Journal of Thermal Vol. 6, No.

2, 2011

Science and
Technology
Visualization and Emission Spectra of Flames in
Combustion Chamber of
Swirling-Oxidizer-Flow-Type Hybrid Rocket
Engines*
Saburo YUASA**, Tatsuya IDE**, Miho MASUGI**, Takashi SAKURAI**,
Noriko SHIRAISHI** and Toru SHIMADA***
**Department of Aerospace Engineering, Tokyo Metropolitan University,
6-6 Asahigaoka, Hino, Tokyo 191-0065, Japan
E-mail: syuasa@sd.tmu.ac.jp
***Institute of Space and Astronautical Science, JAXA,
3-1-1 Yoshinodai, Chuo, Sagamihara, Kanagawa 252-5210, Japan

Abstract
In order to clarify flame behavior in the combustion chamber of
swirling-oxidizer-flow-type hybrid rocket engines, a small hybrid rocket engine
with a large quartz glass window was fabricated to observe the inside of the
combustion chamber from the front directly. PP and PMMA fuel grains were
burned under the swirling and non-swirling oxygen flow conditions. For both PP
and PMMA combustion, the disturbed swirling flames were found to develop closer
to the grain surface than those without swirl, resulting in an increase of the fuel
regression rates. The swirling flames consisted of not a single large flame but an
annular aggregate of small striated flames on the grain surface. Emission spectra of
counter-diffusion flames over the PP and PMMA rods showed that the strong
continuum due to hot carbon particles in the flames, and the weak C2 and CH
emissions were also visible. The filtered photographs which took the continuum
and C2 emissions in the hybrid rocket combustion chamber suggested that the
gas-phase reactions of PMMA with oxygen substantially occurred not in the center
region of the combustion chamber, but near the PMMA gain surface. For PP
combustion, the gas-phase reactions with oxygen may also occur in some degree
near the central region.

Key words: Hybrid Rocket Engine, Swirling Flow, Fuel Regression Rate,
Centrifugal Force, Striated Flame, Visualization, Emission Spectra,
Continuum and C2 Emission, Polymethylmethacrylate(PMMA),
Polypropylene(PP)

*Received 28 Mar., 2011 (No. 11-0222)
[DOI: 10.1299/jtst.6.268]
Copyright © 2011 by JSME
1. Introduction
In hybrid rocket engines, a major drawback is low fuel regression rates. In order to
increase the fuel regression rate, the authors proposed a unique technique using a swirling
oxidizer injection at the head of fuel grain, and showed that the swirling oxidizer injection
increased the overall fuel regression rates of polymethylmethacrylate (PMMA) up to about
3 times in comparison with a non-swirling oxidizer injection(1, 2). The swirling effect on the
fuel regression rate was also found to differ with fuel types; PMMA and polypropylene
(PP)(3). However, the increase mechanism of the fuel regression rate and the different
swirling effect between fuel types are not clear yet.
Visualization and emission spectra measurement of flames are useful to understand the

268
Journal of Thermal Vol. 6, No. 2, 2011
Science and Technology

combustion mechanism in the combustion chamber. Fortunately, the swirling oxidizer

injection may cool the injector head of the combustion chamber of hybrid rocket engine.
Thus, we made a new combustion chamber of the swirling-oxidizer-flow-type hybrid rocket
engine with a large quartz window to observe flames in the combustion chamber with
swirling oxidizer flow. The purpose of this paper is to obtain new information about the
flame behavior and emission spectra in the swirling-oxidizer-flow-type hybrid rocket engine
by direct observation using a visualization window.

2. Experimental Apparatus
Figure 1 shows a schematic of the hybrid rocket engine with a large quartz glass
window at the injector head, and of the emission spectra measuring system. This hybrid
rocket engine system was fundamentally the same in configuration as the previous
engines(2-4). The engine consisted of a swirling-oxidizer-flow-type injector, a fuel grain and
a graphite nozzle. The oxidizer was gaseous oxygen (GOX). In this study, we used two
types of oxygen injector with and without swirl. Each of the injectors had eight oxygen
injection slits. The slits with swirl were turned to the tangential direction to make a swirling
flow (Sg=19.4, where Sg is the geometrical swirl number(5)), and the slits without swirl were
turned radially (Sg=0). Fuel grains made of PMMA and PP were used, and had a single port
with an inner diameter of 40 mm and a length of 200 mm. The nozzle throat diameter was
determined to burn the propellant under predetermined conditions for the equivalence ratio
and the combustion chamber pressure. The rocket engine was ignited by a small amount of
black powder for about 0.5 s after GOX was supplied.
Hybrid rocket engines have the peculiar problem of even when an oxygen flow rate, a
nozzle throat area and a grain configuration are given, the equivalence ratio cannot be
precisely controlled before burning. This is because the fuel regression rate depends
strongly on the heat transfer rate from turbulent diffusion flames in the combustion chamber
that is controlled by the oxygen flow condition on the fuel grain surface and the combustion
efficiency. Therefore, the fuel mass flow rate will be determined by consumed fuel data
after a burning test. Unfortunately, in this study, although we made two nozzles with a
different diameter, the equivalence ratio as predicted was not necessarily controlled.
The experimental feature of this engine had a large quartz glass window to observe
flames in the combustion chamber directly. In this experiment, three types of visualization
technique were tried to investigate the burning behavior in the chamber through the
window; direct visualization of flames was taken by a video camera and a high-speed video
camera (framing rate: 500 fps), three filter photographs of the same flame were taken
simultaneously through filters corresponding to CO, C2 and continuum emissions (CO

Thermocouple Grain (PP,PMMA)

Quartz glass
GOX Refractory Thermocouple
φ7.2, φ8.5

Digital camera (Graphite) Pressure port

High-pass Video
Filter C2 Filter
φ60
φ40

Camera

CO Filter Half mirror

200
Graphite nozzle
Thermocouple
Swirler injector (Sg=0,19.4) Unit : [mm]

Fig. 1 Swirling-oxidizer-flow-type hybrid rocket engine with a quartz window to

observe flames and a three- photographs simultaneously taking system of
flame emission.

269
Journal of Thermal Vol. 6, No. 2, 2011
Science and Technology

emission=501 nm; half-width=9.1 nm, C2 emission=516 nm; half-width=1.7 nm, the

continuum emission >580 nm), and the flame spectra were measured using a quartz fiber
spectroscopic system.
In order to obtain fundamental data regarding oxygen/PMMA and PP combustion, an
additional experiment was carried out using a simple counter-flow diffusion flame using 3
mm diameter PMMA and PP rods, surrounded by Ar flow, situated against an O2 stagnant
flow at room temperature and atmospheric pressure. The detail of the experimental setup
was described elsewhere(1).

3. Results and Discussion

3.1 Burning Behavior
Figure 2 shows typical video photographs of burning PP and PMMA grains under swirl
and non-swirling flow conditions taken from the front at about 1 s after ignition. The
location of the nozzle throat is designated as white dots.

(a) (b) (c) (d)

Fuel PP PMMA PP PMMA

O2 flow Swirling Non-swirling
Goitm
11.9 10.1 12.7 13.5
[kg/(m2･s)]
ϕ [-] 1.62 1.47 0.81 0.42

Fig. 2 Flame appearances with swirl and without swirl in the combustion chamber:
PP and PMMA grains.

(a) PP grain
Figure 2(a) is the appearance of a typical swirling PP combustion. It was observed that
extremely turbulent swirling flames were formed close to the grain surface, but they seemed
not to exist in the center region of the combustion chamber. The luminous flame zone
thickness on the gain surface was in the range of 8~12 mm. A part of the flames,
maintaining swirling motion, flowed into the nozzle throat. Figure 3 shows a video picture

Grain edge (Nozzle side) Throat

Flames

Grain edge (Injector side)

Fig. 3 Burning appearance of PP grain with swirl in the combustion chamber:

Pc=1.0[MPa], Goitm=18.5[kg/(m2･s)], φ=1.19, Exposure time=1/8000[s].

270
Journal of Thermal Vol. 6, No. 2, 2011
Science and Technology

of the swirling PP combustion taken from a slanted angle. It is interesting to note that the
disturbed swirling flames consisted of not a single large flame but an aggregate of small
striated flames on the grain surface. In swirling-oxidizer-flow-type hybrid rocket engines,
this type of small striated flames always appeared. This flame appearance was clearly
confirmed by the side view of the engine made of a transparent PMMA combustion
chamber and a fuel grain as shown in Fig.4. The oblique directions of the striated small
flames coincide with the swirl direction of the O2 flow over the whole combustion chamber,
which suggests that a reverse flow of combustion gas did not occur near the burning fuel
grain surface. The aggregate of small striated flames was also confirmed by the high speed
video pictures shown in Fig.5. The small striated flames, adhering to the burning grain
surface, moved along the swirling flow direction. Corresponding to this behavior, the
post-fired and quenched grain surface was found to be covered with innumerable
depressions developed toward the swirl direction of the O2 flow as a whole. These results
imply that the small striated flames formed near the depressions which acted as small flame
holders while regressing themselves.

Flames

Fig. 4 Flame appearance with swirl in the combustion chamber taken from the side view
using transparent PMMA grain and engine case:
Grain length=150[mm], Sg=19.4, Goitm=15.3[kg/(m2 ･ s)], ϕ=0.77, Exposure
time=1/10000 [s].

Flames
0

Grain edge
(Injector side) 25
[mm]

Fig. 5 Burning appearance of PP grain with swirl in the combustion chamber taken by
high-speed video:
Pc=1.0[MPa], Goitm=11.9[kg/(m2･s)], ϕ=1.44, Exposure time:1/500[s].

(b) PMMA grain and difference from PP grain

Figure 2(b) shows the appearance of a typical swirling PMMA combustion. The
burning behavior of a PMMA grain was observed to be fundamentally similar to that of a
PP grain when the equivalence ratio was almost the same. However, the luminous flame
zone thickness on the grain surface was in the range of 4~8 mm. The thickness was
approximately half as wide as that of PP, and the weak luminous region of PMMA in the

271
Journal of Thermal Vol. 6, No. 2, 2011
Science and Technology

center was wider than that of PP.

There are some differences of burning behavior between PMMA and PP. One is related
with the overall fuel regression rate. Figure 6 shows the regression rate formula as rover =a
Goitm n for PMMA and PP grains in our laboratory using swirling oxidizer flows for all the
experimental parameters tested, where rover is the overall fuel regression rate, Goitm is the
intermediate oxygen mass flux(3), a and n are experimental coefficients. From the analysis
the exponent n is found to depend on the fluid dynamics and reaction kinetics in boundary
layer hybrid combustion(6). The reduction in n occurs for kinetically limited cases, and in
the limit, approaches zero(6). The n value of PMMA was larger than that of PP in our
experiments. In particular, the nth power on Goitm of 0.789 for PMMA grains was close to
the n value of the analytical equation, 0.8, based on the diffusion-limited boundary layer
combustion(6, 7).

Fig. 6 Relation between time-averaged overall fuel regression rate and intermediate
oxygen mass flux for PMMA and PP grains under various conditions.

The second is related with an adhesion of carbon on the grain surface. Figure 7 shows
appearances of PMMA and PP grains after combustion. A smaller amount of carbon adhered
to the PMMA grain than the PP grain. This may be due to the higher decomposition ability
of PMMA than PP which is due to the lower values of the melting and vaporization heats
and the decomposition point of PMMA than those of PP(8, 9). In addition, the equilibrium
species of the combustion gas with oxygen show that PP generates carbon more easily than
PMMA at high equivalence ratios; for example, carbon is formed at the equivalence ratio φ
> 3.1 for PP, and at φ > 4.0 for PMMA(9, 10). In fact, PP flames were observed to be more
luminous than PMMA flames in the combustion chamber due to hot carbon particles.
The third is the C* efficiency. Our previous experiments showed that PMMA grains
burned with the C* efficiencies of about 1.0, and PP grains with the C* efficiencies of
0.8-0.9(3). This suggested that complete combustion proceeded in PMMA-fueled hybrid
rocket engines, but incomplete combustion in PP-fueled engines occurred.
These experimental results with respect to the burning behavior of PP and PMMA gains
imply that the gas-phase reaction rate of PMMA with oxygen was faster than that of PP.

3.2 Effects of Swirling Flow

Figure 2 (c) shows the appearance of a typical non-swirling PP combustion. In the case

272
Journal of Thermal Vol. 6, No. 2, 2011
Science and Technology

Fuel PP PMMA
Grain Length, L [mm] 150
t0 [s] 5.4 5.1
・
m0 [g/s] 25.8 13.0
ϕ [-] 1.19 1.24

Leading edge

Rear end

Fig. 7 Grain appearances of leading edges and rear ends after combustion for PP and
PMMA.

of the non-swirling O2 injection, short diffusion flames appeared which attached to the O2
injection slits. The non-swirling flames connected with these diffusion flames, and formed
further away from the grain surface than the swirling flames. In addition, a stronger
emission than the case of swirling injection appeared in the center region of the grain. This
flame behavior was observed at a lean fuel condition of φ=0.81. Thus, for φ > 1.0, the
emission should increase and become more luminous by the generation of carbon in the
flames due to a lack of oxygen concentration. These results implied that the gas-phase
reactions actively occurred in the center region for the non-swirling flow. In this
photograph, eight black areas appeared on the quartz glass window surface. They might be
attributed to carbon particles, decomposing in the processes of vaporization and combustion
of PP even under the overall fuel lean condition, adhered on the quartz window due to an
entrainment from the non-swirling main reverse flow.
On the other hand, in the case of the non-swirling O2 injection for PMMA shown in
Fig.2 (d), the flames formed in the combustion chamber were found to be closer to the grain
surface than the PP flames without swirl. Even under fuel lean conditions, the flames of
PMMA without swirl (Fig.2 (d)) were also thicker than those with swirl (Fig.2 (b)) as well
as the PP flames without swirl.
The flame observation obtained here clearly showed that the centrifugal force due to the
swirling motion of the O2 injection brought flames close to the grain surface, resulting in an
increase of the fuel regression rates.

3.3 Emission Spectra of PMMA and PP combustion

(a) Emission of counter-diffusion flame

When the counter oxygen flow velocity was sufficiently high at 10 m/s, the spectra of a
counter-diffusion flame in a semi-spherical shape on a PMMA rod with oxygen showed that
the emissions of OH, CH and C2 bands were visible, but the continuum emission of carbon
was not observed obviously(1). The emission of C2 and CH bands is a general feature of the

273
Journal of Thermal Vol. 6, No. 2, 2011
Science and Technology

spectra of hydrocarbons burning with oxygen(11). In a diffusion flame, hydrocarbon is

pyrolysed to carbonaceous soot in a region where the concentration of molecular oxygen is
negligible. On the oxygen side, the C2 and CH emission occurs(11). The author mentions that
the CH emission would come from the C2. It is also reported that the C2 and CH radiation
becomes very weak where hot carbon particles are present. The reason for these phenomena
is not clear yet. However, as a fact in our experiment, when the counter oxygen flow
velocity decreased down to 5 m/s, Fig. 8 (a) shows that the continuous emission of the
counter-diffusion flame on the PMMA rod became very strong due to an increase of hot
carbon particles and the C2 and CH emission became weak, but CO band was not observed.
This feature coincides with the spectra of the flat diffusion flame of C2H4 with oxygen(11). A
counter-diffusion flame of a PP rod shown in Fig.8(b), forming higher carbon concentration
than PMMA, also showed that the C2 and CH emission became significantly weaker than
that for PMMA. One of the reasons for this might be due to that the high amount of carbon
particles condensed at high concentrations in the flame physically blocked emission paths
of the C2 and CH bands.
Emission Intensity [Arbitrary Unit]

500
Na 588.9
Na 588.9
589.5 589.5
400
CH 431.5 C2 563.6 C2 563.6
300
C2 516.5
200 C2 516.5
C2 436.5 C2 473.7
CH 431.5 C2 473.7
100 CH 387.1
C2 436.5
CH 387.1

0
350 400 450 500 550 600 350 400 450 500 550 600
Wave Length [nm] Wave Length [nm]
(a) PMMA/O2 (b) PP/O2
Fig. 8 Emission spectra of PMMA/O2 and PP/O2 counter flow diffusion flame:
XO2=1.0, VO2=5 m/s.

Figure 9 shows the time history of emission spectrum of a burning PMMA grain with
identified characteristic bands taken from the front of the hybrid rocket engine. The CO
bands, the atomic lines of K and Ca, and the continuum emission were visible at 0.1 s after
ignition. The spectra of CO, K and Ca emissions were generated by the compositions of
black powder of the igniter. The continuum emission was attributed to thermal emission of
hot carbon particles produced during reactions of PMMA with oxygen on the fuel surface
Emission Intensity [Arbitrary Unit]

500
t= 0.1 s 550.8
t= 3.0 s CO 553.3
400 555.4
495.9
300 CO 497.9
499.6
K 404.4 Ca 422.7
200

100 Na 588.9
589.5
0
350 400 450 500 550 600
Wave Length [nm]
Fig. 9 Time-history of emission spectra of burning PMMA grain in the combustion
chamber taken from the front:
Pc=0.9 [MPa], Goitm=14.7[kg/(m2･s)], ϕ=1.03.

274
Journal of Thermal Vol. 6, No. 2, 2011
Science and Technology

and in gas phase in the combustion chamber. However, these emissions, except the
continuum, disappeared from the stably burning PMMA grain at t=3.0 s. The characteristic
spectra of CH and C2 bands were also invisible. It is noted that the continuum emission had
two broad peaks at about 450 and 580 nm. Whether these correspond to C2 and CH or not is
unclear. This requires further study.

(b) Emission from flames in combustion chamber

Our previous experiment for PMMA grains showed that the CO emission spectra were
visible in the combustion chamber of hybrid rocket engines. So, in this experiment, we also
used the CO band-pass filter in addition to the continuum and C2 filters. When selecting the
C2 wavelength for the C2 band-pass filter, the wavelength was selected so as not to overlap
with the CH emission. Figure 10 shows the filtered photographs for PMMA and PP
combustion taken from the front. In these photographs, it is noted that the C2 and CO
emission intensities passing through the filter for the C2 and CO bands included the
emission intensity of the continuum emission as shown in Fig.7. In addition, unfortunately,
although we tried to take the photographs at one time at a shutter speed of 1/4000, some
time lag occurred due to the low performance of our camera system.
For PMMA combustion, the continuum emission was very strong near the grain
surface, but almost disappeared in the wide region on the center axis. The emission
appearance composed of an annular aggregate of small striated flames, which was similar in
feature to the video and high speed video pictures. The C2-filtered photograph showed also
an aggregate of small flames. It is clear that the C2 emission extended into the center region
beyond the region of the continuum while swirling, even though the emission became very
weak. Since the oxygen concentration should be high in the center region, this behavior
observed for the C2 emission may be in agreement with diffusion flames with oxygen. For
the CO-filtered photograph, the appearance of the CO emission was similar, not to the C2
emission, but to the continuum. These emissions observed here imply that, for PMMA
combustion, the gas-phase reactions with oxygen occurred substantially, not in the central

PMMA Emissions
Continuum C2 CO PMMA grain
tb = 5.0 [s]
Pc = 1.2 [MPa]
Goitm =8.0
[kg/(m2･s)]
ϕ = 1.67 [-]
Exposure time =
1/4000 [1/s]
PP Emissions
Continuum C2 CO PP grain
tb = 5.0 [s]
Pc = 0.9 [MPa]
Goitm = 11.2
[kg/(m2･s)]
ϕ = 1.84 [-]
Exposure time =
1/4000 [1/s]

Fig.10 Filtered photographs of burning grain in the combustion chamber.

Filter transmission wave length; CO emission=501 nm; half-width=9.1 nm,
C2 emission=516 nm; half-width=1.7 nm, the continuum emission > 580 nm

275
Journal of Thermal Vol. 6, No. 2, 2011
Science and Technology

region of the combustion chamber, but near the PMMA grain surface.
On the other hand, for PP combustion, the continuum was visible over a wide area of
the combustion chamber, and spread out into the center region unlike PMMA combustion.
It is clear that all the emissions of PP combustion were also composed of small aggregate
flames. The flames flowed into the nozzle throat while swirling, which constructed a
swirling structure of the C2 emission. The appearance of the CO-filtered photograph seemed
to be similar to both of the C2 and continuum appearances. The behavior of these emission
photographs suggested that the gas-phase reactions with oxygen may occur in some degree
near the central region of the combustion chamber compared with the PMMA combustion.
The combustion behavior observed from the emission spectra and filtered photographs did
not contradict the fuel regression data shown in Fig.6, and they give important data when
considering the combustion mechanism of the hybrid rocket engines for further study.

4. Conclusions
z Flames in the combustion chamber of swirling-oxidizer-flow-type hybrid rocket
engines did not develop in the center axis region, but near the grain surface for both
PMMA and PP combustion. They consisted of not a single large flame but an annulus
aggregate of small striated flames on the grain surface.
z The annulus swirling aggregate flame on the grain surface for PMMA was thinner and
developed closer to the grain surface than that for PP.
z The centrifugal force by the swirling O2 injection brought the PMMA and PP
aggregate flames closer to the grain surface, resulting in an increase of the fuel
regression rates.
z The strong continuum, due to hot carbon particles, weak C2 and CH emission spectra
were visible in counter-diffusion flames with oxygen over the PP and PMMA rods.
z The filtered photographs corresponding to the emission spectra suggest that the
gas-phase reactions of PMMA with oxygen occurred substantially near the PMMA
grain surface, but for PP combustion, some gas-phase reactions may occur near the
central region of the combustion chamber.

Acknowledgements
This research is supported by the Hybrid Rocket Research Working Group (HRrWG) of
Institute of Space and Astronautical Science, Japan Aerospace Exploration Agency. The
authors express thanks to Mr. Saito, D., Mr. Hirata, K. and other students of Yuasa lab. for
their cooperation in conducting this study.

References
(1) Yuasa, S., Shimada, O., Imamura, T., Yamamoto, K., AIAA Paper 99-2322, 35th AIAA
Joint Propulsion Conference, 1999.
(2) Yuasa, S., Yamamoto, K., Hachiya, H., Kitagawa, K., OOWADA, Y., AIAA Paper
2001-3537, 37th AIAA Joint Propulsion Conference, 2001.
(3) Yuasa, S., Sakamoto, M., Sezaki, C., Sakurai, T., Shiraishi, N., Shimada, T., 8th
International Symposium on Special Topics in Chemical Propulsion, Log#143, 2009.
(4) Sakurazawa, T., Kitagawa, K., Hira, R., Matsuo, Y., Sakurai, T., Yuasa, S., Proceedings of
the Asian Joint Conference on Propulsion and Power, D6-4, 2008.
(5) Beer, J., Chigier, N., Combustion Aerodynamics, Applied Science Pub., London, (1972),
pp.106-117.
(6) Humble, R.W., Henry, G. N., Larson, W. J., Space Propulsion Analysis and Design, The

276
Journal of Thermal Vol. 6, No. 2, 2011
Science and Technology

McGraw-hill Companies, Inc. Primis Custom Publishing, New York, (1995), pp.374-403.
(7) Marxman, G. A., Fundamentals of Hybrid Boundary Layer Combustion, In Wolfhard, H.
G., Glassman. I., Green, L. Jr., Eds., Heterogeneous Combustion, AIAA Progress in
Astronautics and Aeronautics, Vol.15, (1964), pp. 485-521.
(8) Van Krevelen, D. W., Properties of Polymers, Elsevier, (1990), pp.790-797.
(9) Mark, H. F., Gaylord, N. G., Encyclopedia of Polymer Science and Technology,
Interscience Publishers, Vol.13, (1970), pp.790-811.
(10) McBride, B.J., Gordon, S., Computer Program for Calculation of Complex Chemical
Equilibrium Compositions and Applications, NASA RP 1311, 1994.
(11) Gaydon, A. G., The Spectroscopy of Flames, Chapman and Hall, London, (1974),
pp.43-44, 253-255.

277