Thermal decomposition pathways of 1,3,3-trinitroazetidine (TNAZ), related 3,3-dinitroazetidium salts, and 15N, 13C, and 2H isotopomers

Article abstract of DOI:10.1021/jp9700950

The thermal decomposition of 1,3,3-trinitroazetidine (TNAZ) and related 3,3-dinitroazetidium (DNAZ⁺) salts was examined neat and in solution. TNAZ kinetics were found (160-250 °C) to be first-order and nearly identical neat and in benzene, with an activation energy of 46.6 kcal/mol (195 kJ/mol). The DNAZ⁺ salts were less thermally stable than TNAZ, and neat did not decompose in a first-order fashion. However, in aqueous solution the DNAZ⁺ salts did decompose following first-order kinetics; their rates were similar with minor differences apparently related to the strength of the anion as a conjugate base. Like simple nitramines such as dimethylnitramine, TNAZ tended to form N₂O rather than N₂, but unlike other nitramines it formed about as much NO as N₂O. TNAZ isotopomers labeled with ¹³C and with ¹⁵N were prepared and used to identify the origin of the decomposition gases and the identity of the condensed-phase products. Early in the decomposition of TNAZ, most of the NO came from the nitro group attached to the azetidium ring nitrogen. Most of the N₂O was the result of the nitro groups interacting with each other, while the majority of the N₂ contained one nitrogen from the ring. Many condensed products have been identified, but five stand out because they are formed in the thermolysis of TNAZ and the three DNAZ⁺ salts [NO₃^-, Cl^-, N(NO₂)₂^-]. These are 3,5-dinitropyridine (M, always a minor product), 1-formyl-3,3-dinitroazetidine (L), 1,3-dinitroazetidine (K), 1-nitroso-3,3-dinitroazetidine (E), and 1-nitroso-3-nitroazetidine (G); the identity of the first four has been confirmed by use of authentic samples. Of these five, the last four have been shown to interconvert with TNAZ and each other under the conditions of these experiments. This study confirms the presence of two competitive TNAZ decomposition pathways. Under the conditions of this study, N-NO₂ homolysis is slightly favored, but products, such as K, resulting from C-NO₂ scission, are also well represented.

Full text of DOI:10.1021/jp9700950

J. Phys. Chem. A 1997, 101, 4375-4383
4375
Thermal Decomposition Pathways of 1,3,3-Trinitroazetidine (TNAZ), Related
15
13
2
3,3-Dinitroazetidium Salts, and N, C, and H Isotopomers
Jimmie Oxley,* James Smith, Weiyi Zheng, Evan Rogers, and Michael Coburn
Department of Chemistry, UniVersity of Rhode Island, Kingston, Rhode Island 02881-0809
ReceiVed: January 7, 1997; In Final Form: April 15, 1997X
The thermal decomposition of 1,3,3-trinitroazetidine (TNAZ) and related 3,3-dinitroazetidium (DNAZ+) salts
was examined neat and in solution. TNAZ kinetics were found (160-250 °C) to be first-order and nearly
identical neat and in benzene, with an activation energy of 46.6 kcal/mol (195 kJ/mol). The DNAZ+ salts
were less thermally stable than TNAZ, and neat did not decompose in a first-order fashion. However, in
aqueous solution the DNAZ+ salts did decompose following first-order kinetics; their rates were similar with
minor differences apparently related to the strength of the anion as a conjugate base. Like simple nitramines
such as dimethylnitramine, TNAZ tended to form N2O rather than N2, but unlike other nitramines it formed
about as much NO as N2O. TNAZ isotopomers labeled with 13C and with 15N were prepared and used to
identify the origin of the decomposition gases and the identity of the condensed-phase products. Early in the
decomposition of TNAZ, most of the NO came from the nitro group attached to the azetidium ring nitrogen.
Most of the N2O was the result of the nitro groups interacting with each other, while the majority of the N2
contained one nitrogen from the ring. Many condensed products have been identified, but five stand out
because they are formed in the thermolysis of TNAZ and the three DNAZ+ salts [NO3 , Cl-, N(NO2)2 ].
These are 3,5-dinitropyridine (M, always a minor product), 1-formyl-3,3-dinitroazetidine (L), 1,3-
dinitroazetidine (K), 1-nitroso-3,3-dinitroazetidine (E), and 1-nitroso-3-nitroazetidine (G); the identity of the
first four has been confirmed by use of authentic samples. Of these five, the last four have been shown to
interconvert with TNAZ and each other under the conditions of these experiments. This study confirms the
presence of two competitive TNAZ decomposition pathways. Under the conditions of this study, N-NO2
homolysis is slightly favored, but products, such as K, resulting from C-NO2 scission, are also well represented.
-
-
Introduction
We have previously reported the decomposition kinetics of
a number of pure nitramines, both simple nitramines, such as
dimethylnitramine (DMN) and diisopropylnitramine (DIPN),1
and multifunctional nitramines such as 1,3,5-trinitro-1,3,5-
triazocyclohexane (RDX) and 1,3,5,7-tetranitro-1,3,5,7-tetrazo-
cine (HMX).2 For these compounds we found first-order
kinetics and nitrosoamine products formed by at least one
decomposition pathway. Now we examine a new3 multifunc-
tional energetic material, 1,3,3-trinitroazetidine (TNAZ), which
has both N-NO2 and C-NO2 linkages. It is unique among
the energetic nitramines in that it can be held above its melting
temperature (101 °C)4 without immediate decomposition. This
property makes it probable that like 2,4,6-trinitrotoluene (TNT),
TNAZ will be processed in the liquid phase; therefore, a
thorough study of its thermal stability is required. One question
of interest to us was which X-NO2 linkage would be weakest.
Politzer predicted an energy difference in products 1 and 2 of
only 2 kcal (Figure 1).5 Although the melt-phase mixtures of
TNAZ have recently been reported,6 the few reports of its
thermal decomposition are limited to examination of its
decomposition gases.7,8 The methodology of our study was to
determine the kinetics of TNAZ decomposition and compare
them to pure nitramines and to identify TNAZ decomposition
products. To aid in the latter study, salts related to TNAZ, the
salts of 3,3-dinitroazetidine (DNAZ), were examined in terms
of their decomposition kinetics and products.
Figure 1. Theoretical bond dissociation energies for TNAZ (ref 5).
1-formyl-3,3-dinitroazetidine were provided by Drs. Alan Marc-
hand, University of North Texas, and Paritosh Dave, Geo-
Centers, NJ. Synthesis of TNAZ isotopomers is described
elsewhere.9 In sealed glass tubes (2.4 mm i.d. × 60-70 mm
long, volume about 200 µL),1 thermolyses were performed
isothermally in the temperature range 160-280 °C for TNAZ
and 100-200 °C for dinitroazetidium salts. For neat thermoly-
ses 0.2-0.3 mg of TNAZ or 0.4-0.5 mg of dinitroazetidium
salts was used; for solution thermolyses about 50 µL of 1%
TNAZ solution in acetone, benzene, or methanol or 1%
dinitroazetidium (DNAZ+) salts in aqueous solution were used.
After thermolysis, the fraction remaining was assessed by
dissolving samples in acetone or methanol and analyzing by
liquid chromatography. For most samples, first-order rate
constants were determined using at least six data points out to
70% decomposition. The thermolyses of the neat dinitroaze-
tidium salts were the exception; first-order rate constants were
derived from only the first 30% of the decompositions.
Arrhenius parameters were determined using rate constants for
at least four temperatures.
Experimental Section
TNAZ was provided by Eglin Air Force Base. 3,3-Dini-
troazetidium salts were synthesized by Dr. Michael Hiskey of
Los Alamos National Laboratory; 1,3-dinitroazetidine and
A Hewlett Packard (HP) high-performance liquid chromato-
graph (HPLC) Model 1084 B equipped with a 20 µL sample
X Abstract published in AdVance ACS Abstracts, June 1, 1997.
S1089-5639(97)00095-9 CCC: $14.00 © 1997 American Chemical Society

4376 J. Phys. Chem. A, Vol. 101, No. 24, 1997
Oxley et al.
TABLE 1: TNAZ Rate Constants, Neat, in Solution (1%), and with Additives
sample
no. data
7
temp range (°C)
160-250
Ea (kcal/mol)
A (s-1
3.55E+17b
)
R2
solventa/additive
rate const at 240 °C
DKIE
1.31
TNAZ
46.6
0.99
neat
neat
4.84E-03
3.69E-03
5.32E-03
5.03E-03
3.86E-03
8.47E-03
2.28E-03
2.05E-02
1.06E-02
5.40E-03
5.26E-03
4.74E-03
1.52E-02
2.74E-04c
1.36E-04c
1.57E-03
TNAZd4
TNAZ
TNAZ
TNAZd4
TNAZ
TNAZ
TNAZ
TNAZ
TNAZ
TNAZ
TNAZ
TNAZ
TNAZ
TNAZ
NODNAZ
6
160-260
40.3
5.61E+14
1.00
benzene
benzene-d6
benzene
acetone
acetone-d6
methanol
MeOH-d4
+10% DPA
+5% HNO3
+50% H2O
H2O
1.06
1.38
7
4
189-280
180-240
43.6
30.7
2.89E+16
2.62E+11
0.99
0.96
3.71
1.93
+50% RDX
neat
acetone
a 1% wt TNAZ in solvents. b Read as 3.55 × 1017. c At 200 °C.
Figure 4. Arrhenius plots of TNAZ neat and 1% TNAZ in various
solvents.
Figure 2. Arrhenius kinetic plots of TNAZ and other energetic
materials.
The separation, identification, and quantification of gaseous
products was accomplished using a HP 5890 gas chromatograph
equipped with a Alltech Heyesep DB 100/120 (30 ft × 1/8 in.)
packed column and a thermal conductivity detector (GC-TCD).
The oven temperature was held at 35 °C for 6 min and then
ramped at 15 °C/min to 180 °C. Certified standard gases (Scott
Speciality Gas Co.) were used to generate standard curves.
Identification of gases was by retention times and by comparison
to the standard gases. Quantification was achieved using
standard curves of area versus concentration of the standard
gases. A HP 5890 gas chromatograph equipped with an
electronic pressure control system, a poraPLOT Q (Chrompack)
capillary column (25 m × 0.25 mm i.d.), and a HP 5971 electron
impact quadrupole mass spectrometer was used for analysis of
gases produced from the decomposition of labeled TNAZ.10 The
capillary tubes containing samples were evacuated and sealed.
After thermolysis sample tubes were broken in a flexible Teflon
tube connected in line with the GC carrier gas and just before
the injector and electronic pressure control system. The
decomposition gases flow into the injector (split ratio 15:1) and
onto the column. The oven temperature was held at -80 °C
for 5 min and then ramped at 15 °C/min to 150 °C.
Figure 3. Energetic materials used as explosives.
loop and a Waters 486 tunable absorbance detector was
employed to determine the rate constants. An Alltech Econo-
sphere C-18, 5 µm (250 mm × 4.6 mm), reversed phase column
and an eluent of CH3OH/THF/H2O(32:4.8:63.2) was used to
separate TNAZ and its products. A Dionex AS-11 anion
column and a mobile phase of 2 mM NaOH/40% CH3CN (flow
1.0 mL/min) and a UV detector (214 nm) were used to analyze
Differential scanning calorimetry (DSC) and thermal gravi-
metric analysis (TGA) were used to determine the relative
thermal stability of TNAZ and dinitroazetidium salts. DSC
measurements used a Perkin-Elmer (PE) DSC-4 or a TA
Instruments modulated DSC-2910, under N2 flow using an
indium calibration standard. DSC samples were weighed (0.2-
0.5 mg) and sealed into glass capillary tubes (0.058-0.059 in.
o.d. and 0.011-0.012 in. wall). These were placed in aluminum
cradles in the DSC head and heated over the range 40-450 °C
at a scan rate of 20 °C/min. TGA experiments used a TGA-2
the DN-, NO3-, and NO2 anions. The concentration of Cl-
-
did not vary much during the course of the thermolysis; it was
determined using a Dionex Model 2000i/sp ion chromatograph
equipped with a AS-4A column and conductivity detector and
using a mobile phase of 2.2 mM Na2CO3/0.7 mM NaHCO3.
+
The cations DNAZ+ and NH4 were separated and detected
using a Waters IC-Pak cation M/D column with a mobile phase
of 0.1 mM EDTA/3 mM HNO3 (flow, 1.0 mL/min) and a
Waters 431 conductivity detector.

Thermal Decomposition Pathways of TNAZ
J. Phys. Chem. A, Vol. 101, No. 24, 1997 4377
Figure 6. Arrhenius plots of DNAZ+ salts, 1% in water, analyzed for
DNAZ+ remaining.
Results
TNAZ Kinetics. Table 1 lists the TNAZ rate constants
obtained from first-order plots over the temperature range 160-
280 °C. Like the nitramines previously examined,1,2 neat TNAZ
at all temperatures was observed to decompose according to
first-order kinetics out to 90% decomposition. The Arrhenius
plot of these rate constants is shown in Figure 2 along with
plots for other energetic materials showing activation energies
(Ea) and pre-exponential factors (A): TNAZ (Ea ) 46.6 kcal/
mol; A ) 3.55 × 1017 s-1); HMX (Ea ) 52.9 kcal/mol; A )
2.46 × 1018 s-1); TNT (Ea ) 41.0 kcal/mol; A ) 4.44 × 1012
s-1); RDX (Ea ) 37.8 kcal/mol; A ) 1.99 × 1014 s-1).
Compared to other commonly used explosives (Figure 3), the
thermal stability of TNAZ is better than that of RDX but worse
than that of HMX or TNT.
Figure 5. 3,3-Dinitroazetidium (DNAZ+) salts examined by thermal
analysis.
TABLE 2: DSC and TGA of DNAZ Salts and TNAZ
DSC (Scan Rate: 20 °C/min)
mp
(°C)
onset
(°C)
exomax
(°C)
-∆H
(cal/mol)
sample
MW
DNAZDN
DNAZNO3
254
210
128
138
147
148
151
951
886
838
435
652
152
161
180
191
DNAZDNT 306
148(dec) 157
161(dec) 174
183.5 177 188
DNAZNTO 277
DNAZCl
TNAZ
TNAZ
192
101
97(endo) 102(endo) 32(endo)
258
278
1085
In benzene solution TNAZ decompositions were first-order,
with rates and activation energies similar to neat TNAZ (Table
1). (For RDX and HMX a similar observation was made; that
is, there was little difference between the rate of the neat and
the rate of the solution decomposition.) However in methanol
and acetone, the TNAZ decomposition rate increased (Figure
4). The increase in TNAZ decomposition rate and variety of
decomposition products indicated a TNAZ/methanol reaction.
Decomposition kinetics were examined for the presence of a
deuterium kinetic isotope effect (DKIE). A small DKIE (1.31
and 1.38, respectively) was observed when neat TNAZ or
TNAZ-d4 (1% in benzene) was thermolyzed at 240 °C. In the
solvents in which TNAZ/solvent interaction (methanol and
acetone) occurred, a large DKIE was measured (Table 1);
however, no DKIE was observed when TNAZ was heated in
deuterated benzene. From these results, we concluded there
was no intermolecular DKIE and only a small intramolecular
one, a result mirroring our observations with RDX and dim-
ethylnitramine. As is the case with nitramines, nitric acid and
diphenylamine (NO2 scavenger) additives affected the TNAZ
decomposition rate only slightly (Table 1), whereas TNAZ, like
TNT, was reactive with ammonia, immediately forming a brown
residue. TNAZ in admixture with RDX exhibited a 2-fold
increase in decomposition rate at 200 °C (Table 1).
Dinitroazetidium Salt Kinetics. The decomposition kinetics
of dinitroazetidium (DNAZ+) salts (Figure 5) were examined
by DSC and by TGA (Table 2). The DSC exothermic maxima
for all the dinitroazetidium salts appear at lower temperatures
than that for TNAZ. This is in agreement with isothermal
studies of these salts which determined decomposition rate
constants at 150 °C in the range 10-2 s-1, whereas the calculated
rate constant for TNAZ at the same temperature is about 10-6
s-1. With the exception of TNAZ, relative ordering of thermal
stabilities observed by TGA is the same as by DSC: DNAZ
TGA (Ramp @ 10 (°C/min)
mp
(°C)
onset
range
% wtd loss
(range)
sample
(°C)a % wt lossb
(°C)c
DNAZDN
DNAZNO3
128
138
156
160
170
174
96.8
92.4
92.7
94.4
135-166
143-165
144-177
149-206
224-292
162-196
235-318
150-212
107-204
87.2
74.6
67.1
46.0
34.1
28.8
42.3
84.6
91.7
DNAZDNT 148(dec)
DNAZNTO 161(dec)
DNAZDNI
151(dec)
180
96.2
DNAZCl
TNAZ
177
101
200
161
97.5
92.2
a The onset of weight loss. b The total weight loss (40-400 °C).
c The ranges indicate regions of significant weight loss. d The weight
loss over the specified range.
thermogravimetric analyzer, nitrogen flushing gas, about 0.5 mg
size sample, and a scan rate of 10 °C/min from 40-400 °C.
The thermolysis products of TNAZ and dinitroazetidium salts
were analyzed by a Varian 3400 gas chromatograph equipped
with a capillary column (J&W DB-5MS, 30 m × 0.25 mm i.d.)
and interfaced to a Finnigan-MAT TSQ-700 tandem mass
spectrometer (injector temperature 200 °C; transfer line 220 °C;
injection volume 1 µL). The oven temperature was held at 80
°C for 1 min and then ramped at 7.5 °C/min to 210 °C. Electron
impact (EI) and chemical ionization (CI) were at 70 eV with
emission currents of 400 µA. The ion source was operated at
150 °C. Methane was used as the CI reagent gas. Collision-
induced dissociation experiments were carried out at collision
energy 15 eV, with 0.5-1 mTorr (1 Torr ) 133.3 Pa) of argon
as the collision gas. A solids probe (25-35 °C) was also used
to introduce TNAZ into the source of the mass spectrometer.
This latter method of introduction yielded mass spectra identical
to those obtained using the GC interface.

4378 J. Phys. Chem. A, Vol. 101, No. 24, 1997
Oxley et al.
TABLE 3: Rate Constants (s-1) and Kinetics Parameters of DNAZ Salts
Neat
DNAZNO3
DNAZ+
NO3
total gas
ASTMa
DNAZCl ASTMa
DNAZDN total gas
-
T (°C)
100
4.17E-04
1.98E-03
5.38E-03
9.07E-03
120
3.14E-03b
2.73E-02
9.54E-02
2.09E-01
3.61E-01
27.6
1.08E-03
3.02E-03
7.50E-03
4.59E-02
8.33E-02
16.4
1.59E-03
4.48E-03
7.34E-03
140
160
170
180
Ea (kcal/mol)
16.8
21.7
49.8
19.1
A (s-1
)
8.7E+12
0.992
1.5E+06
1.00
3.6E+06
4.1E+9
1.3E+22
6.9E+07
R2
1.00
0.995
0.972
0.993
1% DNAZ Salts in Aqueous Solutions
DNAZCl
DNAZNO3
DNAZDN
DNAZ+ DN-
DNAZNTO
DNAZ+
-
T (°C)
DNAZ+
NO3
DNAZ+
100
2.00E-05
2.05E-04
1.32E-03
7.50E-03
3.55E-02
3.06E-06
2.16E-05
9.21E-05
4.56E-04
1.54E-04
4.65E-03
25.0
1.89E-05
1.86E-04
1.26E-03
6.83E-03
3.05E-02
2.80E-05
2.83E-04
1.97E-03
1.17E-02
4.31E-02
9.53E-07
1.22E-05
1.02E-04
8.98E-04
5.83E-03
3.11E-02
36.5
1.11E-04
9.37E-04
5.79E-03
3.48E-02
120
140
160
180
200
Ea (kcal/mol)
31.5
30.9
32.2
30.6
A (s-1
)
6.4E+13
1.8E+09
0.999
2.8E+13
2.2E+14
2.2E+15
1.00
9.7E+13
R2
0.999
1.00
0.999
0.999
a ASTM method E698-79. b Read as 3.14 × 10-3
.
TABLE 4: Decomposition of Aqueous (1%) DNAZ+X- Salts
at 140 °C
acid
pKa
X-
Cl-
k (s-1) (for DNAZ+)
HCl
-7.0
1.25E-03a
1.97E-03
1.32E-03
5.79E-03
-
HN3O4
HNO3
NTO
-5.62
-1.34
+3.67
N(NO2)2
-
NO3
NTO-
a Read as 1.25 × 10-3
.
TABLE 5: Moles of Gas per Mole of DNAZ Salts Heated
for 10 Half-Lives
sample
N2
N2O CO2
CO
NO
total temp (°C)
Neat (mole gas/mole DNAZ salt)
Figure 7. Arrhenius plots of DNAZ+NO3-, 1% in water, analyzed
for cation and anion remaining.
DNAZDN
DNAZDN
DNAZDN
DNAZDN
DNAZDN
1.06 1.72 1.35 0.10 0.03 4.26
1.12 1.38 1.14 0.25 0.00 3.89
1.15 1.64 1.48 0.25 0.00 4.52
1.33 1.73 1.54 0.33 0.09 5.02
0.72 1.17 1.30 0.49 0.96 4.64
100
120
140
150
240
120
140
150
240
160
240
DNAZNO3 1.04 0.61 1.31 0.12 0.46 3.54
DNAZNO3 1.17 0.51 1.38 0.22 0.10 3.39
DNAZNO3 1.18 0.45 1.36 0.35 0.29 3.63
DNAZNO3 0.70 0.13 1.2
0.46 0.72 3.21
DNAZCl
DNAZCl
0.81 0.16 0.63 0.39 0.14 2.13
0.71 0.03 0.44 0.34 0.47 1.99
1-2% DNAZ Salts in H2O at 160 °C
DNAZDN
1.08 1.25 1.10 0.08 0.0
3.51
2.43
1.08
DNAZNO3 0.82 0.64 0.97 0.00 0.0
DNAZCl
0.41 0.34 0.33 0.00 0.0
DNAZ+(NO3-) as shown in Figures 7 and 8. With thermolysis
of the neat salts [DNAZ+Cl- and DNAZ+(NO3)-], the disap-
pearance of the cation was significantly faster than that of the
anion (Figure 8). While the decomposition kinetics of the
aqueous solutions of the dinitroazetidium salts were first-order,
first-order plots of the decomposition of the neat salts were
nonlinear after the first 30-40% decomposition (Figure 8). A
slight yellowing of the salts indicated the onset of accelerated
loss of the cation. Condensed-phase rate constants shown in
Tables 3 and 4 are derived from the first 30% decomposition.
Products of Dinitroazetidium Salt. For only three of the
dinitroazetidium salts [DNAZCl, DNAZ(NO3), DNAZDN] did
we have sufficient sample to examine the decomposition
products. Table 5 shows the gas composition after complete
(10 half-lives) decomposition as a function of thermolysis
temperature. Only for NO is composition markedly affected
by the thermolysis temperature; it increases with temperature.
Figure 8. First-order plots of DNAZ+NO3- neat at 120 °C, analyzed
for cation and anion remaining.
DN ∼ DNAZ(NO3) ∼ TNAZTGA < DNAZ DNT < DNAZ
NTO < DNAZ Cl < TNAZ
.
DSC
Discrepancy in the TGA we attribute to sublimation of TNAZ,
which leads to early weight loss. Some of the 3,3-dinitroaze-
tidium salts (DNAZ+) were examined in more detail: chloride
(Cl-), nitrate (NO3-), dinitramidate (DN-), and the anion of
NTO. In aqueous solution (1%) disappearance of the dini-
troazetidium cation was first-order, and the rate was very similar
for all four dinitroazetidium salts (Figure 6). The cation of the
DNAZ+ NTO- salt decomposed somewhat faster than the other
salts. For all four salts in solution as well as neat, the
dinitroazetidium cation disappeared faster than the anion, e.g.

Thermal Decomposition Pathways of TNAZ
J. Phys. Chem. A, Vol. 101, No. 24, 1997 4379
TABLE 6: Isotopic Distribution of Labeled Isomers for DNAZNO3 (50% dec) and TNAZ at 240 °C (75% dec)
a
b
DNAZ-3-13C-15NO3
DNAZ-15N-15NO3
TNAZ3-13Cc TNAZ-15Nd TNAZ-N-15Ne TNAZ3-13C-N-15Nf
100 °C 150 °C 200 °C 100 °C 150 °C 150 °Cg 200 °C
240 °Ci
240 °Ci
240 °Ci
240 °Ci
N2
19.7
75.7
4.6
36.2
59.3
4.5
42.9
53.7
3.4
0.5
89.0
10.5
N.D.
N.D.
0
0.3
83.3
16.4
2.2
97.8
0.04
0.3
76
23.7
0.8
99.2
0
2.7
14.8
72.3
12.9
3.0
97.0
0.1
99.0
1.0
0
99.4
0.6
98.9
0.89
0.20
62.6
37.4
77.0
23.0
55.7
44.3
2.8
68.9
28.3
4.4
95.6
0
18.9
81.1
98.9
1.1
99.4
0.6
43.0
51.6
5.4
54.1
45.9
34.8
50.9
14.2
98.8
1.2
40.0
53.5
6.5
N15N
15N15N
NO
N.D.h
N.D.
6.2
45.5
54.5
24.5
48.7
26.8
65.2
34.8
70.5
29.5
31.1
47.1
52.9
32.5
47.5
20.0
74.3
25.7
73.1
26.9
35.0
52.3
47.7
31.5
52.3
16.1
61.6
38.4
75.0
25.0
51.3
15NO
N2O
N15NO
37.4
2.5
97.5
3.4
96.6
11.0
88.9
15N15NO 56.4
97.3
CO2
56.9
43.1
64.8
35.2
26.4
13CO2
CO
99.2
0.8
84.6
13CO
HCN
H13CN
HC15N
32.1
67.9
29.4
70.6
29.2
70.8
34.1
65.9
40.7
50.2
48.3
20.6
52.3
12.7
48.7
59.3
15.4
H13C15N 23.4
-
-
a 13C on C3, 15N on NO3
.
b 15N on all nitro group and NO3
.
c 13C on C3. d 15N on all nitro group. e 15N on nitramine nitro group. f 13C on C3,
15N on nitramine nitro group. g 10 t1/2
.
h N.D. ) not detected. i 75% dec.
TABLE 7: Products of Decomposition of Neat DNAZ Saltsf
peak code
E
G
I
J
K
L
M
N
O
S
S1
176 131 192 130 147 175 169 200 223 214 214
a
b
DNAZNO3
DNAZNO3
DNAZClc
DNAZCld
DNAZDNe
l
l
m
l
l
t
t
l
s
t
s
m
t
t
s
m
l
m
m
s
t
m
m
s
t
t
s
t
t
s
s
m
m
t
a 240 °C, 5 s. b 150 °C, 150 s. c 196 °C, 1 min. d 200 °C, 1 min.
e 150 °C, 1 min. l ) large; m ) medium; s ) small; t ) tiny.
g Molecular weight.
TABLE 8: Decomposition Products of 1% DNAZ Salts in
Solutiona
DNAZNO3
MeOHb MeOHb EtOHb
240 °C/ 150 °C/ 150 °C/
Figure 9. Condense-phase products of neat TNAZ and DNAZ+ salts.
DNAZCl
MeOHb
DNAZDN
MeOHb
peak
code MW
5 s
300 s
300 s 150 °C/300 s 150 °C/300 s
A
B
C
150
149
161
t
t
s
t
s
s
l
m
m
m
s
s
m
l
t
D2 147
E
F
G
H
I
s
m
s
m
l
m
s
s
s
s
t
s
m
s
t
s
l
t
t
s
s
t
s
t
176
191
131
191
192
130
147
175
169
200
223
214
213
264
309
322
l
s
s
l
l
s
t
t
s
s
t
s
J
K
L
M
N
O
S
S1
T
U
V
t
s
t
m
t
Figure 10. TNAZ and DNAZ products in methanol.
s
t
t
l
and DNAZ+(NO3)- can produce gas, we examined the behavior
of DNAZ+Cl- to judge the products formed by dinitroazetidium.
Formation of 1 mol of nitrogen gas per 1 mol of dinitroazetidium
salt is a common feature of the neat decompositions, as is the
appearance of 1 mol of carbonaceous gas (0.6 CO2 and 0.4 CO).
Nitrous oxide was only an important gaseous product in the
decomposition of the dinitramidate salt, and we have already
m
l
t
l
s
m
s
a l ) large; m ) medium; s ) small; t ) tiny. b Solvent.
Our analytical techniques did not allow us to quantify NO2,
which was observed when hot samples were first removed from
the thermolysis baths. DNAZ+Cl- formed the least gas, about
2 mol gas/mol DNAZ+Cl-; DNAZ+(NO3)- formed about 3.5
mol gas/mol DNAZ+(NO3)-; and DNAZ+DN- formed the most
gas, about 5 mol gas/mol DNAZ+DN-. Less gas was formed
in the aqueous decompositions of the salts than was formed in
the neat decompositions: about 1 mol less gas per mol
dinitroazetidium salt. Since the anions of both DNAZ+DN-
shown that dinitramidate decomposes producing 1 mol of N2O
- 11
and 1 mol of NO3
.
Small amounts of hydrogen cyanide and
cyanogen were also observed. Information as to the origins of
these gases can be derived from Table 6, which shows the
isotopic distribution of gases derived from DNAZ+(NO3)-
labeled with 13C on C3 and 15N on the nitrate ion (DNAZ-3-

4380 J. Phys. Chem. A, Vol. 101, No. 24, 1997
Oxley et al.
TABLE 9: Decomposition Gases of Neat Energetic
Materials after Heating 10 Half-Lives
decompositions of the neat dinitroazetidium salts were primarily
examined for organic products. These products are tabulated
in Tables 7 (for neat thermolyses) and 8 (for solution thermoly-
ses) and pictured in Figures 9 and 10. Some of the products
(E, K, L, and M) were identified using authentic samples. Some
are assigned by observing fragmentation patterns and comparing
them to that of isotopically labeled DNAZ+(NO3)-. Due to
insufficient quantities of these decomposition products, no
quantitation is possible. Relative amounts can be inferred from
the size of the chromatographic peak, as indicated in the tables
(l ) large; m ) medium; s ) small, t ) tiny). In the case of
neat DNAZ+(NO3)-, we examined the residue for water-soluble
salts and found almost no ammonium ion, in contrast to
observations in the aqueous decompositions. We assumed that
the counterbalancing cation for the slowly disappearing anion
(e.g. Figure 8) is a proton.
CO2% CO% N2O% N2% NO% total (mol) temp (°C)
DMN
DIPN
RDX
HMX
0.53 0.14 0.02 0.31
0.09 0.04 0.09 0.78
0.20 0.20 0.29 0.31
0.22 0.11 0.52 0.15
0.22 0.17 0.02 0.36 0.23
0.5
0.5
4.4
5.4
2.0
4.6
3.2
4.0
1.9
0.6
240
240
240
240
240
240
240
240
200
200
DNAZCl
DNAZDN 0.28 0.11 0.25 0.16 0.21
DNAZNO3 0.38 0.14 0.04 0.22 0.22
TNAZ
ADN
KDN
0.38 0.16 0.03 0.21 0.22
0.53 0.47
0.86 0.14
TABLE 10: Distribution of Isotopic Labeled TNAZ in
Decomposition Gases
22 s
60 s
143 s
300 s
30 m
(100%)
MW
(10%)
(25%)
(50%)
(75%)
Products of TNAZ. The gaseous products of the condensed-
phase thermolysis of TNAZ were similar to those of
DNAZ+(NO3)- and DNAZ+Cl- and to simple nitramines such
as dimethylnitramine (DMN) in that the amount of nitrogen
produced was much greater than the amount of nitrous oxide
(Table 9). Both TNAZ and related dinitroazetidium salts
produced more carbon dioxide than any of the pure nitramines
examined, and unlike the pure nitramines, they all formed
detectable amounts of NO.1,7,8 Nitrogen was the first nitrogen-
containing gas formed. After 20 s heating at 240 °C (roughly
9% decomposition) the gases observed were 0.13 L of N2, 0.51
L of CO2, and 0.22 L of CO at STP per mol of TNAZ; no N2O
nor NO was detected. Isotopic distributions of the decomposi-
tion gases were determined using TNAZ 15N-labeled only at
the nitro group attached to the azetidine nitrogen (TNAZ-N-
15N), TNAZ 15N-labeled at all three nitro groups (TNAZ-15N),
TNAZ with 13C on C3 (TNAZ-3-13C), and TNAZ with both
13C on C3 and 15N on the nitro group of the azetidine nitrogen
(TNAZ-3-13C-N-15N) (Table VI). Table 10 shows the fraction
of each isotopically labeled gas as a function of the fraction
TNAZ reacted at 240 °C. As in the case of the dinitroazetidium
salts, NO2 was sometimes observed but not quantified.
Neat TNAZ-N-15N (15N on Nitramine Nitro Group, 240 °C)
N2
28
29
30
30
31
44
45
46
27
28
28.4
65.0
6.5
34.1
65.9
27.7
56.1
16.2
90.7
9.3
37.1
57.1
5.8
37.4
62.6
27.0
56.9
16.1
90.7
9.3
38.1
55.7
6.2
45.9
54.1
28.9
53.2
17.9
87.7
12.3
43.0
51.6
5.4
54.1
45.9
34.8
50.9
14.2
84.6
15.4
53.2
42.7
4.1
71.4
28.6
45.3
45.2
9.4
N15N
15N15N
NO
15NO
N2O
N15NO
15N15NO
HCN
HC15N
83.2
16.8
Neat TNAZ-3-13C (13C on C3, 240 °C)
CO
28
29
44
45
27
28
58.5
41.5
53.0
47.0
77.9
22.1
70.5
29.5
57.8
42.2
66.0
34.0
78.9
21.1
64.3
35.7
53.1
46.9
77.0
23.0
62.6
37.4
55.7
44.3
80.0
20.0
63.5
36.5
44.7
55.3
13CO
CO2
13CO2
HCN
H13CN
TABLE 11: GC/MS of Neat TNAZ at 240 °C 5 min, Using
CIc
peak codes
D1
E
G
I
J
K
L
M
N
sample
112d 176 131 192 130 147 175 169 189
The NMR spectrum of the TNAZ in acetone solution,
previously sealed and heated at 240 °C for 4 min (about 70%
conversion), showed only one decomposition product, identified
as 1-nitroso-3,3-dinitroazetidine (NO-DNAZ). This identifica-
tion was confirmed by comparison with an authentic sample
TNAZ 10 s
TNAZ 60 s
TNAZa
s
s
l
l
l
l
l
l
l
l
l
l
l
l
l
l
l
l
l
l
l
l
t
t
t
t
t
s
s
s
t
t
t
s
t
t
s
s
s
s
s
s
s
s
t
s
s
s
TNAZ
t
t
m
m
s
m
m
m
m
m
m
m
m
m
l
TNAZ
t
TNAZa,b
t
1
[TNAZ H 5.47 ppm; 13C 105 and 65 ppm; 15N 212 N ring;
TNAZN-15N
TNAZ-15N
TNAZ-3-13C
TNAZ-313C-N-15N
t
t
t
t
t
t
t
s
t
t
t
t
t
m
m
s
1
362 N-NO2; 366 C-NO2; NO-DNAZ H 6.03 and 5.17 ppm;
13C 108, 68, and 63 ppm]. In the thermolysis of neat TNAZ,
no insoluble residue was observed in partially decomposed
samples; but at long decomposition times, neat TNAZ formed
a black tarlike residue. A detailed study of the decomposition
products was conducted using a tandem mass spectrometer. The
results, using GC/MS, are summarized for neat TNAZ and
various TNAZ solutions in Tables 11 and 12, respectively. The
assignments of molecular weights are based on the CI mass
spectrum of each decomposition product. Some of the products
(E, K, L, and M) were identified using authentic samples.
Others are assigned on the basis of their fragmentation patterns
and comparison to that of isotopically labeled TNAZ.12 Relative
amounts can be inferred from the size of the chromatographic
peak as indicated in the tables (l ) large; m ) medium; s )
small, t ) tiny).
m
t
s
a
TNAZ-d4
a Using electron impact detection. b Under the vacuum. l ) large;
m ) medium; s ) small; t ) tiny. d Molecular weight.
c
Figure 11. DNAZ+ salts, possible decomposition mechanism.
Discussion
13C-15NO3) and labeled with15N on both nitro groups and the
nitrate ion (DNAZ-15N-15NO3).
Decomposition Rates. TNAZ, neat and in benzene, decom-
posed by first-order kinetics at all temperatures examined,
leaving some insoluble residue. In protonic solvents such as
water, acetone, or methanol, TNAZ decomposition rate was
In solution all the dinitroazetidium salts formed ammonium
and nitrite ions. The dinitramidate salt also formed nitrate. No
attempt was made to quantify these ionic products. The

Thermal Decomposition Pathways of TNAZ
J. Phys. Chem. A, Vol. 101, No. 24, 1997 4381
TABLE 12: GC/MS of 1% TNAZ in Different Solvents at 240 °C Using CIa
peak codes
sample
dec (%)
B
D2
E
F
G
H
I
J
K
M
N
P
Q
S
T1
U
TNAZ/MeOH 10 s
TNAZ/CD3OD 10 s
TNAZ/MeOH 40 s
TNAZ/MeOH 80 s
TNAZ/MeOH 163 s
TNAZ/MeOH 5 min
TNAZ/MeOH 13 min
TNAZ/H2O 46 s
19
10
56
81
96
99.8
100
50
75
92
5
s
t
s
s
s
t
t
t
m
l
t
t
m
s
m
s
l
l
l
m
s
s
s
l
m
m
l
l
l
t
t
s
s
t
t
t
s
m
m
l
l
t
s
t
t
t
t
t
t
s
s
t
m
m
m
t
t
t
t
s
s
t
l
s
m
m
s
s
t
TNAZ/acetone 163 s
TNAZ/acetone 5 min
TNAZ/C6D6 10 s
t
s
l
m
l
TNAZ/C6H6 3 min
TNAZ/C6D6 3 min
62
60
l
l
l
l
t
s
t
s
s
t
t
a l ) large; m ) medium; s ) small; t ) tiny.
tidium cation to form DNAZ. DNAZ, regardless of its salt of
origin, decomposes at approximately the same rate (Figure 11).
The autocatalytic nature of the reaction may be a result of the
basic nature of the decomposition products of DNAZ.
Gaseous Products. The TNAZ and the dinitroazetidium salts
formed similar distributions of gaseous decomposition prod-
ucts: 2-3 times as much CO2 as CO, about equal ratios of NO
to N2 (at 240 °C), and with the exception of DNAZ+DN-,
significantly more N2 than N2O (Tables 5 and 9). Experiments
with isotopically labeled DNAZ+(NO3)- indicated that most of
the nitrogen gas contained one azetidium nitrogen, while NO
and N2O were formed almost exclusively from the nitrate and
nitro groups (50% decomposition, 150 °C, Table 6). The CO
and CO2 are formed more or less equally from all three carbons.
At 100 °C, nitrogen-containing gases and CO2 contain more
nitrate nitrogen and C3, respectively.
Analysis of the decomposition gases from the labeled TNAZ
shows the fate of the nitro group attached to the azetidine
nitrogen; it is responsible for a disproportionate share of the
early formation of NO, N2O, and N2 (Table 10). In the early
stages of the reaction (10% decomposition) 66% of the NO
comes from the nitro group on the azetidine, but by completion
of the reaction all three nitro groups contribute equally to the
NO composition. The azetidine nitrogen, at most, accounts for
4% of the total NO (Table 6). The fact that at early decomposi-
tion the contribution from the nitro group on the azetidine is
disproportionately high suggests at 240 °C the N-NO2 bond is
preferentially broken over the C-NO2 bond. Analysis of the
fate of the labeled C3 shows that the label is disproportionately
high in the carbonaceous gases at early stages of decomposition,
but decreases to roughly a third of the CO2 and less than a fifth
of the CO, suggesting this is the first carbon oxidized (Table
6).
Figure 12. Thermolysis of TNAZ in acetone (1%) at 240 °C. The
decomposition of TNAZ (fraction remaining) and the formation of NO-
DNAZ (mol NO-DNAZ/mol TNAZ) are included.
Figure 13. First-order plot of neat TNAZ thermolysis at 240 °C,
monitoring TNAZ loss and gas formation.
Condensed-Phase Products. Mass spectra indicate that
1-nitroso-3,3-dinitroazetidine (NO-DNAZ, E) is the major
decomposition product of neat TNAZ and TNAZ in benzene
or acetone solutions (Tables 11 and 12). It is also the major
decomposition product of DNAZ+Cl-, DNAZ+(NO3)-, and
DNAZ+DN-, neat or in methanol (Tables 7 and 8). These three
dinitroazetidium salts share a number of decomposition products
in common with TNAZ: I, E, L, M, G, J, K, N, O (in order
of abundance) (Figure 9). This is not surprising since TNAZ
(I) is a product of their decompositions (Table 7). When the
dinitroazetidium salts or TNAZ is heated in methanol or ethanol
solution, they form decomposition products that show solvent
attachment to the nitrogen of the azetidium ring as well as
products resulting from ring cleavage (Tables 8 and 12, Figure
10). Even in benzene a product T1 was identified were the
phenyl ring is attached to the azetidium nitrogen. Assignments
for the attachment to the azetidine nitrogen rather than to C3
enhanced 2- to 4-fold, and a DKIE was exhibited in deuterated
solvent (Table 1). These observations and the fact that the
activation energy of decomposition is about 10 kcal/mol lower
in methanol than in benzene (Table 1) suggest homolysis
decomposition pathways, dominant in aprotic solvents, are
overwhelmed by a hydrogen transfer pathway in protonic
solvents. In nitroarenes competition between C-NO2 homolysis
and a pathway where hydrogen transfer to the nitro group assists
C-NO2 bond breakage is well-known.13
Neat dinitroazetidium salts decompose in a first-order fashion
to about 30-40% decomposition and then show accelerated
decomposition, which leaves no insoluble residue. The dini-
troazetidium salts have markedly less thermal stability than
TNAZ; their order of thermal stability in water roughly follows
the acidity of the parent acid (Table 4). These results suggest
a mechanism that involves proton transfer from the dinitroaze-

4382 J. Phys. Chem. A, Vol. 101, No. 24, 1997
Oxley et al.
Figure 14. Principal decomposition products of TNAZ and its decomposition products (MS/MS study).
SCHEME 1: TNAZ Decomposition Mechanism
Figure 15. Proposed formation of 3,5-dinitropyridine.
are based on labeling studies, especially noting the loss of 15
N
from TNAZ- N-15N. This may be interpreted as indicative that
N-NO2 scission is more predominant than C-NO2, although
other interpretations are possible. For example, K is the major
decomposition product when TNAZ is heated in methanol,
ethanol, or water (Table 12). This indicates C-NO2 scission
does occur, but perhaps, unlike the amine radical, the C3 radical
can only be easily capped by H• rather than by a larger
hydrocarbon originating from the solvent.
the first and most abundant product observed under the
conditions of these experiments, next in terms of amount and
order of formation are 1,3-dinitroazetidine (K) and 1-formyl-
3,3-dinitroazetidine (L), respectively. We speculate that 1-ni-
troso-3,3-dinitroazetidine (NO-DNAZ, E) is the first TNAZ
The species observed from TNAZ decomposition are the same
as observed in the decomposition of the dinitroazetidium salts
(Table 11, neat TNAZ, and Table 12, TNAZ solutions). The
compounds coded as E, K, L, and M were identified by
comparison with authentic samples. While NO-DNAZ (E) is

Thermal Decomposition Pathways of TNAZ
J. Phys. Chem. A, Vol. 101, No. 24, 1997 4383
SCHEME 2: General Reaction Scheme for TNAZ and
DNAZ Salts
The reaction scheme of TNAZ is complex because I, E, K,
and probably G can interconvert with each other (Scheme 1).
Although 1-formyl-3,3-dinitroazetidine (L) can form I and E
and vice versa, L differs from the other azetidine products in
that its formation must sacrifice another azetidine to provide a
carbon source. 3,5-Dinitropyridine (M) is a minor product of
all the dinitroazetidine decompositions; it is formed by thermal
ring cleavage and subsequent cycloreversion (Figure 15). An
intertwining scheme of products is shown in Scheme 2.
Products E and K are the first and second products in terms of
which products are formed first and in terms of which products
are formed in greatest abundance. Product K is the more stable
of the two. Products E and K are linked through 1-nitroso-3-
nitroazetidine (G). An authentic sample of K was observed to
form both G, E, and TNAZ, while E has been observed to form
G, K, L and TNAZ.
Conclusions
These results are consistent with two important decomposition
pathways for TNAZ: one via N-NO2 bond homolysis and one
via C-NO2 cleavage. Under the conditions of these experi-
ments, N-NO2 is favored, although its dominance is not huge;
66% of the initial (10% decomposition) NO formed was
observed (by 15N-labeling experiments) to come from the nitro
group on the azetidine nitrogen. The activation energy for
TNAZ decomposition (46.6 kcal/mol) is similar to that of pure
nitramines.2 A small DKIE (1.4) was observed as well as a
sensitivity to ammonia; these may result from the C-NO2
cleavage pathway in which hydrogen transfer to the nitro group
may assist in bond breakage. This conclusion is consistent with
the observed enhanced decomposition of TNAZ in protonic
solvents.
TABLE 13: GC/MS of Neat NO-DNAZ, Ka, and Lb at 240
°C Using CIc
peak code
D1
E
G
I
J
K
L
M
N
sample
112d 176 131 192 130 147 175 169 189
NO-DNAZ(E)/1 min
NO-DNAZ(E)/5 min
K/5 min
t
m
l
s
s
l
m
t
t
t
s
s
m
m
t
s
t
s
t
l
K/10 min
s
l
m
s
s
l
m
s
Acknowledgment. We thank Dr. Richard Miller of the
Office of Naval Research for funding this work.
L/5 min
m
l
a 1,3-Dinitroazetidine (K). b 1-Formyl-3,3-dinitroazetidine (L). c l )
large; m ) medium; s ) small; t ) tiny. d Molecular weight.
References and Notes
(1) Oxley, J. C.; Hiskey, M. A.; Naud, D.; Szekeres, R. J. Phys. Chem.
1992, 96, 2505-2509.
product formed. However, under the conditions of these
experiments, it decomposes at an appreciable rate. Figure 12
shows the rate of TNAZ disappearance (1% in acetone) along
with the rate of NO-DNAZ (E) appearance. The concentration
of the latter quickly reaches a plateau. In the decomposition
of neat TNAZ, gas formation was slower than TNAZ loss, we
believe this is due to the fact that TNAZ forms condensed-
phase products that decompose at different rates to ultimately
form the gases (Figure 13). In contrast to NO-DNAZ (E), 1,3-
dinitroazetidine (K) is quite stable. A comparison of the DSC
(20 °C/min) shows that TNAZ and K (exothermic maxima 275
and 274 °C, respectively) are significantly more stable than L
and NO-DNAZ (exothermic maxima 247 and 246 °C, respec-
tively). The thermal stability of NO-DNAZ relative to TNAZ
is reversed in acetone. While neat NO-DNAZ is definitely less
stable than TNAZ, in acetone NO-DNAZ decomposes more
slowly than TNAZ (TNAZ k ) 8.47 × 10-3 s-1 and NO-DNAZ
k ) 1.57 × 10-3 s-1 at 240 °C). The acetone stabilization of
the nitroso decomposition products of nitramines has previously
been observed for RDX.2
(2) Oxley, J. C.; Kooh, A. B.; Szekeres, Zheng, W. J. Phys. Chem.
1994, 98, 7004-7008.
(3) Archibald, T. G.; Gilardi, R.; Baum, K.; George, C. J. Org. Chem.
1990, 55, 2920-2924.
(4) Lyer, S.; Eng, Y. S.; Joyce, M.; Perez, R.; Alster, J.; Stec, D.
Proceedings of ADPA Manufacture Compatibility of Plastics and other
Materials with ExplosiVes, Propellants, Pyrotechnics and Processing of
ExplosiVes, Propellants, and Ingredients, San Diego, CA, April 1991, p
80.
(5) Politzer, P.; Seminario, J. M. Chem. Phys. Lett. 1993, 207, 27-
30. Murray, J. S.; Seminario J. M.; Lane, P.; Politzer P. J. Mol. Struct.
(THEOCHEM) 1990, 207, 193-200.
(6) Mc Kenney, R. L.; Marchand, A. Submitted J. Energ. Mater.
(7) Anex, D. S.; Allman, J. C.; Lee, Y. T. In Chemistry of Energetic
Materials; Olah, G. A., Squire, D. R., Eds.; Academic Press: New York,
1991; Chapter 2.
(8) Oyumi, Y.; Brill, T. B. Combust. Flame 1985, 62, 225-231.
(9) Coburn, M. D.; Hiskey, M. A.; Oxley, J. C.; Smith, J. L.; Zheng,
W.; Rogers, E. J. Energ. Mater. submitted.
(10) Zheng, W.; Dong, X. X.; Rogers, E.; Oxley, J. C.; Smith, J. L. J.
Chromat. Sci., submitted.
(11) Oxley, J. C.; Smith, J. L.; Zheng, W.; Rogers, E.; Coburn, M. D.
J. Phys. Chem., submitted.
(12) Zheng, W.; Rogers, E.; Coburn, M. D.; Oxley, J. C.; Smith, J. L.
J. Mass Spectrom. submitted.
A tabulation of the decomposition products of E, K, and L
is given in Table 13 and schematically illustrated in Figure 14.
(13) Minier, L.; Brower, K.; Oxley, J. C. J. Org. Chem. 1991, 56, 3306-
3314.