Thermal and Mechanical Hazards of

Nitrocellulose and its Mixture with Nitroglycerin

S. Singh1, Q. Kwok1, R. Turcotte1 and M. Paquet2
1Canadian Explosives Research Laboratory, Natural Resources Canada
2General Dynamics Ordnance and Tactical Systems - Canada Valleyfield

7th International Nitrocellulose Symposium, Montréal, Canada, May 31 – June 1, 2016

 Outline
Introduction
Thermal Study
• Accelerating Rate Calorimetry (ARC)

Quantification of Mechanical Hazards

• Friction
• Impact

Summary
Acknowledgement: Special thanks to Mr. Ian Levac
from GD-OTS Canada Valleyfield for the preparation and
the shipping of the samples used in this work
 Introduction
All modes of ignition of energetic materials are essentially
thermal in nature:

 Fast (millisecond) ignitions are intuitively familiar:

• Striking a match
• Lighting a gas BBQ
• Using paper (kindling) to start a fire

 Slow ignition events (days, years) are less intuitive:

• Most energetic materials are self-heating materials and
therefore can be made to spontaneously ignite

 Mechanical ignition hazards

• instantaneous friction and/or impact stimuli invoking an
unintended reaction (evolution of gas, sound, light…)
 Introduction
Literature isothermal studies attempt prediction of safe storage
periods for NC and NC/NG systems

• Slow ignition (T. Kotoyori 2005):

 Autocatalytic induction times from 1 to 10 days (Tiso : 73 to 91°C)
• Fast ignition (A.S. Shteinberg 2006):
Ignition times from 1 to 15 seconds (Tiso : 212 to 242°C)

Mechanical sensitivity testing addresses probability of initiation

(Ip)
• Ip = probability of local ignition
• Proper test to simulate process events (hazards)
• Unbiased scheme to assign “yes/no” reaction to test
• Appropriate statistical analysis method
• Is there an acceptable level of risk?
 Samples
NC NC/NG (70/30)
Grade N Content (mass %)  Prep: GD–OTS Canada Valleyfield
 NC wetted with alcohol/acetone
C; Type I 13.15 ± 0.05 %
 Acetone desensitized NG added
 Mixed to a granular paste with suitable viscosity
 “as rec’d” 25% water-wet  Transferred onto stainless steel plate and left
 Dried in desiccator to constant >48 h to evaporate at room temperature to a
mass. Stored and used from given residual solvent
desiccator for duration of study  Material peeled off and broken into small pieces

“as rec’d” dried

under vacuum
 Stored in sealed
charge
dissipative vials
Thermal Study - ARC

10 mL spherical Ti vessel

1 mL tube Ti vessel
 ARC (Isothermal)
NC NC/NG
 = 0.9 h
130
 = 1.5 h
130  = 0.5 h (70/30)
 = 2.3 h  = 1.9 h
120 Tiso/°C 120 Tiso/°C
 = 4.6 h 80  = 2.4 h 65
 = 7.4 h 85 70
110 90 110
75
 = 10.0 h 95  = 4.8 h 80
100

Tiso/°C
85
Tiso/°C

100  = 12.7 h 105 100 95

110  = 9.5 h
105
 = 21.4 h 115
90 90  = 19.8 h 110
120 115
 = 37.2 h 125  = 29.3 h

80 86 80  = 36.4 h
84

 = 152.2 h
 = 53.5 h
80
70 79
70
35 45 55 65

60 60
0 20 40 60 80 100 120 140 160 0 20 40 60 80 100 120 140 160
Induction time ()/h Induction time ()/h
 ARC (Isothermal)
9

8 Ea = 105 ± 5 kJ mol-1
ln(A/min-1) = 27.4 ± 1.5
7
r2 = 0.976
ln (/min)

6 Ea = 95 ± 11 kJ mol-1
-1
ln(A/min ) = 25 ± 4
5 r2 = 0.898

4
NC/NG 70/30
NC alone (grade C)
3

2
2.5 2.6 2.7 2.8 2.9 3.0
3
10 K/T
Literature + Isothermal ARC Data
250
Isothermal Temperature/°C

200

150

100

50
10-2 10-1 100 101 102 103 104 105
Induction/ignition time/min
Thermal Study - Conclusions
 Empirical fit can be used as a basis to derive NC shelf-
lives from iterative calculations
• More data would be required at lower temperatures
(micro-calorimetry) to improve shelf-life predictions
• More data would be required to link the fast ignition data
to the slow decomposition regime (10 s to 30 min)
• Therefore in the present work, attempts will be made to
develop a pyrolysis technique to obtain such data
 The slow decomposition behaviour of the NC/NG
mixture and pure NC appears very similar
 The NC/NG mixture shows similar induction at
temperatures ≈10°C lower than pure NC
 Mechanical Hazard: Friction
ABL/ICI
W W
Porcelain peg
Friction Test
Porcelain plate
Sample Pendulum Arm

v v

BAM Friction Test Stainless Steel Plate Heavy Weight

Sample Deposit
Fixed Stainless Steel Disc
ABL/ICI  Hazards quantification via statistical data treatment of measured
parameters such as contact geometry, applied pressure, plate velocity, detection of
induced reactions (i.e., monitor emissions)…

Correlation study on many parameters has shown that D = Pa l <v>2 gives the best
overall correlation for a wide range of EMs [Pa = Apparent Pressure, l = Contact
Length (in the direction of motion), <v> = Average Relative Velocity]
 Friction (ABL/ICI)
Main power switch

Load
Pressure
Pendulum
monitor

CO
analyzer
Sliding
PS2
platform
Photodiode
amplifier

NOx Apparatus
analyzer Start/Stop
 Friction: Gas evolution
2000 2000
Dose Threshold Dose Threshold
6
4.55 x 10 W s
-1 4.17 x 106 W s-1
<R> CO / nL mm-1

1500 1500

-1
NC/NG

<R> CO / nL mm
NC
1000 1000
(70/30)
CO Threshold
-1
500 459 nL mm 500 CO Threshold
274 nL mm-1

0 0
Dose Threshold
10 6
4.63 x 10 W s
-1
8
Dose Threshold
3.85 x 106 W s-1
<R> NOx / nL mm-1

-1
<R> NOx / nL mm
6

6
NOx Threshold 4
4 3.44 nL mm-1

NOx Threshold
2 2
1.39 nL mm-1

0 0
106 107 106 107
2 -1
<D = P l v > W s
<D = P l v2> W s-1
Hazards Quantification - Friction
99 99
98 98

95
CO

NC
NOx
Regression
95 NC/NG
CO
NOx
Regression
90 90

80 80 (70/30)
70 70
Probability of ignition/%

Probability of ignition/%
50 50

30 30
20 20

10 10
5 5

2 2
1 1
0.5 0.5
0.2 0.2
0.1 0.1
0.05 0.05
0.01
0.01

0.001
0.001
105 106 107 108 105 106 107 108
<D>/W s-1 -1
<D>/W s
Hazards Quantification - Impact
Hazards Quantification - Impact
99 99
98

95
NC/NG 98
95
NC/NG
90 70/30 90 70/30
80 80
70 70
Probability of Ignition / %

Probability of Ignition / %
50 50

30 30
20 20
10 10
5 5
2 2
1 1
0.5 0.5
0.2
0.1
0.05
NC 0.2
0.1 NC
0.05
0.01
0.01
10-2 10-1 100 101 102
102 103 104 105 106 107
PMax / GPa
Energy Density / J m-2
Mechanical Hazards-Conclusions
 Small scale friction and impact tests can be instrumented and
calibrated to obtain realistic evaluation of Ip for NC and
NC/NG
• Friction
• Compared to NC, the NC/NG mixture has lower thresholds
of CO/NOx production, corresponding to the onset of
combustion reactions with a potential to propagate
• Pal<v>2 dose parameter is “unphysical” . [Ignition of the NC
and NC/NG in this test is expected to be thermal in origin so
“friction temperature” a better dose parameter]
• Impact
• Compared to NC, the NC/NG mixture demonstrates higher
probability to initiate given the same dose
 What is an acceptable level of risk? These Ip data are useful for
insertion into risk assessment models.
© Her Majesty the Queen in Right of Canada, as represented by the Minister of Natural Resources, 2016