Speciation of the products of and establishing the role of water in the reaction of TNT with hydroxide and amines: Structure, kinetics, and computational results

Article abstract of DOI:10.1021/jp408992n

The reaction of trinitrotoluene (TNT) with bases has been investigated by NMR and visible spectroscopy methods. Hydroxide ion was found to react in one of two ways, either by deprotonation of the methyl group or by nucleophilic attack on the aromatic ring to form a σ adduct. The rate of each mode of reaction depends upon the polarity of the solvent. In tetrahydrofuran (THF), σ adduct formation is rapid and the long-term equilibrium product is deprotonation of the methyl group. When the solvent is methanol (MeOH), the two reactions have similar rates and the σ adduct becomes the majority product. Amines were found to be ineffective in directly deprotonating TNT or in forming σ adducts. Rather, the amines react with ambient water to generate hydroxide ion, which then reacts with TNT. The solvent choice and water content are crucial to understanding the reactivity of bases with TNT. To assist in the interpretation of the experimental results, computational analysis was performed at the B3LYP/6-311+G**//HF/6-311+G** level to determine the thermodynamics of the reactions of TNT. The SM8 implicit solvation model was applied to converged geometries and suggested a strong solvation effect upon product formation. Thermodynamic analysis suggested a significant preference of alkoxide or hydroxide attack versus amine attack in any modeled dielectric, consistent with the experimental observations.

Full text of DOI:10.1021/jp408992n

Article
pubs.acs.org/JPCA
Speciation of the Products of and Establishing the Role of Water in
the Reaction of TNT with Hydroxide and Amines: Structure, Kinetics,
and Computational Results
Christopher A. Latendresse, Syrena C. Fernandes, Sangmin You, and William B. Euler*
Department of Chemistry, University of Rhode Island, 51 Lower College Road, Kingston, Rhode Island 02881, United States
S
* Supporting Information
ABSTRACT: The reaction of trinitrotoluene (TNT) with bases has been investigated by NMR and visible spectroscopy
methods. Hydroxide ion was found to react in one of two ways, either by deprotonation of the methyl group or by nucleophilic
attack on the aromatic ring to form a σ adduct. The rate of each mode of reaction depends upon the polarity of the solvent. In
tetrahydrofuran (THF), σ adduct formation is rapid and the long-term equilibrium product is deprotonation of the methyl group.
When the solvent is methanol (MeOH), the two reactions have similar rates and the σ adduct becomes the majority product.
Amines were found to be ineffective in directly deprotonating TNT or in forming σ adducts. Rather, the amines react with
ambient water to generate hydroxide ion, which then reacts with TNT. The solvent choice and water content are crucial to
understanding the reactivity of bases with TNT. To assist in the interpretation of the experimental results, computational analysis
was performed at the B3LYP/6-311+G**//HF/6-311+G** level to determine the thermodynamics of the reactions of TNT.
The SM8 implicit solvation model was applied to converged geometries and suggested a strong solvation effect upon product
formation. Thermodynamic analysis suggested a significant preference of alkoxide or hydroxide attack versus amine attack in any
modeled dielectric, consistent with the experimental observations.
TNT, forming TNT−, or to attack the ring, forming the σ
INTRODUCTION
■
adduct TNT−OR−.13−19 While alkoxides are believed to attack
The detection of trinitrotoluene (TNT) has become an
increasingly prevalent research topic over the past decade, in
part because of the rise in international terrorism and the toxic
the C3 carbon of TNT, hydroxide is much less sterically
hindered and is able to form the TNT−OH− C1 adduct,
pushing the methyl group on TNT out of plane.3 Since C1 was
nature of TNT in the environment. Many methods of detection
established as the most favorable site for nucleophilic attack by
have been proposed in the literature, including methods that
hydroxide ion, this is commonly attributed as the site of attack
rely on the reactivity of TNT arising from the electron-deficient
aromatic ring. Since these reactions generate species that
by weaker nucleophiles, such as amines.3,4,9−12,20−29 In the
context of a rapid TNT sensor, the possible initial elementary
absorb in the visible range, simple colorimetric sensors have
been developed in the solution phase.1−4 More complex and
steps are the most crucial ones to understand. Because of this,
subsequent reactions such as the “Janovsky complex”14 (TNT−
attacking TNT) will not be discussed in this study.
sensitive fluorometric sensors have been developed,5−8 with
many using the colorful species in a resonance energy transfer
(RET) sensing mechanism.9−12 In such a sensor, a donor
fluorophore is chosen on the basis of significant overlap of its
emission spectrum with the absorbance spectrum of the TNT
product acceptor. In this fashion, a sensor can be tailored for a
species that absorbs in a specific range.
Alkoxide reactions with TNT have been studied for a
number of years, with two major competing reactions resulting.
Scheme 1 shows the possible reactions between TNT and a
generic base. Reports in the literature show that the ability of
hydroxide or alkoxide to deprotonate the methyl group of
In view of the multitude of TNT sensor proposals in the
literature that depend on the interaction or reaction of TNT
with amines, it is perhaps surprising that a thorough
mechanistic study has not been undertaken previously. These
types of sensors have been developed for both solution-phase
and vapor-phase detection of TNT. Amine reactivity has been
Received: September 8, 2013
Revised: October 14, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/jp408992n | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A
Article
explained by a CT complex,28 TNT−,21 or a σ adduct with an
amine.27
a
Scheme 1. Proposed Reactions of TNT with Bases
A recent paper by Olson et al.31 reported the reactions of
amine bases with 2,4-dinitrotoluene. With dimethyl sulfoxide
(DMSO) as the solvent, it was shown that the extent of
reaction with amine bases was small. Further, at equilibrium the
small amount of product formed was best attributed to
deprotonation of the methyl group, not σ adduct formation.
Since TNT has an additional nitro group, which makes the
aromatic ring more electron-deficient, we decided to investigate
whether TNT has similar reactivity to 2,4-dinitrotoluene. We
1
used H NMR spectroscopy to determine the products of the
reaction of bases with TNT in tetrahydrofuran (THF) and
methanol (MeOH). In nearly all cases, the amines did not react
with TNT unless there was sufficient water in the system to
generate hydroxide ion. Then we examined the kinetics of the
reactions in these solvents using visible spectroscopy, which is
more sensitive than NMR methods. We found that the rate
constants depend upon the solvent system, which leads to
different product distributions in different solvents.
To assist in interpreting the experimental results, computa-
tional methods were also employed here. Previous studies have
restricted their structural optimization computations to
molecular mechanics force fields,32 semiempirical methods,33
and DFT methods.34,35 No computational studies comparing
alkoxide, hydroxide, and amine reactivity through TNT−
formation, σ adduct formation, and CT complex formation
using the same high level ab initio method have been
performed, nor has solvation modeling been addressed
previously. By computationally modeling the reactants and
products of these proposed reactions, we were able to calculate
the relative thermodynamics of each reaction and observe a
significant trend in solvation.
a
TNT can be deprotonated at its methyl group to form TNT− as
either free ions or an ion pair, can be attacked by a nucleophile at the
C1 or C3 carbon to form a TNT−B σ adduct, or can have electron
density donated to its electron-deficient ring to form a CT complex,
TNT:B.
MATERIALS AND METHODS
■
TNT was obtained from Drs. Jimmie Oxley and James Smith
and was used without further purification. The solvents used in
this study, HPLC-grade THF (99.9%, inhibitor free) and
MeOH (99.9%), were purchased from Sigma-Aldrich and used
as received. ACS-reagent-grade 0.1 M aqueous NaOH was
purchased from Anachemia. Ethylamine (EA) (70% in water),
EA in THF (2.0 M), n-propylamine (PA), n-butylamine (BA),
diethylamine (DEA), and triethylamine (TEA) were purchased
from Sigma-Aldrich.
proposed to work in a fashion similar to that for anionic bases,
where deprotonation to give TNT−,20−23 the formation of a
zwitterionic σ complex (TNT−NR3),3,9−12,24−27 and charge
transfer (CT) complexes4,28,29 are all possible, as shown in
Scheme 1. A typical approach found in the literature is to
examine the absorption spectrum of TNT in solution with an
amine reactant before incorporating that reactant into a solid-
phase sensor to project the reactivity of TNT vapor. The time
domains of these studies are poorly documented. Fant et al.3
are commonly referenced for the rapid reaction of TNT with
amino acids and other amines, but their reactions were
performed over a 24 h period. The reactions are commonly
done in solvents that may have competing reactions with TNT
as well as solvents from which trace water is not easily removed.
Therefore, it is possible that in this early stage of development,
the reactivity may not be coming directly from the amines but
rather from alternative sources such as OH− generated through
acid/base equilibria. For example, we have recently shown that
residual water in dimethylformamide (DMF) can react with
TNT to form products observable by both absorption and
emission spectra.30 Additionally, the observed absorption
spectra are commonly misattributed to the incorrect major
absorbing species. There is definite confusion as to the actual
product(s) being formed in solution, as numerous accounts of
the same or similar reactions lead to different proposed
products; for example, the same absorption spectrum has been
NMR experiments were conducted in either THF-d8 (99.5%
D) or in CD3OD (99.8% D), both purchased from Cambridge
Isotope Laboratories. D2O (99.8% D), DCl/D2O (35% w/w,
99% D) solution, and NaOD/D2O (40% w/w, 99.5% D)
solution were purchased from Sigma-Aldrich. TNT solutions
1
were quantitatively prepared, and H and 13C spectra were
acquired before and after the addition of base on a Bruker 300
MHz NMR spectrometer. The time delay from the start of the
reaction to data acquisition was approximately 5 min per
spectrum because of solvent locking and shimming.
Solutions containing TNT for absorbance measurements
were made quantitatively using either THF or MeOH. TNT
stock solution was added to a cuvette and observed prior to
addition of base. While the solution was stirred, a measured
amount of water and stock base solution was quantitatively
added using a micropipet. Spectra were acquired with an Ocean
Optics spectrometer using a white-light tungsten lamp source
with the sample placed in a Peltier temperature-controlled
cuvette holder held at 20.00 °C. The spectral resolution was 1
B
dx.doi.org/10.1021/jp408992n | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A
Article
nm, and the temporal resolution was 50 ms. Experiments were
prepared so that one reactant concentration would be in 100×
molar excess of the other. Data were fit to exponential trends
using graphing software (Microsoft Excel, SigmaPlot).
change. This led us to hypothesize that the solvent played a key
role in the reactivity of TNT: the hydroxide (from adventitious
water in the solvents) generated by reaction with the amine was
the active species in the reaction with TNT, and the amine did
not react directly.
Computational modeling was performed using Spartan ’10
software.36 Reactants and products of TNT reactions with
OH−, CH3O−, NH3, EA, DEA, and TEA were modeled; the
generic set of reactions of TNT with base (B) can be found in
Scheme 1. The amines were chosen to cover a range of basicity
and steric bulk. EA is a good nucleophile and is an appropriate
surrogate for the primary amines that are commonly used in the
aforementioned sensors. DEA is more basic than EA but has
greater steric hindrance. TEA is the best electron donor of the
modeled amines but is so sterically bulky that it cannot attack
the TNT ring. The deprotonation reaction was modeled using
TNT− with a subsequent conjugate acid as free molecules/ions
or as an ion pair. σ complexes formed through nucleophilic
attack at either the C1 or the C3 carbon of TNT were modeled
(TNT−OH− C1, TNT−OH− C3, TNT−NH3 C1, etc.; see the
numbering convention for TNT in Scheme 1). All species were
geometrically optimized using the Hartree−Fock (HF) method
with the 6-311+G** basis set and no imaginary frequencies.
This gave a structure for TNT that matched the crystal
structure37 and reasonable geometries for the other species.
Since the calculated structures were reasonable, a single-point
energy calculation with the B3LYP functional and the 6-31G*
basis set was used to determine the enthalpy and entropy for
each species using the frequency calculation. All of the
thermodynamic computations were performed using an input
temperature of 298.15 K and a pressure of 1 atm. In addition,
the SM8 solvation model38 was applied to determine relative
energies in a wide range of dielectrics in order to assist in our
description of the observed solvation dependence. In these
solvation computations, the converged structures from the HF
outputs were used to perform single-point energy calculations
at the B3LYP/6-31G* level; this approach was chosen since it
was used as a fitting method for the development of the SM8
model.38 To cover a wide range of dielectric constants (ϵ), the
following solvent models were used: vacuum (ϵ = 1), hexane (ϵ
= 2.02), tetrahydrofuran (ϵ = 7.52), 1-propanol (ϵ = 20.1),
methanol (ϵ = 33), formic acid (ϵ = 58), water (ϵ = 80), and
formamide (ϵ = 109). 1H NMR chemical shift estimations were
also computed for each species at the same level of theory.
Molecular orbital energies were obtained using the HF/6-
311+G**-derived structure, the B3LYP functional, and the 6-
311+G** basis set. Time-dependent density functional theory
(TDDFT) calculations at this level of theory were used to
determine the maxima and intensities of the six lowest-energy
electronic transitions. Spectra were calculated using Gaussian
line shapes with the full width at half-maximum arbitrarily set at
20 nm.
In general, a base could react with TNT by deprotonation of
the methyl group, addition to the aromatic ring to form a σ
adduct, or formation of a CT complex. In the case of TEA, a σ
adduct with TNT is unlikely because of the steric bulk of the
three ethyl groups. For the other amines, the nucleophilicity of
the amine should determine the extent of the reactivity. Direct
deprotonation of the methyl group on TNT should depend
upon the basicity. Finally, CT complex formation should
parallel the donor number (DN). For example, TEA has one of
the highest DN values,39 so it might be expected to donate
electron density to the electron-poor TNT aromatic ring and
form a CT complex. Trinitrobenzene (TNB) has been shown
to form CT complexes with aliphatic amines (including TEA)
in cyclohexane, with absorbance maxima in the UV range.40
Finally, each of the amine bases can react with water or
methanol to form hydroxide or methoxide, respectively, further
complicating the reactivity in these systems.
A variety of possible structures for the proposed reaction
products were investigated computationally. Converged
structures for the acid/base and σ complex formation reactions
are shown in Figure 1 for hydroxide and methoxide and Figure
Figure 1. Converged equilibrium geometries of TNT and products
from acid/base or nucleophilic addition reactions with hydroxide or
methoxide. Hydroxide or methoxide are proposed to be capable of
deprotonating the methyl group on TNT to form TNT− or adding to
either ring carbon C1 or C3 to form the σ adduct TNT−OR−.
RESULTS AND DISCUSSION
2 for amines. The energies of all of the converged structures can
be found in Table S1 in the Supporting Information. For the
ion pairs, the positive hydrogen atoms on the alkylammonium
cation interact directly with the oxygen atoms from the nitro
group, where the greatest electron density from the TNT anion
is located. This distorts the methylene group much further out
of the plane of the ring than calculated for the isolated TNT
anion. We were able to converge σ adducts for all of the amines
with the exception of TEA. These optimizations pushed TEA
away from the ring and stabilized in an electrostatic interaction
or CT complex type of orientation.
■
Qualitative experiments were run by adding a few crystals of
TNT to solutions of the various bases in THF or MeOH. In the
cases of NaOH or aliphatic amines (EA, PA, BA, DEA, or
TEA), there was an immediate color change indicative of
chemical reaction. In contrast, the aromatic amines 2,6-lutidine,
collidine, and pyridine and the weak primary amine aniline
showed no color change. Further, dropping a few crystals of
TNT into a few milliliters of dry TEA in the absence of
cosolvent generated no visible absorbance features. Addition of
a drop of water to the solution caused an immediate color
C
dx.doi.org/10.1021/jp408992n | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A
Article
have more localized charge and thus larger distortions. This is
due to the lack of a nucleophile that pushes C7 out of the plane
of the ring, so the original ring planarity is largely retained. The
inclusion of a counterion in the alkylammonium ion pairs leads
to minimal changes in the bond lengths throughout the ring
structure. The ion pairing does force a slightly increased ring
bond angle distortion compared with isolated TNT− and causes
the ring planarity to be decreased, as observed through the
increased C1−C2−C3−C4 dihedral angles. This deviation from
planarity is reduced as the size of the counterion increases the
planarity of the ring while further distorting the C7−C1−C2−C3
dihedral angle. The largest of these distortions comes from the
pairing of TNT− with the triethylammonium cation, which is
attributed to the larger steric bulk. However, the distortions of
these parameters for all of the investigated ion pairs are smaller
than the extent of the distortions that arise from the respective
nucleophilic addition.
To identify which of the predicted structures could be
obtained experimentally, NMR spectroscopy was used to
identify the products of reactions of the bases with TNT.
The solvents used, MeOH and THF, afforded polar and slightly
polar surroundings, respectively. In addition to the aliphatic
amine bases, which vary in basicity, nucleophilicity, and steric
bulk, NaOH was also used as a stronger base than the amines.
Reactions of TNT and sodium deuteroxide in deuterium
oxide solution (NaOD/D2O) were carried out in deuterated
1
THF and deuterated methanol. Figure 3 shows the H NMR
spectra as a function of base-to-TNT ratio in THF-d8, while
Figure 4 shows the results of the same experiment in CD3OD.
Figure 3 shows that the intensities of the aromatic and
methyl proton peaks of TNT decrease as a function of NaOD
Figure 2. Converged equilibrium geometries of products proposed for
the reactions of TNT with amines. All of the bases gave a converged σ
adduct structure except for TEA, which is expected to be too sterically
hindered to attack the ring covalently.
Selected structural parameters from each species’ optimized
geometry are given in Table S2 in the Supporting Information.
While deprotonation of the TNT methyl group depends on the
base strength, formation of the σ adduct depends on the
nucleophilicity. The nucleophiles show a general trend of
increasing bond length as their nucleophilic strength decreases
and their steric bulk increases. The largest bond angle
distortions from that of TNT are seen for the TNT−OH−
and TNT−OCH3− adducts, stretching the ring in-plane. These
adducts also distort the dihedral angles to the greatest degree of
all of the nucleophiles, bending the ring out-of-plane. The
amine adducts show a definite trend of increasing steric bulk
resulting in increased distortion. The C1 adducts appear to
distort the ring less than their C3 counterparts. In addition, the
C1 adducts push the electron-donating methyl group out of the
plane of the ring. Although the C3 adducts do not displace an
electron-donating group, their nucleophiles are significantly
more orthogonal to the ring. This may be another stabilizing
advantage of the C1 adduct over the C3 adduct.
Figure 3. 1H NMR spectra for the reaction of 1.3 × 10−2 M TNT with
increasing concentrations of NaOD/D2O in THF-d8: (a) NaOD:TNT
= 0; (b) NaOD:TNT = 0.14; (c) NaOD:TNT = 0.29; (d)
NaOD:TNT = 0.58; (e) NaOD:TNT = 0.94; (f) NaOD:TNT =
1.30; (g) NaOD:TNT = 1.37; (h) after addition of one drop of 30%
DCl/D2O to the solution in (g).
TNT− formation distorts the bond angles and dihedral
angles the least, which is expected because of the enhanced
resonance stabilization of the ring. The most obvious change is
the contraction of the C1−C7 bond length, indicative of vinyl
character. Resonance in TNT− is also supported by the smaller
distortions of the O−N−O bond angles throughout the
structure in comparison with the other anionic products that
addition in THF-d8, as expected. TNT has two equivalent
aromatic protons represented by the singlet at 8.98 ppm as well
as three equivalent methyl protons appearing as a singlet at 2.62
ppm. Both peaks disappear with addition of NaOD as a product
grows in. The growing singlet peaks (8.31 and 5.63 ppm) are
observed to have a 1:1 integration ratio; this is consistent with
D
dx.doi.org/10.1021/jp408992n | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A
Article
the formation of TNT− and inconsistent with the formation of
to a 3:2 ratio, which is consistent with the C1 deuteroxide σ
adduct formation, where the lower-field feature represents the
two ring protons (on C3 and C5) and the higher-field feature
represents the methyl group protons. Deuterium exchange can
be observed with the methylene group on TNT−, as the peak at
5.68 ppm diminishes and subsequently disappears while the
ring protons on TNT− remain visible, ultimately leading to
TNT−OH− or a CT complex. The formation of an OD−
σ
adduct at the C1 position should give two peaks integrating as
3:2 while a σ adduct formed at the C3 position should lead to
three peaks integrating as 3:1:1. A CT complex would also give
two peaks integrating as 3:2 with chemical shifts close to the
spectral positions of unreacted TNT. The two ring protons at
C3 and C5 of TNT− give rise to the singlet at 8.31 ppm, while
the two equivalent vinyl-type protons on C1 appear as a singlet
at 5.63 ppm. With no other features apparent in this spectrum,
TNT− appears to be the exclusive product on the time scale of
1 h. The relative peak positions are consistent with the report
of TNT− formation in DMSO/MeOH using alkoxide bases by
Fyfe et al.13 The small shifts observed in the methyl peaks of
unreacted TNT are caused by H/D exchange.41 To confirm the
assignment of TNT− as the product, excess DCl was added to
the system. This deuterated the methylene group on TNT− to
form a CH2D group (seen as a multiplet near 2.57 ppm),
restored the aromatic peak at 8.98 ppm, and led to complete
loss of the product peaks at 5.63 and 8.31 ppm. The 13C NMR
spectrum of unreacted TNT in THF-d8 (Figure S1 in the
Supporting Information) displays the expected aromatic peaks
at 123.1, 134.3, 146.9, and 152.4 ppm and the methyl peak at
15.4 ppm. The DEPT-135 spectrum (1 h acquisition time)
recorded after the deprotonation with OD− and subsequent
deuteration with DCl shows the aromatic carbons containing H
atoms, C3 and C5, remaining at the same chemical shift (122.9
ppm) while the methyl carbon peak is no longer visible (Figure
S2 in the Supporting Information). This confirms the extensive
H/D exchange in the system.
deuteration to give CHD− and CD2 groups. Finally, at higher
−
NaOD:TNT ratios, another feature becomes apparent at 8.82
ppm that is unassigned. This resonance may arise from the
aromatic protons on a deuterated TNT molecule or may be
from a further decomposition product. Addition of DCl to the
last point in the titration did not completely restore TNT,
which is different than the behavior observed in THF. TNT
was exposed to DCl as a control for this experiment, and no
change to the original TNT features was observed.
Our results differ from Fyfe’s observation of a 1:1 ratio for
the C3 adduct of methoxide,13 where two singlets were detected
at 8.45 and 6.18 ppm in 87.5% DMSO/12.5% MeOH. It is also
important to note that the products make up only a small
percentage of the total integration at equimolar TNT and
NaOD concentration; this suggests that the equilibria lie much
further toward reactants in MeOH than in THF.
The 13C NMR spectrum of unreacted TNT in CD3OD
(Figure S3 in the Supporting Information) is similar to that
observed in THF-d8. After the addition of DCl to the NaOD/
TNT system, a 13C DEPT 135 experiment was undertaken.
The data (Figure S4 in the Supporting Information) also
support the deuteration of the TNT− methylene to create a
mixture of CH2D, CHD2, and CD3 groups. The CHD2 group
appears as a positive feature at 15.5 ppm, while the CH2D
group is a negative feature centered at 15.1 ppm and split into
three peaks (14.8 and 15.3 ppm) by the D atom.
In contrast, TNT− is not the only product in CD3OD
(Figure 4). At a low NaOD:TNT ratio (0.24), only TNT− is
observed, as evidenced by the growth of singlets at 8.30 and
1
Figure 5 shows the H NMR spectra of TNT mixed with
−
5.68 ppm for the C3/C5 protons and the CH2 protons,
various alkylamines in THF-d8. The different amines show four
modes of reactivity: TEA shows no reaction; DEA and BA
respectively. As the amount of NaOD increases, a second
product grows in with peaks at 5.30 and 8.41 ppm integrating
1
Figure 5. H NMR spectra for the reaction of TNT upon addition of
bases in THF-d8: (a) unreacted TNT (1.5 × 10−2 M); (b) TNT (1.4 ×
10−2 M) + NaOD (8.1 × 10−3 M); (c) TNT (2.7 × 10−1 M) + DEA
(1.3 M); (d) TNT (2.7 × 10−1 M) + TEA (0.91 M); (e) TNT (2.7 ×
10−1 M) + EA/H2O (0.92 M); (f) TNT (2.7 × 10−1 M) + PA (1.4
M); (g) TNT (2.7 × 10−1 M) + BA (0.73). Peaks for the unreacted
amines are found upfield from 5 ppm with significantly higher
intensities than in the spectra shown.
Figure 4. 1H NMR spectra for the reaction of 1.5 × 10−2 M TNT with
increasing concentration of NaOD/D2O in CD3OD: (a) NaOD:TNT
= 0; (b) NaOD:TNT = 0.24; (c) NaOD:TNT = 0.49; (d)
NaOD:TNT = 0.97; (e) NaOD:TNT = 1.95; (f) NaOD:TNT =
2.44; (g) after addition of one drop of 30% DCl/D2O to the sample in
(f). The asterisk indicates an impurity in the sample.
E
dx.doi.org/10.1021/jp408992n | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A
Article
show a single product that is not TNT−; PA gives peaks
consistent with the formation of TNT−, albeit exchange-
broadened; and EA displays peaks for TNT− and a second
product. The observations of TNT− do not correlate well with
the expected basicities of the amines in water; however, they do
correlate well to the water content. EA is an aqueous solution
and the PA used was wet. In the cases of DEA, TEA, and BA,
which were dry, no TNT− was observed, but the addition of
water to the sample generated spectra similar to those seen for
EA and PA. This suggests that none of the alkylamines are
strong enough bases to directly deprotonate the methyl group
on TNT.
The products of the reactions of BA and DEA with TNT and
the second product of the reaction with EA are assigned to N-
based σ adducts. BA shows a single feature at 7.60 ppm that is a
broadened triplet, indicating that the resonance originates from
a CH2 group on the BA. There is not enough spectral
information to indicate if the BA adds to the C1 or C3 position.
The second product from EA has peaks at 7.23 and 5.91 ppm
that have no splitting, suggesting that either the product is a
hydroxide σ adduct or that the exchange is too severe to resolve
any coupling. The spectrum resulting from the reaction of DEA
is well-resolved with quartets at 7.22 and 6.55 ppm and singlets
at 5.48, 5.53, and 7.94 ppm, suggesting formation of σ adducts
rather than TNT−. The observation of two quartets, arising
from the CH2 groups on DEA, and two singlets, originating
from the methyl group on TNT, suggests that two σ adducts
form. DEA is expected to be more nucleophilic than the
primary amines but not bulky enough to prevent reaction at
either C1 or C3. However, the observation of only one peak in
the 8 ppm region implies that the DEA attacks only a single
site. The observed spectrum for the DEA is assigned as two
conformers of a C1 addition. The location of the ethyl groups
on the nitrogen atom leads to the two structures: in the lower-
energy case the ethyl groups are located adjacent to the nitro
groups, while in the second case there is rotation by ∼90° so
that one ethyl group is adjacent to the ring and the other is
adjacent to the methyl group. These structures are shown in
Scheme 2. The calculated energy difference (in vacuum, DFT/
B3LYP/HF6-311G*) is ∼35 kJ/mol.
Figure 6. 1H NMR spectra for the reaction of TNT with added bases
in CD3OD: (a) unreacted TNT (2.9 × 10−1 M); (b) TNT + NaOD;
(c) TNT (2.9 × 10−1 M) + DEA (8.7 × 10−2 M); (d) TNT (2.9 ×
10−1 M) + TEA (8.7 × 10−2 M); (e) TNT (2.9 × 10−1 M) + EA/H2O
(8.7 × 10−2 M); (f) TNT (2.9 × 10−1 M) + PA (8.7 × 10−2 M); (g)
TNT (2.9 × 10−1 M) + BA (8.7 × 10−2 M). Peaks for the unreacted
amines are found upfield from 5 ppm with significantly higher
intensities than in the spectra shown.
exchange of the methylene hydrogens that leads to the loss of
the signal at 5.6 ppm relative to the aromatic protons at 8.3
ppm. The addition of the amines to MeOD may also generate
methoxide (which is also capable of deprotonating TNT)
through an acid/base equilibrium. Since the TNT− features are
barely above the level of noise, it is also possible that other
products may be formed in amounts below the detection limits.
The unreacted TNT signals (not shown) are the predominant
features, with intensities approximately 500 times those of the
products. Clearly, the equilibria for these reactions lie much
closer to reactants than products.
NMR chemical shifts were calculated for reactions of TNT
with hydroxide ion, to compare to the experimentally available
data. Table 1 shows the comparison of the computations (in
vacuum) to the experimentally observed features in tetrahy-
drofuran (THF) and methanol (MeOH). Although the values
differ based on the lack of solvation in the calculations, the
differences are in a reasonable range to account for solvent
shifts. There are only small chemical shifts effects when
1
Figure 6 shows the vinyl region of the H NMR spectra for
the reactions of TNT with the amine bases in CD3OD. The
same major features appear here for the TNT− anion at 8.3 and
5.6 ppm, with the “wet” amines showing higher amounts. As
seen in the NaOD titration in MeOD, there is obvious H/D
1
Table 1. Comparison of Computed and Experimental H
NMR Data for TNT and Alkoxide Reactions
Scheme 2. Conformers of the DEA σ adduct to TNT
a
computed
(vacuum)
observed
species
TNT
protons
THF
MeOH
aromatic H
methyl H
aromatic H
methyl H
aromatic H
methyl H
aromatic H
methyl H
8.320
8.97
2.62
8.30
5.64
N.O.
N.O.
N.O.
N.O.
8.96
2.64
8.29
5.68
8.41
5.28
N.O.
N.O.
2.107
TNT−
7.752, 7.733
4.854, 4.829
7.597, 7.482
1.461
TNT−OH− C1
TNT−OH− C3
7.938, 5.895
1.630
a
N.O. indicates that the product was not observed. All of the predicted
and observed peaks were singlets. The computed value for each methyl
group is the average of the three computed chemical shifts.
F
dx.doi.org/10.1021/jp408992n | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A
Article
changing from the nonpolar THF to the more polar methanol
system. The calculated chemical shifts, in comparison, are
found at lower field than their experimental counterparts, as
expected, especially given the relatively small basis set used.
The magnitude of the solvation shifts (from calculated to
experimental) are all approximately the same, at about 0.5 to
0.9 ppm shifted downfield. The one significant outlier here is
the computation for the methyl protons in TNT−OH−; while
the change of the adjacent carbon from sp2 to sp3 hybridization
would be expected to lower the chemical shift, the inclusion of
the hydroxyl group in close proximity would be expected to
raise the chemical shift. With additional reorientation of the
nitro groups, it is not implausible to observe a shift higher than
5 ppm, as observed in methanol.
adducts versus the alkoxide adducts. This corresponds to the
structural analysis, where the amines do not distort the central
ring of TNT as significantly as OH− or CH3O−. Since the
HOMO surfaces are localized on the aromatic ring, larger
distortion raises the relative MO energies.
Figure 8 shows the TDDFT-calculated electronic spectra for
TNT− and the C1 and C3 σ adducts for OH− (Figure 8a),
The electronic structures of the various TNT reaction
products were calculated in order to assist in understanding the
visible spectra. Figure 7 shows the computed energies of the
Figure 8. Absorption spectra computed at the B3LYP/6-311+G**//
HF/6-311+G** level, normalized to individual species’ computed
maxima. (a) TNT−OH− C1 (black solid line), TNT−OH− C3 −(red
dashed line), TNT− (blue dot-dashed line); (b) TNT−OCH3 C1
Figure 7. Computed MO diagram for the explored TNT products
with bases at the B3LYP/6-311+G**//HF/6-311+G** level. Double
arrows indicate the HOMO for each species.
−
(black solid line), TNT−OCH3 C3 (red dashed line); (c) TNT−EA
C1 (black solid line), TNT−EA C3 (red dashed line); (d) TNT−DEA
C1 (black solid line), TNT−DEA C3 (red dashed line).
frontier molecular orbitals of TNT and the discussed products
of the alkoxide and amine reactions. TNT has its occupied
MOs at the lowest energy relative to those of the other species
and has the largest HOMO−LUMO gap, which is consistent
with the experimental observation of no visible absorbance
maxima. The anionic species TNT−, TNT−OH− C1, TNT−
CH3O− (Figure 8b), EA (Figure 8c), and DEA (Figure 8d).
The predicted spectra for the OH− and CH3O− σ adducts
suggest that bonding to C1 should give two spectral peaks
(Figure 8a, black solid line) while bonding to C3 should give
three peaks (Figure 8a, red dashed line). Thus, the electronic
spectra may be able to distinguish the two species. In addition,
TNT− also shows three peaks, with the lowest-energy transition
of all of the computed species (Figure 8a, blue dot-dashed line).
TNT− is expected to have an absorption band in the lower-
energy region of the visible range, allowing for a unique signal
at which only TNT− can be detected. In contrast, the amine
adducts show similar doublets for either C1 or C3 attack.
Because of this, no distinction between amine adduct species is
likely to be observed in the electronic spectra. The wavelengths
and intensities used to calculate the spectra in Figure 8 are
given in Table S5 in the Supporting Information.
OH− C3, TNT−OCH3 C1, and TNT−OCH3 C3 all have
very similar MO energies. TNT− is shown here to have the
smallest HOMO−LUMO gap, which is consistent with the
spectral assignments made previously in absorbance measure-
ments.13,14 Of the two hydroxide adducts, the C1 adduct has a
lower HOMO−LUMO gap and is expected to have an
absorbance maximum at longer wavelength than the C3 adduct.
The energies of the HOMOs and LUMOs are given in Table
S5 in the Supporting Information.
−
−
The amine adducts show a size-dependent trend of MO
energies, where increased steric bulk destabilizes the MOs. This
is true for both the C1 and the C3 adducts, but with little
difference in the MO energies of each C1/C3 adduct pair. There
is significant difference between the energies of the amine
G
dx.doi.org/10.1021/jp408992n | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A
Article
Having good evidence for the products formed by TNT and
hydroxide, we conducted experiments in which we observed the
visible absorption spectrum for 5 min after the reaction was
started. Briefly, we are dealing with the following competing
equilibria:
the second product is clearly the major absorbing species even
before equilibrium is established. The spectrum of the initial
product has the shape of an anionic σ adduct, TNT−OH−, with
λmax ≈ 440 nm, a second maximum at ∼510 nm, and a largely
featureless tail into longer wavelengths. The spectrum of the
second product belongs to TNT−,13,14 typified by a doublet at
about 500 and 525 nm and a broad maximum at 650 nm. This
is consistent with the TDDFT calculations, which predicted
k1
TNT(solv) + OH−(solv) XooY TNT−(solv) + H2O(solv)
k−1
1
two peaks for TNT−OH− and three peaks for TNT−. The H
(1)
NMR spectra in THF showed only the presence of TNT−,
indicating that the amount of TNT−OH− found is small, below
the detection limit of the NMR experiment. Figure 10 shows
TNT(solv) + OH−(solv) Xkoo2Y TNT−OH−(solv)
k−2
(2)
To determine the rate constants for the concurrent formation
of TNT− and TNT−OH−, we derived an expression to fit the
absorbance data as a function of time. The full derivation can be
found in the Supporting Information. To obtain pseudo-first-
order conditions, one of the reactants must be in large excess
over the other and the H2O concentration must be much larger
than the TNT− concentration. In order to meet the latter
condition, water was added to each solvent. Under these
conditions, the concentration of each product can be
monitored as a function of time using a single exponential
rise function:
obst
[product] = a(1 − e−k
)
(3)
where the pre-exponential factor a and the observed rate
constant kobs are functions of intrinsic rate constants and initial
concentrations of reactants. Substituting the absorbance for the
concentration using Beer’s law yields
Figure 10. Initial 5 min absorbance experiment observing the products
of 2.2 × 10−3 M TNT reacting with 2.2 × 10−5 M NaOH in THF: (a)
0.00 s; (b) 0.38 s; (c) 0.75 s; (d) 5.33 s; (e) 14.35 s; (f) 61.57 s; (g)
108.63 s. The inset shows fits of the absorbance data from 520 and 700
nm to eq 4.
Aprod = εprodba(1 − e−k
)
obst
(4)
where Aprod is the absorbance of the product, εprod is the molar
absorptivity of the product at the monitored wavelength, and b
is the path length (1 cm for all of the experiments reported
here). Under pseudo-first-order conditions, the time evolution
can be treated independently (see the Supporting Information),
so in cases where both TNT− and TNT−OH− absorb at the
same wavelength, eqs 4 for the two species (i.e., with different
εprod, a, and kobs values) are summed to account for the total
absorbance.
spectra from the reaction of excess TNT with NaOH in THF.
A similar, but not identical, spectral shape compared to the
excess NaOH experiment is seen. The relative intensity of the
doublet at 500 and 525 nm is reversed, indicating TNT−OH−
formation early but eventually disappearing. A control experi-
ment was conducted to determine the effect of a possible
scattering component of NaOH in THF on the kinetic fits. We
observed a reasonably significant baseline increase with the
addition of NaOH to THF, which dissipated within a few
seconds. This led us to omit some initial data when finding rate
constants when the absorbance signal was on a similar order of
magnitude as the scattering. Figures S5−S10 in the Supporting
Information show the data for different initial concentrations of
TNT.
In Figure 9, we observe that the reaction of TNT with excess
NaOH in THF forms two products, one that grows in
immediately and the other that forms more slowly. However,
When MeOH was used as the solvent, different results were
observed, as shown in Figure 11. When hydroxide ion is in
excess, TNT− is the predominant absorbing species early on, as
indicated by the broad features at 515 and 650 nm and the low
absorbance at 440 nm. After some time, TNT−OH− evolves, as
evidenced by the increase in absorbance at 440 nm. This is
consistent with the order of appearance of these species in the
1H NMR spectra. Figure 12 shows the experiment with excess
TNT in MeOH. These spectral features are consistent with
TNT−OH− formation, while there is still slight evidence to
suggest a small TNT− component given the small shoulder in
the absorption spectrum at 650 nm. Also, it should be noted
that the relative absorbances of the feature at 450 nm and the
feature at 520 nm change slightly over the course of the
observation; this implies that more than one species is being
observed. Figures S11−S16 in the Supporting Information
show the data for several concentrations of NaOH and TNT.
Figure 9. Initial 5 min absorbance experiment observing the products
of 4.2 × 10−3 M NaOH reacting with 4.2 × 10−5 M TNT in THF: (a)
5.21 s; (b) 22.96 s; (c) 95.13 s; (d) 153.70 s; (e) 214.10 s; (f) 289.78
s. Note the spectral shape change early on, consistent with the OH− σ
adduct giving way to TNT− as the major absorbing species. The inset
shows fits of the absorbance data from 520 and 700 nm to eq 4.
H
dx.doi.org/10.1021/jp408992n | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A
Article
Equation 4 was used to find rate constants for all of the data
in both THF and MeOH. In cases where both TNT− and
TNT−OH− were observed, a sum of two exponentials was
used with the assumption that reactions shown in eqs 1 and 2
could be treated independently. Only the early time data were
used, before any decays in absorption were observed. The
observed rate constants and pre-exponential factors can be
found in Table 2. The observed rate constants found at the two
different wavelengths are reproducible within the anticipated
errors of the kinetic method. As discussed above, in THF, the
major absorbing species is TNT−. In MeOH, the excess NaOH
experiments show TNT− as the major absorbing species, while
the excess TNT experiments show TNT−OH− as the major
absorbing species. To find the intrinsic rate constants, the
molar absorptivities of each absorbing species need to be found
at each wavelength.
Figure 11. Initial 5 min absorbance experiment observing the products
of 2.5 × 10−2 M NaOH reacting with 2.5 × 10−4 M TNT in MeOH:
(a) 0.14 s; (b) 0.79 s; (c) 4.09 s; (d) 9.02 s; (e) 55.69 s; (f) 172.51 s;
(g) 268.17 s. The inset shows fits of the first 15 s of absorbance data
from 520 and 700 nm to eq 4; in this case, the data at 520 nm required
two exponential rise functions for a proper fit.
The molar absorptivity of TNT− (εTNT ) was calculated in
−
THF under the assumption that all of the TNT was
deprotonated in excess OH− with the sole product being
1
TNT−, which is appropriate on the basis of the H NMR
titration data. With the extrapolated absorbance at equilibrium
and the initial TNT concentration, the resulting calculations
yielded εTNT = 4000 and 11 200 L mol−1 cm−1 at 700 and 520
−
nm respectively. Unfortunately, the same reaction in MeOH
was shown to be more complicated by NMR spectroscopy, so
the same approach was not appropriate for calculating the
molar absorptivity of TNT− in MeOH. The pre-exponential
factors can be used to find the molar absorption coefficient at
520 nm for TNT−OH−. Each pre-exponential factor found in
the kinetic fits represents the fraction of the absorption at
infinite time attributable to each species, TNT− and TNT−
OH−. If it is assumed that the limiting reagent is completely
Figure 12. Initial absorbance experiment observing the products of 4.1
× 10−2 M TNT reacting with 4.1 × 10−4 M NaOH in MeOH: (a) 0 s;
(b) 0.50 s; (c) 4.83 s; (d) 5.10 s; (e) 5.39 s; (f) 5.67 s; (g) 6.23 s. The
spectrum increases then decreases, with no additional features
appearing during the observation. The inset shows fits of the first 10
s of absorbance data from 520 and 700 nm to eq 4.
−
−
−
consumed in these reactions, then aTNT = εTNT f TNT [L]0,
where fTNT is the fraction of reaction that ends up as TNT−
−
and [L]0 is the initial concentration of the limiting reagent.
−
−
−
−
Likewise, aTNT−OH = εTNT−OH f TNT−OH [L]0 and f TNT
+
−
−
f TNT−OH = 1, so we find the average εTNT−OH at 520 nm to be
∼400 L mol−1 cm−1 in THF.
Table 2. Observed rate constants and pre-exponential factors for pseudo-first order TNT + NaOH experiments in THF and
MeOH calculated by fitting absorbance data at 520 or 700 nm to eq 4. The values in parentheses are the standard errors of the
fit for the last digit of the reported constant
TNT−
TNT−OH−
700 nm
520 nm
520 nm
[TNT]0 (M)
[OH−]0 (M)
[H2O]0 (M)
kobs (s−1
)
a (a.u.)
kobs (s−1
)
a (a.u.)
kobs (s−1
)
a (a.u.)
Reactions in THF
0.186(1)
4.4 × 10−5
3.3 × 10−5
2.2 × 10−5
2.2 × 10−2
1.7 × 10−2
1.3 × 10−2
4.4 × 10−3
3.3 × 10−3
2.2 × 10−3
2.2 × 10−4
1.7 × 10−4
1.3 × 10−4
2.3
7.6(1) × 10−3
7.1(2) × 10−3
7.8(1) × 10−3
7.3(2) × 10−3
7.0(2) × 10−3
8.0(3) × 10−3
6.7(1) × 10−3
6.1(1) × 10−3
7.0(1) × 10−3
3.9(4) × 10−3
5.7(2) × 10−3
5.3(5) × 10−3
0.516(3)
0.343(3)
0.2310(9)
0.503(7)
0.312(2)
0.166(3)
0.18(2)
0.27(2)
0.064(4)
0.059(3)
1.7
0.123(1)
1.1
0.0825(5)
0.14
0.11
0.091
0.123(1)
0.026(3)
0.04(1)
0.12(2)
0.0832(7)
0.018(4)
0.034(7)
0.0489(6)
0.033(8)
Reactions in MeOH
0.185(2)
3.3 × 10−4
2.2 × 10−4
1.5 × 10−4
4.1 × 10−2
2.9 × 10−2
2.2 × 10−2
1.4 × 10−2
2.5 × 10−2
1.8 × 10−2
1.4 × 10−2
4.0 × 10−4
2.9 × 10−4
2.2 × 10−4
1.4 × 10−4
13
0.73(2)
0.85(2)
0.75(2)
0.9(2)
0.3(2)
0.33(8)
0.4(2)
9.3
6.8
1.1
0.7
0.6
0.4
0.12(2)
0.83(1)
0.90(3)
0.401(4)
0.185(3)
0.060(2)
1.22(5)
2.1(3)
1.7(1)
2.8(2)
1.32(3)
0.82(9)
0.33(1)
0.074(3)
I
dx.doi.org/10.1021/jp408992n | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A
Article
The intrinsic rate constants for the reactions between TNT
and OH− can be estimated using the observed fits and the
absorption coefficients. The forward rate constants k1 and k2 are
found from the slopes of plots of akobs versus [TNT]0[OH−]0,
which are equal to εki (i = 1, 2). The reverse rate constants k−1
and k−2 are found from kobs. The slope of a plot of kobs versus
spectrum was normalized to match the absorbance at 563 nm
in each spectrum, as shown in Figure 13B2−G2.
Again, the band shapes from 550 to 800 nm were identical,
supporting the assumption that this spectral region mainly
arises from absorption by TNT−OH−. Finally, the TNT−OH−
component was removed to give the spectra shown in Figure
13B3−G3. A similar process was done for the spectra taken in
excess TNT, where the experimental spectra and the final
difference spectra are shown in Figure 13B4−G4. In excess
base, the primary amines, EA, PA, and BA have few or no
spectral components for which TNT− and TNT−OH− do not
account, with most of the absorbance attributed to TNT−.
There is little difference between the spectra using EA or EA/
H2O, indicating that there is sufficient water in “dry” THF to
allow the reaction to occur. With excess TNT, all of the amines
show a feature that can be assigned to a nitrogen-based σ
adduct, TNT−B, and this accounts for most of the absorbance
in the visible region (the large tail below 425 nm is the edge of
the TNT UV absorption). For DEA, EA, PA, and BA, these
features are consistent with the NMR spectra. For the case of
excess TEA, there is no peak in the 400−500 nm region, and
only small contributions from TNT− and TNT−OH− are
observed. Consistent with the NMR spectroscopy results, the
extent of reaction for the amines with TNT is small, only a few
percent in each case. Estimation of the percent formation of the
N-based σ adduct is not possible because the molar absorption
coefficient is not known at any wavelength. However, in excess
base, DEA had the largest feature assigned to TNT−DEA,
consistent with the NMR spectroscopy data. Other than for the
70% aqueous solution of EA, no additional water was added to
these samples. Thus, the reactivity arises from hydroxide
generated by residual water from the solvents or absorbed from
humidity in the air. This in part limits the extent of reaction.
However, even for the aqueous EA sample, the percent
conversion to the anions is low. It should be recalled that when
neat amine is used as the solvent, no reaction is observed unless
water is added to the system, supporting the need for water for
any reaction to occur.
[H2O]0 gives k−1 when TNT− is the observed species and kobs
=
k−2 when TNT−OH− is monitored. All of these plots are
shown in Figures S17−S20 in the Supporting Information. The
intrinsic rate constants for each solvent are given in Table 3,
Table 3. Intrinsic Rate Constants and Thermodynamic
a
Parameters for the Reaction of TNT with OH− at 20 °C
in THF
1.3(1)
in MeOH
k1 (M−1 s−1
)
3(2) × 105* (520 nm)
2.0(8) × 105* (700 nm)
5(3) × 10−3
1.1(5) × 105* (520 nm)
1.9(6)
k−1 (M−1 s−1
k2 (M−1 s−1
)
2.2(3) × 10−4
2(1)
)
k−2 (s−1
)
0.03(2)
Keq[1]
6(2) × 103
7(6) × 101
−22(2)
Keq[2] (M−1
)
ΔG°[1] (kJ/mol)
ΔG°[2] (kJ/mol)
−10(2)
a
Each value in parentheses is the uncertainty in the last digit of the
reported constant. Values labeled with a * are εki, where the molar
absorption coefficients in MeOH are unknown.
except for those cases where the absorption coefficient is not
known. The reverse rate constant for the deprotonation
reaction is about 10 times larger in MeOH than in THF,
accounting for the more complicated speciation in MeOH. Also
given in Table 3 are the equilibrium constants for eqs 1 and 2
(Keq[1] and Keq[2], respectively) and ΔG° for each reaction at
20 °C in THF. The equilibrium constants were estimated from
the rate constants as Keq[1] = k1/k−1 and Keq[2] = k2/k−2.
The NMR experiments and qualitative tests indicated that
the amine reacted with TNT only in the presence of water.
Furthermore, additional products were observed in the NMR
spectra for some amines, indicating additional complexity in the
reaction between TNT and the amine bases. To attempt to
understand the kinetics of the reaction between the amine bases
and TNT, the visible spectra of the new products needed to be
determined. This was done using THF as the solvent, as shown
in Figure 13. As noted above, the spectrum of the equilibrium
product of TNT reacting with excess NaOH is that of nearly
pure TNT−, while when TNT is in excess both TNT− and
TNT−OH− spectral features are observed.
On the basis of the TDDFT calculations, we made the
assumption that TNT−OH− has minimal absorption at 700 nm
and normalized the TNT− spectrum to match the mixed
spectrum at that wavelength. This assumption is justified by
noting that the band shape of the normalized TNT− spectrum
and the mixed spectrum match identically from 680 nm to the
800 nm baseline. The difference spectrum, shown in Figure
13A3, then represents the spectrum of TNT−OH−. The same
process was followed for the spectra of the amine reaction
products. The spectrum for TNT− was first normalized to 700
nm, as shown in Figure 13B1−G1. Again, the band shapes from
∼670 to 800 nm were identical for all of the amines, showing
that this region in all of the spectra is dominated by TNT−
absorption. Figure 13B2−G2 shows the difference spectra with
the TNT− component removed. Then the TNT−OH−
The N-based σ adduct is the predominant product under
excess TNT conditions, which is counterintuitive. This
demonstrates the role of the hydrolysis reaction of the amine.
In the experiments with excess TNT there is negligible OH−
available to react, perhaps as much as 4 or 5 orders of
magnitude less hydroxide than TNT. This allows for the amine
base to attack the aromatic ring to form the N-based σ adduct.
In the case of excess base, the amine reacts with water,
presumably with a higher rate, to generate hydroxide ion, which
then reacts with the TNT. This drives the hydroxide equilibria
until all of the TNT is consumed, leaving little TNT to form
TNT−B.
Figure 14 shows the same spectral analysis in MeOH. To
estimate the pure TNT− spectrum, a spectrum taken after 1 s of
reaction was used. This is consistent with the NMR
spectroscopy data (Figure 4b) showing that at modest
TNT:OH− ratios only the deprotonated anion was observed.
The TNT−OH− spectrum was estimated from the same
reaction after 300 s and found by difference. Once the TNT−
and TNT−OH− spectra in MeOH were determined, the same
procedure as used for THF was followed to deconvolute the
spectra for the amine bases. The conclusions about the
reactivity are similar for THF and MeOH on the basis of the
visible spectra.
J
dx.doi.org/10.1021/jp408992n | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A
Article
Figure 13. Spectra of the reaction products of various bases with TNT in THF after 300 s. Row A: (A1) 2.2 × 10−5 M TNT + 2.2 × 10−3 M OH−,
giving exclusively TNT−; (A2) black line 2.2 × 10−3 M TNT + 2.2 × 10−5 OH−, red dashed line TNT− contribution; (A3) TNT−OH− spectrum
found as the difference of the spectra from A2. Row B: EA. Row C: EA/H2O. Row D: PA. Row E: BA. Row F: DEA. Row G: TEA. For the amine
bases, the columns show the following spectra: Column 1: black line, product with 100× excess base; red dashed line, TNT− contribution. Column
2: black line, TNT− contribution removed; red dashed line, TNT−OH− contribution. Column 3: black line, residual absorbance from a σ adduct or
CT complex. Column 4: black line, product with 100× excess TNT; blue dashed line, residual absorbance after removal of TNT− and TNT−OH−
contributions.
TNT−(solv) + HB+(solv) ⇄ TNT−B(solv)
On the basis of the experimental observations, the
mechanism of the reaction of TNT with the amines must
(6)
include hydrolysis to generate hydroxide ion and the
ammonium cation, as shown in eq 5:
TNT−OH−(solv) + B(solv)
⇄ TNT−B(solv) + OH−(solv)
(7)
B(solv) + H2O(solv) ⇄ HB+(solv) + OH−(solv)
(5)
TNT−OH−(solv) + HB+(solv)
⇄ TNT−B(solv) + H2O(solv)
The equilibrium constant for this reaction is on the order of
10−5−10−6, depending upon the base. Since no reaction is
(8)
observed in neat amine bases, direct deprotonation of TNT by
the amine and attack of the amine to form the σ adduct can be
ruled out. Under those conditions where a σ adduct can be
observed, it must occur by reaction of TNT− or TNT−OH−
with either the amine or the ammonium cation (eqs 6−8):
We have no evidence to establish which of these equilibria, if
any, are dominant. Even under excess reactant conditions, the
mechanism is too complicated to solve the rate equations in a
simple closed form. In the case of MeOH, we did not consider
K
dx.doi.org/10.1021/jp408992n | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A
Article
Figure 15. Absorbance evolution for the reaction of TNT with DEA in
(left) THF and (right) MeOH: (top left) 8.68 × 10−3 M TNT + 0.90
M DEA in THF from t = 0 to 150 s; (bottom left) 0.352 M TNT + 3.5
× 10−3 M DEA in THF from t = 0 to 253 s; (top right) 3.44 × 10−3
M
TNT + 0.34 M DEA in MeOH from t = 0 to 229 s; (bottom right)
0.101 M TNT + 1.0 × 10−3 M DEA in MeOH from t = 0 to 52 s.
primary product in THF after 5 min of reaction is the N-
centered σ adduct, which is consistent with the visible spectra.
The time-dependent spectra for all of the amines were used
to determine the observed rate constants for the reactions with
these bases, which are reported in Table 4. The data were fit to
biexponentials in most cases, with the fast process being
assigned to formation of TNT− and the slow process being
assigned to formation of the σ adducts TNT−OH− and/or
TNT−B. In the case of TEA, there also may be a contribution
from a CT complex.40 The observed rate constants showed no
pattern as the initial concentration of either TNT or base was
changed by small amounts (factors of 2 or 3), so the reported
values are averages over all measurements. The effect of the
solvent is modest, but generally the reaction to form TNT−
proceeds 2−3 times faster in THF than in MeOH, although the
reaction of DEA is 10 times faster in THF under excess TNT
conditions. In excess base, the observed rate constants are
similar in the two solvents for the primary amines, while DEA
and TEA react more rapidly in MeOH. Formation of the σ
adducts gives similar rate constants for the two solvents with
either reactant in excess. The key finding, though, is that the
amines require the presence of water to react under any set of
conditions.
In an attempt to understand the different reactivities of TNT
with bases in the two solvents, thermodynamic parameters were
estimated using the DFT results. Table 5 shows the parameters
ΔEr°xn, ΔGr°xn, ΔHr°xn, and ΔSr°xn calculated for acid/base and
nucleophilic addition reactions between TNT and several bases
in vacuum. All of the reactions, both acid/base and nucleophilic
addition, with the anionic bases are calculated to be favorable,
while none of these reactions are favorable for the amine bases.
For the reaction of TNT with hydroxide or methoxide, ΔG° is
similar for the acid/base and nucleophilic addition reactions for
each anion (although all of the reactions with OH− are more
favorable by more than 100 kJ/mol compared with CH3O−),
which suggests that a mixture of products should be observed
experimentally. There is a distinction in the entropy change,
which is slightly favorable for the acid/base reactions but
unfavorable for the nucleophilic addition reactions. This implies
that the speciation in the reaction should change considerably
depending upon temperature.
Figure 14. Spectra of the reaction products of various bases with TNT
in MeOH. All of the spectra were measured after 300 s of reaction.
The spectra on the left are under 100-fold excess base conditions, and
the spectra on the right are under 100-fold excess TNT conditions.
For each amine base, the red dotted line shows the contribution of
TNT− to the experimental spectrum, the green dotted line shows the
contribution of TNT−OH− to the experimental spectrum, and the
blue dashed line is the remaining component, assigned to the N-based
σ adduct. The same procedure as described in Figure 13 was used to
subtract the TNT− and TNT−OH− components from each of the
observed spectra.
the possibility that hydroxide deprotonates the alcohol and the
methoxide ion contributes to the overall reactivity because the
equilibrium constants are small.42
As an example of the temporal evolution of the reaction,
Figure 15 shows the absorbance as a function of time for the
reaction of TNT with DEA under conditions of excess TNT
and excess DEA in both THF and MeOH. The prominence of
the TNT− spectrum is evident early in the reaction for all of the
spectra except that for excess TNT in THF. As the reaction
proceeds, the product(s) predominantly become either TNT−
OH− or the nitrogen-centered σ adduct TNT−DEA. In the
case of excess TNT in THF, the primary products are the σ
adducts, although detailed analysis did show a small TNT−
component. On the basis of the NMR spectroscopy results, the
L
dx.doi.org/10.1021/jp408992n | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A
Article
Table 4. Observed rate constants for the reactions of amine bases with TNT. The reported values are averages over at least
three different trials. Values in parentheses are uncertainties determined from the standard error of the mean for the last digit of
the reported rate constant. For EA in THF, the reported values include data from both EA in THF and EA/H2O in THF
100-fold excess TNT
kobs (s−1) TNT−OH−/TNT−B
Reactions in THF
100-fold excess amine
kobs (s−1) TNT−OH−/TNT−B
base
kobs (s−1) TNT−
kobs (s−1) TNT−
EA
0.8(1)
0.0047(5)
5(1)
0.02(1)
PA
4(2)
0.0027(9)
0.015(2)
0.012(2)
0.012(2)
BA
1.6(7)
10(5)
0.9(2)
0.0038(4)
0.0045(8)
0.0027(2)
4(2)
DEA
TEA
2.0(5)
1.1(4)
Reactions in MeOH
EA
0.47(2)
0.52(2)
0.41(1)
0.61(2)
1.3(2)
0.005(2)
0.02(2)
2.9(4)
5.8(8)
3.9(4)
10(4)
7(2)
0.008(3)
0.014(7)
0.006(2)
0.0065(2)
0.024(4)
PA
BA
DEA
TEA
0.02(2)
0.014(8)
Table 5. Thermodynamic Parameters for Each Explored Reaction between TNT and Bases Calculated at the B3LYP/6-31G*//
HF/6-311+G** Level in Vacuum at T = 298.15 K (Data for the Calculations Are Given in Tables S3 and S4 in the Supporting
Information)
reaction
ΔEr°xn (kJ/mol)
ΔGr°xn (kJ/mol)
ΔHr°xn (kJ/mol)
ΔSr°xn (J mol−1 K−1
)
TNT + OH− → TNT− + H2O
TNT + OH− → TNT−OH− C1
TNT + OH− → TNT−OH− C3
−411.3
−487.1
−460.7
−412.4
−435.6
−408.7
−409.8
−480.1
−453.1
8.7
−149.2
−149.0
TNT + CH3O− → TNT− + CH3OH
−293.8
−348.1
−351.5
−289.9
−281.6
−283.4
−286.1
−335.3
−337.8
12.8
−180.2
−182.4
TNT + CH3O− → TNT−OCH3 C1
−
TNT + CH3O− → TNT−OCH3 C3
−
TNT + NH3 → TNT− + NH4
487.3
65.7
16.8
29.9
499.3
113.5
77.8
495.1
69.8
28.0
41.0
−14.2
−146.5
−167.0
−165.8
+
TNT + NH3 → [TNT−][NH4 ]
+
TNT + NH3 → TNT−NH3 C1
TNT + NH3 → TNT−NH3 C3
90.5
TNT + EA → TNT− + H+−EA
TNT + EA → [TNT−][H+−EA]
TNT + EA → TNT−EA C1
TNT + EA → TNT−EA C3
431.7
42.2
−6.1
6.9
440.4
106.9
68.0
438.8
47.4
4.4
−5.5
−199.5
−213.3
−213.8
81.4
17.7
TNT + DEA → TNT− + H+−DEA
TNT + DEA → [TNT−][H+−DEA]
TNT + DEA → TNT−DEA C1
TNT + DEA → TNT−DEA C3
391.0
23.5
36.9
30.1
399.8
94.0
398.0
29.1
47.4
40.2
−5.8
−217.7
−227.8
−229.0
115.4
108.4
TNT + TEA → TNT− + H+−TEA
TNT + TEA → [TNT−][H+−TEA]
371.8
56.1
381.8
128.6
380.0
63.8
−6.2
−217.2
The deprotonation reactions with the amines are all
calculated to be strongly disfavored, consistent with the
experimental observations. When the product is allowed to
form an ion pair, the electrostatic interaction reduces ΔE°,
ΔG°, and ΔH° by more than 300 kJ/mol. The nucleophilic
addition reactions are also disfavored for all of the amines. The
nucleophilic addition and ion pair formation reactions have
similar ΔGo values and are disfavored both enthalpically and
entropically.
To assess the effect of solvation, the SM8 solvation model
was used. Since this model assumes no change to the
vibrational modes upon which the thermodynamic calculations
rely,38 it is possible to use only the energy changes to assess the
influence of a solvent. Figure 16 shows the energy change,
ΔErxn, for each explored TNT reaction as a function of solvent
dielectric constant. The results again show the contrast between
the anionic bases and the amine bases: as the dielectric constant
increases, the reactions with hydroxide and methoxide ions
become less favorable (but always still have ΔErxn < 0), while
the reactions of TNT with amines become more favorable with
increasing dielectric constant, with ΔErxn becoming negative for
most cases at high enough dielectric constant. In all cases, the
proton transfer reaction is calculated to be energetically less
favorable than nucleophilic addition.
The relative reactivities of hydroxide and methoxide are
independent of solvation modeling, as the energy difference
M
dx.doi.org/10.1021/jp408992n | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A
Article
Figure 16. ΔErxn plotted as a function of dielectric constant for TNT reactions with bases. The plots indicate the solvation dependence of TNT
reactions with hydroxide, methoxide, and selected amines. The ordinate for each plot shows the reactants, and the legend for each plot shows the
products. Energy calculations were performed at the B3LYP/6-31G*//HF/6-311+G** level of computation using the SM8 solvation model. Some
points from the free ion computations at lower dielectric constant are off scale and thus have been omitted from this figure. It should be noted that
the ΔErxn data for the methoxide adducts overlap such that the data points cannot be discerned. T = 298.15 K.
between the computed species remains similar for any dielectric
addition, the change in relative energy for the NH3 and EA σ
adduct formations is about an order of magnitude less than for
the hydroxide adducts in methanol (ε = 33). In vacuum, the
computation indicates a nonfavorable reaction for the ammonia
adduct and a barely favorable adduct formation with EA. This
does not project well for the use of TNT/primary amine
reactions for solid−vapor applications to detect TNT.
The size difference of DEA relative to EA changes the
relative favorability of action as a Brønsted−Lowry base versus
a nucleophile. The calculations predict the formation of the
[H+−DEA][TNT−] ion pair to be the most favorable reaction,
with the free ions still more favorable than either σ adduct at
dielectric constants higher than that of 1-propanol (ϵ = 20.1). It
is interesting to note that the C3 carbon is expected to be the
more energetically favorable addition site for the larger DEA
nucleophile, but not by so much that we would not expect to
see evidence of both adducts at equilibrium. Since TNT−TEA
adducts did not converge because of the steric bulk of TEA, the
only comparison to be made is that of TEA’s capacity as a
Brønsted−Lowry base. The TEA deprotonation reactions are
not expected to be energetically favorable.
constant. The C1 TNT−OH− adduct is predicted to be the
−
more energetically favorable product. The TNT−OCH3
adduct isomers show very similar energies for all modeled
dielectric constants, as the larger nucleophile is sterically
hindered at the C1 attack site. This also suggests kinetic
hindrance of the adduct formation, as TNT− is typically
observed prior to the adduct’s appearance.
To compare to experimental results, the ΔErxn values
calculated for TNT + OH− in THF and MeOH were found
to be the following: deprotonation in THF, −250.4 kJ/mol;
deprotonation in MeOH, −132.3 kJ/mol; C1 σ adduct
formation in THF, −310.5 kJ/mol; and C1 σ adduct formation
in MeOH, −185.8 kJ/mol. Quantitatively, the calculated values
are quite different than the ΔG° (which should be similar to,
but not identical to, ΔErxn) values estimated from the rate
constants for the competing reactions (−22 kJ/mol for
deprotonation and −12 kJ/mol for σ adduct formation; Table
3). The differences can be attributed in part to the limitations
of the solvation model used and to errors in the experimental
parameters. In contrast, the calculations are successful in
predicting that the extent of reaction should be significantly
smaller in MeOH than in THF.
The calculations predict that the reaction of the amines may
become favorable in solvents with a sufficiently large dielectric
constant, but by only a few tens of kJ/mol. In contrast, even in
solvents with high dielectric constants, the reaction between
TNT and either hydroxide or methoxide is expected to be
favorable by 100 kJ/mol or more. This implies that in any wet
environment, where amines react with water to generate
hydroxide, there is a significant energetic preference for TNT to
The NH3 and EA products follow a similar trend, where the
σ adduct formation reactions are of lower relative energy than
the formation of an ion pair or free ions. The free ion
formations are calculated to be the least energetically favorable,
especially in solvents with low dielectric constants, which
cannot stabilize the charges through electrostatic forces. In
N
dx.doi.org/10.1021/jp408992n | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A
Article
̈
(2) Uzer, A.; Erc
̧
ag
̆
, E.; Apak, R. Selective Colorimetric
react with hydroxide rather than with the amine. This is
consistent with experimental results.
Determination of TNT Partitioned between an Alkaline Solution
and a Strongly Basic Dowex 1-X8 Anion Exchanger. Forensic Sci. Int.
2008, 174, 239−243.
CONCLUSION
■
́
(3) Fant, F.; De Sloovere, A.; Matthijsen, K.; Marle, C.; El Fantroussi,
We have shown through related 1H NMR and visible
absorbance spectroscopy, kinetic analysis, and computational
modeling that amines do not play a direct role in the initial
formation of decomposition products of TNT. Amines are not
basic enough to directly deprotonate the methyl group of TNT,
nor are they strong enough nucleophiles to compete with the
small quantity of anionic species that they form through other
acid/base reactions in solution. The reaction of an amine with
water to generate hydroxide is required for a reaction with
TNT to be observed, and then both deprotonation and σ
adduct formation can be observed. The speciation in solution
depends upon a combination of factors, including the specific
base, the solvent, and the relative initial concentrations of
reactants. When the base is hydroxide rather than an amine, the
deprotonation reaction to give TNT− is slightly more favorable
in both THF and MeOH compared with formation of the σ
adduct, TNT−OH−. For the amine bases, the primary
component leading to the color change is TNT−, but the σ
adducts, either TNT−OH− or TNT−B, may be present in
larger concentration.
S.; Verstraete, W. The Use of Amino Compounds for Binding 2,4,6-
Trinitrotoluene in Water. Environ. Pollut. 2001, 111, 503−507.
(4) Jiang, Y.; Zhao, H.; Zhu, N.; Lin, Y.; Yu, P.; Mao, L. A Simple
Assay for Direct Colorimetric Visualization of Trinitrotoluene at
Picomolar Levels Using Gold Nanoparticles. Angew. Chem., Int. Ed.
2008, 47, 8601−8604.
(5) Kumar, M.; Vij, V.; Bhalla, V. Vapor-Phase Detection of
Trinitrotoluene by AIEE-Active Hetero-oligophenylene-Based Carba-
zole Derivatives. Langmuir 2012, 28, 12417−12421.
(6) Yang, J. S.; Swager, T. M. Porous Shape Persistent Fluorescent
Polymer Films: An Approach to TNT Sensory Materials. J. Am. Chem.
Soc. 1998, 120, 5321−5322.
(7) Algarra, M.; Campos, B. B.; Miranda, M. S.; Esteves da Silva, J. C.
G. CdSe Quantum Dots Capped PAMAM Dendrimer Nano-
composites for Sensing Nitroaromatic Compounds. Talanta 2011,
83, 1335−1340.
(8) Olley, D. A.; Wren, E. J.; Vamvounis, G.; Fernee, M. J.; Wang, X.;
Burn, P. L.; Meredith, P.; Shaw, P. E. Explosive Sensing with
Fluorescent Dendrimers: The Role of Collisional Quenching. Chem.
Mater. 2011, 23, 789−794.
(9) Gao, D.; Wang, Z.; Liu, B.; Ni, L.; Wu, M.; Zhang, Z. Resonance
Energy Transfer-Amplifying Fluorescence Quenching at the Surface of
Silica Nanoparticles toward Ultrasensitive Detection of TNT. Anal.
Chem. 2008, 80, 8545−8553.
(10) Fang, Q.; Geng, J.; Liu, B.; Gao, D.; Li, F.; Wang, Z.; Guan, G.;
Zhang, Z. Inverted Opal Fluorescent Film Chemosensor for the
Detection of Explosive Nitroaromatic Vapors through Fluorescence
Resonance Energy Transfer. Chem.Eur. J. 2009, 15, 11507−11514.
(11) Zhang, K.; Zhou, H.; Mei, Q.; Wang, S.; Guan, G.; Liu, R.;
Zhang, J.; Zhang, Z. Instant Visual Detection of Trinitrotoluene
Particulates on Various Surfaces by Ratiometric Fluorescence of Dual-
Emission Quantum Dots Hybrid. J. Am. Chem. Soc. 2011, 133, 8424−
8427.
The observation of visible-range absorbances, which are
commonly used in sensor development, result from TNT− or
TNT−B depending on the solvent identity and water content.
This finding may indicate the need for reassessment of existing
sensors for better mechanistic clarification as well as caution
regarding the choice of solvent in similar systems. Any sensor
system using an amine transducer will operate under excess
base conditions, which will require the presence of water to
generate hydroxide to allow reaction. This implies that control
of the humidity in a gas-phase sensor system will be important.
(12) Xia, Y.; Song, L.; Zhu, C. Turn-On and Near-Infrared
Fluorescent Sensing for 2,4,6-Trinitrotoluene Based on Hybrid
(Gold Nanorod)−(Quantum Dots) Assembly. Anal. Chem. 2011,
83, 1401−1407.
ASSOCIATED CONTENT
* Supporting Information
■
S
13C NMR spectra of TNT in THF-d8 and CD3OD; DEPT-135
spectra of TNT reacted with NaOD and then neutralized with
DCl/D2O; derivation of the kinetic model; UV−vis absorbance
spectra as a function of time for all reactions between TNT and
the different bases; and tabulated thermodynamic parameters of
individual modeled reactants and products, including CT
complexes, bases, and conjugate acids. This material is available
free of charge via the Internet at http://pubs.acs.org.
(13) Fyfe, C. A.; Malkiewich, C. D.; Damji, S. W. H.; Norris, A. R.
Flow Nuclear Magnetic Resonance Investigation of the Transient and
Stable Species Formed by the Attack of Alkoxide Ions on 2,4,6-
Trinitrotoluene. J. Am. Chem. Soc. 1976, 98, 6983−6988.
(14) Bernasconi, C. F. Kinetic and Spectral Study of Some Reactions
of 2,4,6-Trinitrotoluene in Basic Solution. I. Deprotonation and
Janovsky Complex Formation. J. Org. Chem. 1971, 36, 1671−1679.
(15) Buncel, E.; Norris, A. R.; Russell, K. E.; Tucker, R. Kinetic Study
of the Reaction of Ethoxide Ion with 2,4,6-Trinitrotoluene and 2,4,6-
Trinitrotoluene-d3. J. Am. Chem. Soc. 1972, 94, 1646−1649.
(16) Shipp, K. G.; Kaplan, L. A. Reactions of α-Substituted
Polynitrotoluenes. II. The Generation and Reactions of 2,4,6-
Trinitrobenzyl Anion. J. Org. Chem. 1966, 31, 857−861.
(17) Caldin, E. F.; Long, G. The Rates and Equilibria of Some
Reactions of Trinitrotoluene and Related Substances from −80 to +20
degrees C. Proc. R. Soc. A 1955, 228, 263−285.
(18) Norris, A. R. Kinetic Studies of the Reactions of 2,4,6-
Trinitrotoluene and 2,4,6-Trinitrotoluene-d3 With Ion-Paired KOBut
and Free tert-Butoxide Ion in tert-Butanol. Can. J. Chem. 1980, 58,
2178−2182.
(19) Buncel, E.; Norris, A. R.; Russell, K. E.; Wilson, H. Kinetic and
Equilibrium Studies of the Reactions of 2,4,6-Trinitrotoluene and
2,4,6-Trinitrotoluene-d3 with Sodium and Potassium t-Butoxide in t-
Butanol. Can. J. Chem. 1974, 52, 2306−2315.
AUTHOR INFORMATION
Corresponding Author
*E-mail: weuler@chm.uri.edu.
Notes
■
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We gratefully acknowledge Drs. Jimmie Oxley and James Smith
for providing us with TNT. We also thank the Department of
Homeland Security ALERT Center for funding.
REFERENCES
■
(1) Lu, Q.; Collins, G. E.; Smith, M.; Wang, J. Sensitive Capillary
Electrophoresis Microchip Determination of Trinitroaromatic Ex-
plosives in Nonaqueous Electrolyte following Solid Phase Extraction.
Anal. Chim. Acta 2002, 469, 253−260.
(20) Tu, R.; Liu, B.; Wang, Z.; Gao, D.; Wang, F.; Fang, Q.; Zhang,
Z. Amine-Capped ZnS−Mn2+ Nanocrystals for Fluorescence Detec-
tion of Trace TNT Explosive. Anal. Chem. 2008, 80, 3458−3465.
O
dx.doi.org/10.1021/jp408992n | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A
Article
(21) Xie, C.; Liu, B.; Wang, Z.; Gao, D.; Guan, G.; Zhang, Z.
Molecular Imprinting at Walls of Silica Nanotubes for TNT
Recognition. Anal. Chem. 2008, 80, 437−443.
(22) Zhou, H.; Zhang, Z.; Jiang, C.; Guan, G.; Zhang, K.; Mei, Q.;
Liu, R.; Wang, S. Trinitrotoluene Explosive Lights Up Ultrahigh
Raman Scattering of Nonresonant Molecule on a Top-Closed Silver
Nanotube Array. Anal. Chem. 2011, 83, 6913−6917.
(23) Wang, Y. Q.; Zou, W. S. 3-Aminopropyltriethoxysilane-
Functionalized Manganese Doped ZnS Quantum Dots for Room-
Temperature Phosphorescence Sensing Ultratrace 2,4,6-Trinitroto-
luene in Aqueous Solution. Talanta 2011, 85, 469−475.
(24) Chen, Y.; Chen, Z.; He, Y.; Lin, H.; Sheng, P.; Liu, C.; Luo, S.;
Cai, Q. L-Cysteine-Capped CdTe QD-Based Sensor for Simple and
Selective Detection of Trinitrotoluene. Nanotechnology 2010, 21,
No. 125502.
(39) Gutmann, V. Empirical Parameters for Donor and Acceptor
Properties of Solvents. Electrochim. Acta 1976, 21, 661−670.
(40) Foster, R.; Mackie, R. K. Intermolecular Charge-Transfer
Complexes of 1,3,5-Trinitrobenzene with Aliphatic Amines. J. Chem.
Soc. 1962, 3843−3844.
(41) Buncel, E.; Russell, K. E.; Wood, J. Hydrogen Exchange in 2,4,6-
Trinitrotoluene. Chem. Commun. 1968, 252−253.
́
(42) Rived, F.; Canals, I.; Bosch, E.; Roses, M. Acidity in Methanol−
Water. Anal. Chim. Acta 2001, 439, 315−333.
(25) Guan, G.; Liu, R.; Wu, M.; Li, Z.; Liu, B.; Wang, Z.; Gao, D.;
Zhang, Z. Protein-Building Molecular Recognition Sites by Layer-by-
Layer Molecular Imprinting on Colloidal Particles. Analyst 2009, 134,
1880−1886.
(26) Zou, W. S.; Sheng, D.; Ge, X.; Qiao, J. Q.; Lian, H. Z. Room-
Temperature Phosphorescence Chemosensor and Rayleigh Scattering
Chemodosimeter Dual-Recognition Probe for 2,4,6-Trinitrotoluene
Based on Manganese-Doped ZnS Quantum Dots. Anal. Chem. 2011,
83, 30−37.
(27) Dasary, S. S. R.; Singh, A. K.; Senapati, D.; Yu, H.; Ray, P. C.
Gold Nanoparticle Based Label-Free SERS Probe for Ultrasensitive
and Selective Detection of Trinitrotoluene. J. Am. Chem. Soc. 2009,
131, 13806−13812.
(28) Xie, C.; Zhang, Z.; Wang, D.; Guan, G.; Gao, D.; Liu, J. Surface
Molecular Self-Assembly Strategy for TNT Imprinting of Polymer
Nanowire/Nanotube Arrays. Anal. Chem. 2006, 78, 8339−8346.
(29) Gao, D.; Zhang, Z.; Wu, M.; Xie, C.; Guan, G.; Wang, D. A
Surface Functional Monomer-Directing Strategy for Highly Dense
Imprinting of TNT at Surface of Silica Nanoparticles. J. Am. Chem. Soc.
2007, 129, 7859−7866.
(30) Latendresse, C. A.; Fernandes, S. C.; You, S.; Zhang, H. Q.;
Euler, W. B. Fluorescent Species Formed by Reaction of
Trinitroaromatics with N,N-Dimethylformamide and Hydroxide. J.
Phys. Chem. A 2013, 117, 324−332.
(31) Olson, E. J.; Xiong, T. T.; Cramer, C. J.; Buhlmann, P.
̈
Interaction of a Weakly Acidic Dinitroaromatic with Alkylamines:
Avoiding the Meisenheimer Trap. J. Am. Chem. Soc. 2011, 133,
12858−12865.
(32) Qasim, M. M.; Gorb, L.; Magers, D.; Honea, P.; Leszczynski, J.;
Moore, B.; Taylor, L.; Middleton, M. Structure and Reactivity of TNT
and Related Species: Application of Spectroscopic Approaches and
Quantum-Chemical Approximations toward Understanding Trans-
formation Mechanisms. J. Haz. Mater. 2009, 167, 154−163.
(33) Qasim, M. M.; Moore, B.; Taylor, L.; Honea, P.; Gorb, L.;
Leszczynski, J. Structural Characteristics and Reactivity Relationships
of Nitroaromatic and Nitramine ExplosivesA Review of Our
Computational Chemistry and Spectroscopic Research. Int. J. Mol.
Sci. 2007, 8, 1234−1264.
(34) Politzer, P.; Murray, J. S.; Koppes, W. M.; Concha, M. C.; Lane,
P. Computational Investigation of Amine Complexes of 2,4,6-
Trinitrotoluene. Cent. Eur. J. Energ. Mater. 2009, 6, 167−182.
(35) Hill, F. C.; Sviatenko, L. K.; Gorb, L.; Okovytyy, S. I.; Blaustein,
G. S.; Leszczynski, J. DFT M06-2X Investigation of Alkaline
Hydrolysis of Nitroaromatic Compounds. Chemosphere 2012, 88,
635−643.
(36) Spartan ’10, version 1.1.0; Wavefunction, Inc.: Irvine, CA, 2010.
(37) Golovina, N. I.; Titkov, A. N.; Raevskii, A. V.; Atovmyan, L. O.
Kinetics and Mechanism of Phase Transitions in the Crystals of 2,4,6-
Trinitrotoluene and Benzotrifuroxane. J. Solid State Chem. 1994, 113,
229−238.
(38) Marenich, A. V.; Olson, R. M.; Kelly, C. P.; Cramer, C. J.;
Truhlar, D. G. Self-Consistent Reaction Field Model for Aqueous and
Nonaqueous Solutions Based on Accurate Polarized Partial Charges. J.
Chem. Theory Comput. 2007, 3, 2011−2033.
P
dx.doi.org/10.1021/jp408992n | J. Phys. Chem. A XXXX, XXX, XXX−XXX