by the photo ignition of suspended nanomaterials. Photo ignitionwas achieved by injecting 2 mg of single-wall carbon nanotubes(SWCNTs) containing 70% Fe impurity by weight through a diffusiveair-blast style injector located in the tube side wall 3 cm from the

Fig. 2. Detonation tube

end wall and exposing the nanomaterials to a Xe flash. The as-pro-duced SWCNTs with Fe impurity (no purification) were synthesizedby Nano-C in a pre-mixed combustion process where the introduc-tion of a Fe catalyst precursor allows for the SWCNT growth. Theresulting Fe ‘‘impurity’’ is specified by Nano-C as 70% by weightwhich is approximately 10% Fe by volume. The air-blast injectionresulted in a distributed suspension of SWCNT-Fe clumps through-out approximately the first 6 cm of the detonation tube. The unifor-mity of the suspension could not be quantified due to insufficientoptical access to the ignition zone, where high pressures areachieved, but we have previously shown that quasi-homogenousphoto ignition of gaseous fuel/oxidizer can be achieved within aspherical volume with a diameter of approximately 3–4 cm usingthis relatively crude particle injection method [21]. The injectednanomaterials are suspended in aggregate clumps of lm dimen-sions due to adhesion and entanglement of the flame synthesizedSWCNTs. The C2H4/O2/N2 mixtures were used for ‘‘air-blast’’ injec-tion to ensure homogeneity of the gas mixture. Following injection,the suspended SWCNTs were exposed to a Xe camera flash(�300 mJ of visible light, �1 ms flash duration), housed withinthe tube and located at the head end, resulting in a quasi-distrib-uted ignition phenomenon, as illustrated in Fig. 1. Both the sparkand photo ignition hardware and experimental procedures usedin this study were identical to those previously detailed by Berko-witz and Oehlschlaeger [21].

Following the initiation of combustion at the head end of thetube by either spark or photo ignition, the acceleration of the com-bustion wave from high-speed flame to detonation was monitoredusing the ion sensors (Fig. 3). The distance versus time trajectoriesprovided by the ion sensors can be converted into wave velocityversus time, of course the resulting wave velocity is averaged overthe sensor axial spacing. Example wave trajectory and wave velocity

experimental setup.

Fig. 4. Example wave trajectories (top) and wave velocity profiles (bottom) for thespark and photo ignition of stoichiometric C2H4/O2 mixtures at 30.5 and 30.3 kPa,respectively.

D.J. Finigan et al. / Combustion and Flame 159 (2012) 1314–1320 1317

measurements are shown in Fig. 4 for both spark and photo ignitionexperiments. In all experiments a sharp rise in wave velocity corre-sponding to DDT was observed at some axial location in the tube. Atthe axial location of DDT the transition in measured wave velocitywas typically from 500–1000 m/s to greater than 2000 m/s withinone sensor spacing or tube diameter (7.62 cm). The Chapman–Joug-uet (CJ) detonation velocities for the studied mixtures and initialconditions are 2140–2360 m/s, per calculations carried out usingthe STANJAN thermochemical equilibrium routine [30]. The DDTrun-up distance was defined as the axial location, from the ignitionend of the tube, where the measured combustion wave velocityequaled or surpassed the CJ detonation velocity. Because velocitymeasurements (e.g., Fig. 4) are reported at the mid-point betweensensor locations, the DDT run-up distance is defined as the axiallocation of the upstream sensor for the first sensor pair where themeasured wave velocity equaled or surpassed the CJ velocity. Usingthe measured wave velocity profiles, DDT run-up distances weremeasured for C2H4/O2/N2 mixtures, selected because they providedDDT within the 1 m tube length.

Fig. 5. Velocity–distance measurements for the spark and photo ignition of threestoichiometric C2H4/O2/N2 mixtures (0%, 20%, and 40% N2) at varying initialpressure.

3. Results and discussion

Wave velocity measurements and corresponding DDT run-updistances for spark and photo ignition experiments are shown inFigs. 5 and 6, respectively; the measured DDT run-up distancesare also given in Tables 1 and 2 and all measured velocity profilesand calculated CJ detonation velocities are given in Table 3.Measurements of both wave velocity and DDT run-up distanceswere highly reproducible, as illustrated by the overlapping datafor repeated conditions in Fig. 6 and Tables 1–3. The high level ofapparent reproducibility is partly due to the 7.62 cm limit in reso-lution of DDT run-up distance imposed by the ion sensor spacing.The measurements show a reduction of DDT run-up distance withincreasing pressure and an increase in DDT run-up distance withincreasing N2 dilution, consistent with trends found in the litera-ture [31,32].

The results shown in Figs. 5 and 6 also illustrate shorter DDTrun-up distances for distributed photo ignition compared tosingle-point spark ignition. The measured reductions in DDTrun-up distance are approximately a factor of 1.5� to 2� for photoignition, with respect to spark ignition. Greater relative reductions

for photo ignition were generally observed for longer DDTdistances, occurring at lower initial pressures and greater N2 dilu-tion. The magnitude of reduction in DDT run-up distance is consis-tent with the factor of 2� reduction in ignition delay and ignitionrise times reported by Berkowitz and Oehlschlaeger for the photo

Fig. 6. DDT run-up distance as a function of initial pressure for both the spark andphoto ignition of three stoichiometric C2H4/O2/N2 mixtures (0%, 20%, and 40% N2).Calculated Chapman–Jouguet (CJ) detonation velocities also illustrated.

Table 2Measured DDT run-up distances for the photo ignition of the three stoichiometricC2H4/O2/N2 mixtures studied.

Photo, 0% N2 Photo, 20% N2 Photo, 40% N2

P (kPa) DDT distance (m) P (kPa) DDT distance (m) P (kPa) DDT distance (m)

24.9 0.4572 50.0 0.4572 129.6 0.685825.1 0.4572 50.1 0.4572 148.9 0.38130.1 0.2286 50.2 0.4572 150.3 0.38130.3 0.2286 70.0 0.3048 169.6 0.304840.1 0.1524 70.1 0.304840.1 0.1524 84.9 0.152440.3 0.1524 85.0 0.152450.2 0.0762 101.5 0.076250.2 0.0762 101.6 0.076250.3 0.076270.0 0.076270.1 0.076270.1 0.0762

1318 D.J. Finigan et al. / Combustion and Flame 159 (2012) 1314–1320

ignition of stoichiometric C2H4/air mixtures in comparison to sparkignition [21].

We hypothesize that the promotion of DDT observed for photoignition occurs because the early-time heat release resulting fromphoto ignition is distributed volumetrically and therefore is greaterin magnitude than that from a single-point spark ignition [21]. Thispresumably results in larger early-time flame area, increased insta-bility and wrinkling of the early flame, faster transition to turbu-lence, and higher rates of turbulent flame acceleration, all ofwhich will lead to faster leading shock formation, increased shockamplification, and accelerated DDT. This hypothesis is in concert

Table 1Measured DDT run-up distances for the spark ignition of the three stoichiometric C2H4/O

Spark, 0% N2 Spark, 20% N2

P (kPa) DDT distance (m) P (kPa)

25.0 0.762 50.125.0 0.762 50.129.6 0.381 50.630.1 0.381 69.930.5 0.381 70.139.9 0.2286 84.939.9 0.2286 85.040.0 0.2286 101.449.9 0.1524 101.750.1 0.152450.1 0.152470.0 0.152470.0 0.152470.1 0.1524

with the experimental observations of Sinibaldi et al. [33] whoshowed that the location of ignition in a tube influences theearly-time flame area which correlates with flame accelerationand DDT run-up distance.

In addition to the observed reduction in DDT run-up distance,the measured wave velocity profiles for the highest pressure photoignition cases for 0% and 20% N2 dilution show what appear to beextremely overdriven detonation waves at the ignition end of thedetonation tube. In these cases the magnitude of the measuredwave velocities are far greater than that measured for any of theDDT cases; in some experiments the measured detonation veloci-ties at DDT were in excess of 2500 m/s but not greater than3000 m/s. In the case of experiments performed for stoichiometricC2H4/O2 at 50 kPa and 70 kPa, the wave velocities measured at thefirst sensor pair (sensors located 7.62 and 15.24 cm from theignition end wall) were in the range of 6000–7000 m/s, followedby decay within one ion senor location (one tube diameter) to avelocity near that of a CJ detonation. These extremely high wavevelocities might indicate that multiple combustion fronts havebeen formed by the distributed photo ignition and/or that a deto-nation wave may have been directly initiated by the distributedphoto ignition. Berkowitz and Oehlschlaeger [21] showed thatthe photo ignition event can be quasi-homogenous, which underhighly-sensitive fuel/O2 conditions could lead to quasi-volumetricexplosion at the ignition end of the tube, directly following theflash, resulting in direct initiation of a detonation due to the highenergy release. Matsui and Lee [9] reported a critical energy ofaround 100 mJ for the direct single-point initiation of a detonation

2/N2 mixtures studied.

Spark, 40% N2

DDT distance (m) P (kPa) DDT distance (m)

0.8382 130.2 0.83820.8382 132.4 0.83820.8382 148.9 0.68580.4572 149.6 0.68580.4572 169.6 0.45720.30480.30480.07620.0762

Table 3Combustion wave velocity profile measurements and calculated Chapman–Jouguet (CJ) detonation velocities. All experiments performed at an initial temperature of 297 ± 2 K.

Spark ignition (left) Flash ignition (right)

P (kPa) CJ vel. (m/s) Axial location (cm) (sensor mid-points) P (kPa) CJ vel. (m/s) Axial location (cm) (sensor mid-points)Wave velocity (m/s) Wave velocity (m/s)

/ = 1.0 C2H4/O2, 0% N2 / = 1.0 C2H4/O2, 0% N2

25.0 2307 49.53 57.15 64.77 72.39 80.01 87.63 24.9 2307 26.67 34.29 41.91 49.53 57.15 64.77560 680 635 896 2822 2822 1058 828 918 2628 2458 2309

25.0 2307 49.53 57.15 64.77 72.39 80.01 87.63 25.1 2308 26.67 34.29 41.91 49.53 57.15 64.77491 635 615 866 2309 2822 680 712 662 2309 2930 2627

29.6 2315 11.43 19.05 26.67 34.29 41.91 49.53 30.1 2316 11.43 19.05 26.67 34.29 41.91 49.53377 432 501 552 2721 2930 680 686 2721 2822 2309 2309

30.1 2316 11.43 19.05 26.67 34.29 41.91 49.53 30.3 2316 11.43 19.05 26.67 34.29 41.91 49.53334 403 582 692 2241 2822 630 837 2721 2721 2309 2309

30.5 2317 11.43 19.05 26.67 34.29 41.91 49.53 40.1 2330 11.43 19.05 26.67 34.29 41.91 49.53416 448 591 770 2309 2540 965 2721 2381 2381 2309 2309

39.9 2329 11.43 19.05 26.67 34.29 41.91 49.53 40.1 2330 11.43 19.05 26.67 34.29 41.91 49.53401 686 2241 2381 2309 2241 865 2309 2381 2309 2309 2309

39.9 2329 11.43 19.05 26.67 34.29 41.91 49.53 40.3 2330 11.43 19.05 26.67 34.29 41.91 49.53336 526 2241 2621 2309 2241 940 2721 2309 2309 2309 2309

40.0 2330 11.43 19.05 26.67 34.29 41.91 49.53 50.2 2340 11.43 19.05 26.67 34.29 41.91 49.53433 646 2241 2381 2241 2241 5861 2721 2458 2381 2381 2381

49.9 2340 11.43 19.05 26.67 34.29 41.91 49.53 50.2 2340 11.43 19.05 26.67 34.29 41.91 49.53540 2931 2540 2931 2005 2309 5080 2540 2381 2309 2381 2381

50.1 2340 11.43 19.05 26.67 34.29 41.91 49.53 50.3 2340 11.43 19.05 26.67 34.29 41.91 49.53595 2822 2458 2381 2241 2309 4233 2721 2540 2241 2458 2381

50.1 2340 11.43 19.05 26.67 34.29 41.91 49.53 70.0 2356 11.43 19.05 26.67 34.29 41.91 49.53629 2822 2458 2241 2458 2241 4010 2540 2381 2381 2381 2309

70.0 2356 11.43 19.05 26.67 34.29 41.91 49.53 70.1 2356 11.43 19.05 26.67 34.29 41.91 49.53940 2721 2458 2458 2309 2381 6927 2930 2381 2005 3048 2309

70.0 2356 11.43 19.05 26.67 34.29 41.91 49.53 70.1 2356 11.43 19.05 26.67 34.29 41.91 49.53802 2540 2309 2381 2458 2381 4010 2628 2381 2381 2381 2381

70.1 2356 11.43 19.05 26.67 34.29 41.91 49.53953 2540 2381 2627 2117 2381

/ = 1.0 C2H4/O2/N2, 20% N2 / = 1.0 C2H4/O2/N2, 20% N2

50.1 2230 49.53 57.15 64.77 72.39 80.01 87.63 50.0 2330 41.91 49.53 57.15 64.77 72.39 80.01569 582 866 726 819 2117 856 2059 2822 2381 2241 2309

50.1 2230 49.53 57.15 64.77 72.39 80.01 87.63 50.1 2230 41.91 49.53 57.15 64.77 72.39 80.01485 488 615 668 693 2721 856 2721 2381 2241 2309 2241

50.6 2231 49.53 57.15 64.77 72.39 80.01 87.63 50.2 2230 34.29 41.91 49.53 57.15 64.77 72.39501 610 712 762 712 2822 786 847 2177 2721 2309 2241

69.9 2244 34.29 41.91 49.53 57.15 64.77 72.39 70.0 2244 11.43 19.05 26.67 34.29 41.91 49.53640 907 2721 2540 2381 2177 907 674 819 2241 2721 2381

70.1 2244 34.29 41.91 49.53 57.15 64.77 72.39 70.1 2244 11.43 19.05 26.67 34.29 41.91 49.53651 620 2540 2721 2458 2177 540 573 876 2241 2822 2381

84.9 2252 19.05 26.67 34.29 41.91 49.53 57.15 84.9 2252 11.43 19.05 26.67 34.29 41.91 49.53610 605 2381 2721 2309 2309 828 2458 2540 2241 2241 2241

85.0 2252 11.43 19.05 26.67 34.29 41.91 49.53 85.0 2252 11.43 19.05 26.67 34.29 41.91 49.53615 540 699 2628 2628 2458 886 2309 2540 2540 2309 2241

101.4 2259 11.43 19.05 26.67 34.29 41.91 49.53 101.5 2259 11.43 19.05 26.67 34.29 41.91 49.533436 2822 2458 2241 2309 2241 3464 2540 2628 2381 2241 2309

101.7 2260 11.43 19.05 26.67 34.29 41.91 49.53 101.6 2260 11.43 19.05 26.67 34.29 41.91 49.533313 2458 2627 2309 2309 2309 4763 3175 2309 2540 2309 2241

u = 1.0 C2H4/O2/N2, 40% N2 u = 1.0 C2H4/O2/N2, 40% N2

130.2 2142 49.53 57.15 64.77 72.39 80.01 87.63 129.6 2141 49.53 57.15 64.77 72.39 80.01 87.63540 560 646 699 747 2458 635 625 591 2177 2721 2627

132.4 2143 49.53 57.15 64.77 72.39 80.01 87.63 148.9 2147 34.29 41.91 49.53 57.15 64.77 72.39778 467 581 718 645 2381 929 2241 2721 2381 2177 2116

148.9 2147 49.53 57.15 64.77 72.39 80.01 87.63 150.3 2148 26.67 34.29 41.91 49.53 57.15 64.77482 657 867 2540 2628 2540 640 740 2381 2721 2309 2241

149.6 2147 49.53 57.15 64.77 72.39 80.01 87.63 169.6 2152 11.43 19.05 26.67 34.29 41.91 49.53548 552 657 2117 2381 2458 540 600 556 2259 2721 2628

169.6 2152 19.05 26.67 34.29 41.91 49.53 57.15582 548 504 595 2822 2381

D.J. Finigan et al. / Combustion and Flame 159 (2012) 1314–1320 1319

in stoichiometric C2H4/O2 at an initial pressure of 1 atm. While the105 mJ spark does not provide direct initiation for stoichiometricC2H4/O2 mixtures at initial pressures of 50 kPa or 70 kPa, theexposure of the suspended nanomaterials to the Xe flash (approx-imately 300 mJ of visible optical energy with a 1 ms flash duration)results in unique initial velocity profiles, perhaps indicative ofdirect detonation initiation resulting from a volumetric explosion.Further investigation of the dynamics of the photo ignition and

detonation formation phenomena, through high-speed opticalimaging and multiple dynamic pressure measurements, is neededto determine if direct detonation initiation is possible and furtherunderstand flame acceleration and DDT from distributed photoignition for cases where it is not.

The two-fold reduction in DDT run-up distances demonstratedhere using the distributed photo ignition of suspended nanomate-rials in gaseous fuel/oxidizer mixtures are similar in magnitude to

1320 D.J. Finigan et al. / Combustion and Flame 159 (2012) 1314–1320

those realized using distributed transient plasma ignition [19]. Thephoto ignition method has advantages for detonation engines, inthat it provides DDT promotion with a very low optical power([22] reported photo ignition with as little as 100 mW/cm2 flashpower, here we used a Xe flash with 300 mJ of visible energy anda duration of 1 ms) and therefore could be implemented with sim-ple light sources (e.g., flash lamp, light-emitting diode, diode laser)requiring low-mass low-voltage power supplies. However, compli-cating its application to detonation engines, the photo ignitionmethod utilizes photo-sensitive nanomaterials that would haveto be distributed into either the oxidizer or fuel stream.

Further research is needed to evaluate the photo ignition meth-od demonstrated here for DDT promotion and ignition in general.For this study, exposing SWCNTs with Fe impurity with to the flashfrom a Xe camera flash was implemented simply because it wasknown to produce the photo ignition phenomenon. However, othermetal and carbon-based nanomaterials are also known to exhibitphoto ignition (e.g., graphene oxide [34]) and initial experimentalstudies performed in our laboratory suggest that nanomaterialselection is important for optimizing ignition, flame acceleration,and DDT promotion. Similarly, the optimization of the light sourcehas yet to be considered and could provide further gains. Impor-tantly, the photo ignition demonstrations presented here were car-ried out for sensitive C2H4/O2/N2 mixtures. These studies need tobe extended to fuel/air conditions where the photo ignition tech-nique may need to be combined with other fluid mechanic meansof promoting DDT (e.g., orifice plates, helical spirals, ramps) for suf-ficiently short DDT run-up distances for engine applications.

4. Summary

The promotion of deflagration-to-detonation transition (DDT)using the distributed photo ignition of photo-sensitive nanomate-rials suspended in fuel/oxidizer mixtures has been demonstratedfor the first time. Single-wall carbon nanotubes (SWCNTs) with70% Fe impurity by weight were suspended at one end of a closeddetonation tube filled with C2H4/O2/N2 mixtures. The SWCNTswere exposed to a Xe camera flash causing them to photo igniteand subsequently produce a volumetrically distributed ignition ofthe C2H4/O2/N2 mixture. The distributed photo ignition leads toenhanced flame acceleration and deflagration-to-detonation tran-sition (DDT). Combustion wave velocity measurements made withion sensors show that photo ignition provides DDT run-updistances that are around a factor of 1.5� to 2� shorter than fortraditional single-point spark ignition. We hypothesize that the in-creased volumetric energy release rate resulting from distributedphoto-ignition enhances DDT due to greater early-time flame areaand turbulence levels, resulting in accelerated formation andamplification of the leading shock and accelerated DDT. For themost sensitive C2H4/O2 mixtures studied, photo ignition yieldsextremely high combustion wave velocity measurements immedi-ately following ignition, with velocities approximately 3� greaterthan the Chapman–Jouguet detonation velocity, suggesting thatthe detonation was either directly initiated or that multiple com-bustion fronts are formed by the distributed ignition. Further study

is needed to understand the mechanism for detonation formationfor these cases.

Acknowledgments

We are grateful for the support of the US Office of Naval Re-search with Dr. Gabriel Roy as technical monitor and to HeesikYoo, Aaron Ide, Stephen Kim, and Garrett Ellsworth for initial setupof the experiment.

References

[1] J.H.S. Lee, I.O. Moen, Prog. Energy Combust. Sci. 6 (1980) 359–389.[2] E.S. Oran, V.N. Gamezo, Combust. Flame 148 (2007) 4–47.[3] G. Ciccarelli, S. Dorofeev, Prog. Energy Combust. Sci. 34 (2008) 499–550.[4] J.E. Shepherd, Proc. Combust. Inst. 32 (2009) 83–98.[5] K. Kailasanath, AIAA J. 38 (2000) 1698–1708.[6] G.D. Roy, S.M. Frolov, A.A. Borisov, D.W. Netzer, Prog. Energy Combust. Sci. 30

(2004) 545–672.[7] F.K. Lu, Prospects for detonations in propulsion, in: Proc. of the 9th Int. Symp.

on Experimental and Computational Aerothermodynamics of Internal Flows(ISAIF9), 2009.

[8] V.E. Tangirala, A.J. Dean, D.M. Chapin, P.F. Pinard, B. Varatharajan, Combust. Sci.Technol. 176 (2004) 1779–1808.

[9] H. Matsui, J.H. Lee, Proc. Combust. Inst. 17 (1979) 1269–1280.[10] Ya.B. Zel’dovich, V.B. Librovich, G.M. Makhviladze, G.I. Sivashinsky, Astronaut.

Acta 15 (1970) 313–321.[11] J.H.S. Lee, R. Knystautas, N. Yoshikawa, Acta Astronaut. 5 (1978) 971–972.[12] S.-Y. Lee, J. Watts, S. Saretto, S. Pal, C. Conrad, R. Woodward, R. Santoro, J.

Propul. Power 20 (2004) 1026–1036.[13] V. Bychkov, D. Valiev, L.-E. Eriksson, Phys. Rev. Lett. 101 (2008) 164501.[14] C.T. Johansen, G. Ciccarelli, Combust. Flame 156 (2009) 405–416.[15] C.M. Brophy, Initiation Improvements for Hydrocarbon/Air Mixtures in Pulse

Detonation Applications, AIAA Paper 2009-1611, 2009.[16] C.M. Brophy, R.K. Hanson, J. Propul. Power 22 (2006) 1155–1161.[17] S.M. Frolov, V.Y. Basevich, V.S. Aksenov, J. Propul. Power 19 (2003) 573–580.[18] G. Ciccarelli, C. Johansen, M.C. Hickey, J. Propul. Power 21 (2005) 1029–1034.[19] F. Wang, J.B. Liu, J. Sinibaldi, C. Brophy, A. Kuthi, C. Jiang, P. Ronney, M.A.

Gundersen, I.E.E.E. Trans, Plasma Sci. 33 (2005) 844–849.[20] V.P. Zhukov, A.Yu. Starikovskii, Combust. Explosion Shock Waves 42 (2006)

195–204.[21] A.M. Berkowitz, M.A. Oehlschlaeger, Proc. Combust. Inst. 33 (2011) 3359–

3366.[22] P.M. Ajayan, M. Terrones, A. de la Guardia, V. Huc, N. Grobert, B.Q. Wei, H.

Lezec, G. Ramanath, T.W. Ebbesen, Science 296 (2002) 705.[23] B. Bockrath, J.K. Johnson, D.S. Sholl, B. Howard, C. Matranga, W. Shi, D. Sorescu,

Science 297 (2002) 192.[24] N. Braidy, G.A. Botton, A. Adronov, Nano Lett. 2 (2002) 1277–1280.[25] J. Smits, B. Wincheski, M. Namkung, R. Crooks, R. Louie, Mater. Sci. Eng. A 358

(2003) 384–389.[26] B. Chehroudi, S.A. Danczyk, A Novel Distributed Ignition Method Using Single-

Wall Carbon Nanotubes (SWCNTS) and a Low-Power Flash Light, GlobalPowertrain Congress – World Powertrain Exposition, 2006.

[27] M.R. Manaa, A.R. Mitchell, R.G. Garza, P.F. Pagoria, B.E. Watkins, J. Am. Chem.Soc. 127 (2005) 13786–13787.

[28] B. Chehroudi, G.L. Vaghjiani, A.D. Ketsdever, Method for Distributed Ignition ofFuels by Light Sources, US Patent 7517,215 B1, April 14, 2009.

[29] B. Chehroudi, G.L. Vaghjiani, A.D. Ketsdever, Apparatus for Distributed Ignitionof Fuels by Light Sources, US Patent 7665,985 B1, February 23, 2010.

[30] W.C. Reynolds, STANJAN Interactive Computer Programs for ChemicalEquilibrium Analysis, Stanford University, 1986.

[31] M. Kuznetsov, V. Alekseev, I. Matsukov, S. Dorofeev, Shock Waves 14 (2005)205–215.

[32] M. Cooper, S. Jackson, J. Austin, E. Wintenberger, J.E. Shepherd, J. Propul. Power18 (2002) 1033–1041.

[33] J.O. Sinibaldi, C.M. Brophy, J.P. Robinson, Ignition Effects on Deflagration-to-detonation Transition Distance in Gaseous Mixtures, AIAA Paper 2000-3590,2000.

[34] S. Gilje, S. Dubin, A. Badakhshan, J. Farrar, S.A. Danczyk, R.B. Kaner, Adv. Mater.22 (2010) 419–423.