Ballistic properties and burning behaviour of an ammonium perchlorate/guanidine nitrate/sodium nitrate airbag solid propellant


Ulas, A.; Risha, G.A.; Kuo, K.K.

Fuel 85(14-15): 1979-1986

2006


Available
online
at
www.sciencedirect.com
SCIENCE
dDIRECT.
ELSEVIER
Fuel
85
(2006)
1979-1986
www.fuelfirst.com
Ballistic
properties
and
burning
behaviour
of
an
ammonium
perchlorate/guanidine
nitrate/sodium
nitrate
airbag
solid
propellant
A.
Ulas
a
'*,
G.A.
Risha
b
,
K.K.
Kuo
b
a
Department
of
Mechanical
Engineering,
The
Middle
East
Technical
University,
06531
Ankara,
Turkey
b
Department
of
Mechanical
and
Nuclear
Engineering,
The
Pennsylvania
State
University,
PA,
USA
Received
26
January
2006;
received
in
revised
form
22
March
2006;
accepted
23
March
2006
Available
online
24
April
2006
Abstract
An
experimental
investigation
on
the
determination
of
ballistic
properties
and
burning
behavior
of
a
composite
solid
propellant
for
airbag
application
was
conducted.
The
experimental
results
were
obtained
using
a
high-pressure
optical
strand
burner.
Steady-state
burn-
ing
rates
were
determined
for
a
pressure
range
of
20.8-41.5
MPa
and
initial
propellant
temperatures
of
—30
to
+80
°C.
For
the
pressure
and
temperature
ranges
tested,
the
temperature
sensitivity
was
on
the
order
of
1
x
10
-3
K
-1
.
The
pressure
exponent
was
found
to
be
a
function
of
the
initial
propellant
temperature
and
was
0.75
at
25
°C.
The
activation
energy
and
the
pre-exponential
factor
of
the
Arrhe-
nius
equation
are
2.735
kcal/mol
and
15.06
cm/s,
respectively.
The
pressure
deflagration
limit
for
this
propellant
was
found
to
be
in
the
range
of
8.37-8.72
MPa.
During
combustion,
small
condensed-phase
spherical
particles
were
ejected
from
the
burning
surface.
The
size
of
the
particles
decreased
with
either
increasing
the
pressure
or
the
initial
propellant
temperature.
For
pressures
below
41.4
MPa,
average
particle
size
was
on
the
order
of
900
gm,
and
at
84.4
MPa,
the
bead
size
was
much
smaller,
on
the
order
of
300
gm.
A
chemical
analysis
on
these
particles
using
both
the
ESEM
and
the
X-ray
diffraction
method
indicated
that
the
material
of
the
beads
was
mostly
sodium
chloride
with
a
small
amount
of
silicon-containing
compounds.
©
2006
Elsevier
Ltd.
All
rights
reserved.
Keywords.•
Airbag;
Solid
propellant;
Ballistic
properties
1.
Introduction
The
presence
of
airbags
in
today's
automobiles
provides
increased
protection
to
the
vehicle
operator
and
passen-
gers.
Crash
tests
showed
that
for
an
airbag
to
be
useful
as
a
protective
device,
the
bag
must
deploy
and
inflate
within
40
ms
[1].
Therefore,
a
rapid
pressurizing
system
must
be
employed
in
airbags.
Research
on
airbag
inflator
design
for
automobiles
has
been
progressing
since
the
1960's
and
different
airbag
inflation
techniques
have
been
considered
over
these
years
including
stored
gas
system,
solid-propellant
gas
generators,
and
hybrid
systems
[2].
The
stored
gas
system
is
simply
a
compressed
gas
chamber
that
is
equipped
with
a
valve
that
opens
during
collision.
In
Corresponding
author.
Tel.:
+90
312
210
5260;
fax:
+90
312
210
1266.
E-mail
address.•
aulas@metu.edu.tr
(A.
Ulas).
0016-2361/$
-
see
front
matter
©
2006
Elsevier
Ltd.
All
rights
reserved.
doi:10.1016/j.fue1.2006.03.026
the
solid-propellant
system,
a
solid-propellant
grain
is
ignited
upon
collision
and
generates
gaseous
combustion
products
to
inflate
the
airbag.
The
hybrid
airbag
inflator
system
is
a
combination
of
the
two.
It
is
initiated
with
a
pyrotechnic
igniter
to
combust
the
reactive
stored
gas
to
produce
combustion
product
gases
to
inflate
the
airbag.
Among
these,
solid-propellant
gas
generator
is
the
most
common
and
the
most
reliable.
Over
the
past
30
years,
different
solid-propellant
gas
generator
technologies
have
been
explored
to
meet
the
stringent
requirements
for
motor
vehicle
airbag
applica-
tions
[3].
Sodium-azide
based
inflators
are
still
the
most
common
in
airbag
systems
[1].
Typically,
the
grain
is
approximately
60%
sodium
azide
and
there
are
about
75-
100
g
of
propellant
in
the
driver's
side
inflation
system
[1].
This
paper
considers
the
combustion
of
an
Ammo-
nium
Perchlorate/Guanidine
Nitrate/Sodium
Nitrate
Silicon
Fluid
r
O
O
O
O
0
Thermocouple
A
1980
A.
Ulas
et
al.
1
Fuel
85
(2006)
1979-1986
solid-propellant
formulation
to
produce
combustion
prod-
uct
gases
to
inflate
an
automobile
airbag.
The
chemical
ingredients
(by
weight)
of
the
propellant
include
23.5%
Ammonium
Perchlorate
(NH
4
C10
4
),
58.5%
Guanidine
Nitrate
(CH6N403),
17.8%
Sodium
Nitrate
(NaNO
3
),
and
0.2%
silicon
dioxide
(SiO
2
).
In
this
study,
burning
behavior
of
this
propellant
was
investigated
at
different
operating
conditions.
In
addition,
ballistic
properties
such
as
the
pres-
sure
deflagration
limit
(PDL),
the
burning
rate
as
a
function
of
pressure
and
initial
propellant
temperature,
the
tempera-
ture
sensitivity,
and
the
activation
energy
of
this
propellant
were
determined
experimentally.
2.
Experimental
Experiments
were
conducted
in
a
high-pressure
optical
strand
burner.
The
operating
pressure
and
the
initial
pro-
pellant
temperature
were
varied
from
20.8
to
41.5
MPa
and
from
—30
to
+80
°C,
respectively.
A
schematic
dia-
gram
of
the
setup
is
shown
in
Fig.
1.
In
this
investigation,
nitrogen
was
used
as
the
purge
gas.
The
initial
temperature
of
the
propellant
strand
was
controlled
by
preconditioning
the
purge
gas
initial
temperature
using
either
a
cryogenic
heat
exchanger
or
an
electric
heating
unit
for
low-
and
high-temperature
experiments,
respectively.
The
initial
pro-
pellant
strand
temperature
was
determined
by
either
an
S-
or
R-type
thermocouple
whose
junction
was
embedded
in
the
propellant
strand.
A
Pulnix
CCD
camera
and
a
Super
VHS
Panasonic
VCR
were
used
for
the
real-time
recording
of
the
propellant
burning
process.
A
Setra
pres-
sure
transducer
was
used
to
record
the
instantaneous
bur-
ner
pressure.
To
have
a
constant
pressure
during
test
runs,
a
computer-controlled
solenoid
valve
was
installed
in
the
exhaust
line.
The
solenoid
valve
maintained
constant
pres-
sure
in
the
burner
with
less
than
±1%
variation.
A
Nicolet
data
acquisition
system
was
used
to
record
the
thermocou-
ple
signals
and
the
pressure
transducer
output.
The
propellant
was
received
in
the
form
of
powder
and
it
was
pressed
into
cylindrical
pellets
using
a
computer-con-
trolled
presser.
The
propellant
strands
were
pressed
using
an
extremely
detailed
pressing
procedure
to
meet
stringent
density
specifications.
Before
the
pressing,
the
propellant
powder
was
dried
at
90
°C
for
removing
any
moisture.
For
PDL
determination
tests,
short
pellets
(about
6
mm
in
length)
were
pressed
according
to
the
following
proce-
dure.
First,
the
pressure
was
increased
to
138
MPa
and
kept
constant
at
this
level
for
one
minute.
This
duration
was
sufficient
for
the
powder
to
relax
at
this
pressure.
The
pressure
was
then
increased
to
241
MPa
and
again
kept
constant
at
this
level
for
another
minute.
Finally,
the
pressure
was
increased
to
345
MPa
and
kept
constant
for
ten
more
minutes.
During
this
period,
the
pressed
pow-
der
was
fully
relaxed,
and
the
density
became
uniform
throughout
the
pellet.
For
the
burning
rate
measurements,
longer
pellets
(20-23
mm
long)
were
required
to
suppress
the
effect
of
the
transients
at
the
beginning
and
end
of
10,000
psi
Volidyne
Pressure
Gauge
Tronducer
Pressure
soin'd
0
H.V.6
H.V.5
0 0
Optical
Strand
Burner
plc
--.11111
Combustion
Product
Exhaust
Data
Acquisition
Sys
em
FTS
Constant
Temperature
Bath
TC
PID
TC
Controlle
Charge
AMP
Ca
ries
Demodulator
I.B.M.
P.C.
KZ=
O
O
O
Surge
Tank
Nichrome
wire
30,000
psi
Pressure
Gauge
TINS
30,000
psi
Reservoir
or
Gas
Bottle
Air
Purc
Dryer
Check
Valve
HA/2
Power
Supply
Purge
Exhaust
Resistance
Heater
TC
H.V.3
15,000
psi
Pressure
Gouge
Cryogenic
Heat
Exchanger
SCR
Power
Supply
Safety
Relief
Valve
H,V,4
Fig.
1.
Schematic
of
the
high-pressure
optical
strand
burner
experimental
setup.
Cup
Cone
CO
2
Laser
Beam
11
NOSOL-362
Initiator
B35
mm
Nichrome
Wire
6.35
mm
TC
Junction
Extension
Wire
38.1
mm
Propelled
Sample
Copper
Sample
Holder
A.
Ulas
et
al.
1
Fuel
85
(2006)
1979-1986
1981
the
tests.
The
same
pressing
procedure
was
used
to
press
these
long
pellets,
except
that
after
reaching
138
and
241
MPa
pressures,
it
was
kept
at
these
levels
for
5
min.
Also,
after
reaching
345
MPa,
a
30-min
interval
was
required
for
the
relaxation
of
the
pellets
based
upon
the
force
gauge
reading.
The
density
of
each
pellet
was
determined
by
measuring
the
geometry
(i.e.,
diameter
and
length)
and
mass
of
the
pellet.
A
micrometer
with
a
resolution
of
0.025
mm
and
a
micro-scale
(XE-50
Denver
Instrument)
with
a
resolution
of
ten
thousandth
of
a
gram
was
used
to
measure
the
geom-
etry
and
mass
of
the
pellet,
respectively.
Only
pellets
with
a
density
of
1.62
g/cc
or
greater,
as
specified
by
the
propel-
lant
manufacturer,
were
used
for
the
strand
burner
tests.
For
measuring
the
sub-surface
temperature
profile
and
the
surface
temperature
of
the
propellant
strand,
the
pro-
pellant
powder
was
pressed
into
a
cone
and
cup
configura-
tion
to
embed
the
micro-thermocouple
inside
the
propellant.
Following
a
procedure
given
in
[4],
thermocou-
ple
wires
25-µm
in
diameter
were
welded
and
then
rolled
into
a
flat
foil
with
a
thickness
of
approximately
8
gm.
The
thermocouple
bead
was
placed
at
the
top
of
the
cone,
and
the
cup
portion
was
then
slowly
positioned
on
the
cup
to
squeeze
the
thermocouple
bead
between
the
two
propel-
lant
surfaces.
Force
was
exerted
on
the
top
of
the
propel-
lant
strand
in
the
vertical
direction
to
slightly
deform
the
thermocouple
bead
to
match
the
contour
of
the
propellant
interface.
A
schematic
of
the
sample
configuration
is
shown
in
Fig.
2a.
For
the
PDL
experiments,
a
small
piece
of
NOSOL
363
(a
double-base
gun
propellant)
was
attached
to
the
top
of
the
propellant
strand
and
a
CO
2
laser
was
used
as
the
igni-
ter
(Fig.
2b).
NOSOL
363
ignition
initiator
provided
uniform
ignition
of
the
test
sample,
also
it
prevented
in-
depth
radiation
absorption
by
the
sample,
and
flame
extinction
that
could
occur
when
the
radiative
ignition
source
was
abruptly
removed.
For
the
steady-state
burning
rate
experiments,
the
propellant
strand
was
ignited
by
a
conventional
nichrome
wire
(Fig.
2a).
2.1.
Burning
surface
observation
The
close-up
observation
of
the
burning
surface
of
a
propellant
is
very
useful
in
understanding
the
combustion
behavior
of
the
propellant.
In
order
to
observe
the
burning
surface
of
propellant
with
a
high
resolution,
a
copper-
vapor
laser
was
incorporated.
A
schematic
diagram
of
the
propellant
sample
configuration
is
shown
in
Fig.
3.
Optical
lenses
were
used
to
convert
the
cylindrical
laser
beam
into
a
thin
laser
sheet,
which
has
a
width
around
8-
10
mm
at
the
location
of
the
burning
sample.
A
Panasonic
SVHS
VCR
with
a
filming
rate
of
60
fields
per
second
was
used
in
conjunction
with
a
Pulnix
CCD
camera
having
a
90-mm
macro
lens
to
record
the
burning
surface
phenom-
ena.
The
top
surface
of
the
pellets
was
slanted
at
45°
prior
to
the
tests
for
a
better
view
of
the
surface
(Fig.
3).
For
burning
surface
visualization
tests,
the
propellant
sample
was
ignited
with
the
CO
2
laser
without
any
NOSOL-363
propellant
initiator.
3.
Results
and
discussion
3.1.
Pressure
deflagration
limit
The
results
of
PDL
tests
show
that
the
PDL
limit
is
sen-
sitive
to
the
density
of
the
pressed
pellets.
The
PDL
limit
for
pellets
with
densities
larger
than
1.62
g/cc
ranged
from
7.34
to
8.03
MPa,
whereas
the
PDL
limit
for
pellets
with
densities
in
the
range
of
1.56-1.61
g/cc
ranged
from
8.37
to
8.72
MPa.
The
PDL
data
was
used
as
a
lower
pressure
boundary
for
steady-state
strand
burner
experiments.
3.2.
Steady-state
burning
rate
The
burning
rate
was
deduced
using
a
linear
curve
fit
to
data
obtained
from
recorded
images.
The
time
vs.
distance
data
were
deduced
by
measuring
the
distance
burned
and
the
corresponding
time.
Fig.
4
shows
typical
time
vs.
dis-
tance
trajectories
of
the
propellant
for
the
various
pressures
(a)
(b)
Fig.
2.
Sample
configurations
for
the:
(a)
steady-state
strand
burner
experiments,
(b)
PDL
experiments.
-4
10
-
T.
-
25
°
C
T.=
80
°
C
I.
-
T
i
=
-30
°
C
r
b
=0.1996P°.635
R
2
=
0.9992
r
b
=0.1452P
°354
R
2
=
0.9910
3
-
Burn
ing
ra
te
[cm
/s
]
2
-
r
b
=0.2133P
a665
R
2
=
0.9941
1
-
0.9
-
0.8
Increasing
Pressure
,
20.8
MPa
27.7
MPa
34.6
MPa
41.5
MPa
111
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
08
Time
[s]
Fig.
4.
Burning
surface
trajectories
at
various
pressures.
1982
A.
Ulas
et
al.
1
Fuel
85
(2006)
1979-1986
Exhaust
Plume
Jet
Hi-speed
camera
Re-circulating
eddy
flame
holder
Solid-propellant
Purge
flow
Propellant
holder
Fig.
3. Propellant
sample
configuration
for
burning
surface
observation.
12.5
10
Burne
d
dis
tance
[cm]
7.5
5
2.5
0
tested.
The
linearity
of
the
curves
suggests
that
the
1-D
lin-
ear
burning
assumption
is
valid.
Fig.
5
is
a
series
of
video
images
taken
from
a
test
con-
ducted
at
a
pressure
of
P
=
41.5
MPa
to
illustrate
the
"layer-by-layer"
regression
process.
Frame
zero
shows
the
onset
of
the
propellant
strand
ignition
using
nichrome
wire.
Frame
four
(0.13
s
later),
the
beginning
of
layer-by-
layer
burning
was
observed.
The entire
strand
was
consumed
by
frame
24,
which
corresponds
to
0.8
s
after
ignition.
From
all
the
video
recordings,
it
is
observed
that
the
flame
is
always
attached
to
the
burning
surface
of
the
propellant
strand,
which
is
a
typical
phenomenon
of
AP-containing
propellants.
Fig.
6
shows
the
burning
rate
as
a
function
of
pressure
and
initial
propellant
temperature.
Although
the
propellant
powder
supplied
by
the
manufacturer
was
very
limited,
great
care
was
taken
to
conduct
multiple
burning
rate
mea-
surements
at
the
same
test
conditions.
In
Fig.
6,
one
can
see
the
multiple
data
points
(two
or
even
three
data
points)
at
the
same
pressure.
Sometimes
the
data
points
are
very
close
to
each
other
and
they
are
on
the
top
of
each
other.
Show-
ing
the
multiple
data
points
at
each
test
condition
does
the
same
purpose
as
the
error
bars.
By
looking
at
the
multiple
test
results
at
each
pressure
in
the
plots,
the
variations
in
the
burning
rate
measurements
can
be
easily
seen.
The
results
showed
that
the
burning rate
data
at
the
same
test
conditions
are
very
reproducible,
indicating
that
the
error
bars
would
be
very
small.
The
burning
rate
law
expressions
as
a
function
of
chamber
pressure
at
three
different
initial
propellant
temperatures
are
also
given
in
Fig.
6.
For
the
Pressure
[psig]
3000
4000
5000
6000
20
30
40
Pressure
[MPa]
Fig.
6.
Pressure
and
initial
propellant
temperature
dependency
of
the
burning
rate.
4
10
Frame:
0
16
20
24
Time:
0.00
0.13 0.33
0.53
0.67
0.80
Fig.
5.
Series
of
selected
frames
of
strand
burner
experiment
at
P
=
41.5
MPa.
r
b
[cm/s]
=
0.191
P[MPai
°
7339
R
2
=
0.9994
r
b
[cm/s]
=
0.21334
P[MPa]
°
665
-v
-R
2
=
0.9996
-
T,
=
80
°
C
T,
=
100
°
C
r
b
[cm/s]
=
0.1368
P[MPa]
°78°
R
2
=
0.9996
r
b
[cm/s]
=
0.1996
P[MPa]
°
1335
R
2
=
0.9997
-
1
1
=
-
30
°
C
V
=
-10
°
C
1.2
Ca
lcu
la
te
d
bu
rn
ing
ra
te
[cm
/s]
2.8
2.6
2.4
2.2
2
1.8
1.6
1.4
10
Burn
ing
ra
te
[cm
/s
]
0.1
0.5
1.1
1
0.9
0.8
0.7
0.6
[s/
u
]
al
ai
6ul
ui
nq
p
al
ei
noi
eo
10
E
U
.
a)
io
0)
c
m
Burn
ra
te
[cm
/s
]
0.1
A.
Ulas
et
al.
1
Fuel
85
(2006)
1979-1986
1983
I
'
I I
+4%
o
f
-4%
r
b
[cm/s]
=
0.0121
P[MPa]
o.
652
T[Kr
467
-•
'
t
t
I I
t
t
I I
1
2
1.4
1.6
1.8
2
2.2
2.4
2.6
28
Experimental
burning
rate
[cm/s]
Fig.
7.
Correlation
of
operating
parameters
to
predict
burning
rate.
room
temperature
case,
the
pressure
exponent
was
0.75
based
on
the
pressure
range
tested.
An
empirical
expression,
Eq.
(1),
was
developed
to
cor-
relate
the measured
burning rate
to
operating
parameters
(pressure
and
initial
temperature).
Fig.
7
shows
the
exper-
imentally
measured
burning
rate
vs.
the
calculated
burning
rate
using
Eq.
(1).
This
correlation
can
be
used
to
within
4%
accuracy.
rb
[cm/S]=
0.0121P
[MP4
1652
T
1
[4
.497
In
Ref.
[5],
high-pressure
burning
rate
data
obtained
from
an
underwater
strand
burner
using
an
ultrasound
sys-
tem
for
this
propellant
is
given
at
initial
temperatures
-10,
23,
and
100
°C.
Figs.
8-10
present
combined
low-pressure
Pressure
[psig]
2000
4000
6000
8000
10
4
r
b
[cm/s]
=
0.1335
P[MPa]
°°°2
:
R
2
=
0.9907
10
100
Pressure
[MPa]
Fig.
8.
Combined
low-
and
high-pressure
burn
rate
data
at
room
temperature.
Pressure
[psig]
4000
6000
8000
10000
20
30
40
50
60
70
80
90100
Pressure
[MPa]
Fig.
9.
Combined
low-
and
high-pressure
burn
rate
data
at
elevated
temperatures.
Pressure
[psig]
4000
6000
8000
10000
20
30
40
50
60
70
80
90100
Pressure
[MPa]
Fig.
10.
Combined
low-
and
high-pressure
burn
rate
data
at
cold
temperatures.
and
high-pressure
data.
In
spite
of
the
difference
in
burning
rate
measurement
techniques,
the
two
sets
of
room
temper-
ature
data
can
be
fitted
by
a
single
power
law
expression
as
shown
in
Fig.
8.
For
other
than
room
temperature
test
runs,
each
set
has
its
own
power
law
correlation
as
shown
in
Figs.
9
and
10.
The
burning
rate
results
in
Figs.
9
and
10
are
not
correlated
with
a
single
power
law
due
to
different
initial
temperatures
for
the
two
sets
of
data.
In
Fig.
9,
the
initial
temperatures
are
+80
and
+100
°C,
and
in
Fig.
10
they
are
-30
and
-10
°C.
(
1
)
10
-
-•-
-
P
=
86.8
MPa
-
-X
-
-
P
=
46.3
MPa
-V-
P
=
34.6
MPa
-
-V-
-
P
=
59.8
MPa
-
-•
-
-
P
=
127.8
MPa
-0-
P
=
27.7
MPa
-
-
-
-
P
=
114.2
MPa
-0-
P
=
41.5
MPa
P
=
20.8
MPa
---------
--
----
A
r
----
___
_
------
-------
- -
- - - -
----
-
_3
--------------
-
I
----------------
+20%
-
------
-
'
-20%
a
p
[e]
=
7.701x1e+
2.071x10
-5
P[MPa]
-
1.032x10
7
P[MPat
20
40
60
80
100
120 140
Pressure
[MPa]
1984
A.
Ulas
et
al.
1
Fuel
85
(2006)
1979-1986
3.3.
Temperature
sensitivity
Pressure
[psig]
5000
10000
15000
20000
Temperature
sensitivity
is
an
important
characteristic
of
a
solid
propellant.
It
is
desirable
to
have
a
solid
propellant
with
a
low
temperature
sensitivity,
so
that
the
burning
rate
of
the
solid
propellant
is
not
significantly
affected
by
the
temperature
of
the
environment.
Solid-propellant
tempera-
ture
sensitivity,
0p,
is
defined
by
Eq.
(2),
which
represents
the
change
of
the
burning
rate
per
degree
change
in
initial
temperature
at
constant
pressure.
a
in
rb
Cfp
aT
i
The
temperature
sensitivity
at
various
pressures
was
determined
by
constructing
a
semi-natural
log
plot
of
burn-
ing
rate
versus
initial
temperature.
In
this
plot,
the
slope
of
the
linear
curve
fit
represents
the
temperature
sensitivity
of
the
propellant
at
the
specified
pressure.
A
plot
illustrating
this
method
is
shown
in
Fig.
11.
After
o
was
determined
at
each
pressure,
a
plot
showing
the
pressure
dependency
of
temperature
sensitivity
was
constructed
(Fig.
12).
The
o
is
0.0011
K
-1
at
P
=
20.8
MPa
and
gradually
increases
up
to
0.00208
K
-1
at
P
=
114
MPa.
Fig.
12
also
shows
a
2nd
order
polynomial
fitted
to
the
experimental
data.
The
propellant
temperature
sensitivity
values
are
quite
low,
which
means
that
the
propellant
performance
would
not
be
significantly
influenced
by
the
temperature
of
its
envi-
ronment.
The
low
temperature
sensitivity
characteristic
of
the
propellant
is
also
apparent
from
Fig.
11,
which
shows
almost
a
20%
change
in
the
propellant
burning
rate
when
the
propellant
initial
temperature
changes
from
-30
to
+100
°C.
In
the
ballistic
field,
20%
increase
in
the
propel-
lant
burning
rate
for
an
initial
temperature
range
as
wide
as
the
one
considered
in
this
study
(-30
to
+100
°C)
is
not
considered
a
significant
increase.
Low
temperature
-80 -40
0
40
80
Initial
temperature
[
°
C]
Fig.
11.
Burning
rate
as
a
function
of
initial
temperature
at
various
pressures.
0.0025
k
0.002
ti"
.5
0.0015
Co
0.001
r
.Ls
N
Fig.
12.
Pressure
dependency
of
temperature
sensitivity.
sensitivity
is
a
highly
desirable
characteristic
of
a
propel-
lant.
When
the
burning
rate
is
not
significantly
influenced
with
the
propellant
initial
temperature,
the
airbag
inflation
time
(which
is
related
to
the
propellant
burning
rate)
is
also
not
significantly
effected
with
the
initial
temperature.
The
temperature
sensitivity
information
and
its
pressure
depen-
dency
are
useful
for
design
predictions
of
the
performance
of
airbag
inflators
using
this
propellant.
3.4.
Burning
surface
temperature
The
subsurface
temperature
profile
of
a
solid
propellant
is
governed
by
an
exponential
profile
until
it
reaches
the
surface
[6].
Therefore,
to
determine
the
surface
tempera-
ture,
the
temperature
profile
was
plotted
on
a
semi-log
plot
versus
distance
from
the
propellant
surface.
The
departure
point
on
the
semi-log
plot
represents
the
surface
tempera-
ture
(T
s
).
A
typical
temperature
profile
is
shown
in
Fig.
13.
The
sub-surface
thermal
wave
penetration
depth
is
on
the
order
of
20
µm.
The
dashed
vertical
line
at
x
=
0
in
Fig.
13
represents
the
burning
surface
location.
The
positive
x
domain
corresponds
to
the
gas-phase
region
and
the
negative
x
domain
corresponds
to
the
unburned
solid-phase
region.
The
surface
temperature
for
the
pres-
sure
range
of
20.8-41.5
MPa
varied
from
580
to
755
K.
The
gas-phase
flame
temperature
could
not
be
measured
because
the
thermocouple
was
burnout
in
the
high-temper-
ature
flame
region.
In
the
Arrhenius-type
burning
rate
law,
Eq.
(3),
the
activation
energy
and
the
pre-exponential
factor
were
deter-
mined
as
E
a
=
2.735
kcal/mol
and
A
=
15.06
cm/s,
respec-
tively,
by
plotting
measured
burning
rates
vs.
surface
temperatures
(Fig.
14)
and
curve
fitting
the
measured
data.
r
b
A
exp
\
R
TS
/
(3)
(
2
)
1.6
1.2
0.8
••••••••
0.4
0
-0.4
-0.8
1000
900
800
700
600
----
500
----
P
=
27.7
MPa
r
b
=
1
78
cm/s
T
8
=641
K
400
300
Gas-Phase
1
200
'
-40
-20
20
40
1
0
A.
Ulas
et
al.
1
Fuel
85
(2006)
1979-1986
1985
Distance
from
surface
[gm]
Fig.
13.
A
typical
plot
of
temperature
vs.
distance
from
the
burning
surface.
Surface
temperature,T
8
[K]
800
700
600
10
A
=
15.06
cm/s
E
a
=
2.735
kcal/mol
0.1
0.0012
0.0013
0.0014
0.0015
0.0016
0.0017
Inverse
surface
temperature,
1/1",
[K
-1
]
Fig.
14.
Arrhenius
plot.
3.5.
Burning
surface
observations
At
low-pressures
(P
<
17
MPa),
the
propellant
burns
in
a
highly
non-one-dimensional
manner;
i.e.,
there
was
a
dynamic
melt
zone
on
the
burning
surface,
which
made
it
difficult
to
achieve
"layer-by-layer"
burning.
Fig.
15
shows
several
instantaneous
pictures
taken
from
the
recorded
video
images.
Most
of
the
time,
this
thick
melt
zone
ejected
large
particles
(beads)
from
the
burning
surface,
with
diam-
eters
ranging
from
400
to
2000
gm.
Sometimes,
either
these
ejected
beads
or
drippings
from
the
liquid
melt
layer
ignited
the
exposed
side
surfaces
of
the
strand.
It
was
observed
that,
as
the
pressure
increased,
the
thickness
of
this
melt
layer
decreased,
resulting
in
a
lesser
effect
on
the
surface
burning
behavior
of
the
propellant.
Spherical
beads,
shown
in
Fig.
16,
were
recovered
after
every
test
runs.
A
particle
size
analysis
was
conducted
on
the
recovered
beads.
At
a
pressure
of
86.3
MPa,
the
aver-
age
diameter
of
beads
was
found
to
be
approximately
300
gm.
For
pressures
below
41.5
MPa,
the
average
diam-
eter
was
1000
gm.
3.6.
Chemical
analysis
of
"bead-like"
residue
Peculiar
combustion
residues
(beads)
were
observed
after
every
test
run.
The
beads
were
very
small
for
the
ambient
and
elevated
temperature
test
runs,
but
slightly
Fig.
16.
Image
of
the
burning
surface
(P
=
7.68
MPa).
Temp
era
ture
[
K]
Burn
ing
ra
te
[cm
/s
]
(1,10:28.39
INCREASING
TIME
Fig.
15.
Photograph
showing
the
burning
surface
(Laser-assisted
burning
at
P
=
2.86
MPa).
1986
A.
Ulas
et
al.
1
Fuel
85
(2006)
1979-1986
I9KU
28.8x
5000
0601
TAL02
I
i
,
48KX
IOU
0064
TAIN2
(a)
(b)
Fig.
17.
Magnified
view
of
beads
under:
(a)
29
times,
(b)
1500
times
magnification.
larger
for
the
low
temperature
tests.
The
bead-like
residue
was
analyzed
using
an
environmental
scanning
electron
microscope
(ESEM)
to
construct
the
elemental
map
of
the
bead
and
an
X-ray
diffraction
to
determine
its
com-
pound
composition.
The
X-ray
diffraction
results
agreed
well
with
the
ESEM
results
and
indicated
that
the
material
of
the
beads
was
mostly
sodium
chloride,
NaCl,
with
a
small
amount
of
silicon-containing
compounds.
Fig.
17
is
a
photograph
of
the
beads
at
a
magnification
of
28.8
and
1500
times
of
their
original
size.
The
diameters
of
particles
in
this
photograph
are
approximately
900
µm.
4.
Summary
and
conclusions
An
experimental
study
was
conducted
to
determine
the
steady-state
burning
characteristics
and
some
important
ballistic
properties
of
an
Ammonium
Perchlorate/Guani-
dine
Nitrate/Sodium
Nitrate
solid-propellant
formulation
for
airbag
applications.
The
pressure
deflagration
limit
of
the
propellant
was
found
to
be
in
the
range
of
8.37-
8.72
MPa.
The
pressure
exponent
was
0.802
at
room
tem-
perature.
The
burning
rate
at
room
temperature
can
be
represented
by
an
Arrhenius
burning rate
law
with
activa-
tion
energy
of
2.739
kcal/mol
and
pre-exponential
factor
of
15.06
cm/s.
The
temperature
sensitivity
of
the
propellant
is
relatively
low
(0.001-0.002
K
-1
).
The
low
7p
characteristic
represents
a
desirable
feature
of
an
airbag
propellant.
The
temperature
sensitivity
dependency
on
pressure
was
deter-
mined
for
pressures
ranging
from
20.8
to
128
MPa.
At
low-pressures
(P
<
17.3
MPa),
the
propellant
burns
in
a
highly
non-one-dimensional
manner,
due
to
the
dynamic
melt
zone
and
particle
formation
and
ejection
phenomena
near
the
burning
surface
region.
However,
for
pressures
beyond
20.8
MPa,
it
burns
in
a
nearly
"layer-by-layer"
form.
The
propellant
generates
some
residual
spherical
beads
on
the
order
of
900
µm
for
pressures
below
41.5
MPa.
For
pressures
around
86.3
MPa,
the
spherical
beads
are
on
the
order
of
300
µm
in
diameter.
The
bead
diameter
decreases
with
increasing
chamber
pressure
and/
or
initial
propellant
temperature.
Both
the
ESEM
and
the
X-ray
diffraction
method
indicate
that
the
composition
of
the
bead
residue
was
mostly
sodium
chloride.
Acknowledgements
The
authors
would
like
to
thank
Talley
Defense
Systems
for
the
sponsorship
of
this
research
program.
The
interest
and
support
of
Mr.
Christopher
P.
Ludwig
and
Dr.
Robert
Glick
are
greatly
appreciated.
Also,
the
authors
would
like
to
acknowledge
the
micro-thermocouple
and
pellet
press-
ing
effort
of
Professor
Baoqi
Zhang
at
PSU.
Special
thanks
go
to
Mr.
Bruce
Anderson
of
Talley
Defense
Systems
for
Talley's
high-pressure
burning rate
data.
References
[1]
Struble
DE.
Airbag
Technology:
what
it
is
and
how
it
came
to
be.
In:
proceedings
of
the
1998
international
congress
and
exposition,
Feb
23-
26,
1998,
SAE
special
publications
Airbag
Technology,
vol.
1333,
p.
73-92,
Paper
980648
SAESA2.
[2]
Deppert
TM,
Barnes
MW,
Mendenhall
IV,
Taylor
RD.
Development
of
gas
generants
for
passive
automobile
restraint
systems,
proceedings
of
the
2nd
international
symposium
on
sophisticated
car
occupant
safety
systems:
AirBag
2000
Conference,
Karlsruhe,
Germany,
November
1994.
p.
10-1-17.
[3]
Sherman
D.
The
rough
road
to
airbags,
Invention
and
Technology,
Summer
1995.
[4]
Zenin
A.
Thermal
wave
profiling
measurements
using
fine-wire
thermocouples.
An
Invited
Presentation
at
Penn
State
University
1994;10(August).
[5]
Risha
GA,
Ulas
A.
Kuo
KK,
Koch
DE,
Ludwig
CP,
Glick
R.
Combustion
behavior
of
TAL-1308
composite
solid
propellant
for
airbag
applications,
AIAA
Paper
No.
99-2632,
35th
AIAA/ASME/
SAE/ASEE
joint
propulsion
conference
and
exhibit,
Los
Angeles,
CA,
June
20-23,
1999.
[6]
Kubota
N.
Survey
of
rocket
propellants
and
their
combustion
characteristics.
Fundamentals
of
Solid
Propellant
Combustion.
In:
Kuo
KK,
Summerfield
M,
editors.
Progress
in
Astronautics
and
Aeronautics,
vol.
90.
AIAA;
1994.
p.
1-52
[chapter
1].