Polymorphic Phase of RDX Report
Published on: Mar 4, 2016
Transcripts - Polymorphic Phase of RDX Report
Methods to Control the Polymorphic Phase of RDX-Based Explosives
DHS-HS STEM Summer Intern, Transportation Security Laboratory, William J. Hughes
Technical Center, Atlantic City, NJ 08405
The effect of mass loading and solvent deposition on the polymorphic phase of RDX was
examined using Raman spectroscopy. Direct depositions of quantified RDX and C4 solutions in
THF and ACN were made onto PTFE, or Teflon®, substrates at various mass loadings and imaged
with Raman microscopy to determine the initial phase. The effect of dry transfer was examined
with Raman measurements. An additional set of samples of C4 were made with aerosol spray
deposition at known masses and imaged with Raman microscopy to determine the effect of sample
Raman spectra of the samples revealed minimal differences between C4 and RDX deposits,
as well as solvent choice for RDX. Spectra of bulk RDX and C4 showed that α-RDX is the
dominant phase observed in bulk material. The solution deposits confirm that α- and β-RDX were
present at higher mass loadings, while only β-RDX was present at low mass loadings. Results
showed that direct deposit alone is incapable of producing α-RDX. However, the data indicate that
to successfully create the desired bulk phase (α-RDX), the depositions of RDX or C4 must be dry
transferred onto a substrate.
The Transportation Security Lab (TSL) aims to certify detection equipment by making
standards for fingerprints of explosives. Creating accurate and precise standards from fingerprints
is impossible due to the variability in mass and concentration from consecutive fingerprints. With
this in mind, the trace laboratory at the TSL uses techniques to mimic the different metrics of the
fingerprint, including aspects, such as particle distribution, particle size, adhesion, mass loading,
and isomorphic state.
The bulk component of the plastic C4 explosive is 1,3,5-trinitro-1,3,5-triazine (RDX). This
explosive exhibits polymorphism. Specifically, RDX has two polymorphs, α and β, that can exist
in ambient conditions as conformational polymorphs of one another. The α polymorph commonly
is found in bulk, or visible, quantities of C4
and has been determined through previous
studies to have an axial, equatorial
conformation of the NO2 groups around the
triazine ring (Figure 1).1,2 The α-phase is stable
at room temperature and atmospheric
conditions. The β-phase is more kinetically stable than the α-phase. β-RDX can remain on a surface
for long periods of time but will revert back to α-RDX if subjected to mechanical disturbances.
The β-phase contains all axial NO2 groups around the triazine ring.3
Characterizing the conversion from α- to β-RDX polymorphs using various solvents, mass
loadings, and deposition techniques will provide insight toward accurate imitation of C4
fingerprints. For example, if a specific deposition technique only produces β-RDX, use of those
artificial fingerprints for the test and evaluation of detection technologies, such as those based on
Raman emission, which is sensitive to polymorphs, will not provide realistic test results. The β-
phase may alter additional metrics important in trace test standards production, such as adhesion,
particle size, and distribution, thereby decreasing the standards validity. Additionally, it is
imperative that all field detection devices are able to identify all trace amounts of RDX and/or C4,
not just the more common α form.
Figure 1. AAE and AAA conformations of α and β-
Previous studies have concluded that the
polymorphic state of RDX is affected by deposition
technique and mass loading.4,5 To further such
discoveries, this study aims to explore various
solvent depositions at different mass loadings and
their effects on the polymorphic state of RDX via
The techniques to be used are “direct
deposit,” “dry transfer,” and aerosol spray deposition. Direct deposit involves depositing a known
volume and concentration of the threat in solvent onto a substrate. The solvent is allowed to air
dry until completely evaporated, mimicking the random distribution of particles in a fingerprint.
Dry transfer uses the direct deposit method onto PTFE (Teflon®) strips. Subsequently, the residue
is rubbed onto a new substrate to illustrate the transfer of a fingerprint onto a surface. Finally,
aerosol spray deposition involves a constant spray of a known volume and concentration of an
explosive in solvent onto a substrate.
Two RDX stock solutions were made in acetonitrile and tetrahydrofuran. One C4 stock
solution was made in tetrahydrofuran at a concentration of 20 mg/mL. Solution concentrations
underwent quality control procedures to verify concentrations and then diluted down to 10, 1, and
0.1 μg/μL. Deposits of 10 μL of each concentration were made in triplicate onto PTFE strips using
a calibrated micropipette (Eppendorf, P/N: 1691078) to provide samples at 100, 10, and 1 µg. The
solvent then was allowed to evaporate at room temperature.
Figure 2. Photograph of a RamanRXN™
Microprobe with an optical microscope.
The C4 stock solution was deposited onto PTFE strips in triplicate at 100, 10, and 1 μg
using an Aerosol Spray Deposition System (ASDS) in 1 cm2 deposits.
The measured Teflon® strips were used to dry transfer onto three-square-inch aluminum
test substrates (ACT, Batch: 50829811). This was accomplished by rubbing the dried PTFE strip
onto the test substrate to transfer the particles.
Raman spectra of the samples were collected via a RamanRXN™ Microprobe with optical
imaging capabilities (Figure 2). The excitation source used for the measurements was an Invictus™
785-nm NIR Laser, powered at 151 mW. The integration time was 5 s/spectra. Additionally,
Raman measurements and white light optical images were taken of each sample at a magnification
3. Results and Discussion
Raman spectroscopy was chosen to differentiate between α- and β-RDX in this study due
to availability in the lab. The laser detects the inelastically scattered light through the specimen.
The majority of the scattered light is the same energy as the laser source, but when the energy
interacts with the individual vibrations of a molecule, the energy frequency may shift, and the
difference in the energy is detected. The intensity of the shifted energy can be plotted with regard
to frequency, yielding a Raman spectrum of the sample. The spectrum provides information about
the chemical structure of the specimen. Theoretical spectra of RDX show multiple characteristic
peaks that distinguish α- from β-RDX; however, at low mass loadings, the height of most of these
peaks is too low to distinguish. For that reason, the most prominent of the distinct α- and β-RDX
peaks analyzed is characteristic to the ring breathing mode, or strain, which is located at 883 cm-1
and 877 cm-1, respectively. 4
The sensitivity of the Raman microscope was proven to display accurate measurements at
the predicted shifts of the peak of interest for β- and α-RDX. The literature value of the ring
breathing mode for α-RDX was verified by comparing it to collected bulk RDX and C4 Raman
spectra (Figure 4.a and Figure 4.b). The collected values were in agreement with the literature values
at 883 cm-1.4
C4 and RDX:
Raman measurements showed minimal differences between the C4 and pure RDX
frequency shifts using the direct deposit method. Results displayed in Table 1 and Table 2 show
that β- and α-RDX are both present at 100 and 10 μg on PTFE strips, and only β-RDX is present
at 1 μg. Minimal difference between C4 and RDX
data is expected. About 91 percent of the
composition of military-grade C4 in the United
States is RDX, giving it the characteristics of pure
RDX.6 The amount of pure RDX directly deposited
with respect to polymorphic change is consistent
with previous studies conducted on glass
β / α α
100.b β / α α
100.c β / α α
β / α α
100.b β / α α
100.c β / α α
β / α α
10.b β / α α
10.c β / α α
β / α α
10.b β / α α
10.c β / α α
1.b α α
1.c α α
1.b α α
1.c α α
Table 1. Polymorphic form of RDX at various
mass loadings for pure RDX.
Solvent did not appear to have a significant impact on the polymorphic form of the RDX
crystals when the same threat was deposited. Results compiled in Table 1 conclude that β- and α-
RDX are present at 100 and 10 μg on PTFE strips containing deposited RDX in acetonitrile
and tetrahydrofuran, with the
exception of one replicate
of 100 μg of RDX in THF.
At 1 μg, only β-RDX was
observed in acetonitrile and
Raman spectra showed that direct deposit produced β- and α-RDX at 100 and 10 μg and
only β-RDX at a mass of 1 μgfor pure RDX. C4direct deposit results showed no significant trends,
with a mixture of β and α-RDX at all mass loadings. Results for pure RDX direct deposit displayed
uniformly for all solvents and mass loadings. The presence of both β- and α-RDX can be explained
by first-order kinetics of the β-RDX isomer and the low energy barrier for interconversion of
Results for C4 direct deposit were sporadic, showing no trends. This could be due to the
selectivity of the areas imaged by the Raman microscope; in actuality, β- and α-RDX could have
Teflon Metal Teflon Metal
100.a β α α α
100.b β / α α β / α α
100.c β α α α
10.a β / α α β α
10.b β α β α
10.c β α n/o α
1.a β / α α β α
1.b β / α α β α
1.c β α β α
Table 2. Polymorphic form of RDX at various mass loading for C4 in
THF. Cells containing n/o were not observed.
been present when only β-RDX was observed. Due to time constraints, further research is required
to confirm this conclusion.
Aerosol spray deposition produced a combination of β- and α-RDX and only β-RDX at
100 and 10 μg and only β-RDX at 1 μg. Table 2 summarizes these findings. Observing only β-RDX
below 10 μg proves that the RDX polymorphic phase is sensitive to mass loading. The effect of
mass loading on the polymorphic state of RDX is theorized to be related to the kinetic stability of
trace amounts of β-RDX on the substrate surface.4 Substrate effects were not examined in this
Figure 3. 40x40 point Raman image of 1 µg RDX in ACN pre-dry transfer (a) and post-dry transfer (b) at
877 cm-1; 40x40 point Raman image of 1 µg RDX in ACN pre-dry transfer (c) and post-dry transfer (d) at
study but should be incorporated into future research to understand fully the effect of mass loading
on the metastable form of RDX.
Dry transfer uniformly produced only α-RDX after deposition with RDX and C4 at all
mass loadings at the locations that were imaged (Figure 4). This can be observed in Table 1 and
Table 2. It is apparent that only one isomorph is present at a time for the 1 µg deposit of RDX in
ACN pre- and post-transfer (Figure 3). Dry-transfer data for C4 and pure RDX displayed uniform
results at all mass loadings at the locations imaged. Dry transfer converted any β-RDX to α-RDX.
The energy barrier for the conversion from β- to α-RDX has been determined through previous
work to be less than 1 kcal/mol at room temperature.7 The low activation energy explains the
conversion to α-RDX after a small mechanical disturbance has been forced onto the β-RDX during
the dry transfer.
Deposits of C4 and RDX in various solutions under several deposition techniques were
examined with Raman microscopy. Results showed that direct deposit will not produce a
polymorphically accurate artificial fingerprint with solely α-RDX at a mass as high as 100 µg.
Additionally, spectra indicated that a switch from α- and β-RDX to only β-RDX can occur below
10 µg. The data conclude that to successfully create a standard for fingerprint production in only
α-RDX, depositions of RDX or C4 must be dry transferred onto a substrate. Substrate effects on
RDX polymorphism should be incorporated into future research to understand fully the effect of
mass loading on RDX phases.
I would like to thank the Department of Homeland Security (DHS) Science and
Technology (S&T) Directorate Office of University Programs, as well as the Oak Ridge Institute
for Science and Education (ORISE) for supporting my research by appointment with the HS-
STEM Summer Internship Program. Additionally, I would like to express a vast amount of
gratitude toward Dr. Jason Stairs, Dr. John Brady, Dr. Alex Gordon, and all other TSL employees
for being exemplary mentors and for graciously taking time from their busy schedules to support
my research. I have taken away a remarkable expanse of knowledge, experience, and appreciation
for the significance of the work that the TSL accomplishes. I thank you for inviting me into your
workplace and laboratories and for encouraging my enthusiasm as a scientist.
(1) Hakey, P.; Ouellette, W.; Zubieta, J.; Korter, T. “Redetermination of cyclo-
trimethylenetrinitramine”. Acta Cryst. E 2008, 64, 1428.
(2) Choi, C. S.; Prince, E. “The crystal structure of cyclotrimethylenetrinitramine”. Acta Cryst.
B 1972, 28, 2857.
(3) Millar, D. I. A et al. “The crystal structure of beta-RDX-an exclusive form of an explosive
revealed”. Chem. Comm. 2009, 2009, 562-564.
(4) Torres, P et al. “Vibrational Spectroscopy Study of β and α RDX deposits”. J. Phys. Chem.
B 2004, 108, 8799-8805.
(5) Emmons, E. D et al. “Characterization of Polymorphic States in Energetic Samples of 1,3,5-
Trinitro-1,3,5-Triazine (RDX) Fabricated Using Drop-on-Demand Inkjet Technology”. Appl. Sci.
2012, 66, 628-635.
(6) Reardon, Michelle R.; Bender, Edward C. “Differentiation of Composition of C-4 Based
on the Analysis of the Process Oil”. J. Forensic Sci. 2005, 50, 1–7.
(7) Vladimiroff, T.; Rice, B. M. “Predictions of Crystal Structures from First Principles”. J.
Phys. Chem. A 2002, 106, 10437.
Figure 4. Representative Raman spectra of bulk RDX (a) and C4 (b); 1 μg direct deposit of RDX in ACN on
Teflon pre-dry transfer (c); 1 μg direct deposit of C4 in THF on Teflon pre-dry transfer (d); 1 μg RDX in ACN
on bare aluminum post-dry transfer (e); 1 μg C4 in THF on bare aluminum post-dry transfer (f).