Thermal Risk Analysis Based on Reaction Mechanism: Application to the 2,6-Diaminopyrazine-1-oxide Synthesis Process

Article abstract of DOI:10.1021/acs.oprd.0c00529

Thermal risk analysis is essential for the development of chemical reactions. This work should be carried out on the basis of a thorough comprehension of the reaction mechanism. In this article, the synthesis process for 2,6-diaminopyrazine-1-oxide (DAPO), an important intermediate in the synthesis of the famous explosive 2,6-diamino-3,5-dinitropyrazine-1-oxide (LLM-105), was employed to show the importance of understanding the reaction mechanism for thermal risk analysis. First, we investigated the reaction mechanism of DAPO synthesis. The reaction mechanism was divided into two stages on the basis of the amount of triethylamine dosed: the first half of triethylamine dosing and the second half of triethylamine dosing until the end of the reaction. Then the thermal properties of DAPO synthesis and the thermal stability of the materials involved were experimentally studied using a reaction calorimeter (RC1) and a differential scanning calorimeter (DSC), respectively. The results show that the temperature corresponding to the maximum reaction rate reached in a time of 24 h under adiabatic conditions (TD24) is higher than the maximum temperature of the synthesis reaction (MTSR) for both of these stages, indicating that once cooling failure occurs, immediately stopping addition of triethylamine could prevent the occurrence of secondary decomposition.

Full text of DOI:10.1021/acs.oprd.0c00529

pubs.acs.org/OPRD
Article
Thermal Risk Analysis Based on Reaction Mechanism: Application to
the 2,6-Diaminopyrazine-1-oxide Synthesis Process
Published as part of the Organic Process Research & Development joint virtual special issue “Process Safety from
Bench to Pilot to Plant” in collaboration with ACS Chemical Health & Safety and Journal of Loss Prevention in
the Process Industries.
Yuehua Li, Yang Wang, Longyu Liao, Zichao Guo,* Liping Chen, and Wenqian Wu
Cite This: Org. Process Res. Dev. 2021, 25, 849−857
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: Thermal risk analysis is essential for the development of chemical reactions. This work should be carried out on the
basis of a thorough comprehension of the reaction mechanism. In this article, the synthesis process for 2,6-diaminopyrazine-1-oxide
(DAPO), an important intermediate in the synthesis of the famous explosive 2,6-diamino-3,5-dinitropyrazine-1-oxide (LLM-105),
was employed to show the importance of understanding the reaction mechanism for thermal risk analysis. First, we investigated the
reaction mechanism of DAPO synthesis. The reaction mechanism was divided into two stages on the basis of the amount of
triethylamine dosed: the first half of triethylamine dosing and the second half of triethylamine dosing until the end of the reaction.
Then the thermal properties of DAPO synthesis and the thermal stability of the materials involved were experimentally studied using
a reaction calorimeter (RC1) and a differential scanning calorimeter (DSC), respectively. The results show that the temperature
corresponding to the maximum reaction rate reached in a time of 24 h under adiabatic conditions (TD24) is higher than the
maximum temperature of the synthesis reaction (MTSR) for both of these stages, indicating that once cooling failure occurs,
immediately stopping addition of triethylamine could prevent the occurrence of secondary decomposition.
KEYWORDS: 2,6-diaminopyrazine-1-oxide (DAPO), cyclization reaction, reaction mechanism, thermal risk assessment
storage and use of intermediates, and high product purity. The
cyclization reaction takes N-nitrosobis(cyanomethyl)amine (1)
as the raw material to synthesize the intermediate 2,6-
diaminopyrazine-1-oxide (DAPO). As one important step to
synthesize LLM-105, several predecessors have conducted
research on this process. Zhao16 speculated about the
cyclization reaction process of DAPO. Wang17 optimized the
synthesis process of DAPO and replaced the triethylamine with
sodium hydroxide to increase the DAPO yield to 78%. For
research on the DAPO cyclization process, there are few
related literature reports, and research on the thermal risk of
this process is hardly involved.
This work evaluated the thermal risk of the cyclization
process based on a thorough understanding of the cyclization
mechanism. First, the cyclization mechanism of DAPO was
studied by means of small-scale tests of reactants and reference
to relevant literature. Then the exothermic properties were
experimentally investigated using a reaction calorimeter
(RC1e), and finally, the thermal stability of the materials
involved in the cyclization was investigated. This work studied
1. INTRODUCTION
2,6-Diamino-3,5-dinitropyrazine-1-oxide (LLM-105) is a
highly heat-resistant explosive developed by the Livermore
Energetic Materials Laboratory in the United States.1−3 LLM-
105’s energy is 20% higher than that of trinitrotriaminoben-
zene (TATB), and it has relatively low sensitivity to impact,
sparks, friction, and shock waves. It has shown certain
application prospects in the fields of petroleum perforating
bombs and boosters.4−8
To synthesize LLM-105, Bellamy9−12 used 2,6-diaminopyr-
azine as the starting material to obtain the target product LLM-
105 through a two-step process of oxidation and nitration.
However, the side product of oxidation in this method has
properties similar to those of the main product, which is not
conducive to separation and purification, and the nitration
reaction process is difficult to control. Another traditional
synthesis method uses 2,6-dichloropyrazine as the starting
material and completes it through a four-step reaction (two
methoxyl oxidation, nitration, amination, and oxidation).13
However, this process has the disadvantages of high cost and
use of a highly corrosive material, namely, trifluoroacetic
acid.14
Received: December 7, 2020
Published: March 2, 2021
Pagoria15 used iminodiacetonitrile as the starting material to
obtain LLM-105 through a three-step reaction of nitrosation,
cyclization, and nitration (the synthesis route is illustrated in
Figure 1). This synthesis route has the advantages of cheap and
readily available raw materials, simple process steps, easy
https://dx.doi.org/10.1021/acs.oprd.0c00529
Org. Process Res. Dev. 2021, 25, 849−857
© 2021 American Chemical Society
849

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Figure 1. Synthetic method for LLM-105 through a three-step reaction of nitrosation, cyclization and nitration.
Table 1. Specific Operating Conditions for the RC1e tests
mass of initially charged mixture
(g)
mass of hydroxylamine hydrochloride
(g)
mass of triethylamine
(g)
mole
a
run
initially charged mixture
ratio
1
2
3
4
compound 1/anhydrous methanol
compound 1/anhydrous methanol
compound 1/anhydrous methanol
anhydrous methanol
15/118.5
15/118.5
15/118.5
118.5
9
9
18
9
26.2 (6.55/6.55/13.1)
26.2
26.2
26.2 (14.2/12)
1:1:2
1:1:2
1:2:2
1:2
a
Molar ratio for runs 1, 2, and 3: compound 1:hydroxylamine hydrochloride:triethylamine. Molar ratio for run 4: hydroxylamine
hydrochloride:triethylamine.
Figure 2. Synthesis mechanism of DAPO.
the thermal risk of the reaction process from the perspective of
the reaction mechanism, which can provide a more accurate
understanding of the risk of the reaction and improve the
accuracy of thermal risk assessment of chemical processes.
In run 4, anhydrous methanol was added to the reactor
without the starting material 1. At 5 °C, hydroxylamine
hydrochloride was added, and then triethylamine was dosed in
two portions (14.2 and 12 g).
2.2.2. Differential Scanning Calorimeter. The differential
scanning calorimeter (DSC) was manufactured by Mettler
Toledo. The samples were sealed in gold-plated high-pressure
crucibles (40 μL). For dynamic tests, the heating rates were 2,
4, 8, and 10 K/min, and the temperature range was 30−300
°C. All of the DSC tests were carried out under a high-purity
nitrogen atmosphere, and the nitrogen flow rate was 50 mL/
min. The DSC crucibles were weighed before and after the
tests, and no significant mass loss was observed.
2. EXPERIMENT AND APPARATUS
2.1. Reagents. N-Nitrosobis(cyanomethyl)amine (98%
purity) was purchased from Shanghai Yichun Chemical
Technology Co., Ltd. Hydroxylamine hydrochloride (purity
99%) and triethylamine (purity 99%) were purchased from
Shanghai Minrell Chemical Technology Co., Ltd. Anhydrous
methanol (AR, ≥99.5% purity) was purchased from
Sinopharm Chemical Reagent Company. All of the reagents
were used without further purification.
2.2. Apparatus. 2.2.1. Reaction Calorimeter. The RC1e
reaction calorimeter used in this work was manufactured by
Mettler Toledo and had a 0.5 L glass reactor and an automatic
feeding device. The specific operating conditions for the RC1e
tests are listed in Table 1.
A total of four runs in the isothermal mode were carried out
to characterize the thermal properties of the cyclization
reaction. The operation procedures of runs 1, 2, and 3 were
as follows: First, anhydrous methanol and compound 1 were
added to the reactor at 5 °C. Then hydroxylamine hydro-
chloride was added, after which triethylamine was slowly dosed
with stirring. The reaction mixture was incubated at 5 °C for
30 min and then heated up to 25 °C and allowed to react for 2
h. Postprocessing included filtration, washing, and drying.
Triethylamine was dosed through a feeding pump. For run 1,
triethylamine was dosed in three portions (6.55, 6.55, and 13.1
g). For runs 2 and 3, triethylamine was continuously dosed to
the reactor at one time.
2.2.3. 1H NMR Spectroscopy. 1H NMR spectra were
obtained using a Bruker Avance III 500 MHz spectrometer
1
with an H sensitivity of ≥650:1 (0.1% EB). The chemical
shifts (δ) were reported in parts per million downfield from
tetramethylsilane as an internal standard. The solvent was
1
DMSO-d6. In the H NMR spectrum, the characteristic peaks
of DMSO and H2O appear at δ 2.5 and δ 3.3, respectively.
2.2.4. High-Performance Liquid Chromatography. The
composition of the product obtained from the RC1e
experiment was quantitatively determined by HPLC. The
HPLC instrument was produced by Shandong Wukong
Instrument Company. The manufacturer of HPLC column is
. The chromatographic column used was a Venusil MP C18
column (4.6 mm × 250 mm, 5 μm) (Agela Techonologies Inc.,
serial number 035843). The solvent system was CH3OH/
CH3COONH4 + 0.1% HCOOH, and the flow rate was 1 mL/
min. The column temperature was 30 °C, and the absorption
wavelength was 254 nm.
2.2.5. Mass Spectrometry Analysis. A Finnigan TSQ
Quantum Ultra AM mass spectrometer manufactured by
Thermo Fisher Scientific was used to analyze the molecular
850
https://dx.doi.org/10.1021/acs.oprd.0c00529
Org. Process Res. Dev. 2021, 25, 849−857

Organic Process Research & Development
pubs.acs.org/OPRD
Article
weight of the sample. Before the test, the sample was dissolved
in anhydrous methanol, and the dissolved sample was filtered
with a 0.22 μm microporous membrane.
3. RESULTS AND DISCUSSION
3.1. Research on the Mechanism of the Cyclization
Reaction. With respect to the synthesis mechanism of DAPO,
on the basis of Harrington’s18 research, Zhao16 speculated that
the synthesis mechanism of DAPO includes three steps in
sequence: elimination of −NO, condensation of hydroxyl-
amine, and proton transfer. Triethylamine plays two roles in
this reaction: one is to react with hydroxylamine hydrochloride
to generate free hydroxylamine, and the other is to remove
−NO from compound 2. Wang17 experimentally proved that
the synthesis of DAPO starts with the condensation reaction of
compound 1 and hydroxylamine, not elimination of −NO.
Wang proposed the synthesis pathway shown in Figure 2.
To further understand the cyclization processes, three small-
scale tests with different molar ratios were carried out. The
molar ratios and experimental results are shown in Table 2.
1
Figure 4. H NMR spectrum of DAPO.
Table 2. Experimental Results with Different Molar Ratios
of Reactants
a
b
expt
molar ratio
product
yield (%)
1
2
3
30:1:1:2
30:1:1:1
30:1:2:2
DAPO
2
4
60
−
91
a
Molar ratio: anhydrous methanol:compound 1:hydroxylamine
b
1
hydrochloride:triethylamine. The product yields were calculated as
Figure 5. H NMR spectrum of compound 2.
100% × (molar amount of product)/(molar amount of compound 1).
filtration of the reaction mixture, we also used HPLC to test
the reaction filtrate. As shown in Figure 6b, this indicated that
compound 4 still remained in the reaction filtrate. From the
structure of compound 4, one can conclude that both of the
cyano groups in compound 1 condensed with free hydroxyl-
amine, and no elimination reaction occurred. On this basis, the
synthesis mechanism of compound 4 shown in Figure 7 is
proposed.
On the basis of the above results, one can conclude that the
reaction involves two stages. In the first stage, triethylamine
reacts with hydroxylamine hydrochloride to produce free
hydroxylamine, which then condenses with compound 1 to
produce compound 2. In the second stage, triethylamine
catalyzes the removal of the −NO group from compound 2. If
the addition of triethylamine stops before half of the
When the compound 1:hydroxylamine hydrochloride:triethyl-
amine molar ratio was 1:1:2, a light-yellow precipitate (Figure
1
3a) was obtained after the reaction. The H NMR results for
the precipitate in Figure 4 proved that the precipitate was
DAPO. At a molar ratio of 1:1:1, the reaction produced a
1
brown precipitate (Figure 3b). According to the H NMR
(Figure 5) and mass spectrometry results, this precipitate was
verified to be compound 2. When the molar ratio was 1:2:2, a
white precipitate (Figure 3c) was obtained after the reaction.
The product was analyzed by HPLC and 1H NMR
1
spectroscopy, as shown in Figure 6. According to H NMR
analysis, the product was N-nitrosoiminodiethylamine oxime
(4). The yield of compound 4 was ∼91%. The HPLC results
indicated that the purity of compound 4 was about 95%. After
Figure 3. Solid precipitates: (a) DAPO; (b) compound 2; (c) compound 4.
851
https://dx.doi.org/10.1021/acs.oprd.0c00529
Org. Process Res. Dev. 2021, 25, 849−857

Organic Process Research & Development
pubs.acs.org/OPRD
Article
1
Figure 6. (a) H NMR spectrum and (b) HPLC results for compound 4.
−14.05 kJ (based on compound 1, the normalized heat was
−116.2 kJ/mol).
3.2. Reaction Calorimetry Results. In this section, the
thermal features of DAPO synthesis reaction are discussed.
The heat flow profiles for runs 2, 3, and 4 are shown in Figures
9−11, respectively, and summarized in Table 3.
Figure 7. Synthetic mechanism of compound 4.
triethylamine has been dosed into the reactor, the removal of
the −NO group and the cyclization reaction will not happen.
This is the case because the triethylamine added will
immediately be transformed to triethylamine hydrochloride,
which cannot catalyze the subsequent removal of the −NO
group and cyclization reaction.
To characterize the reaction rate in the first dosing stage,
RC1e run 1 was conducted. The results are shown in Figure 8.
Figure 9. Heat flow and dosing profiles for run 2.
Figure 8. Heat flow profile for run 1.
It can be seen that when the dosing of triethylamine was
stopped when half of the triethylamine had been dosed, both
the heat flow and the Tr profile gradually decreased. In
combination with the result of laboratory-scale experiment 2 in
Table 2, one can reasonably expect that in the first dosing stage
only the condensation of free hydroxylamine and compound 1
occurs and that the subsequent reactions such as removal of
the −NO group and cyclization hardly occur.
Furthermore, after the first dosing stage, the reactions of
−NO group removal, proton transfer, cyclization, and
rearrangement occur in a continuous way, which means that
it is impossible to separate these reactions in the second
reaction stage. By integration of the heat flow profile in Figure
8, the heat released in the first dosing stage was obtained as
Figure 10. Heat flow and dosing profiles for run 3.
Through integration of the heat flow profiles (Figures
9−11), the total heats (Qtot) generated in runs 2, 3, and 4 were
−34, −29, and −2.4 kJ, respectively. In run 3, the compound
1:hydroxylamine hydrochloride:triethylamine molar ratio was
1:2:2. On the basis of the reaction mechanism discussed in the
previous section, both of the cyano groups in compound 1
condensed with the free hydroxylamine to generate compound
4.
852
https://dx.doi.org/10.1021/acs.oprd.0c00529
Org. Process Res. Dev. 2021, 25, 849−857

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Figure 11. Heat flow and dosing profiles for run 4.
Table 3. RC1 Results for Runs 1−4
Figure 12. DSC test curves for the related materials in stage 1 at 10
K/min.
a
b
run
cp (J g−1 K−1
)
Qtot (kJ)
ΔTad (K)
Table 4. DSC Test Results for the Related Materials in Stage
1
1
2
3
4
2.44
2.44
2.41
2.46
−34
−34
−29
−2.4
82.6
82.6
67.7
6.3
a
b
sample
mass (mg)
To (°C)
Q (J/g)
compound 1
sample1
sample 2
sample 3
compound 2
0.77
3.70
2.65
3.55
0.34
163.2
205.3
105.2
90.6
−2899.0
−117.2
−73.5
−198.5
−5217.1
a
cp is the specifical heat capacity of the reaction mixture and was
obtained by the calibration produced of RC1. For Q, the exotherm is
expressed by the minus sign.
b
130.1
a
With respect to the heat flow profiles in run 4, one can find
that the exothermic signals sharply decreased to zero once the
dosing event stopped. At first, we considered that the heat
released was caused by the neutralization reaction between
hydroxylamine hydrochloride and triethylamine. However, this
contradicted with the fact that heat was still released after a
stoichiometric amount of triethylamine had been added. Then
we speculated that the heat released was ascribed to the mixing
between triethylamine and methanol. This speculation was
experimentally proven by another RC1e run in which
triethylamine was dosed into methanol (see the Supporting
Information for more details).
Onset temperature (i.e., the temperature at which the exothermic
b
peak deviates from the baseline). Decomposition heat.
tested by dynamic DSC experiments at four different heating
rates. The dynamic DSC heat flow profiles and results for
sample 3 are shown in Figure 13 and Table 5.
3.3. Thermal Stability Analysis of Materials. Thermal
stability analysis of the materials involved in the DAPO
synthesis process is an essential task for thermal risk
assessment. In this section, we discuss the dynamic DSC
tests that we employed to quantitatively characterize the
thermal stabilities of the materials in stage 1 (condensation of
free hydroxylamine) and stage 2 (−NO elimination,
cyclization, proton transfer, and rearrangement).
3.3.1. Stage 1. In stage 1, the following materials are
involved: compound 1, a mixture of compound 1 and
anhydrous methanol (sample 1), a mixture of compound 1,
anhydrous methanol, and hydroxylamine hydrochloride
(sample 2), the final reaction mixture in the first stage (sample
3), and compound 2. The dynamic DSC results are shown in
Figure 12 and Table 4.
Figure 13. DSC test results for sample 3 at different heating rates.
In Figure 13, one can see that the decomposition of sample
3 contains two overlapped processes. The onset temperature of
the first process ranged from 65.5 to 90.6 °C. The average
From Table 4, one can see that compared with compound 1,
compound 2 has a lower onset temperature (130.1 °C) and
higher decomposition heat (−5217.1 J/g), indicating that
compound 2 is more thermally risky than compound 1.
Consequently, as shown in Table 4, the onset temperature of
sample 3, which contains compound 2, is lower than those of
samples 1 and 2. Furthermore, the decomposition heat of
sample 3 is higher than those of samples 1 and 2. Since sample
3 is the most thermally unstable mixture in Table 4, we studied
the thermal stability of sample 3 in more detail. Sample 3 was
Table 5. Dynamic DSC Test Results for Sample 3
a
b
c
β (K/min)
mass (mg)
To (°C)
Q (J/g)
2
4
8
3.52
3.54
3.50
3.55
65.5
75.6
83.2
90.6
−176.0
−206.3
−186.8
−198.5
10
a
b
Heating rate. Onset temperature (i.e., the temperature at which the
c
exothermic peak deviates from the baseline). Decomposition heat.
853
https://dx.doi.org/10.1021/acs.oprd.0c00529
Org. Process Res. Dev. 2021, 25, 849−857

Organic Process Research & Development
pubs.acs.org/OPRD
Article
decomposition heat of sample 3 was −191.9 J/g. To conduct
kinetic analysis of the decomposition process of sample 3 in
Figure 13, the Friedman method was employed.19 The
Friedman method is a model-free method whose basic
equation is shown in eq 1:
The comparison between the experimental and simulated
reaction progress profiles is shown in Figure 14. The
Eα
idα y
j
j
j
z
z
z
ln
= ln[f(α)Aα] −
dt
RT
k
{
(1)
α
α
where dα/dt is the reaction rate (in s−1), α is the conversion,
Aα is the pre-exponential factor at reaction conversion α (in
s−1), f(α) is the reaction model,; Eα is the activation energy (in
kJ/mol), R is the ideal gas constant (8.314 J mol−1 K−1), and T
is the temperature (in K).
In the adiabatic system, the heat released by the sample is
completely used for the temperature rise of the sample itself,
that is,
Figure 14. Comparison between the experimental and simulated
reaction progress profiles for sample 3.
dT
dt
dα
dt
′
cpm
= Q
(2)
correlation coefficient R in Figure 14 is higher than 0.99,
which proves the validation of the simulated results. On the
basis of the above kinetic analysis results, we then predicted
the trend of TMRad,20−22 as shown in Figure 15. From Figure
15, one can directly read that TD24, which is the temperature
corresponding to TMRad for 24 h, is equal to 31.1 °C.
where m is the sample mass and c′p is the specific heat capacity
of the sample. Because
Q
mcp
′
ΔTad
=
′
(3)
where ΔTa′d is the adiabatic temperature rise of the material
when the thermal inertia factor (φ) is equal to 1, it follows that
for φ = 1,
dT
dt
dα
′
= ΔT
ad dt
(4)
For φ ≠ 1, eq 4 can be expressed as
dT
dt
1
φ
dα
′
=
ΔT
ad dt
(5)
Equation 1 for the reaction rate can be rearranged to
i
y
Eα
dα
dt
j
j
z
z
z
= f(α)A exp −
j
α
j
j
z
z
RT
(6)
α
k
{
Figure 15. TMRad prediction curve for sample 3.
By the use of eqs 4 and 5, eq 6 can be transformed into
i
y
Eα
dT
dt
j
j
z
z
z
′
3.3.2. Stage 2. Dynamic DSC experiments were also
performed on three materials involved in stage 2: the final
reaction mixture (sample 4), DAPO, and compound 4. The
results are shown in Figure 16 and summarized in Table 6. The
heat flow curve for sample 4 has two separate exothermic
peaks. The first exothermic peak starts at 121.2 °C, and the
decomposition heat for this peak was obtained as −203.8 J/g.
From Table 6, one can find that compound 4, as a side product
in the DAPO synthesis process, presents a more risky nature
than the target product DAPO. The exothermic peak of
compound 4 in Figure 16 appears in the same temperature
range as the first peak of sample 4, indicating that the first
exothermic peak of sample 4 can be ascribed to the
decomposition of compound 4 (the presence of compound 4
was confirmed by HPLC analysis). The second exothermic
peak of sample 4 corresponds to the decomposition of the
target compound DAPO.
= ΔT f(α)A exp −
j
ad
α
j
j
z
z
RT
(7)
(8)
α
k
{
i
y
z
z
z
Eα
dT
dt
1
φ
j
j
′
=
ΔT f(α)A exp −
j
ad
α
j
j
z
z
RT
α
k
{
respectively. The time under adiabatic conditions can be
obtained using the following equation:
tα
tα
dα
t =
dt =
∫
∫
α
Eα
0
0
f(α)Aα exp −
(
)
RT
(9)
α
for α ∈ [0, 1]. Since the Friedman method gives the
relationship between ln[f(α)Aα] and Eα at conversion α, the
relationship between the temperature rise rate at a certain
conversion rate can be obtained according to eq 7 or eq 8.
Combining this with eq 9, one can obtain the temperature rise
rate versus time curve at any starting temperature, from which
the relationship between the time to maximum rate under
adiabatic conditions (TMRad) and the starting temperature T
can be obtained.
Since sample 4 presented the lowest exotherm onset
temperature in Table 6, we conducted more dynamic DSC
tests on sample 4. The results are shown in Figure 17 and
summarized in Table 7.
854
https://dx.doi.org/10.1021/acs.oprd.0c00529
Org. Process Res. Dev. 2021, 25, 849−857

Organic Process Research & Development
pubs.acs.org/OPRD
Article
TMRad prediction curve for sample 4 is shown in Figure 19
and gives a TD24 of 69.4 °C.
Figure 16. DSC test curves for the related materials in stage 2 at 10
K/min.
Figure 18. Comparison between the experimental and simulated
reaction progress profiles for sample 4.
Table 6. DSC Test Results for the Related Materials in Stage
2
a
b
sample
mass (mg)
To (°C)
Q (J/g)
sample 4
DAPO
compound 4
3.82
0.72
0.88
121.2
248.2
121.6
−649.28
−1353.3
−3565.9
a
Onset temperature (i.e., the temperature at which the exothermic
b
peak deviates from the baseline). Decomposition heat.
Figure 19. TMRad prediction curve for sample 4.
3.4. Thermal Risk Analysis. In order to better understand
the consequences of insufficient cooling capacity, it is
necessary to calculate the maximum temperature that the
reaction system can reach under adiabatic conditions (MTSR)
by calculating the heat accumulated after the target reaction
runs away during the dosing process.23−25 The acquisition of
the MTSR is related to the maximum temperature that can be
reached at any time if the reaction runs away (Tcf):26
Figure 17. DSC test results for sample 4 at different heating rates.
Table 7. Dynamic DSC Test Results for Sample 4
Mrf
Mr,max
MTSR = max(T ) = T + ΔT Xac,max
cf
r
ad
a
b
c
β (K/min)
mass (mg)
To (°C)
Q′ (J/g)
(10)
2
4
8
3.82
3.79
3.78
3.82
101.4
110.8
119.6
121.2
−205.5
−205.2
−196.4
−203.8
where Tr is the temperature of the reaction system before
cooling failure (which can be considered as the process
temperature), ΔTad is the adiabatic temperature rise caused by
heat accumulation, Xac,max is the maximum thermal accumu-
lation of the unreacted reactant, Mrf is the mass of the reaction
mixture at the end of dosing, and Mr,max is the mass of the
reaction mixture at the maximum accumulation.
10
a
b
Heating rate. Onset temperature (i.e., the temperature at which the
exothermic peak deviates from the baseline). Decomposition heat for
the first exothermic peak of sample 4.
c
3.4.1. Stage 1. From the RC1e result for run 1, the heat
generated in stage 1 was obtained as 14.05 kJ. The adiabatic
temperature rise ΔTad at this stage can be calculated as follows:
The two exothermic peaks in Figure 17 are separated
completely. One can reasonably expect that when a thermal
runaway event occurs, the first exothermic peak will be
triggered first. Hence, to calculate the safety parameters of
TMRad, the kinetic calculation focused only on the first
exothermic peak. The value of the correlation coefficient R in
Figure 18 illustrates the validity of the fitting results. The
Q tot
ΔT =
ad
mcp
(11)
where m is the mass of the entire reaction mixture added to the
reactor and cp is the specific heat capacity of the reaction
mixture. For stage 1,
855
https://dx.doi.org/10.1021/acs.oprd.0c00529
Org. Process Res. Dev. 2021, 25, 849−857

Organic Process Research & Development
pubs.acs.org/OPRD
Article
14.05 × 103
155.6 × 2.44
and could surpass the TD24 of 69.4 °C, potentially leading to
higher safety risks.
ΔT =
= 37 K
ad
(12)
Furthermore, by integration of the heat flow profile for run 2,
the heat produced before the addition of half of the
triethylamine was found to be 11.2 kJ. Thus, the heat
accumulated in the reaction mixture for stage 1 was
4. CONCLUSION
A thermal risk assessment of DAPO synthesis based on a study
of the reaction mechanism has been carried out. The following
conclusions were obtained:
Q ac = 14.05 − 11.2 = 2.85 kJ
(13)
(1) In the first dosing stage, only the condensation of free
hydroxylamine and compound 1 occurs; in the second
dosing stage, the subsequent reactions such as removal
of the −NO group and cyclization occur. If the dosing
stopped in the first stage, the reactions involved in the
second stage hardly occur.
The thermal accumulation for run 3 was:
2.85
14.05
Xac,max
=
= 20.3%
(14)
Consequently, using eq 10 we can calculate the MTSR value
for stage 1, which is equal to 12.5 °C.
The above analysis shows that the MTSR (12.5 °C) is less
than the TD24 (31.1 °C), which means that in the case of
cooling failure, immediately stopping the addition of triethyl-
amine will not cause secondary decomposition of the reaction
mixture.
3.4.2. Stage 2. For stage 2, the reaction temperature is
increased from 5 to 25 °C. This process is non-isothermal, and
the MTSR is calculated as follows:
At the end of the triethylamine addition, the heat
accumulation of the material is the largest, and by integration
of the heat flow profile of run 2, the heat conversion rate at this
time is about 55.8%. Thus, the MTSR at this time point is
calculated to be
(2) The main intermediate product in the first dosing stage
is compound 2, and the main side product in the DAPO
synthesis process has been proven to be compound 4.
(3) Dynamic DSC tests revealed that both intermediate 2
and side product 4 are more thermally risky than the
target compound DAPO.
(4) Thermal risk analysis showed that a thermal runaway
incident can be avoided as long as the dosing is stopped
immediately after the cooling failure event occurs.
ASSOCIATED CONTENT
■
sı
* Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acs.oprd.0c00529.
Q
mcp
MTSR5°C = T +
Xac
p
Mass spectrum profiles for 2, calorimetry test result to
measure the heat of mixing for trimethylamine and
methanol, profiles of E and ln[f(α)Aα] versus α for
samples 3 and 4, and HPLC profiles of side product 4
and the reaction filtrate obtained in run 2 (the product
liquid after DAPO was filtered out) (PDF)
34 × 103
168.7 × 2.44
= 5 +
× (1 − 0.558)
(15)
= 41.5 °C
When the temperature increased to Tp = 25 °C, about 70.3%
of the heat was released. Thus, the MTSR at 25 °C can be
obtained as
AUTHOR INFORMATION
■
Q
mcp
Corresponding Author
MTSR25°C = T +
Xac
p
Zichao Guo − Department of Safety Engineering, School of
Chemical Engineering, Nanjing University of Science and
Technology, Nanjing, Jiangsu 210094, China; orcid.org/
0000-0002-2288-1174; Phone: +86 025 84315526;
Email: gzc319@njust.edu.cn; Fax: +86 025 84315526
34 × 103
168.7 × 2.44
= 25 +
× (1 − 0.703)
(16)
= 49.5 °C
Authors
On the basis of the above calculations, the MTSR of stage 2
is determined to be 49.5 °C. The MTSR (49.5 °C) is lower
than TD24 (69.4 °C), which means that in the case of a cooling
failure, immediately stopping the addition of triethylamine will
not cause the secondary decomposition. However, if all of the
triethylamine is accidentally dosed to the reactor in a short
time, the MTSR can be approximated as the sum of Tp and
ΔTad, given by
Yuehua Li − Department of Safety Engineering, School of
Chemical Engineering, Nanjing University of Science and
Technology, Nanjing, Jiangsu 210094, China
Yang Wang − Department of Safety Engineering, School of
Chemical Engineering, Nanjing University of Science and
Technology, Nanjing, Jiangsu 210094, China
Longyu Liao − Institute of Chemical Materials, China
Academy of Engineering Physics, Mianyang 621999, China
Liping Chen − Department of Safety Engineering, School of
Chemical Engineering, Nanjing University of Science and
Technology, Nanjing, Jiangsu 210094, China
Wenqian Wu − Department of Safety Engineering, School of
Chemical Engineering, Nanjing University of Science and
Technology, Nanjing, Jiangsu 210094, China
34 × 103
168.7 × 2.44
MTSR = T + ΔT = 5 +
= 87.6 °C
p
ad
In this case, the MTSR is higher than TD24 (69.4 °C),
indicating that decomposition of the reaction mixture will be
triggered. Therefore, from the safety point of view, the dosing
rate of triethylamine must be effectively controlled to avoid
accumulation of reactants. Otherwise, the practical value of the
MTSR could be higher than the MTSR obtained in this work
Complete contact information is available at:
https://pubs.acs.org/10.1021/acs.oprd.0c00529
856
https://dx.doi.org/10.1021/acs.oprd.0c00529
Org. Process Res. Dev. 2021, 25, 849−857

Organic Process Research & Development
pubs.acs.org/OPRD
Article
(cyanomethyl)amine and Alkoxides. J. Org. Chem. 1979, 44, 1128−
1130.
(19) Friedman, H. L. Kinetics of Thermal Degradation of Char-
forming Plastics from Thermogravimetry. J. Polym. Sci., Part C: Polym.
Symp. 1964, 6, 183−195.
(20) Wu, D. J.; Qian, X. M. Experimental Study on the Thermal
Runaway of Hydrogen Peroxide with In-/Organic Impurities by a
Batch-Reactor. J. Loss Prev. Process Ind. 2018, 51, 200−207.
(21) Wang, Z.; Richter, S. M.; Gates, B. D.; Grieme, T. A. Safety
Concerns in a Pharmaceutical Manufacturing Process Using Dimethyl
Sulfoxide (DMSO) as a Solvent. Org. Process Res. Dev. 2012, 16,
1994−2000.
(22) Vernieres-Hassimi, L.; Dakkoune, A.; Abdelouahed, L.; Estel,
L.; Leveneur, S. Zero-Order versus Intrinsic Kinetics for the
Determination of the Time to Maximum Rate under Adiabatic
Conditions (TMRad): Application to the Decomposition of Hydrogen
Peroxide. Ind. Eng. Chem. Res. 2017, 56, 13040−13049.
(23) Guo, Z. C.; Li, S. R.; Zhou, P.; Chen, L. P.; Chen, W. H.
Insights into Maximum Temperature of Synthesis Reactions in
Isothermal Homogeneous Semi-Batch Reactors. Thermochim. Acta
2018, 668, 103−109.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
1
The authors thank Dr. Shi Mingyang for the H NMR tests
and Dr. Li Xin for his help in formation heat calculation. This
work was financially supported by the National Key R&D
Program of China (2017YFC0804701-4) and the Opening
Project of the Key Laboratory of Special Energy Materials
(Nanjing University of Science and Technology), Ministry of
Education, China.
REFERENCES
■
(1) Pagoria, P. F. Synthesis, Scale-Up, and Characterization of 2,6-
diamino-3,5-dinitropyrazine-1-oxide (LLM-105). Propellants, Explos.,
Pyrotech. 1998, 23, 156−160.
(2) Cutting, J. L.; Chau, H. H.; Peng, H. P. A Small-Scale Screening
Test for HE Performance: Application to the New Explosive LLM-
105. Pyrotech. Intell. 2001, 2, 45−53.
(24) Shah, S.; Fischer, U.; Hungerbuhler, K. Assessment of Chemical
̈
Process Hazards in Early Design Stages. J. Loss Prev. Process Ind. 2005,
18, 335−352.
(25) Zhang, L.; Yu, W. D.; Pan, X. H.; Fang, J. J.; Hua, M.; Chen, F.
M.; Jiang, J. C. Thermal Hazard Assessment for Synthesis of 3-
Methylpyridine-N-Oxide. J. Loss Prev. Process Ind. 2015, 35, 316−320.
(26) Stoessel, F. Thermal Safety of Chemical Processes: Risk Assessment
and Process Design; Wiley-VCH: Weinheim, Germany, 2008.
(3) Pagoria, P. F. Synthesis of high explosive. UCID 220622297,
1997.
(4) Alós, M. A.; Strozzi, F.; Zaldívar, J. M. A New Method for
Assessing the Thermal Stability of Semi-Batch Processes Based on
Lyapunov Exponents. Chem. Eng. Sci. 1996, 51, 3089−3094.
(5) Zaldívar, J. M.; Molga, E.; Alós, M. A.; Hernández, H.;
Westerterp, K. R. Aromatic Nitrations by Mixed Acid. Fast Liquid-
Liquid Reaction Regime. Chem. Eng. Process. 1996, 35, 91−105.
(6) Zaldívar, J. M.; Barcons, C.; Hernandez, H.; Molga, E.; Snee, T.
J. Modelling and Optimization of Semi-Batch Toluene Mononitration
with Mixed Acid from Performance and Safety Viewpoints. Chem.
Eng. Sci. 1992, 47, 2517−2522.
(7) Zaldívar, J. M.; Molga, E.; Alós, M. A.; Hernandez, H.;
Westerterp, K. R. Aromatic Nitrations by Mixed Acid. Slow Liquid-
Liquid Reaction Regime. Chem. Eng. Process. 1995, 34, 543−559.
(8) D’angelo, F. A.; Brunet, L.; Cognet, P.; Cabassud, M. Modeling
and Constraint Optimization of an Aromatic Nitration in Liquid-
Liquid Medium. Chem. Eng. J. 2003, 91, 75−84.
(9) Bellamy, A. J.; Golding, P. The Study of Some Potential New
Synthetic Routes to LLM-105 (2,6-Diamino-3,5-dinitropyrazine 1-
oxide). Cent. Eur. J. Energy Mater. 2007, 4, 33−57.
(10) Shaw, J. T.; Brotherton, C. E.; Moon, R. W.; Winland, M. D.;
Anderson, M. D.; Kyler, K. S. The Preparation of 2,6-Diaminopyr-
azine, 2,6-Diazidopyrazine and Some of Their Derivatives. J.
Heterocycl. Chem. 1980, 17, 11−16.
(11) Ritter, H.; Licht, H. H. Synthesis and Reactions of Dinitrated
Amino and Diaminopyridines. J. Heterocycl. Chem. 1995, 32, 585−
590.
(12) Hollins, A.; Merwin, L. H.; Nissan, R. A.; Wilson, W. S.;
Gilardi, R. Aminonitropyridines and their N-Oxides. J. Heterocycl.
Chem. 1996, 33, 895−904.
(13) Yu, X. Y.; Lu, M.; Nie, F. D. Improvement of the Synthesis
Process of 2,6-Diamino-3,5-Dinitropyrazine-1-Oxide. Chin. J. Explos.
Propellants 2012, 35, 10−10.
(14) Li, Y.; Huang, F. C.; Wang, Y. B.; Hu, L. L.; Zhang, M. M.;
Zhou, W. J. Research Progress on Synthesis and Application of 2,6-
Diamino-3,5-dinitropyrazine-1-oxide. Chem. World 2017, 1, 57−64.
(15) Pagoria, P. F.; Zhang, X. M. Synthesis of Pyrazines Including
2,6-Diaminopyrazine-1-oxide (DAPO) and 2,6-Diamino-3,5-dinitro-
pyrazine-1-oxide (LLM-105). WO 2010/123806 A1, 2010.
(16) Zhao, X. F.; Liu, Z. L.; Yao, Q. Z.; Cheng, J.; Dong, Y. Synthesis
of 2,6-Diamino-3,5-dinitropyrazine-1-oxide. Chin. J. Explos. Propellants
2012, 35, 15−17.
(17) Wang, J. M.; Du, Y. Optimization of Synthetic Method of 2,6-
Diamino-pyrazine-1-oxide. Chin. J. Explos. Propellants 2017, 40, 25−
27.
(18) Vohra, S. K.; Harrington, G. W.; Zacharias, D. E.; Swern, D.
New Synthesis of 2-Amino-6-Alkoxypyrazines from N-Nitrosobis-
857
https://dx.doi.org/10.1021/acs.oprd.0c00529
Org. Process Res. Dev. 2021, 25, 849−857