Flow-Tube Investigations of Hypergolic Reactions of a Dicyanamide Ionic Liquid Via Tunable Vacuum Ultraviolet Aerosol Mass Spectrometry

Article abstract of DOI:10.1021/acs.jpca.6b06289

The unusually high heats of vaporization of room-temperature ionic liquids (RTILs) complicate the utilization of thermal evaporation to study ionic liquid reactivity. Although effusion of RTILs into a reaction flow-tube or mass spectrometer is possible, competition between vaporization and thermal decomposition of the RTIL can greatly increase the complexity of the observed reaction products. In order to investigate the reaction kinetics of a hypergolic RTIL, 1-butyl-3-methylimidazolium dicyanamide (BMIM⁺DCA^-) was aerosolized and reacted with gaseous nitric acid, and the products were monitored via tunable vacuum ultraviolet photoionization time-of-flight mass spectrometry at the Chemical Dynamics Beamline 9.0.2 at the Advanced Light Source. Reaction product formation at m/z 42, 43, 44, 67, 85, 126, and higher masses was observed as a function of HNO₃ exposure. The identities of the product species were assigned to the masses on the basis of their ionization energies. The observed exposure profile of the m/z 67 signal suggests that the excess gaseous HNO₃ initiates rapid reactions near the surface of the RTIL aerosol. Nonreactive molecular dynamics simulations support this observation, suggesting that diffusion within the particle may be a limiting step. The mechanism is consistent with previous reports that nitric acid forms protonated dicyanamide species in the first step of the reaction.

Full text of DOI:10.1021/acs.jpca.6b06289

Subscriber access provided by Heriot-Watt | University Library
Article
Flow-Tube Investigations of Hypergolic Reactions of a Dicyanamide
Ionic Liquid Via Tunable Vacuum Ultraviolet Aerosol Mass Spectrometry
Steven D Chambreau, Christine J. Koh, Denisia Maria Popolan-Vaida, Christopher J.
Gallegos, Justin B. Hooper, Dmitry Bedrov, Ghanshyam L. Vaghjiani, and Stephen R Leone
J. Phys. Chem. A, Just Accepted Manuscript • DOI: 10.1021/acs.jpca.6b06289 • Publication Date (Web): 22 Sep 2016
Downloaded from http://pubs.acs.org on September 26, 2016
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
The Journal of Physical Chemistry A is published by the American Chemical Society.
1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 41
The Journal of Physical Chemistry
1
2
3
4
5
6
7
8
9
Flow-Tube Investigations of Hypergolic Reactions of a Dicyanamide Ionic Liquid
Via Tunable Vacuum Ultraviolet Aerosol Mass Spectrometry
1
2
2,3
4
Steven D. Chambreau, Christine J. Koh, Denisia PopolanꢀVaida, Christopher J. Gallegos,
5,6
5,6
4,*
2,3
Justin B. Hooper, Dmitry Bedrov, Ghanshyam L. Vaghjiani Stephen R. Leone
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
2
.
.
ERC, Inc., Edwards Air Force Base, California, 93524, U.S.A.
Departments of Chemistry and Physics, University of California, Berkeley, California, 94720, U.S.A.
3. Chemical Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, California, 94720, U.S.A.
4. Propellants Branch, Rocket Propulsion Division, Aerospace Systems Directorate, Air Force Research
Laboratory, AFRL/RQRP, Edwards Air Force Base, California, 93524, U.S.A.
5
.
Department of Materials Science & Engineering, University of Utah, 122 South Central Campus Drive,
Room 304, Salt Lake City, Utah 84112, U.S.A.
Wasatch Molecular Inc., 825 North, 300 West, Salt Lake City, Utah 84103, U.S.A.
6
.
Abstract
The unusually high heats of vaporization of roomꢀtemperature ionic liquids (RTILs) complicate
the utilization of thermal evaporation to study ionic liquid reactivity. Although effusion of
RTILs into a reaction flowꢀtube or mass spectrometer is possible, competition between
vaporization and thermal decomposition of the RTIL can greatly increase the complexity of the
observed reaction products. In order to investigate the reaction kinetics of a hypergolic RTIL, 1ꢀ
butylꢀ3ꢀmethylimidazolium dicyanamide (BMIM DCA ) was aerosolized and reacted with
gaseous nitric acid, and the products were monitored via tunable vacuum ultraviolet
photoionization timeꢀofꢀflight mass spectrometry at the Chemical Dynamics Beamline 9.0.2 at
the Advanced Light Source. Reaction product formation at m/z 42, 43, 44, 67, 85, 126, and
higher masses was observed as a function of HNO exposure. The identities of the product
species were assigned to the masses based on their ionization energies. The observed exposure
profile of the m/z 67 signal suggests that the excess gaseous HNO initiates rapid reactions near
the surface of the RTIL aerosol. Nonꢀreactive molecular dynamics simulations support this
observation, suggesting that diffusion within the particle may be a limiting step. The mechanism
is consistent with previous reports that nitric acid forms protonated dicyanamide species in the
first step of the reaction.
+
ꢀ
3
3
*
Email: ghanshyam.vaghjiani@edwards.af.mil; Tel: 661ꢀ275ꢀ5657; Fax: 661ꢀ275ꢀ5471
1
ACS Paragon Plus Environment

The Journal of Physical Chemistry
Page 2 of 41
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
INTRODUCTION
The emergence of room temperature ionic liquids (RTILs) in the last decade has stemmed
mostly from their extremely high enthalpies of vaporization, driving chemical technologies
where reduction of vapors is important, such as replacement of volatile organic compounds
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2-7
(VOC) as solvents and highly toxic and volatile liquid propellants. Additional interest in
8
RTILs is based on their conductive properties for use as electrolytes in fuel cells and on their
9
,10
high thermal stabilities for high temperature applications such as heat exchange.
The
significant enthalpies of vaporization are due to the ionic nature of the liquids, where the
intermolecular attractive Coulombic force is significantly stronger than those found in molecular
liquids that are held together by van der Waals, dispersion and possibly hydrogen bonding.
Therefore, the energy required to separate ions, ion pairs or clusters of ions from the surface of
the ionic liquid is much higher than for molecular liquids. In fact, competition between
vaporization of intact RTILs and thermal decomposition to form molecular species that
1
1,12
subsequently vaporize can both contribute to mass loss upon heating of RTILs.
While it has been previously demonstrated that RTILs can be evaporated intact and
1
3
introduced into a mass spectrometer (MS) for detection, differentiating between MS signals of
the intact RTIL, the thermal decomposition products and molecular impurities in the RTIL is
rather difficult. The resulting number densities in the gas phase from the effusive source are
quite low, and highly sensitive MS techniques are needed, usually with extensive signal
averaging required.
1
4
An alternative to effusion sources that has been recently demonstrated is to produce
RTIL aerosols, which can be introduced into the MS, flash vaporized and detected via
photoionization techniques. Another advantage to this approach is that the RTIL aerosols can be
2
ACS Paragon Plus Environment

Page 3 of 41
The Journal of Physical Chemistry
1
2
3
4
5
6
7
8
9
reacted with oxidizers and the resulting processed aerosol particles can be probed for product
species. Previous attempts to study the reactivity of hypergolic RTILs via supersonic reactive
scattering of HNO with a RTIL surface and probing via angleꢀresolved quadrupole MS resulted
3
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
†
in detection only of reactive halide impurities present in the RTILs from their synthesis.
The interest in understanding the hypergolic ignition mechanism of RTILs stems from the
desire to improve the performance of the stateꢀofꢀtheꢀart bipropellants used currently, i.e.,
monomethylhydrazine and nitrogen tetroxide. It has been estimated that the number of ionic
1
8 15
combinations of cations and anions that could potentially be ionic liquids is ~ 10 . Since it
would be impossible to synthesize all of these potential ionic liquids and test them for
hypergolicity, understanding the hypergolic ignition mechanisms of RTILs will greatly support
the efficient development of nextꢀgeneration RTIL propellants. Previous work on the hypergolic
ꢀ
ꢀ
ignition of dicyanamide (DCA : N≡CꢀNꢀC≡N ) based ionic liquids with HNO as the oxidizer
3
indicates the reaction occurs primarily on the anion, proceeding through a highꢀenergy
intermediate that is highly thermally unstable and decomposes to create the “chemical spark”
3
needed for ignition. Heat released from the reaction can cause thermal decomposition of the
RTIL and oxidizer, and can contribute to the proper conditions leading to ignition. The ignition
2
delay of these RTILs has also been modeled using chemical kinetics modeling. More recently,
other hypergolic RTILs have been discovered with different anions (e.g., nitrocyanamides,
6
,16
4
cyanoborohydrides, aluminoborohydrides, etc.), and different oxidizers (primarily H O ), or
2
2
17,18
even with hypergolic cations.
This paper presents the experimental investigation of the reactivity of dicyanamideꢀbased
hypergolic RTIL aerosols with HNO via tunable vacuum ultraviolet (VUV) photoionization
3
timeꢀofꢀflight mass spectrometry (TOFMS) and discusses how the results give insight into
†
Unpublished work carried out with the Minton Group at Montana State University, Bozeman, MT.
3
ACS Paragon Plus Environment

The Journal of Physical Chemistry
Page 4 of 41
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
hypergolic reactivity. To our knowledge, this is the first study investigating the reactivity of
ionic liquid aerosols. This method could ultimately prove useful in investigating reaction
mechanisms for hypergolic combinations with new RTILs and other oxidizers. Nonꢀreactive
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
molecular dynamics simulations indicate that diffusion of HNO within the RTIL aerosol particle
3
+
may be a limiting factor in hypergolic product formation. Additional insight on the role of NO2
in the hypergolic mechanism of RTILs using FTIR spectroscopy is also discussed.
EXPERIMENTAL
The ionic liquid used in these experiments is 1ꢀbutylꢀ3ꢀmethylimidazolium dicyanamide,
+
ꢀ
BMIM DCA (98%, EMD Chemicals), which was used without further purification. White
fuming nitric acid was synthesized by reacting sulfuric acid (oleum) with sodium nitrate or
15
15
isotopicallyꢀlabeled potassium nitrate (K NO , Aldrich: 98 atom % N) and the nitric acid was
3
distilled off from the reaction vessel.
The aerosol reaction kinetics flowꢀtube apparatus is depicted in Figure 1. The ionic
+
ꢀ
liquid aerosol was generated using solutions of BMIM DCA in deionized water varying in
concentration from 0.2 to 4.6 grams per liter and spraying these solutions through an atomizer
1
4
with 10 psi nitrogen backing pressure. The resulting particles were dried through a 1ꢀmeterꢀ
long drying tube containing silica gel desiccant, but some residual water likely remains in the
aerosols. The resulting ionic liquid aerosol size distributions (Figure 2) were measured using a
scanning mobility particle sizer (SMPS), and typical aerosol size distributions ranged from 50 to
500 nm with an optimal peak at 100 nm as determined by controlling the backing pressure in the
atomizer. The nitric acid was introduced using a glass bubbler, and the bubbler temperature was
maintained at 20 °C in a water bath, yielding 50 Torr of HNO . The HNO concentration was
3
3
4
ACS Paragon Plus Environment

Page 5 of 41
The Journal of Physical Chemistry
1
2
3
4
5
6
7
8
9
monitored by observing the optical absorbance at wavelength, λ = 213.9 nm, using an
absorbance cell of length = 10.0 cm equipped with a Zn lamp and a photodetector attached to a
ꢀ19
2
suitable bandpass filter. An absorption cross section value of σ(ΗΝΟ3) = 4.39 x 10 cm
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ꢀ
1
19
molecule at 213.9 nm was used. The nitric acid was diluted in nitrogen to the appropriate
concentrations using a calibrated mass flow controller. The ionic liquid aerosol flow was rapidly
mixed with the diluted nitric acid stream and the mixture was flowed through a 1.0ꢀinchꢀID
Pyrex tube. Reaction times were calculated using the volume of the tube (length = 25 or 50 cm)
3
ꢀ1
divided by the total flow (500 cm min , at STP) through the tube. The resulting [HNO ] range
3
1
3
ꢀ3
5
was (1ꢀ50) x 10 molec cm , while the aerosol number density was fixed at ~ 1 x 10 particle
ꢀ
3
cm . The reaction was quenched at the end of the flowꢀtube by running the reacted flow through
a denuder containing sodium carbonate (Na CO ) pellets to absorb and neutralize the unreacted
2
3
nitric acid.
The chemical composition of the aerosols exiting the reactor is measured using a custom
built, softꢀionization vacuum ultraviolet aerosol mass spectrometer as described by Gloaguen et
2
0
al. Briefly, the reactor mixture is sampled through a 500 ꢀm orifice coupled to an aerodynamic
lens to select aerosol particles ≥ 50 nm for the purpose of focusing the particle beam and
reducing the gas phase molecular concentration by means of pinhole plates described by Liu et
2
1,22
al.
Subsequently, inside the mass spectrometer, the particles are thermally vaporized on a
heated copper block at ~ 115 °C and photoionized by tunable VUV radiation produced by the
Chemical Dynamics Beamline at the Advanced Light Source. Mass spectra were taken as scans
of the photon energy in the range from 7.4 to 15.0 eV, typically in 0.1 eV steps and averaged for
5
00,000 repeller pulses, to improve the signalꢀtoꢀnoise ratio of the measured photoionization
energy curves of the detected products. A part of the remaining flow exiting the reactor is
5
ACS Paragon Plus Environment

The Journal of Physical Chemistry
Page 6 of 41
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
sampled by the SMPS to simultaneously determine the particle size distribution (see Figure 2)
and number concentration.
Additionally, a rapidꢀscan Fourier transform infrared spectroscopy apparatus (described
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
previously ) was utilized to perform drop tests with mixtures of varying ratios from 0ꢀ100 wt%
+
ꢀ
of H SO in HNO onto BMIM DCA . The purpose of mixing H SO and HNO is to generate
2
4
3
2
4
3
+
+
NO and investigate the influence of NO on the hypergolic reaction. Several drops of the
2
2
+
ꢀ
BMIM DCA were put into a cuvette in the apparatus, and the H SO /nitric acid mixtures were
2
4
+
ꢀ
introduced by dropping from a syringe into the cuvette containing the BMIM DCA . The IR
ꢀ1
ꢀ1
beam probed the volume directly above the cuvette to record spectra from 500ꢀ3850 cm (4 cm
resolution) every 10 msec.
THEORY
Ab initio Calculations
Theoretical values of the ionization energy (IE) were determined by calculating the gasꢀ
phase free energy difference between the cationic and neutral species at 298 K and corrected for
zeroꢀpoint energies using the M06/6ꢀ31+G(d,p) level of theory. Condensed phase simulations
2
3
11
using the generic ionic liquid model (SMDꢀGIL), described previously, were performed at the
M06/6ꢀ31G+(d,p) level of theory to determine the free energy profiles of proposed initial
+
−
reactions. The SMDꢀGIL solvent descriptor input parameters used for BMIM DCA as a solvent
are e = 11.50, n = 1.4300, γ = 61.24, Σα H
= 0.229, Σβ
H
= 0.265, φ = 0.2000 and ψ = 0.0000. An
2
2
estimate of the uncertainty in the calculated values is ± 0.2 eV for the ionization energies (mean
unsigned error) and 7.9 kJ/mol for the SMDꢀGIL model calculations (average mean unsigned
24
error).
6
ACS Paragon Plus Environment

Page 7 of 41
The Journal of Physical Chemistry
1
2
3
4
5
6
7
8
9
APPLE&P Molecular Dynamics
To obtain molecular level insight into the factors contributing to the experimental product
+
ꢀ
curves, we conducted atomistic molecular dynamics (MD) simulations of a BMIM DCA film
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
exposed to HNO vapor. The objective of these simulations is to investigate the interplay
3
between the adsorption of HNO vapor to the surface of the RTIL droplet and the diffusion of
3
those molecules inside the droplet. Each system contained 300 RTIL pairs and 100 HNO3
molecules. In these simulations, the simulation cell dimensions were 35 x 35 x 140 Å with the
larger dimension being perpendicular to the RTIL film interface. Simulations were conducted at
333 K. MD simulations utilized the atomistic polarizable APPLE&P force field that has been
2
5
tuned for description of energetic RTILs and their mixtures with nitric acid. Covalent bond
26
lengths were constrained using the velocityꢀVerlet form of the SHAKE algorithm. The Ewald
summation method was used for treatment of longꢀrange electrostatic forces between partial
atomic charges and between partial charges and induced dipoles. A tapering function was used
to drive the induced dipole/induced dipole interactions to zero at a cutoff of 10 Å, with scaling
starting at 9.3 Å. Induced dipoles were calculated using an iterative procedure at every time
step. A cutoff of 10 Å was used for all van der Waals interactions and the real part of
electrostatic interactions in the Ewald summation. A multiple time step reversible reference
27
system propagator algorithm was employed. A time step of 0.5 fs was used for bonding,
bending, dihedral, and outꢀofꢀplane deformation motions, while a 1.5 fs time step was used for
nonꢀbonded interactions within a cutoff radius of 6.0 Å. Finally, the nonꢀbonded interactions in
the range between 6.0 and 10.0 Å and the reciprocal part of the electrostatic interactions were
updated every 3 fs.
7
ACS Paragon Plus Environment

The Journal of Physical Chemistry
Page 8 of 41
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
RESULTS AND DISCUSSION
Vacuum Ultraviolet Photoionization Time-of-Flight Mass Spectrometry
I. Product Identification
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
+
ꢀ
Representative mass spectra of the BMIM DCA aerosols alone (black) and upon
reaction with HNO (red) at 11.0 and 13.5 eV photoionization energies and 115 °C copper block
3
+
ꢀ
temperature can be seen in Figure 3. Upon reaction with HNO gas, the BMIM DCA aerosol
3
leads to several product species characterized by m/z 42, 43, 44, 67, 85 and 126. By measuring
the photoionization efficiency (PIE) curves and their ionization energy (IE) values, of these
peaks, see Figure 4, we have attempted to identify their structures by comparison of their
experimental IE values with M06ꢀcalculated IE values and literature IE values, if available.
Photoion (PI) appearance energies were determined for the major photoions by fitting the PIE
2
14,28,29
curves to the function α(EꢀE ) near the photoionization threshold, as described previously.
o
These are reported along with their 298 K calculated (M06/6ꢀ31+G(d,p)) adiabatic ionization
energies in Table 1.
The formation of melamine (C H N , m/z 126) has been seen previously in hypergolic
3
6
6
3
0,31
reactivity experiments and in dicyanamide salts thermal decomposition.
Melamine has a
theoretical IE value of 8.5±0.2 eV (M06) that matches well with our m/z 126 experimental IE
value of 8.7±0.2 eV. Alternative m/z 126 species are discussed in the Supporting Information.
Although the melting point of melamine is above 250 °C, it could have a detectable vapor
ꢀ6
pressure at the copper heater block temperature of 115 °C at reduced pressure of 10 Torr. The
3
0
32
melamine cation is also known to fragment to m/z 85, 43 and 42 (see Table 1), and this is also
discussed further in the Supporting Information.
8
ACS Paragon Plus Environment

Page 9 of 41
The Journal of Physical Chemistry
1
2
3
4
5
6
7
8
9
Table 1. Photoproduct ionization energies (IE) and appearance energies (AE).
REACTION
PI mass
126
85
IE (expt., eV)
8.7
IE (theo, eV)
8.5
+
ꢀ
126 + hv ꢀ 126 + e
+
ꢀ
85 + hv ꢀ 85 + e
10.2
13.2
13.8
11.4
10.3
10.2
10.0
+
ꢀ
CO + hv ꢀ CO + e
44
44
43
42
2
2
+
ꢀ
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
N O + hv ꢀ N O + e
2
2
+
ꢀ
HNCO + hv ꢀ HNCO + e
11.7
10.4
10.3
+
ꢀ
H NCN + hv ꢀ H NCN + e
2
2
+
ꢀ
HNCNCN + hv ꢀ HNCNCN + e
67
REACTION
PI mass
85
AE (expt., eV)
10.2
AE (theo, eV)
11.6
+
ꢀ
ꢀ
1
1
1
26 + hv ꢀ H NCNC(NH)NH + HNCN + e
2
2
+
26 + hv ꢀ H NCNH + HNCNC(N)NH + e
43
42
44
42
43
67
11.4
10.3
10.5
10.3
11.4
10.2
11.96
13.23
10.45
10.6
11.9
11.2
2
2
+
ꢀ
26 + hv ꢀ H NCN + (H NCN) + e
2
2
2
+
ꢀ
85 + hv ꢀ H NCO + HNCN + e
85 + hv ꢀ H NCN + HNCO + e
85 + hv ꢀ HNCO + H NCN + e
2
+
ꢀ
2
+
ꢀ
2
+
ꢀ
+
ꢀ
BMIM DCA + hv ꢀBMIM: + HNCNCN + e
The formation of cyanourea (m/z 85) has been proposed as the product of the acidꢀ
3
catalyzed reaction product of dicyanamide anion and water, and it could occur in this work due
to the water retained in the RTIL aerosol generation process. Cyanourea has a theoretical IE
value of 10.0±0.2 eV (M06) that is consistent with the experimental IE of m/z 85 of 10.2 eV.
Alternative m/z 85 species are discussed in the Supporting Information. Fragmentation of the
cyanourea cation has also been considered in the Supporting Information.
As the product masses get smaller, the number of possible species is also reduced,
making identification less complicated. For m/z 42, the experimental IE value of 10.3±0.2 eV
matches very well with that for cyanamine, H NCN (calculated IE value of 10.4±0.2 eV, M06,
2
3
3
and literature IE value of 10.4 eV ), and we identify m/z 42 as cyanamine. M/z 42 alternate
33
species include the closedꢀshell H N C species diazomethane (IE value of 8.999±0.001 eV ) and
2
2
3
3
3
ꢀH diazirine (IE value of 10.3 and 10.18±0.05 eV ) but 3ꢀH diazirine is highly thermally
unstable and is not suspected in this system. Also, ionization of ketene, H C O, with known IE
2
2
33
value of 9.617±0.003 eV, can effectively be eliminated as a possible source for m/z 42. In
9
ACS Paragon Plus Environment

The Journal of Physical Chemistry
Page 10 of 41
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
order to exclude the formation of m/z 42 as a dissociative photoionization fragment from larger
m/z 85 or 126 products, the appearance energies (AEs) of m/z 42 from these larger species were
calculated (see Table 1). The calculated value of AE (M06) for m/z 42 resulting from
dissociative ionization of cyanourea (m/z 85) is 10.6 eV and the AE from melamine (m/z 126) is
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
+
13.2 eV, although any m/z 42 resulting from fragmentation of 126 would be expected to be
+
significantly less than 10% of the 126 peak intensity when compared to the electronꢀimpact
3
2
ionization (~ 70 eV) mass spectrum of melamine. The experimental value of 10.3 eV is less
than the 10.6 eV AE of m/z 42 from fragmentation of cyanourea. On the other hand, cyanourea
can thermally decompose to H NCN + HNCO with a thermal decomposition onset temperature
2
3
4
ꢀ6
of Tdec = 100 °C. However, with the copper heater block temperature = 115 °C at ~ 10 Torr
pressure, evaporation of cyanourea would be expected to be more significant than thermal
decomposition, but we cannot entirely eliminate the latter possibility as a source for m/z 42 (with
10.4 eV M06 IE value).
Previous hypergolic ignition experiments indicate the formation of HNCO (m/z 43) as
3
+
ꢀ
evidenced by IR spectroscopy, and the appearance of m/z 43 in the reaction of BMIM DCA
aerosols + HNO supports this. The experimental IE value for m/z 43 is 11.4±0.2 eV, and while
3
the theoretical IE value (M06) is slightly higher at 11.7±0.2 eV, the literature value is known to a
3
5
high degree of accuracy to be at an IE value of 11.595±0.005 eV, which matches reasonably
well with the experimental value. In addition, the shape of the PIE curve is very similar to the
35
literature photoelectron spectrum. Attempts were also made to identify possible dissociative
photoionization fragmentation of the m/z 85 and m/z 126 species as sources for m/z 43 using
M06 methods (see Table 1) and by examining the PIE curves carefully to look for features that
10
ACS Paragon Plus Environment

Page 11 of 41
The Journal of Physical Chemistry
1
2
3
4
5
6
7
8
9
could be resolved and demonstrate fragmentation is occurring. Unfortunately, the signalꢀtoꢀ
noise ratio in the measured m/z 43 signal was not sufficient to discern this.
The experimental photoion appearance energy of m/z 67 has a value of 10.2±0.2 eV, and,
as seen in Table 1, the M06 IE of the protonated DCA (HDCA = HNCNCN) has a value of
10.3±0.2 eV, which suggests that the m/z signal is from HDCA. (Species like C H with much
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
5
7
3
3
lower IE values are ruled out as alternate candidates for m/z 67 as were C NH and C HNO. )
4
5
3
+
ꢀ
15
By reacting BMIM DCA aerosols with isotopically labeled H NO , the expected
3
3
products ꢀ as predicted by the hypergolic reaction mechanism ꢀ at m/z 44, namely CO and N O,
2
2
15 14
can be separated, as the isotopically labeled product N NO appears at m/z 45, as seen in
Figure 5. Alternately, the contribution to the m/z 44 peak could be separated by increasing the
resolution of the light source by reducing the slit size the light passes through. Highꢀresolution
PIE curves for pure CO , pure N O, and a mixture of CO /N O gases leaked into the MS for
2
2
2
2
calibration purposes can be seen in Figure 6. Note that the CO peak centered at 13.5 eV can be
2
baselineꢀresolved from the N O peak centered at 14.3 eV. The separation of CO and N O at
2
2
2
high resolution allows for the measurement of the CO and N O yields as a function of HNO
2
2
3
ꢀ
exposure. The overall proposed hypergolic reaction mechanism is: [N(CN) ] + 4HNO
→
2CO2
2
3
ꢀ
+
3N O + 2H O + NO with a ratio of N O/CO = 3/2 for the complete reaction. The CO and
2
2
3
2
2
2
N O response curves as a function of HNO exposure can be seen in the Supporting Information.
2
3
Although the shapes of the response curves are consistent with each other and with those of m/z
2, 43, 85 and 126, the intensity ratio of the N O/CO response of approximately 5:1 does not
4
2
2
match the 3/2 ratio mentioned above, likely due to interferences from other photoion fragments
such as from fragments of m/z 85 and m/z 126. Unfortunately, due to this, we cannot get an
accurate stoichiometric ratio from this data.
11
ACS Paragon Plus Environment

The Journal of Physical Chemistry
Page 12 of 41
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
II. Species Profile
The decay of HDCA (m/z 67) signal as a function of HNO exposure was investigated to
3
ꢀ
determine the second order rate coefficient for the reaction of DCA with HNO . A detailed
3
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
analysis of this work is presented in the Supporting Information. Also, an attempt to vary
aerosol particle size to vary surface area to volume (SA/V) was made in order to see how the
decay of m/z 67 signal is impacted with exposure. The aerosol size distribution as measured by
the average particle diameter was able to be varied from ~ 110 nm to ~ 150 nm by varying the
backing pressure and the RTIL concentration in the solution in the aerosol generator, yielding a
relative difference in SA/V of 27%. However, this was not significant enough to detect changes
in the shapes of the exposure curves for m/z signals for HDCA, CO , N O, etc. An alternate
2
2
method to generate RTIL aerosols of highly varied sizes or to select monodisperse portions of
the aerosol distributions of different radii would be beneficial to investigate the impact on the
proposed kinetics mechanism.
Molecular Dynamics Modeling
Initially, the systems were equilibrated with repulsive biasing interactions that prevented
aggregation of HNO in the gas and its adsorption onto the RTIL film. As the RTIL liquid
3
interface and the HNO gas phase have equilibrated the biasing potential was turned off, and
3
processes of HNO adsorption and diffusion through RTIL film were monitored as seen in Figure
3
7. Figure 8 shows the density profiles illustrating molecular distributions in the system for
several timeframes. As expected, the density profiles of the RTIL phase (illustrated by the
+
density of BMIM cations) are barely changing upon interaction with HNO gas. However, the
3
density profiles of HNO show more significant changes. First, one can observe that within 20
3
12
ACS Paragon Plus Environment

Page 13 of 41
The Journal of Physical Chemistry
1
2
3
4
5
6
7
8
9
ps a pronounced peak in the HNO density has developed at the interface with RTIL, indicating
3
rapid adsorption of HNO molecules that were in the vicinity of the RTIL film. Interestingly this
3
peak remains strong after 4 ns and 6 ns of simulations. The figure illustrates that diffusion of
HNO through RTIL film is relatively slow. A region of low HNO can be observed in the
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
3
3
middle of the RTIL film even after 6 ns of simulations. We would like to point out that in our
previous MD simulation study of liquidꢀliquid mixing of the same RTIL and HNO using similar
3
sized RTIL film, complete intermixing of HNO and RTIL phase had occurred after 6 ns of
3
25
simulation. In contrast, gas interpenetration appears to be much slower in the simulations,
likely due to the much lower concentration gradient in the gasꢀliquid system. This result is
consistent with other experimental observations of solubility of gases like CO in ionic liquids,
2
where the diffusion coefficient is found to be 1ꢀ2 orders of magnitude smaller than that in
36
ꢀ
molecular liquids. Hence, it is likely that much of the reaction, e.g., protonation of DCA , and
subsequent reactions leading to product formation would occur near the surface; For 150ꢀnm
particles, this surface depth varies from 8 to 30 nm, respectively, for S=0.28 and 0.78, where S is
the (outer) fractional volume of the total RTIL particle volume that is available for HNO3
reaction to take place, when the corresponding (kdecomp/(kdecomp+k ) values where 0.1 and 0.9, as
vap
discussed in the Supporting Information. For 110ꢀnm particles, the corresponding depth varies
from 6 to 22 nm. Therefore, the measured phenomenological rate coefficient k = 4.14±0.32 x
2
ꢀ16
ꢀ1
3 ꢀ1
10 molec cm s (see Supporting Information for how this value was obtained) would have
contributions from both the intrinsic reaction rate and HNO diffusion rate within the aerosol.
3
The rate coefficient k is expected to be independent of the diffusion rate of gaseous HNO onto
2
3
ꢀ
the particle surface, which is fast, with an equivalent rate coefficient calculated to be ~ 1.2 x 10
11
ꢀ1
3
ꢀ1
37
molec cm s for our experimental flowꢀtube conditions. The heterogeneous nature of this
1
3
ACS Paragon Plus Environment

The Journal of Physical Chemistry
Page 14 of 41
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
+
ꢀ
system and the phobic behavior of the very dissimilar HNO solute in the BMIM DCA solvent
3
appear to have contributed to the observed m/z product profiles. Previous studies on aerosol/gas
reactions have also observed interfacial reactivity behavior; where the reaction is either limited at
the surface face or is limited by the slow diffusion of the gaseous reactant into the aerosol
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
particle, and our results appear to rule out rapid diffusion within the RTIL so that [HNO ] would
3
3
8
be constant throughout. The competing forces which ultimately control the HNO diffusion
3
rate and distribution within the aerosol particle likely include particle surface energy and
chemical potential, and diffusion cannot be simply described as driven by the entropy of mixing.
Generic Ionic Liquid Model
In a recent paper, it was proposed that the relative acidity of HNO and HDCA would
3
11
reverse from the gas phase to the condensed phase in the ionic liquid. This was simulated
2
3
using the soꢀcalled generic ionic liquid (SMDꢀGIL) model. In the gas phase, a proton would
ꢀ
ꢀ
favor binding with NO over DCA , but in the simulated condensed phase, the proton would
3
ꢀ
prefer binding to the DCA . This has important implications in the hypergolic reaction, as proton
ꢀ
3
transfer from HNO to DCA has been proposed as the initial step in hypergolic ignition. In
3
light of this, it is informative to estimate the equilibrium constant, K , in the condensed phase
eq
ꢀ
ꢀ
o
for the reaction: HNO + DCA ↔ NO + HDCA using the relationship ꢁG = ꢀRTlnK . The
3
3
eq
o
acid
+
ꢀ
+
ꢀ
SMDꢀGIL ꢁG
values for the reactions, HNO → H + NO and HDCA → H + DCA are
3
3
5
66.6 and 578.0 kJ/mol, respectively. The average mean unsigned error for thermochemical
o
kinetics using M06 is 7.9 kJ/mol. Combining these reactions gives an overall SMDꢀGIL ꢁG rxn
ꢀ
=
ꢀ11.4 kJ/mol for the reaction HNO + DCA ↔ NO + HDCA, yielding a value of K = 97,
3
3
eq
indicating the forward reaction is highly preferred over the reverse reaction:
1
4
ACS Paragon Plus Environment

Page 15 of 41
The Journal of Physical Chemistry
1
2
3
4
5
6
7
8
9
K = k /k = 97
eq
f
rev
3
The proposed second and third steps in the hypergolic ignition reaction were the
ꢀ
nucleophilic attack of NO on the protonated DCA, followed by NO migration to the terminal
3
2
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
nitrogen. The enthalpy profile for this mechanism has been calculated recently at the
3
9
B3LYP/augꢀccꢀPVDZ and MP2/6ꢀ311++G(d,p) levels of theory in the gas phase. The
enthalpies and free energies for the first three steps of the mechanism are calculated here using
the M06/6ꢀ31+G(d,p) and SMDꢀGIL methods using the input geometries from Ref 39 and the
resulting free energy diagram can be seen in Figure 9 (298 K, 1 atm). Here, the proton transfer
ꢀ
ꢀ
from the HNO to DCA is exoergic by 11.4 kJ/mol. The subsequent NO nucleophilic attack on
3
3
the HDCA is endoergic by 59.0 kJ/mol (+47.6 kJ/mol relative to reactants) with a free energy
‡
barrier of ꢁG ≤ 89.7 kJ/mol (+78.3 kJ/mol relative to reactants). This free energy barrier is
assigned as an upper limit due to the tendency of the SMDꢀGIL/M06 method to return to
geometry 2A (Ref 39) rather than converge to a transition state (i.e., the free energy was
determined by fixing the geometry and performing a frequency calculation (freq, Gaussianꢀ09),
4
0
to yield the upper limit to the free energy of the transition state and a single imaginary frequency
corresponding to the nucleophilic attack on the carbon, which was confirmed by internal reaction
coordinate analysis). The next step in the proposed mechanism, the migration of NO from the
2
CꢀO oxygen to the terminal nitrogen, has a calculated free energy barrier of 76.5 kJ/mol and the
product formed is highly exoergic with a reaction free energy of ꢀ152.9 kJ/mol for this step.
Although the mechanism depicted in Figure 9 proceeds through a series of anionic
species, the reactions proposed in Equations (1) and (2) in the Supporting Information imply
reactions between neutral species. SMDꢀGIL free energy calculations for the neutral reactions
were also explored and reported here (Figure 10). While the relative free energy of the
15
ACS Paragon Plus Environment

The Journal of Physical Chemistry
Page 16 of 41
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
stationary state of the nitrated, protonatedꢀdicyanamide species (40.1 kJ/mol) and the free energy
‡
barrier for the nitro migration (ꢁG = 59.5 kJ/mol, +99.6 kJ/mol relative to reactants) are lower
in value than that for the corresponding species in the anionic mechanism (Figure 9), the overall
free energy release for this process is also substantially less (ꢀ39.2 vs ꢀ105.3 kJ/mol) relative to
the reactants. Figure 11 shows the free energy profile resulting due to proton transfer to the
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
terminal nitrogen on the HDCA. A search for the transition state for the nucleophilic attack of
ꢀ
NO on the diprotonated dicyanamide (similar to the first transition state in Figure 9) was
3
1
1
unsuccessful and is not included in Figures 11 and 12. It should be noted that previous work
indicated that species with ionic character are solvated more readily in the highly polar ionic
liquid medium, whereas neutrals or species with little ionic character are solvated less, resulting
in potentially significant differences in the SMDꢀGIL mechanism versus the corresponding gas
phase mechanism, for example the reverse in free energies of acidity for HDCA and HNO as
3
mentioned earlier.
The recent work by Nichols and coworkers also investigated the contribution of a second
HNO molecule (acting as a spectator) on the reaction in order to determine the possible
3
3
9
solvation effect on the (gasꢀphase) energy profile. The addition of a second HNO molecule
3
has a strong stabilizing effect, which lowers the enthalpies of all intermediates and transition
states below the energy of the reactants. These calculations were repeated here using the input
geometries from the Supplemental Information from Nichols’ work, shown in Figure 12. In
contrast to the gas phase calculations, using the condensed phase SMDꢀGIL model, the addition
ꢀ
of a second, neutral HNO molecule has a destabilizing effect on the DCA ꢀHNO adduct relative
3
3
to the anionic route in Figure 9 of ꢀ2.4 – (ꢀ18.7) = +16.3 kJ/mol. This could perhaps be
explained if the additional, neutral HNO perturbs the solvation shell of the polar ionic liquid
3
16
ACS Paragon Plus Environment

Page 17 of 41
The Journal of Physical Chemistry
1
2
3
4
5
6
7
8
9
ꢀ
ꢀ
solvent surrounding the anionic DCA ꢀHNO , thereby destabilizing the DCA [HNO ] relative to
3
3 2
ꢀ
DCA ꢀHNO .
3
The differences in the Nichols results (enthalpies) and our results (free energies) indicate
that there is a significant entropic contribution to the reaction profile. Therefore, we report the
Gaussian calculated (1 atm, 298 K) enthalpy (ꢁHº), free energy (ꢁGº) and entropy (TꢁSº)
contributions for both the gas phase (M06) and for the condensed phase (SMDꢀGIL) in order to
better understand the reaction energetics involved (see Supporting Information, Table S3). The
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ꢁ
Hº values determined by M06 for the addition of both one and two HNO molecules have
3
similar trends to those in Ref 39 and the enthalpies match reasonably well. However, due to the
entropy loss in combining two or three molecules into one adduct, the ꢀTꢁSº entropic
contributions can shift the energetics of the products higher relative to the reactants. For the
ꢀ
addition of one HNO molecule to DCA , the average M06 gas phase shift due to ꢀTꢁSº is +11.3
3
ꢀ
kcal/mol (47.2 kJ/mol), and for the addition of two HNO molecules to DCA , the average M06
3
gas phase shift due to ꢀTꢁSº is +22.5 kcal/mol (94.1 kJ/mol), which largely tends to cancel out
the potential solvation effect of addition of a second HNO molecule (see Table S3 in the
3
Supporting Information). In fact, in the SMDꢀGIL condensed phase, there is almost no
difference in the free energy profiles for one or two HNO molecules as seen in Table S3 of the
3
Supporting Information.
ꢀ
The overall energetics in Figure 9 indicate that the proton transfer from HNO to DCA is
3
ꢀ
exoergic, the subsequent nucleophilic addition of NO is endoergic with a barrier, there is a
3
barrier for the NO migration and the NO migration reaction is highly exoergic. This sequence
2
2
of reactions is consistent with Figure 10 and with the experimental kinetics results in this paper,
and it is possible that the NO migration is the rateꢀlimiting step in the hypergolic ignition
2
17
ACS Paragon Plus Environment

The Journal of Physical Chemistry
Page 18 of 41
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
mechanism. The energy release of NO migration could lead to the temperature increase
2
+
ꢀ
observed in the hypergolic ignition of BMIM DCA + HNO . The entropic contributions to the
3
reaction energetics is significant, and these should be included when describing the reaction
profile as the free energy of reaction.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Rapid-scan Fourier Transform Infrared Spectroscopy
The wellꢀknown acidꢀcatalyzed catalyzed conversion of nitriles to amides proceeds via
protonation on the nitrile nitrogen followed by nucleophilic addition of water to the nitrile carbon
and finally ends with proton migration from water to the terminal nitrogen:
(
1)
In the proposed hypergolic ignition reaction mechanism, an analogous reaction to the acidꢀ
ꢀ
catalyzed conversion of nitrile to amide via nucleophilic addition of NO followed by NO
3
2
migration to the terminal N in the dicyanamide occurs (see Figure 9).
(2)
1
8
ACS Paragon Plus Environment

Page 19 of 41
The Journal of Physical Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(3)
(4)
ꢀ
The presence of water in the system would compete as a nucleophile with NO , leading to the
3
formation of the isocyanoureate (Equation 1). In order to investigate the contribution of the
isocyanoureate species in the hypergolic ignition mechanism, separate rapidꢀscan FTIR drop
+
ꢀ
tests with HNO and BMIM DCA and sodium isocyanoureate were performed, and the results
3
are seen in Figure 13. While the dicyanamide leads to the formation of CO , HNCO and N O as
2
2
seen in the previous ignition testing, reaction of HNO with the isocyanoureate only leads to the
3
formation of CO and does not yield the HNCO and N O species needed to lead to ignition. This
2
2
would explain the high sensitivity of ignition delay time to the presence of water in the system,
either in the HNO or in the ionic liquid fuel.
3
One suspected reactive intermediate in the hypergolic ignition mechanism of
+
ꢀ
+
BMIM DCA + HNO is NO . The investigation into the sensitivity of ignition delay time to
3
2
+
NO concentration utilized the commonly used nitration chemistry:
2
+
ꢀ
+
HNO + 2 H SO ↔ NO + 2HSO + H O
(5)
3
2
4
2
4
3
19
ACS Paragon Plus Environment

The Journal of Physical Chemistry
Page 20 of 41
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
It is wellꢀknown and widely utilized in nitration reactions that the addition of H SO to HNO
3
2
4
+
can shift the equilibrium in favor of NO between about 80ꢀ90% H SO in HNO . The ratio of
2
2
4
3
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
H SO to HNO was varied in steps of 10 wt%, and the results were measured as a function of
2
4
3
the normalized ratio of N O to CO (peak areas), versus increasing H SO percent, seen in Figure
2
2
2
4
1
4a below. As the concentration of nitrate decreases, the N O/CO ratio also decreases. Ignition
2 2
is observed between 0ꢀ60% H SO /HNO , and although no ignition is observed between 70ꢀ88%
2
4
3
H SO /HNO , reaction is observed in the form of smoke evolution above the cuvette. Above
2
4
3
+
88%, no reaction is observed. Superimposed on the graph in Figure 14b is the normalized NO2
+
ꢀ
+
concentration versus H SO /HNO ratio, and the shift towards NO + 2HSO + H O
2
4
3
2
4
3
products is essentially coincident with the loss of hypergolic reactivity. This result would imply
+
that NO is not an active species in the hypergolic ignition mechanism for dicyanamideꢀbased
2
RTIL + HNO3.
The use of the nitration reaction was further utilized to see if the isocyanoureate could be
converted back to a reactive species, and a series of drop tests with high H SO /HNO ratios (in
2
4
3
+
the range where NO would be favored) and sodium isocyanoureate were performed, and the
2
+
results are seen in Figure 15. At 85 wt% H SO /HNO , approximately the middle of the NO
2
2
4
3
curve seen in Figure 14b, a vigorous reaction is observed and the reꢀemergence of N O appears
2
in the IR spectrum (Figure 15b). This implies that the conversion back to nitroisocyanoureate
can promote reactivity, which would be consistent with the originally proposed ignition
mechanism:
2
0
ACS Paragon Plus Environment

Page 21 of 41
The Journal of Physical Chemistry
1
2
3
4
5
6
7
8
9
(6)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Lastly, the sensitivity of the ignition to the presence of N O should be noted, as N O could
2
2
accelerate oxidation and combustion in the ignition process.
CONCLUSIONS
In this paper, we demonstrate the ability of aerosol generation coupled with tunable VUV
photoionization mass spectrometry to identify many of the products of the reaction of a
hypergolic dicyanamideꢀbased ionic liquid with HNO , such as CO , N O and HNCO, and is
3
2
2
consistent with the previously proposed reaction mechanism. To our knowledge, this is the first
study on the chemical reactivity of ionic liquid aerosols. Measurement of the product profiles as
a function of HNO exposure allows for insight into the kinetics of the initial reaction steps in the
3
+
ꢀ
hypergolic reaction of BMIM DCA + HNO . Kinetics analysis indicates the initial reaction of
3
+
ꢀ
BMIM DCA with HNO is much faster than the subsequent reaction of HDCA with HNO , and
3
3
that the reactivity is limited to near the surface of the aerosol particle. Nonꢀreactive molecular
dynamics simulations support the idea that the HNO diffusion is slow in the particle and that
3
reactivity would be limited to regions near the surface of the aerosol particle. Additional FTIR
+
experimental results indicate that NO appears not to be an important species in the hypergolic
2
ignition mechanism.
21
ACS Paragon Plus Environment

The Journal of Physical Chemistry
Page 22 of 41
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
ACKNOWLEDGEMENTS
The authors gratefully acknowledge funding from the U.S. Air Force Office of Scientific
Research for supporting C.J.K., D. P.ꢀV., and S.R.L. (Grant Nos. FA9550ꢀ10ꢀ1ꢀ0163 and
FA9550ꢀ14ꢀ0154), and for S.D.C. and G.L.V. (Grant No. FA9300ꢀ06ꢀCꢀ0023). This work at the
ALS was supported by the Director, Office of Energy Research, Office of Basic Energy
Sciences, Chemical Sciences Division of the U.S. Department of Energy under Contract No. DEꢀ
AC02ꢀ05CH11231 (D. P.ꢀV. and S.R.L.). The authors thank Dr. KevinWilson of the Chemical
Dynamics beamline for help in the measurements.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
22
ACS Paragon Plus Environment

Page 23 of 41
The Journal of Physical Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
+
ꢀ
Figure 1. Schematic of the BMIM DCA aerosol + HNO reaction kinetics flowꢀtube setup. The
3
denuder is filled with Na CO that removes and neutralizes HNO , effectively ending the
2
3
3
reaction.
23
ACS Paragon Plus Environment

The Journal of Physical Chemistry
Page 24 of 41
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
+
ꢀ
Figure 2. Size distribution of aerosols from 0.2 g/L, 0.5 g/L and 4.3 g/L BMIM DCA in water
from the aerosol source (10 psi N backing pressure). Average diameter of the aerosol particles
2
increases with increasing concentration.
2
4
ACS Paragon Plus Environment

Page 25 of 41
The Journal of Physical Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
+
ꢀ
Figure 3. VUV photoionization TOFMS spectra of BMIM DCA + HNO at a) full spectrum,
3
1
1.0 eV and b) m/z 40ꢀ48, 13.5 eV (CO PIE maximum).
2
25
ACS Paragon Plus Environment

The Journal of Physical Chemistry
Page 26 of 41
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 4. Photoionization efficiency (PIE) curves for m/z 42, 43, 44, 67, 85 and 126, with IE of
10.3±0.1, 11.4±0.3, 10.5±0.3, 10.2±0.2, 10.2±0.2 and 8.7±0.2 eV, respectively.
26
ACS Paragon Plus Environment

Page 27 of 41
The Journal of Physical Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5. Mass spectra at photoionization energy of 14.4 eV between m/z 40ꢀ50 for a)
+
ꢀ
+
ꢀ
15
BMIM DCA + HNO and b) BMIM DCA + H NO .
3
3
27
ACS Paragon Plus Environment

The Journal of Physical Chemistry
Page 28 of 41
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 6. High resolution PIE curves for N O, CO and a 1:1 mixture of CO to N O.
2
2
2
2
28
ACS Paragon Plus Environment

Page 29 of 41
The Journal of Physical Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
+
ꢀ
Figure 7. Snapshots from simulations of BMIM DCA liquid/HNO gas interface: initial (top)
and after 6 ns of simulation (bottom).
3
29
ACS Paragon Plus Environment