Int. J. Alternative Propulsion, Vol. 1, No. 2/3, 2007
Flame dynamics in sub-millimetre combustors
Shaurya Prakash
Department of Mechanical and Industrial Engineering,
University of Illinois at Urbana-Champaign, 1206 W Green Street,
Urbana, IL 61801, USA
Fax: +1-217-244-6534
E-mail: [email protected]
Adrian D. Armijo
Department of Mechanical and Industrial Engineering,
University of Illinois at Urbana-Champaign,
981 Grand Canyon Parkway #316,
Hoffman Estates, IL 60194, USA
E-mail: [email protected]
Richard I. Masel
Department of Chemical and Biomolecular Engineering,
University of Illinois at Urbana-Champaign, 600 S Mathews,
Urbana, IL 61801, USA
Fax: +1-217-333-5052
E-mail: [email protected]
Mark A. Shannon*
Department of Mechanical and Industrial Engineering,
University of Illinois at Urbana-Champaign, 1206 W Green Street,
Urbana, IL 61801, USA
Fax: +1-217-333-1942
E-mail: [email protected]
*Corresponding author
Abstract: This paper describes experimental observations of flame dynamics
within an alumina non-premixed combustor with a sub-millimetre (0.75 mm)
minimum size. These flame dynamics occur within a total combustor volume of
130 mm3 and are imaged through a sapphire window. The confined flames are
surveyed using visual still-frames, high-speed flame imaging, measurement of
external wall temperatures, and recordings of acoustic emissions from the
combustor. The observed flame dynamics are characterised by an oscillating
edge-flame that is anchored at the combustor inlet and is accompanied by
ignition/extinction events leading to the eventual formation of a stable edgeflame and distinct cellular structures along the reaction channel. This paper
investigates the unstable and unsteady flame dynamics within the nonpremixed sub-millimetre combustors from ignition to the formation of stable
flame structures.
Keywords: alumina Al2O3; microcombustion; edge-flame; non-premixed;
flame cells; methane; instability; transient; flame structure; non-stabilised
flames.
Copyright © 2007 Inderscience Enterprises Ltd.
325
326
S. Prakash et al.
Reference to this paper should be made as follows: Prakash, S., Armijo, A.D.,
Masel, R.I. and Shannon, M.A. (2007) ‘Flame dynamics in sub-millimetre
combustors’, Int. J. Alternative Propulsion, Vol. 1, No. 2/3, pp.325–338.
Biographical notes: Shaurya Prakash graduated with a bachelor’s degree in
Mechanical Engineering from the University of Arkansas, Fayetteville in 2001
and began his foray into the world of research and graduate studies. He earned
his master’s degree in 2003 at the University of Illinois at Urbana-Champaign
also in Mechanical Engineering, working on laser-induced fluorescence
measurements of hydroxyl radicals over various inert surfaces. He is currently
pursuing a PhD degree under the supervision of Prof. Mark A. Shannon at
the University of Illinois. His areas of interest and expertise include
microfabrication, microscale combustion, and surface-mediated transport in
micro- and nano-systems.
Adrian D. Armijo graduated with a Bachelor of Science degree in Mechanical
Engineering with a minor in Mathematics from the University of Oklahoma in
2003. He obtained a Master of Science degree in Mechanical Engineering at the
University of Illinois at Urbana-Champaign in 2006. His graduate work
included describing the steady-state flame structure observed in alumina nonpremixed microburners under the supervision of Prof. Mark A. Shannon. He is
currently working for Motorola Inc. as an Automotive Mechanical Engineer.
Prof. Richard I. Masel received his MS in Chemical Engineering from Drexel
University in 1973 and his PhD from The University of California, Berkley in
1977. He joined the University of Illinois in 1978 and is now Professor of
Chemical and Biomolecular Engineering and of Electrical and Computer
Engineering. Over the years, Masel has worked on kinetics and catalysis, and
more recently fuel cells, microchemical systems and MEMS. In 2003, he help
found Tekion, a developer of fuel cells for portable electronics. Masel served
for 3 years as Chief Technical Officer of Tekion, before stepping down to the
position of Chief Technical Advisor. More recently, Masel and Shannon
founded Cbana Laboratories, a developer of microfluidic products.
Prof. Mark A. Shannon graduated with a bachelor’s, master’s and PhD in 1989,
1991 and 1993, respectively, in Mechanical Engineering from the University of
California at Berkeley. His goal is to advance the state-of-knowledge in
coupled phenomenon at the micro- to nano-scale, with application to hightemperature combustion reactors, water purification, microfabrication, nanoand micro-electromechanical systems (NEMS and MEMS, respectively),
micro- and nano-fluidic devices and mesoscale machines that bridge the nanoand micro-scales to the normal scales. His areas of expertise and interest
include energetics of reactor and separation systems; the development of new
microfabrication methods utilising lasers, electric fields, plasma and chemistry;
and creating new NEMS, MEMS and mesoscale energetic devices, and water
purification systems. Students in his group conduct research related to
microfabrication process development, experimental measurements and
computational research.
1
Introduction
Since the 1990s, rapid development of microfabrication technologies has led to
tremendous growth in miniaturisation of several systems and devices. Development of
micro- and miniature systems has led to an increased power requirement at the
microscale to operate these devices and systems. Recently, there has been growing
Flame dynamics in sub-millimetre combustors
327
interest in employing combustion for replacement of conventional batteries to meet these
power requirements (Kyritsis et al., 2004a,b), leading to advancements in the field of
micro-power generation (MPG). In addition, micro- and meso-scale combustion systems
are being developed as heat sources for many different applications (Dunn-Rankin et al.,
2005; Fernandez-Pello, 2002; Vican et al., 2002), including high-temperature reactors
(Schaevitz et al., 2001; Shannon et al., 2002). Several researchers have enhanced both the
science and technology in MPG through development of microcombustors, fuel cells,
engines, heaters and reactors (Aichlmayr, 2002; Kim et al., 2005; Masel and Shannon,
2001; Mehra et al., 2000; Miesse et al., 2004; Prakash et al., 2005a; Ronney, 2003; Yeom
et al., 2005). This interest in combustion-based systems is due to the high energy and
power densities often achieved with combustion-based heat engines. For instance,
hydrocarbon fuels can provide 40−50 MJ/kg of heat energy; in comparison to 0.4−0.7
MJ/kg of electrical energy for state-of-the art Li-ion batteries. Thus, even with a 2%
conversion of that heat energy to usable energy (giving 0.8 to 1 MJ/kg) will exceed the
best battery energy densities currently achieved. Despite several advancements in the
field of micro- and meso-scale combustion, many questions remain unanswered with
regard to constructing systems that provide predictable and controllable heat output with
uniform wall temperatures that can be operated for long times to meet the promise of
portable and personal power sources.
Over the last few years, efforts have focused on developing homogeneous
non-premixed and catalytic combustors to circumvent the problem of combustor wall
material failure (Kyritsis et al., 2004a,b; Miesse et al., 2004, 2005a,b; Norton and
Vlachos, 2005; Prakash et al., 2005a,b). The importance of surface chemistry and wall
temperature in sub-millimetre or microcombustors (Jensen, 2000; Miesse et al., 2004;
Moore, 2000; Prakash et al., 2005a,b; Vican et al., 2002), and the effect that treated
alumina (Al2O3) walls have on maintaining high hydroxyl (OH*) radical concentration in
homogeneous flames (Prakash et al., 2005a) have been reported. In the same work
(Prakash et al., 2005a) it was shown that in contrast to platinum traditionally inert
materials such as quartz and alumina walls support a gas-phase OH* radical
concentration higher by almost an order of magnitude at elevated temperatures. Thus,
radical quenching phenomena in alumina combustors are expected to be less dominant
than in combustors made from some other materials. One of the major challenges in these
high surface area to volume ratio combustors is to maintain an optimum balance between
sustaining homogeneous combustion and maximising the heat output as the combustion
process is highly sensitive to heat losses (Aichlmayr, 2002; Prakash et al., 2005b;
Ronney, 2003). It has been previously shown that there is little correlation between the
flame quenching distances and the thermal conductivity of a material, especially at high
temperature (Miesse et al., 2004). For instance, at 1000 K the thermal conductivity of
alumina (Miesse et al., 2004) is higher than that of quartz by almost 3 times but the
chemical action (Prakash et al., 2005a) is almost an order of magnitude lower. The
problem is further complicated by the fact that transient effects during start up and in
operation can change the combustion process greatly from steady-state operation
(Fernandez-Pello, 2002). Therefore, understanding the effect of various parameters that
govern sustained combustion and flame dynamics at these length scales is an important
step toward developing and building combustion-based power sources.
In the past, several investigations have shown that enclosing any heat source or flame
in a chamber can cause heat-driven oscillations and possible acoustic emission (Raun
et al., 1993; Rayleigh, 1878; Tyndal, 1897). Acoustic emissions in combustion systems
328
S. Prakash et al.
have been investigated for over a century (Feldman, 1968a,b; Gonzalez, 1996; Miesse
et al., 2004; Miller et al., 2005; Nicoud and Poinsot, 2005; Rayleigh, 1878; Richecoeur
and Kyritsis, 2005) beginning with the work of Higgins and ‘singing flames’ in 1777
(Tyndal, 1897). There have been recent reports (Richecoeur and Kyritsis, 2005) of
oscillatory flame behaviour leading to sound emission with periodic extinction/re-ignition
phenomena in mesoscale curved quartz ducts. Some numerical studies have focused on
investigating the onset of oscillations, stability criteria and oscillatory flame propagation
in non-premixed configurations (Kukuck and Matalon, 2001; Thatcher and Dold, 2000;
Thatcher et al., 2002).
In this paper, experimental observations of flame dynamics in a sub-millimetre
alumina combustor are presented. The purpose of this paper is to describe the observed
flame dynamics that are characterised by rapid extinction–ignition events accompanied
by acoustic emissions. Still-frame and high-speed imaging along with the measurement
of acoustic emissions are used to identify the origin of these observed flame dynamics.
2
Experimental description
Material preparation and burner fabrication have been reported previously (Miesse, 2005;
Miesse et al., 2004, 2005a,b; Prakash et al., 2005a,b; Prakash et al., 2006). A brief
description follows. The alumina non-premixed sub-millimetre or microcombustors are
machined from 1 mm thick α-phase polycrystalline alumina sheets (96% purity from
McMaster-Carr). The sheets are machined into two identical Y-shaped pieces, with each
leg having a width of 10 ± 0.5 mm and length of 35 ± 0.5 mm. A 5 ± 0.5 mm wide by
0.375 mm deep channel is machined into the centre of the pieces, so when assembled
the total depth is 0.75 ± 0.1 mm. In one of the alumina pieces, a slot is machined to
fit a 36 mm × 7 mm C-plane sapphire piece to provide a window for imaging of the
flame dynamics. The gases come into the reaction channels right angles with respect
to each other.
The machined alumina pieces are cleaned using a standard degrease procedure with
acetone and isopropyl alcohol followed by the standard RCA-2 clean (DI H2O/HCl/H2O2
in 4:1:1 ratio by volume) at 60 °C to remove metallic contaminants. Next, a two-step
anneal process is carried out. First, the polycrystalline alumina and sapphire pieces
are annealed in an argon atmosphere at 1550 °C for 1 h followed by a second anneal at
1100 °C in an oxygen atmosphere for 10 h. The two-step anneal process is thought to
reduce surface defects and minimises oxygen vacancies that can exist in alumina
(Henrich and Cox, 1994), helping to reduce the effects of radical quenching (Prakash
et al., 2005a,b). All components are then manually aligned and sealed along the outside
with an Aremco 569 high-temperature ceramic adhesive (Aremco Inc., Valley Cottage,
NY). Type R thermocouples (TCs) obtained from Omega Engineering Inc. (Stamford,
CT) are located along the outside wall of the alumina combustor and connected to a data
acquisition system to obtain external wall temperature profiles as a function of time. High
purity methane (CH4) and oxygen (O2) gases are controlled by M100B MKS mass flow
controllers, and flow through attached hoses. The entire assembly is then packed in
between two approximately 1 inch thick insulation layers of fibrous alumina. The exhaust
and unburnt gases flow freely into the ambient air (Figure 1).
The flame dynamics are imaged via still-frame and high-speed imaging. The
still-frames are captured by a Canon EOS Mark II digital camera with an attached
Flame dynamics in sub-millimetre combustors
329
infrared filter, and the high-speed images are acquired by using a Vision Research Inc.
camera (model no. Phantom v7.1). The emitted sound is recorded using a Knowles
Acoustic (Itasca, IL) microphone (model no. FG-3329) at a sampling rate of 20 kHz to
give a 2.5 times over Nyquist sampled maximum frequency of 4000 Hz reported here.
The microphone is connected to a pre-amplifier and a Hewlett-Packard (model no.
54645D) oscilloscope to monitor the acoustic signal. In addition, the measured signal is
recorded using a computer-controlled data acquisition system.
Figure 1
A digital photograph of the combustor with the sapphire window is shown
TCs to DAQ
O2
CH4
Burner with
sapphire window
Reaction
channel
Fibrous insulation
Note: The fuel and oxidiser come in at the two legs of the Y-shaped and the combustion occurs in the
reaction channel as shown. The reaction channel is 5 mm wide and 35 mm long with a total depth of 0.75
mm. TCs are attached to the outside wall of the combustor at 0.5, 1.5 and 2 cm from the combustor
exhaust. The sapphire window is not insulated.
3
Results and discussion
The flame dynamics under investigation here evolve in time before reaching a steadystate configuration comprising an edge-flame anchored at the combustor inlet and flame
cells. The individual cellular flame structures in the combustor are referred to as flame
cells in this paper to be consistent with past publications (Miesse et al., 2004, 2005a,b;
Prakash et al., 2005b; Prakash et al., 2006). A sapphire window without insulation is
required to conduct flame imaging and hence the acoustic and imaging data discussed
and described here are obtained from different combustors. This leads to different
boundary conditions for the combustors and can cause slight changes in observed flame
structure, as has been reported previously (Prakash et al., 2005b; Prakash et al., 2006).
However, the general behaviours observed are reproducible and repeatable. Two different
types of combustors are used in imaging and acoustic measurements due to the higher
degree of reproducibility of data in a combustor that is well insulated on all sides.
The fuel (CH4) and oxidiser (O2) flow into the reaction channel at right angles with
respect to each other. Flow rates are maintained such that the flow is laminar (with
Re < 100), and previous experiments do not demonstrate any apparent turbulent or eddy
mixing along the reaction channel (Miesse et al., 2005a,b). Thus, the mixing between the
fuel and oxidiser is primarily dominated by inter-diffusion.
330
S. Prakash et al.
3.1 Imaging
Ignition is carried out at the exhaust by means of an external spark. Immediately after
ignition, a continuous edge-like flame forms in the combustor channel that is anchored at
the base of the combustor inlet, where the fuel and oxidiser streams meet initially, along
with an unconfined diffusion flame that exists at the combustor exit. The continuous
edge-like flame exists along the middle of the reaction channel. Any unburnt fuel is
consumed by the unconfined diffusion flame at the combustor exit. Extinguishing the
unconfined methane–air diffusion flame before stabilisation, which will be discussed
later, leads to extinction of the combustion process within the reaction channel. Shortly
after ignition, the continuous edge-like flame, approximately 1 mm wide, is observed to
physically oscillate axially (lengthwise down the channel) and wiggle laterally (across the
width of channel) within the combustor channel displaying a wavy nature. To better
understand the observed flame dynamics, high-speed imaging (Figure 2) was carried out
at 1000 frames/s with an exposure time of 900 µs to capture the flame dynamics. Figure 2
shows that the flame structure is not a continuous flame as observed but rather a sequence
of several individual flames. In addition, these flames appear to be moving rapidly from
the edge-flame anchored at the combustor inlet toward the external methane–air flame at
the combustor exit at a frequency of approximately 142 Hz.
Figure 2
Flame structure images at early times after ignition in a sub-millimetre alumina
combustor
Direction of
bulk flow
Individual flame
moving toward exhaust
Edge-flame anchored
at combustor inlet
Increasing time
Note: Images from a high-speed camera show several individual flames travelling rapidly toward the
exhaust. The external methane–air diffusion flame is located at the top of the photos. Flow rates of the
fuel and oxidiser are equal at 150 sccm each. The image was digitally filtered in Adobe Photoshop® to
enhance the contrast and provide greater clarity to the image. The original image shows very low
intensity of the flame with respect to the ambient and background light due to the short exposure times
required for high-speed imaging.
The stable and steady-state flame structure is characterised by an edge-flame anchored at
the combustor inlet and distinct flame cells existing in the reaction channel. Before
formation of this steady-state flame structure, still-frame imaging shows (Figure 3A) a
stable edge-flame and two flame cells oscillating axially along the length of the reaction
channel (Regime B, as discussed later). Closer investigation of the flame dynamics using
Flame dynamics in sub-millimetre combustors
331
high-speed imaging reveals that the flame cells are not oscillating along the length of
the combustor reaction channel but are characterised by rapid extinction–ignition events.
The ‘oscillatory’ flame structure observed in still-frame imaging is due to the relatively
large integration time of the still-frame camera. The high-speed camera images presented
in Figure 3B were captured at 500 frames/s at an exposure time of approximately 1800
µs. The image sequence in Figure 3B shows that the flame structure actually consists of
a dynamic structure, a ‘travelling’ flame cell that marks the extinction–ignition event.
This cell propagates along the length of the combustor channel beginning with an ignition
event near the exhaust, travels toward the inlet and undergoes extinction near the
anchored edge-flame. This ignition–extinction phenomenon of the flame cell repeats
itself at a frequency of approximately 62 Hz. High-speed imaging also shows lateral
oscillations of the edge-flame at a frequency of approximately 80−90 Hz. Once the stable
flame cells form (Figure 3A2); the external diffusion flame at the combustor exhaust can
be extinguished and the edge-flame and flame cells remain.
Figure 3
Still-frame images
A1
CH4 O2
A2
CH4 O2
B
Traveling flame cell shows
ignition-extinction event
Edge-flame anchored
at combustor inlet
Note: (A1 and A2) show a flame structure that apparently depicts an oscillating flame cell. High-speed
camera images (B) at 500 frames/s show that the flame cells are not oscillating but are marked by rapid
extinction–ignition events with the ignition event occurring near the exhaust and the extinction near the
anchored edge-flame.
332
S. Prakash et al.
It has been shown previously (Feldman, 1968a,b; Majdalani et al., 2001; Raun et al.,
1993) that for Rijke tubes, locating a heat source at a quarter of the length of the tube can
lead to instabilities with maximum amplitude of the pressure waves (i.e. sound waves) in
accordance with the Rayleigh criterion (Rayleigh, 1878). Comparing the observed stable
flame structure (Figure 3A2) within these combustors to Rijke tubes, it can be observed
from Figure 3 that the edge-flame at the inlet extends from the bottom of the reaction
channel to a length of approximately 8.9 mm. The nominal dimension of the combustor
channel is 35 mm. This flame acts as a continuous heat source close (~2% difference
from the exact location) to the quarter-length criterion for ideal Rijke tubes. For other
combustors tested, a maximum variation up to 7% from the quarter-length criterion was
observed. It appears that the instabilities observed in terms of the emitted acoustics are
similar to those observed in traditional Rijke tubes. Furthermore, the Strouhal number
(St) is less than 0.1, suggesting that the oscillatory instabilities observed in the flame are
hydrodynamically driven (Crow and Champagne, 1971).
3.2 Acoustic emission
The sub-millimetre combustors investigated here also emit notable acoustics during the
initial start-up. The voltage signal obtained as a function of time from the microphone
shows distinct acoustic regimes. Two different sounds are heard. Shortly after ignition, a
‘buzzing’ type sound is heard (Regime A) which transitions to a sharp whistle-like sound
(Regime B) after a few seconds of combustor operation. As the fuel continues to burn
within the combustor, a brief transition back to the buzzing sound occurs just before the
formation of the steady-state flame structure. One caveat with respect to the acoustic
emission is that the time in which each regime starts and/or stops changes with different
combustors, flow rates of reactants, and thermal boundary conditions, indicating that the
instabilities in these sub-millimetre combustors can be functions of several different
parameters. However, these acoustic regimes occur in all cases with the general
behaviours being reproducible and repeatable.
For a better understanding of the flame dynamics, the frequencies emitted in Regimes
A and B, which contained notable acoustic emission, were evaluated. Regime A is
distinguished by a ‘buzzing’ type sound indicating the presence of low-frequency
components, while Regime B is a characterised by a whistle-like sound with a dominant
frequency near 700 Hz. The acoustic data is obtained on a windowless all-alumina
combustor that is insulated on all sides. Hence, the time taken for flame stabilisation is
different from that of the sapphire window combustors used for imaging the flame
dynamics. However, the general characteristics exhibited by the combustors are the
same as has been discussed previously. Figure 5 shows the fast Fourier transform (FFT)
power spectrum of the acoustic signal. Figure 5A shows a sharp impulse followed by the
ringing structure with a frequency of approximately 65 Hz. Earlier, the high-speed
imaging data (Figure 3) had shown that the frequency of the ignition–extinction
phenomenon of the travelling flame cell is on the order of 62 Hz. These two frequencies
agree within 3 Hz. Thus, the acoustic data supports the high-speed imaging data with the
extinction–ignition event frequency showing good agreement (within 5 Hz) between
the two characterisation techniques employed here to evaluate the flame dynamics.
Figures 3B and 5A correspond to the time event in Regime B. Figure 5A also shows the
1/f drop-off at frequencies less than about 10 Hz. The FFT power spectrum for the higher
frequency components observed during Regime A can be observed in Figure 5B. A broad
Flame dynamics in sub-millimetre combustors
333
decaying distribution, with a peak near 2500 Hz, shows the strongest signal in this
Regime for the higher frequency components (>500 Hz). In addition, significant structure
in the power spectrum from 500 to 1500 Hz is also seen in Figure 5B.
In Regime A, it can be seen in Figure 4 that large temperature gradients can exist
along the combustor wall in the axial direction. For instance, approximately 60 s after
ignition, the wall temperature difference between the top TC and bottom TC (2 cm gap)
is greater than 200 K (gradients on the order of 100 K/cm). In Regime B, the dominant
frequency is near 712 Hz and the harmonics of this frequency are also observed. To
verify that the higher frequencies are indeed harmonics, and are not aliases of other
higher frequency modes, data collected up to 10000 Hz (20 kHz sampling rate) does not
show any other frequencies other than harmonics between 5000 and 10000 Hz.
Furthermore, this dominant frequency in Regime B (~700 Hz) could not be observed
in the high-speed images up to 1000 frames/s. Higher frame rates could not be employed
as the intensity of the flame was too low at the low exposure times required for
faster imaging.
Figure 4
A representative plot of the combustor external wall temperature profile
1000
Flame Stabilization
Top TC
Middle TC
Bottom TC
900
Onset of Regime B
Temperature (K)
800
700
600
Ignition, Onset
of Regime A
500
400
300
0
30
60
90
120
150
180
210
240
270
300
Time (sec)
Note: Thermocouples are attached approximately 0.5 cm (top TC), 1.5 cm (middle TC) and 2.5 cm
(bottom TC) from the combustor exhaust. The total flow rate is 300 sccm with 100 sccm each of CH4 and
200 sccm of O2. Different acoustic regimes are demarcated in the plot. Regime A: buzzing sound.
Regime B: whistle-like sound.
Higher frequency components that are not directly observed from flame imaging data are
seen in the FFT analysis (Figure 5B). One hypothesis that explains the existence of these
higher frequency (>500 Hz) components is that there is some complex interaction
between the flame and the combustor structure. This interaction can be explained as
follows. Consider the combustor channel as a pipe closed at one end. In this mode,
similar to an organ pipe, the fundamental frequency for the combustor reaction channel is
on the order of 2400 Hz for only oxygen flowing in the channel and on the order of
2800 Hz for only methane flowing at room temperature. During the combustion process
within the combustor gas temperatures are much higher than room temperature and since
the speed of sound is proportional to the square root of gas temperature, at the higher gas
334
S. Prakash et al.
temperatures the average fundamental mode should be on the order of 3700−4500 Hz.
Furthermore, the speed of sound in CH4 can be greater by almost 25% than that in O2, it
is expected that multiple modes could be present due to the speed of sound in methane,
oxygen and products, all potentially at different temperatures. In addition, it is possible
that some interference may occur between the different acoustic modes. As the higher
frequency components are not a direct outcome of the flame dynamics, these are not
observed in flame dynamics imaging. Figure 5C shows the FFT for time event after
ignition in Regime A. This FFT shows that the dominant frequencies are near 90 and
140 Hz. These frequencies appear to be those of the lateral oscillation frequency of the
edge-flame and the individual flames at early times after ignition, as discussed earlier,
which appear to be moving rapidly from the edge-flame anchored at the combustor inlet
toward the external unconfined diffusion flame at the combustor exit (Figure 2).
Figure 5A The FFT of the acoustic signal in Regime B after ignition is seen
(A)
1/f drop-off
Extinction ignition
frequency
Note: The important frequency to note is approximately at 65 Hz corresponding to the extinction–ignition
event of the flame cell.
Figure 5B Higher frequency components arising due to the complex interaction between the flame
and the combustor structure are shown
(B)
Combustor
structural
frequencies
Flame dynamics in sub-millimetre combustors
335
Figure 5C The lateral oscillation of the edge-flame anchored at the combustor inlet is also seen
(C)
Frequency of individual
flame in Fig. 2
Lateral oscillation of the
anchored edge- flame
Note: Total flow rate: 300 sccm with equal CH4 and O2
3.3 Inlet flow conditions
Different flow conditions are evaluated to elucidate the effect of flow rate and inlet
stoichiometry on flame dynamics. Two different flow conditions are considered. In the
first set of experiments, the total flow rate is kept constant and the inlet ratio of the fuel
and oxidiser, that is the inlet equivalence ratio, φ (in this paper φ = 0.67−4 are
investigated), are altered with respect to one another. The same trends are displayed by
the flame dynamics and external wall temperatures for all experimental conditions.
However, the one observed difference was in the time taken for the formation of
the stable flame structures. As the time taken for flame stabilisation increases, the
measured wall temperatures also increase. Comparing conditions with total flow
rate of 250 sccm ( ReCH4 ~ 34 and ReO2 ~ 52 ), 300 sccm ( ReCH4 ~ 41 and ReO2 ~ 62 ),
and 400 sccm ( ReCH4 ~ 55 and ReO2 ~ 82 ) with methane–oxygen flow rates being
equal, the time taken for the subsequent formation of flame cells decreases with
increasing total flow rates. The Reynolds (Re) estimates are based on a characteristic
length of 0.75 mm and gas properties are estimated at room temperature. For instance, as
the total flow rate increases from 250 to 400 sccm, the time taken for formation of flame
cells decreases from about 455 to 86 s.
4
Conclusions
The experimental results presented in this paper demonstrate that a physically oscillating
flame accompanied by ignition/extinction events and which emits sound, leads to
formation of stable flames comprised of distinct burning zones in the sub-millimetre
combustors. High-speed imaging and acoustic data show that the flame dynamics are
distinguished by a frequency near 62 Hz due to the extinction/ignition phenomenon of the
flame cells. The inlet equivalence ratio and relative shear between fuel and oxidiser
streams do not appear to be the primary factors affecting flame stabilisation but the flame
336
S. Prakash et al.
dynamics are affected by the total flow rate, where increasing flow rates of the reactants
leads to faster formation of stable flame structures. Both low-frequency components
(<200 Hz) characteristic of the flame dynamics to several higher frequency components
(>500 Hz) probably due to the interaction between reacting flow and the combustor
structure are found to exist in these non-premixed sub-millimetre combustors.
Acknowledgements
The current work was supported by the Department of Defense Multidisciplinary
University Research Initiative (MURI) programme administered by the Army Research
Office under grant DAAD19-01-1-0582. The authors thank Jeahyong Han for help with
acoustic data collection, Jessica Mullen, Robert Covderdill, Prof. C.F. Lee and Prof. D.C.
Kyritsis for assistance during high-speed imaging. The authors also thank Prof. D.C.
Kyritsis for several fruitful discussions. The authors acknowledge Dillon Yoshitomi for
assistance during combustor fabrication.
References
Aichlmayr, H.T. (2002) Design Considerations, Modeling, and Analysis of Micro-Homogeneous
Charge Compression Ignition Combustion Free-Piston Engines, University of Minnesota.
Crow, S.C. and Champagne, F.H. (1971) ‘Orderly structure in jet turbulence’, Journal of Fluid
Mechanics, Vol. 48, No. 3, pp.547–591.
Dunn-Rankin, D., Leal, E.M. and Walther, D.C. (2005) ‘Personal power systems’, Progress in
Energy and Combustion Science, Vol. 31, pp.422–465.
Feldman Jr., K.T. (1968a) ‘Review of literature in Rijke thermoacoustic phenomena’, Journal of
Sound and Vibration, Vol. 7, pp.83–89.
Feldman Jr., K.T. (1968b) ‘Review of the literature on Sondhauss thermoacoustic phenomena’,
Journal of Sound and Vibration, Vol. 7, pp.71–82.
Fernandez-Pello, A.C. (2002) ‘Micro-power generation using combustion: issues and approaches’,
Proceedings of the Combustion Institute, Vol. 29, pp.883–899.
Gonzalez, M. (1996) ‘Acoustic instability of a premixed flame propagating in a tube’, Combustion
and Flame, Vol. 107, No. 3, pp.245–259.
Henrich, V.E. and Cox, P.A. (1994) The Surface Science of Metal Oxides, Cambridge University
Press, Cambridge.
Jensen, C.D. (2000) The Dependence of Flame/Wall Interactions on the Composition of the Walls
for Determining the Material Composition of a Micro-Combustor, Department of Chemical
and Biomolecular Engineering, University of Illinois, Urbana.
Kim, N.I., Kato, S., Kataoka, T., Yokomori, T., Maruyama, S., Fujimori, T. and Maruta, K. (2005)
‘Flame stabilization and emission of small swiss-roll combustors as heaters’, Combustion and
Flame, Vol. 141, No. 3, pp.229–240.
Kukuck, S. and Matalon, M. (2001) ‘The onset of oscillations in diffusion flames’, Combustion
Theory and Modelling, Vol. 5, No. 2, pp.217–240.
Kyritsis, D.C., Coriton, B., Faure, F., Rouchoudhury, S. and Gomez, A. (2004a) ‘Optimization of a
catalytic combustor using electrosprayed liquid hydrocarbons for mesoscale power
generation’, Combustion and Flame, Vol. 139, Nos 1–2, pp.77–89.
Kyritsis, D.C., Roychoudhury, S.C., McEnally, E.S., Pfefferle, L.D. and Gomez, A. (2004b)
‘Mesoscale combustion: a first step towards liquid fueled batteries’, Experimental Thermal
and Fluid Science, Vol. 28, No. 7, pp.763–770.
Flame dynamics in sub-millimetre combustors
337
Majdalani, J., Entezam, B. and Moorhem, W.K.V. (2001) ‘The Rijke tube revisited via laboratory
and numerical experiments’, 35th AIAA Thermophysics Conference, Anaheim, CA.
Masel, R.I. and Shannon, M.A. (2001) Microcombustor Having Submillimeter Critical Dimensions,
USA. 6193501.
Mehra, A., Zhang, X., Ayon, A.A., Waitz, I.A., Schmidt, M.A. and Spadaccini, C.M. (2000) ‘A
six-wafer combustion system for a silicon micro-gas turbine engine’, Journal of
Microelectromechanical Systems, Vol. 9, No. 4, pp.517–527.
Miesse, C.M. (2005) Development of Microcombustors and Characterization of Confined SubMillimeter Laminar Diffusion Flames, Department of Chemical and Biomolecular
Engineering, University of Illinois, Urbana.
Miesse, C.M., Masel, R.I., Jensen, C.D., Shannon, M.A. and Short, M. (2004) ‘Submillimeter-scale
combustion’, AIChE Journal, Vol. 50, No. 12, pp.3206–3214.
Miesse, C.M., Masel, R.I., Short, M. and Shannon, M.A. (2005a) ‘Diffusion flame instabilities in a
0.75 mm non-premixed microburner’, Proceedings of the Combustion Institute, Vol. 30,
pp.2499–2507.
Miesse, C.M., Masel, R.I., Short, M. and Shannon, M.A. (2005b) ‘Experimental observations
of methane–oxygen diffusion flame structure in a sub-millimetre microburner’, Combustion
Theory and Modelling, Vol. 9, No. 1, pp.77–92.
Miller, F.J., Dietrich, D.L., Struk, P., Tien, J.S. and Mellish, B.P. (2005) ‘Premixed flames
stabilized on or spreading inside microtubes’, Proceedings of the Fourth Joint Meeting of the
US Sections of the Combustion Institute, Philadelphia, PA, 20–23 March.
Moore, G. (2000) Failure of a Thin-Film Mullite Synthesis Due to Carbothermic Reduction by the
SiC, Department of Mechanical and Industrial Engineering, University of Illinois, Urbana.
Nicoud, F. and Poinsot, T. (2005) ‘Thermoacoustic instabilities: should the Rayleigh criterion be
extended to include entropy changes?’, Combustion and Flame, Vol. 142, pp.153–159.
Norton, D.G. and Vlachos, D.G. (2005) ‘Hydrogen assisted self-ignition of propane/air mixtures in
catalytic microburners’, Proceedings of the Combustion Institute, Vol. 30, pp.2473–2480.
Prakash, S., Glumac, N.G., Shankar, N. and Shannon, M.A. (2005a) ‘OH concentration profiles
over alumina, quartz, and platinum surfaces using laser-induced fluorescence spectroscopy in
low-pressure hydrogen/oxygen flames’, Combustion Science and Technology, Vol. 177, No. 4,
pp.793–817.
Prakash, S., Miesse, C.M., Masel, R.I. and Shannon, M.A. (2005b) ‘Flame structure variations in
sub-millimeter combustors due to heat transfer through combustor walls’, Proceedings of the
Fourth Joint Meeting of the US Sections of the Combustion Institute, Philadelphia, PA,
20–23 March.
Prakash, S., Armijo, A.D., Masel, R.I. and Shannon, M.A. (2006) ‘Unsteady flames in
microcombustion’, Proceedings of the ASME International Mechanical Engineering Congress
and Exposition, Chicago, IL, 5–10 November.
Raun, R.L., Beckstead, M.W., Finlinson, J.C. and Brooks, K.P. (1993) ‘A review of Rijke tubes,
Rijke burners and related devices’, Progress in Energy and Combustion Science, Vol. 19,
No. 4, pp.313–364.
Rayleigh, J.W.S. (1878) ‘The explanation of certain acoustical phenomena’, Nature, Vol. 18,
pp.319–321.
Richecoeur, F. and Kyritsis, D.C. (2005) ‘Experimental study of flame stabilization in low
Reynolds and Dean number flows in curved mesoscale ducts’, Proceedings of the Combustion
Institute, Vol. 30, pp.2419–2427.
Ronney, P.D. (2003) ‘Analysis of non-adiabatic heat-recirculating combustors’, Combustion and
Flame, Vol. 135, No. 4, pp.421–439.
Schaevitz, S.B., Franz, A.J., Jensen, K.F. and Schmidt, M.A. (2001) ‘A combustion-based MEMS
thermoelectric power generator’, The 11th International Conference on Solid-State Sensors
and Actuators, Munich, Germany, 10–14 June, Vol. 1, pp. 30–33.
338
S. Prakash et al.
Shannon, M.A., Moore, G.V., Ganley, J., Miesse, C., Rice, C., Seebauer, E.G. and Masel, R.I.
(2002) ‘High-temperature microcombustion-based ammonia microchemical hydrogen
generator reactors for PEM fuel cells’, Technical Digest Solid-State Sensors and Actuators
Workshop, Hilton Head Island, SC, 3–7 June, pp.27–30.
Thatcher, R.W. and Dold, J.W. (2000) ‘Edges of flames that do not exist: flame-edge dynamics in a
non-premixed counterflow’, Combustion Theory and Modelling, Vol. 4, pp.435–457.
Thatcher, R.W., Omon-Arancibia, A.A. and Dold, J.W. (2002) ‘Oscillatory flame edge
propagation, isolated flame tubes and stability in a non-premixed counterflow’, Combustion
Theory and Modelling, Vol. 6, pp.487–502.
Tyndal, J. (1897) Sound, D. Appleton and Company, New York.
Vican, J., Gajdeczko, B.F., Dryer, F.L., Milius, D.L., Aksay, I.A. and Yetter, R.A. (2002)
‘Development of a microreactor as a thermal source for microelectromechanical systems
power generation’, Proceedings of the Combustion Institute, Vol. 29, pp.909–916.
Yeom, J., Mozsgai, G.Z., Flachsbart, B.R., Choban, E.R., Asthana, A., Shannon, M.A. and Kenis,
P.J.A. (2005) ‘Microfabrication and characterization of a silicon-based millimeter scale, PEM
fuel cell operating with hydrogen, methanol, or formic acid’, Sensors and Actuators B, Vol.
107, No. 2, pp.882–891.