You are on page 1of 6

Instrumented pressing of inert powders to

study the effect of particle size


Cite as: AIP Conference Proceedings 1979, 060008 (2018); https://doi.org/10.1063/1.5044805
Published Online: 03 July 2018

Jamie A. Stull, Darla Graff Thompson, Caitlin Woznick, et al.

ARTICLES YOU MAY BE INTERESTED IN

Synthesis, formulation and performance evaluation of reduced sensitivity explosives


AIP Conference Proceedings 1979, 100005 (2018); https://doi.org/10.1063/1.5044877

Measurement and reactive burn modeling of the shock to detonation transition for the HMX
based explosive LX-14
AIP Conference Proceedings 1979, 100022 (2018); https://doi.org/10.1063/1.5044894

Modeling a high explosive cylinder experiment


AIP Conference Proceedings 1979, 030008 (2018); https://doi.org/10.1063/1.5044778

AIP Conference Proceedings 1979, 060008 (2018); https://doi.org/10.1063/1.5044805 1979, 060008

© 2018 Author(s).
Instrumented Pressing of Inert Powders to Study the Effect
of Particle Size
Jamie A. Stulla), Darla Graff Thompson, Caitlin Woznick, and Racci DeLuca

M-7, High Explosives Science and Technology, Los Alamos National Laboratory,
Los Alamos New Mexico 87545, USA
a)
Corresponding author: jamie.stull@lanl.gov
Abstract. It is well-known that the detonation and mechanical properties of pressed high explosives (HEs) depend on density.
Furthermore, specific particle size distributions have been shown to compact more easily and to effect some shock/detonation
properties. Theoretically, particle size distributions can be optimized for compaction and performance. Here, in anticipation of
future experiments on HE powders, we explore the role of inert particle-size characteristics on compaction properties and pellet
integrity. An instrumented compaction instrument was used to press inert powders that have similar particle size distributions to
1,3,5,7-tetranitro-1,3,5,7-tetrazocane (HMX), as well as compounds that are being developed as mock materials for HMX. The
measured force and displacement measurements provide information in the form of compaction curves and Heckel plots. The
results of the compaction measurements and subsequent Instron compression tests identify 5-iodo-2'-deoxyuridine (IDOX) as a
potential HMX mock for density and mechanical strength. Relevant measurements were also made towards calibration and
implementation of the modified Drucker-Prager Cap model for compaction of HE powders. LA-UR-17-28406

INTRODUCTION
The aim of these studies is to explore how particle size variations affect compaction of powders and to
understand the response of a Measurement Control Corporation-1 (MCC-1) compaction instrument. These
experiments will provide the foundation for future work involving HE powders. By measuring various compaction
variables we hope to better understand the effect of particle size on the compaction process and how these correlate
with features of the final compact. We have used three classes of inert material in order to test various particle
characteristics: 1) coarse and fine sugar blends to simulate HMX particle sizes, 2) various high density inert
materials selected to simulate the density and mechanical properties of HMX, and 3) microcrystalline cellulose
(Avicel®), to demonstrate the utility of the MCC-1 instrument in parameterizing the Drucker-Prager Cap (DPC)
model for powder compaction.
With similarities to HMX particle size, sugar has previously been used for experiments measuring the shock
response of a compact [1]. However, sugar has a lower theoretical maximum density (TMD) than HMX and absorbs
water, causing it to be of limited use as an HMX density or mechanical properties mock [2]. Ideally, a new mock
material could replace several others currently being used to simulate specific properties of HMX and here we
present preliminary work on some candidate powders. Progress has also been made towards implementing the DPC
model for the compaction of explosive materials [3]. The DPC model has been applied to HE using the experimental
results of inert materials that are typically used in the pharmaceutical industry. Here, we demonstrate that the MCC-
1 instrument can provide critical parameters for the DPC model.

EXPERIMENTAL PROCEDURES
A variety of inert materials were chosen for this study. Although sugar has a much lower TMD (1.59 g/cm3) than
HMX (1.902 g/cm3), a mixture of granulated and powdered sugar (C&H®) simulates typical HMX bimodal particle
size distributions [1]. Three other inert materials studied in this work were selected to have similar TMD values to
HMX, but particle size and morphology were not controlled. IDOX, 5-iodo-2'-deoxyuridine (TMD = 2.001 g/cm3)

Shock Compression of Condensed Matter - 2017


AIP Conf. Proc. 1979, 060008-1–060008-5; https://doi.org/10.1063/1.5044805
Published by AIP Publishing. 978-0-7354-1693-2/$30.00

060008-1
and PFBA, 2,3,4,5,6-pentafluorobenzamide (TMD = 1.884 g/cm3) were purchased from Sigma-Aldrich, while N-
BPFPO, N,N'-bis(2,3,4,5,6-pentafluorophenyl)oxamide (TMD = 1.932 g/cm3) was synthesized following the
method of Yamaguchi et al. [4]. Avicel® PH101 was purchased from Sigma-Aldrich and 1 wt% of calcium stearate
was subsequently added as a lubricant.
An MCC-1 Scientific Compaction Research tablet press (MCC, East Hanover, NJ) used a single load cycle to
die-press the powders into 10 mm diameter pellets; pellet heights varied from 5-8 mm, depending on the target
density and instrument settings. The unique nature of the compaction instrument does not allow the user to define a
precise end pressure or pellet height. The ultimate pellet dimensions are a product of the position of the upper and
lower punches and the weight of powder added to the die. The MCC-1 press has two linear variable differential
transformers (LVDTs) to measure the displacement of the upper and lower punch, and three transducers to measure
the transmitted, radial and ejection loads. Compression tests of compacted pellets were conducted to failure on an
Instron 8862 workstation. All compression experiments were performed at ambient temperature using a crosshead
rate of 0.0675 in/min.

SUGAR PRESSING STUDIES

Granulated sugar has an average particle size of approximately 420 μm [1]. Powdered sugar has particle sizes
between 10-20 μm. Granulated sugar closely matches the coarse classes of HMX (Classes I and III) and powdered
sugar matches the fine classes of HMX (Classes II and V) as defined by MIL-DTL-45444 [5]. We also pressed
bimodal mixtures of the granulated and powdered sugar to represent particle size distributions in common HMX
explosive formulations. PBX 9501 combines HMX classes I and II in a ratio of 3:1. Another formulation combines
HMX classes I and V in a ratio of 1.2:1. These distributions will be referred to as Bimodal 1, which combines the
granulated and powdered sugar in a ratio of 3:1, and Bimodal 2, which combines granulated and powdered sugar in
a ratio of 1.2:1. All of the sugar pellets in this study were pressed neat with no additional binders.
A large number of granulated and powdered sugar pellets were pressed in order to understand the response of the
compaction instrument, shown by the data in Figure 1. The solid lines represent the volume fractions
(φ = density/TMD) of the different powders as they were compacted. These volume fractions were calculated using
the upper and lower LVDTs on the MCC-1 press. The pellets were approximately 10 mm in diameter and 8 mm tall;
the height of the pellets varied somewhat with powder weight, due to the compliance in the press. We observed a
significant deviation between the final measured densities of the pellets and the data from the press, exacerbated at
higher densities. We believe this is due to a combination of the elastic recovery in the material and unaccounted for
compliance of the instrument; further studies are underway to understand this relationship.

1.05

1
Solid Limit φ = 1
0.95

0.9

0.85 Granulated Pellets


Granulated Pellets
Granulated Press
Granulated Press Data
Data
φ

0.8 Powdered Pellets


Powdered Pellets
Powdered Press
Powdered Press Data
Data
0.75 Bimodal
Mock PBX1 Pellets
9501 Pellets
Bimodal
Mock PBX1 Press Data Data
9501 Press
0.7 Mock PBXN−9
Bimodal Pellets
2 Pellet
Mock PBXN−9
Bimodal Press
2 Press DataData
0.65
Φο  0.64 
0.6
0 1 2 3 4 5
Pressure (psi) x 10
4

FIGURE 1. Compaction curves of sugar pellets. All symbols are individually pressed pellets, plotted versus their maximum
compaction pressure. Solid traces are the compaction curves obtained from load cells and LVDTs on the instrumented press.

The granulated sugar pellets require lower pressure to achieve equal volume fractions (i.e. densities) compared
with the powdered sugar (Figure 1). The granulated material is easier to compact. Early in the compaction process,
both of the bimodal distributions require significantly lower pressures to reach equivalent volume fractions. At

060008-2
higher volume fractions near 0.85, the bimodal mixtures and granulated sugar require similar pressures for a given
volume fraction. These observations are consistent with the general understanding of space filling efficiency, which
has shown that a mixture of particle sizes can eliminate voids more easily, especially during the early stages of
compaction involving particle rearrangement and packing.
The data from the press can also be plotted in the form of a Heckel plot, which gives a semi-quantitative
evaluation of the compaction properties of the different materials [6, 7]. The Heckel equation (1) is given as:

⎛ ⎞
⎜ 1 ⎟
ln ⎜ ⎟= k*P+ A (1)
⎜ 1− D ⎟
⎝ TMD ⎠

When the logarithmic function on the left is plotted versus pressure, P, k is the slope and A is the y-intercept.
The density of the pellet is represented as D. This relationship allows us to derive a crystal aggregate yield pressure
based on the compaction response. Figure 2 shows the compaction response of granulated sugar, powdered sugar,
and the two bimodal distributions. A solid black line was fit to the linear portion of the compaction curves. The
reciprocal of the slope, known as Heckel’s parameter (1/k), is considered an approximation of the yield pressure,
and the y-intercept (A) represents the degree of packing due to particle rearrangement [6, 7]. Heckel plot results are
tabulated in the inset in Figure 2. Both bimodal distributions have similar yield pressure and particle rearrangement
during the compaction process. Bimodal yield pressures are higher than granulated sugar, lower than powdered
sugar. This demonstrates particle size effects on compressibility. Additionally, both of the bimodal sugar
distributions undergo more particle rearrangement during the compaction process than either the granulated or
powdered sugar (larger value of A). This is consistent with the expected compaction advantages of a bimodal
powder distribution.

Granulated
Powdered
3.0 Bimodal 1
Bimodal 2

2.5
ln(1/(1-D/TMD))

2.0
Material k 1/k A
Granulated Sugar 5.86e-5 17050 1.33

1.5 Powdered Sugar 4.52e-5 22099 1.28

Bimodal 1 5.27e-5 18967 1.58

1.0 Bimodal 2 4.98e-5 20081 1.57

10 20 30 3
40x10
Axial Pressure (psi)

FIGURE 2. Heckel plots of granulated, powdered, Bimodal 1 and Bimodal 2 sugar. The inset shows the Heckel plot results.

MOCK HMX PRESSING STUDIES


Existing mocks for HMX are designed to mimic a specific material property such as density, strength, or thermal
response. Ideally, an HMX mock would be able to model as many properties as possible. Yeager et al. have taken
the first steps to identify a new HMX mock to match both density and mechanical properties, among other
considerations such as melting point and water solubility [2]. They identified three materials as likely candidates and
here we performed a compaction study on them. The mock materials were pressed to a variety of densities to
produce compaction curves (as shown in Figure 3a) and then were mechanically tested in compression (Figure 3b).
Each symbol in Figure 3a represents an individually pressed pellet; a log fit has been included as a guide only
(dashed trace). The sugar Bimodal 1 data has also been included for a reference point. Two variations of the IDOX
material were included; one was the as-received material (IDOX) and the second was the recrystallized material

060008-3
(RIDOX). Both IDOX variants behave somewhat similarly to Bimodal 1 overall, but IDOX requires less pressure to
compact to the same volume fraction as RIDOX. This difference is likely a function of particle size distribution, as
inspection of as-received IDOX revealed a larger distribution of sizes than RIDOX. N-BPFPO required the lowest
pressures overall to compact to higher volume fractions, performing better than the other mocks tested. PFBA
showed the worst performance, with a poor maximum volume fraction and issues on ejection resulting in cracks and
fractures in the compact. No effort has yet been made to optimize particle size distribution of the mock candidates
towards desired properties. The compressive strength of the pellets were measured using uniaxial testing. The results
are compared to the Bimodal 1 sugar distribution in Figure 3b.

40
0.97
IDOX
1.880 g/cc (φ = 0.940)
Recrystallized IDOX
30 1.875 g/cc (φ = 0.937)
0.87 N-BPFPO
1.808 g/cc (φ = 0.940)

Stress (MPa)
Bimodal
Mock PBX1 Pellets
9501
ϕ

PFBA
IDOX 20 1.645 g/cc (φ = 0.870)
IDOXPellets Bimodal Sugar
RIDOX
RIDOXPellets 1.474 g/cc (φ = 0.927)
0.77
N-BPFPO
N-BPFPOPellets
10
BFBAPellets
PFBA

0.67
0 20000 40000 60000 0
Max Pressure (psi) 0 2 4 6 8
Strain (%)
(a) (b)

FIGURE 3. (a) Compaction curves of the mock candidates; symbols are from individually pressed pellets and dashed lines are
log fits to the pellet data. (b) Compression curves of mock materials. Each of the materials was pressed to a similar TMD except
for the PFBA, which could not be pressed to an equivalent TMD.

IDOX is much stronger than N-BPFPO and PFBA materials, with RIDOX having the highest compressive
strength. We suspect that IDOX/RIDOX strength differences, as with the slight variations in compaction curves,
could be due to the different particle sizes distributions of the two materials. To confirm this, we pressed pellets
using powder-sieved material to the same particle size distributions, and tested the compressive strength (not
shown). These tests revealed that if the particle size distribution was consistent, the compression strength of the two
materials was nearly equivalent. This suggests it might be possible to optimize the particle size distribution to fine-
tune the mechanical properties. Overall, IDOX approximately matches the compressive strength of Bimodal 1 sugar,
which has been shown to mimic HMX compressive properties fairly well [8]. The other two mocks show less initial
promise as an HMX mechanical mock.

MODIFIED DRUCKER-PRAGER CAP MODEL

Finally, we have performed MCC-1 measurements towards parameterizing the modified DPC model for
compaction of HE powders. The modified DPC model has been used to simulate the compaction properties of an
inert powder [9]. In this work, model parameters came from an instrumented press that measured the displacement
of the punches during compaction, the radial pressure in the die, and the axial load during compaction [9]. Han et al.
used the applied load for these measurements (note that our MCC-1 press only measures the transmitted load). There
are eight variables in the DPC model [9]. These variables are density dependent and are determined from both the
loading and unloading data collected by the instrument.
Prior to testing HE, and in an effort to reproduce the results of Han et al., we similarly compacted the inert
material Avicel® [9]. Our axial stress versus strain curves match those reported by Han et al. reasonably well,
Figure 4a. We do observe significant deviations in our data when the radial stress versus the axial stress are plotted,
Figure 4b (compare to Han [9] Figure 2). We are currently exploring the cause of these deviations in our
measurements. For the Han et al. studies, material weight was kept constant and the aspect ratio was varied in order
to obtain higher density pellets. For our study, the aspect ratio was kept relatively constant, within the error of the
MCC-1 press, and the weight of the pellets was varied to obtain higher or lower density pellets. This could relate to

060008-4
the observed differences. Avicel® compaction data, in advance of HE compaction data, will be used to inform
current application of the DPC model to HE powders [3].

100 Av7
300 Av7 1.45 g/cm3 Av7

Radial Stress (MPa)


Axial Stress (MPa)

90 Av5
Av5 1.35 g/cm3 80 Av5
Av4
3 70
200 Av4 1.15 g/cm Av3
60 Av2
Av2 0.90 g/cm3 50 Av2

40
100
30
20

0 10
0 0.2 0.4 0.6 0.8
Axial Strain 0 100 200 300
Axial Stress (MPa)
(a) (b)

FIGURE 4. a) Axial stress versus strain curves of Avicel® compacted to various densities, and b) the corresponding radial stress
versus axial stress curves.

CONCLUSIONS
We have pressed pellets of inert materials including sugar, new mock HMX candidates, and Avicel® in an
instrumented press. The compaction properties for different particle sizes of sugar have been compared, including
two bimodal particle size distributions. Heckel plot analysis showed the expected differences for changes in particle
size distribution. Bimodal particle size distributions optimize packing. Compaction properties of three newly-
proposed HMX mock materials were measured and compared. This study included one recycled material and
investigated the effect of particle size changes. Based on the compaction curves and the compression properties of
these three materials, we have concluded that IDOX shows the most potential as a new mock for HMX. Finally, we
have made progress demonstrating that the MCC-1 compaction instrument can help to parameterize the modified
DPC model for the compaction of HE.

ACKNOWLEDGMENTS
We want to thank the Q-6 group at LANL for purchasing the instrument and initial funding, Marvin Shorty for
synthesizing N-BPFPO, John Yeager and Amanda Higginbotham-Duque for providing the high TMD mock
materials and technical guidance for portions of this project. The G. T. Seaborg Institute Postdoctoral Fellowship
Program, the DOE/NNSA WSEAT Program and the Joint Munitions Program also provided funding for this work.

REFERENCES
1. S. A. Sheffield, R. L. Gustavsen and R. Alcon, Shock Compression of Condensed Matter Proceedings (1997).
2. J. D. Yeager, A. Higginbotham-Duque, M. Shorty, P. Bowden and J. Stull, Journal of Energetic Materials,
Journal of Energetic Materials, Accepted (2017).
3. R. Stevens and S. Ouellette, Los Alamos National Laboratory, Internal Report LA-CP-16-20218 (2016).
4. K. Yamaguchi, G. Matsumura, N. Haga and K. Shudo, Acta Crystallographica Section C 48 (3), 558 (1992).
5. Detail Sepcification: HMX (Cyclotetramethylenetetranitramine) MIL-DTL-45444C, (1996).
6. R. Ramberger and A. Burger, Power Technology 43, 1-9 (1985).
7. A. Hassanpour and M. Ghadiri, Powder Technology 141, 251-261 (2004).
8. D. J. Funk, G. W. Laabs, P. Peterson and B. W. Asay, AIP Conference Proceedings 370 (1), 145-148 (1996).
9. L. Han, J. Elliott, A. Bentham and A. Mills, International Journal of Solids and Structures 45, 3088 (2008).

060008-5

You might also like