Catalytic decomposition of ammonium nitrate in superheated aqueous solutions

Article abstract of DOI:10.1021/ja971618k

The decomposition of aqueous ammonium nitrate at elevated temperatures and pressures is examined as function of chloride, nitrate, and total acidity. Catalysis requiring both chloride and acid was observed in solutions containing 20% (w/w) NH₄N

Full text of DOI:10.1021/ja971618k

9738
J. Am. Chem. Soc. 1997, 119, 9738-9744
Catalytic Decomposition of Ammonium Nitrate in Superheated
Aqueous Solutions
Joseph H. MacNeil, Hai-Tao Zhang, Polly Berseth, and William C. Trogler*
Contribution from the Department of Chemistry and Biochemistry, UniVersity of California at San
Diego, La Jolla, California 92093-0358
X
ReceiVed May 19, 1997
Abstract: The decomposition of aqueous ammonium nitrate at elevated temperatures and pressures is examined as
a function of chloride, nitrate, and total acidity. Catalysis requiring both chloride and acid was observed in solutions
containing 20% (w/w) NH4NO3 at 180 °C. Nitrous oxide and dinitrogen were generated in a 4:1 ratio below 0.2 M
+
H . Dinitrogen formation correlated with the production of additional acidity by the reaction 5NH4NO3 f 4N2 +
+
2
HNO3 + 9H2O. The second-order dependence of the decomposition reaction on [H ] is consistent with the reversible
+
18
formation of NO2 . Incorporation of O into the N2O product, as well as the inverse deuterium isotope effect,
supports this conclusion. A novel mechanism based on the intermediacy of NO2Cl is proposed for the chloride
catalysis and contrasted to the radical-based pathways operational in molten NH4NO3 decompositions. Isotope-
labeling experiments using NH4NO3 lead to the formation of NdNdO-labeled nitrous oxide and the dinitrogen
products Nt N and Nt N in a 1:3 ratio. Decomposition of NH4 NO3 produces only Nd NdO and Nt N.
15
15
1
5
15
15
15
15
15
+
+
+
-
+
This agrees with the reaction sequences: NO2Cl + NH4 f {O2N-NH3} + H + Cl , {O2N-NH3} f N2O +
+
-
+
+
-
+
-
+
H3O and 3NO2Cl + 3NH3 f 3NH2Cl + 3NO2 + 3H , 3NH2Cl f N2 + NH4 + 3Cl + 2H , 3NO2 + 3NH4
f 3N2 + 6H2O. These results bear on the industrial preparation of NH4NO3 and suggest conditions under which
nitrous oxide emissions might be important to the global N2O budget.
Introduction
source of nitrous oxide that accounted for 10-15% of the annual
increase in N2O levels and about 45% of the excess observed
The environmental implications of increasing atmospheric
nitrous oxide (N2O) are well recognized.1 Atmospheric N2O
is the predominant source of stratospheric NO, which catalyzes
11
in the northern hemisphere. This realization led to a voluntary
12
phase-out of the industrial emissions.
As part of an effort to identify various mechanisms of nitrous
2
the O3 + O f 2O2 reaction (a stratospheric ozone sink). In
11,13,14
oxide formation,
we have examined the stability of hot
the lower stratosphere, NO2, formed by the reaction of NO with
ozone, may also sequester ClO as ClONO2.3 This can diminish
the effect of NO and ClO radicals on ozone loss. Nitrous oxide
aqueous solutions of ammonium nitrate. Nitrous oxide emis-
sions from soils treated with ammonium nitrate have been
15
studied in some detail; however, N2O releases during fertilizer
4
has an atmospheric lifetime of 120-150 years, and it has a
greenhouse forcing factor 315 times as great as that of CO2.
manufacture have not been considered. In one widely used
method of NH4NO3 manufacturing, hot aqueous NH3 and HNO3
are combined in a titanium sparger; temperatures of 180 °C may
arise from the exothermicity of this neutralization reaction.
Global capacity for producing ammonium nitrate was 66 billion
5
Accumulation of atmospheric N2O has been estimated at 0.3%
per year over the past several decades, and N2O is currently
1
0-15% above its preindustrial value.6 About 30% of the
7,8
sources in the global nitrous oxide budget are uncertain. Most
16
kg/yr in 1985, about half of which was marketed as a solid.
9
N2O evolution involves terrestrial or oceanic bacterial action;
however, global distribution patterns suggest that there are also
significant anthropogenic sources. For example, industrial
production of adipic acid (a precursor to nylon-66) is a point
Prilling is a drying method which sprays an ammonium nitrate
solution into a countercurrent air stream heated to 180 °C. Water
evaporates as the solution droplets fall. Some nitrogen loss is
1
0
17
known to occur during this step.
X
Molten ammonium nitrate decomposes at temperatures above
200 °C according to eqs 1 and 2. During NH4NO3 synthesis,
low pH is often employed to suppress the second reaction. Pure
ammonium nitrate is known to present an explosion hazard when
it is contaminated with chloride or organics. Previous mecha-
nistic studies of NH4NO3 decomposition have focused on the
Abstract published in AdVance ACS Abstracts, September 15, 1997.
(
(
(
1) Trogler, W. C. J. Chem. Educ. 1995, 72, 973-976.
2) Crutzen, P. J.; Schmailzl, U. Planet Space Sci. 1983, 31, 1009.
3) Stimpfle, R. M.; Koplow, J. P.; Cohen, R. C.; Kohn, D. W.;
Wennberg, P. O.; Judah, D. M.; Toohey, D. W.; Avallone, L. M.; Anderson,
J. G.; Salawitch, R. J.; Woodbridge, E. L.; Webster, C. R.; May, R. D.;
Proffitt, M. H.; Aiken, K.; Margitan, J.; Loewenstein, M.; Podolske, J. R.;
Pfister, L.; Chan, K. R. Geophys. Res. Lett. 1994, 21, 2543-2546.
(
4) Cicerone, R. J. J. Geophys. Res. 1989, 94, 18265.
(11) Thiemens, M. H.; Trogler, W. C. Science 1991, 251, 932-934.
(12) Eur. Chem. News 1991, 41.
(5) Albritton, D.; Derwent, R.; Isaksen, I.; Lal, M.; Wuebbles, D. In
Climate Change 1995: The Science of Climate Change; Houghton, J. T.,
Meira Filho, L. G., Callander, B. A., Harris, N., Kattenberg, A., Maskell,
K., Eds.; Cambridge University Press: Cambridge, U.K., 1995; p 572.
(13) MacNeil, J. H.; Gantzel, P. K.; Trogler, W. C. Inorg. Chim. Acta
1995, 240, 299-304.
(14) MacNeil, J. H.; Berseth, P. A.; Bruner, E. L.; Perkins, T. L.; Wadia,
Y.; Westwood, G.; Trogler, W. C. J. Am. Chem. Soc. 1997, 119, 1668-
1675.
(
6) Lacis, A. A.; Hansen, J.; Lee, P.; Mitchell, T.; Lebedeff, S. Geophys.
Res. Lett. 1981, 8, 1035-1038.
7) Watson, R. T.; Filho, L. G. M.; Varney, S. K. Sources and Sinks;
Cambridge University Press: New York, 1992.
8) Bouwman, A. F.; Fung, I.; Matthews, E.; John, J. Global Biogeochem.
Cycles 1993, 7, 557-597.
9) Delwiche, C. C. Denitrification, Nitrification and Atmospheric Nitrous
Oxide; Wiley & Sons: New York, 1981.
10) Khalil, M. A. K.; Rasmussen, R. A. Tellus 1983, 35B, 161.
(
(15) Matthews, E. Global Biogeochem. Cycles 1994, 8, 411-439.
(16) Weston, C. W. In Encyclopedia of Chemical Technology; Howe-
Grant, M., Ed.; John Wiley & Sons: New York, 1992; Vol. 2, pp 698-
708.
(17) McKetta, J. J.; Cunningham, W. A. Encyclopedia of Chemical
Processing and Design; Marcel Dekker: New York, 1989; Vol. 21, pp 261-
271.
(
(
(
S0002-7863(97)01618-1 CCC: $14.00 © 1997 American Chemical Society

NH4NO3 in Superheated Aqueous Solutions
J. Am. Chem. Soc., Vol. 119, No. 41, 1997 9739
NH NO (l) f N O(g) + 2H O(g)
(1)
(2)
4
3
2
2
NH NO (l) f NH (g) + HNO (g)
4
3
3
3
solid and its anhydrous melt, while comparatively little is known
of ammonium nitrate’s reactivity in solution. The tragic
1
8
explosion at the Terra Industries ammonium nitrate plant in
Port Neal, Iowa, underscored the fact that superheated, aqueous
NH4NO3 can also be violently reactive. When fundamental
information is obtained about the solution mechanisms underly-
ing one of the largest scale industrial processes, its potential
contribution to the nitrous oxide budget may be evaluated. In
addition, a study of solution decomposition reactions may
provide new insight into the explosion hazards present during
the manufacture and use of this important fertilizer.
Figure 1. Synergistic catalysis of 20% (w/w) of ammonium nitrate in
water. Solutions were sealed within a Ti reactor, purged with agron,
and heated to 180 °C. Time zero for each reaction was defined as the
Experimental Section
+
time when the target temperature was first reached. H was added as
All reactions were performed within a 300 mL Parr titanium pressure
reactor, which was purged and pressurized to 50 psi with argon. Gas
samples were extracted from the headspace periodically and analyzed
on a Hewlett Packard 5890 gas chromatograph with a thermal
conductivity detector and a 3392A integrator. Calibration curves were
used to determine the quantities of nitrogen, nitrous oxide, and argon
present in a 50 µL injection, from which the composition of the bulk
headspace was calculated. The volume and mass of solution remaining
at the end of the reaction were recorded, and aliquots were titrated
with a standard NaOH solution to determine total acidity. All solutions
-
HNO
3
, and Cl was introduced as NH
4
Cl.
a solution of 0.20 M DCl/D
NH NO (1.27 g) was also decomposed in a 0.20 M solution (5.60 g)
of HCl prepared with H
the degree of 18O exchange as determined from the N
O/N O product
2
2
O to investigate deuterium isotope effects.
4
3
18
2
O. GC/MS analysis was used to calculate
16
18
2
ratio.
Results and Discussion
were prepared with HPLC-grade H
2
O. Ammonium nitrate, ammonium
1
5
Decomposition of 20% (w/w) Solutions of Ammonium
Nitrate at 180 °C. Solutions containing 20% by weight
certified ACS-grade ammonium nitrate in HPLC purity water
were sealed into a 300 mL titanium pressure reactor. No
detectable decomposition occurred after heating at 180 °C for
chloride, nitric acid, hydrochloric acid, 12 M DCl, NH
4
NO
O (70% O) were all used as
was prepared by triple recrystallization of NH
2
O (99.9%).
3
(99%
1
5
15
15
18
18
N), NH
received. ND
NO from D
4
NO
3 2
(99% N), and H
4
NO
3
4
-
3
Reactions of 20% (w/w) Aqueous Ammonium Nitrate. Solutions
of ammonium nitrate (18.0 g) in water (75 g) were sealed within the
titanium reactor and heated to 180 °C. The decomposition of the
ammonium nitrate solution was monitored both by the total headspace
pressure and the changing headspace gas composition. The influences
of free chloride and total acidity on the rate and course of the
decomposition were evaluated independently. Ammonium chloride was
added to ammonium nitrate solutions prepared in 0.191 M nitric acid
to vary the chloride concentration between 0.000 and 0.298 M.
Conversely, the ammonium chloride concentration was kept fixed at
+
-
1
0 h. Neither 0.20 M H (as HNO3) nor 0.20 M Cl (as NH4-
Cl) catalyzed decomposition individually; however, when they
were added together catalysis was observed. Figure 1 depicts
the headspace pressure profile for these experiments. A plateau
above 600 psi was coincident with complete depletion of
ammonium nitrate. The gaseous products evolved consisted of
78% N2O and 22% N2.
+
Influence of [H ]. Reactions were performed in which the
0
.20 M as the acid concentration was adjusted from 0.000 to 0.512 M.
chloride concentration was maintained at 0.20 M while the
acidity was varied. In these experiments, the amount of N2
produced was proportional to the increase in solution acidity; a
The nitrate dependence was established by the addition of NaNO
3
.
Ammonium concentration studies were not conducted due to the
difficulty of finding a suitable anion, which neither buffered the solution
pH nor was chemically incompatible under the reaction conditions.
+
ratio approaching 2:1 (N2/produced H ) was observed consis-
-
+
tently (Table 1). Decomposition rates were sensitive to the
Ammonium nitrate solutions containing 0.20 M Cl and 0.20 M H
+
+
were also examined at temperatures between 170 and 200 °C to
determine the effects of temperature on the decomposition profile.
Reactions at High Ammonium Nitrate Concentrations. Am-
monium nitrate (18 g) solutions ranging from 15% (w/w) to 50% (w/
initial acid concentration below 0.2 M H , where product H
acceleration dominated the reaction profile (Figure 2). Since
nitrogen generation was quantified throughout each reaction,
the acidity of the solution at any time could be calculated by
-
+
w) in water were prepared with 0.20 M Cl and 0.20 M H . Above
0% (w/w), the chloride loading had to be substantially lowered, and
+
using the 2:1 N2/H ratio to estimate how much additional acid
5
had been created. Each of the decomposition profiles of Figure
no initial acid was added. Warning: ammonium nitrate, either dry or
as a concentrated aqueous solution, presents a significant explosion
hazard when contaminated by chloride and/or organics.
2
were broken down into a series of smaller segments, such
that the headspace pressure increase rate was linear over each
segment. The total solution acidity was determined at the first
and last point for each segment and averaged. Rate data from
each of the experiments agree within experimental error and
show that the reproducible apparent induction times seen at low
initial acid concentrations reflect rate acceleration by the
Isotope-Labeling Experiments. Ammonium nitrate solutions (20%
1
5
15
15
4 3 4 3
w/w) labeled with N (either as NH NO or as NH NO ) and
-
+
containing 0.20 M Cl and 0.20 M H were decomposed at 180 °C.
After the reactor was allowed to cool, the nitrous oxide and dinitrogen
were analyzed on a Hewlett Packard gas chromatograph/mass spec-
trometer (GC/MS) to determine the nature and the extent of 15
N
+
generation of additional H during the reaction. Above 0.2 M
incorporation into both products. Water and CO
with a dry ice/acetone bath and NaOH, respectively. As N
spectral fragmentation product of N O, the pretreated gases were then
separated by freezing out the N O at 77 K and Toepler pumping the
on a vacuum line. ND NO (18 g) was decomposed at 180 °C in
2
were first removed
+
H , the decomposition rates are linear throughout approximately
2
is a mass
80-90% of the reaction. This originates in the offsetting effects
2
of rate acceleration (from additional acid generation) and rate
inhibition (as the ammonium nitrate is consumed). Rate data
derived from the initial linear portion of five experiments with
acid concentrations at or above 0.2 M are plotted in Figure 3.
2
N
2
4
3
(18) Kirschner, E. Chem. Eng. News 1994, Dec. 19 (72), 4-5.

9740 J. Am. Chem. Soc., Vol. 119, No. 41, 1997
MacNeil
Table 1. Product Distributions from the Complete Decomposition
of 20% Ammonium Nitrate Solutions
+
a
[
H ]
(init (final
M) M) produced produced N2:H produced N2O:N2
moles of moles of ratio moles of
-
a
+
[
Cl ]
H
N2
of
N2O
ratio of
+
(
M)
0
0
0
0
0
0
0
0
0
0
0
.200 0.0300 0.363 0.027
.200 0.0616 0.383 0.026
.200 0.0900 0.377 0.025
.200 0.124 0.407 0.024
.200 0.166 0.425 0.022
.199 0.191
0.056
0.046
0.045
0.042
0.042
0.040
0.043
0.039
0.033
0.031
0.031
2.1:1
1.8:1
1.8:1
1.8:1
1.9:1
0.146
0.138
0.137
0.141
0.153
0.142
0.158
0.162
0.165
0.16
2.6:1
3.0:1
3.0:1
3.4:1
3.6:1
3.6:1
3.7:1
4.2:1
5.1:1
5.1:1
5.7:1
.206 0.192 0.446 0.021
.206 0.282 0.498 0.020
.206 0.410 0.572 0.017
.205 0.40
2.0:1
2.0:1
1.9:1
.207 0.512 0.637 0.017
1.8:1
0.174
0
0
0
0
0
0
0
.0502 0.191 0.431 0.021
.0773 0.191 0.430 0.020
.100 0.191 0.436 0.021
.124 0.191 0.434 0.021
.150 0.191 0.439 0.022
.199 0.191
Figure 3. Plot of the second-order rate dependence on the acid
concentration, as determined by the method of initial rates. Initial rate
data for five experiments are plotted against the square of the initial
acid concentration. All reactions were performed in 0.20 M chloride
at 180 °C.
0.032
0.038
0.037
0.040
0.040
0.045
1.6:1
1.8:1
1.8:1
1.8:1
0.113
0.129
0.129
0.144
0.142
0.166
3.5:1
3.4:1
3.5:1
3.6:1
3.6:1
3.7:1
.298 0.191 0.457 0.023
2.0:1
a
Chloride and acid concentrations were determined at room tem-
perature and atmospheric pressure.
Figure 4. Plotting the linear correlation between the overall rate of
decomposition and the chloride concentration used. Reactions were
+
carried out at 180 °C in the presence of 0.191 M H .
Figure 2. Reaction profiles for 20% ammonium nitrate solutions in
0
.20 M chloride at 180 °C, plotted as a function of time for various
+
+
initial acid concentrations: (0) 0.030 M H , (b) 0.062 M H , (4)
0
+
+
+
+
.124 M H , (9) 0.192 M H , (O) 0.282 M H , and (1) 0.512 M H .
The leveling of the curves at extended times represents complete
decomposition.
This shows a second-order rate dependence on initial acid
concentration.
-
Influence of [Cl ]. When the initial acid concentration was
held constant at 0.191 M and the chloride concentration was
+
varied, both the N2O:N2 ratio and the total moles of H
generated remained invariant (Table 1). Plotting the rate of
increase in total headspace pressure versus the chloride con-
centration (Figure 4) shows an approximate first-order depen-
dence. Gas chromatography data were used to quantify first-
order relationships for the rates of production of both N2 and
N2O in 0.191 M HNO3 as a function of the chloride concentra-
tion, from which the following empirical rate equations were
derived at 180 °C:
Figure 5. Reaction profiles for experiments with elevated nitrate levels,
in 0.20 M chloride and 0.20 M acid, at 180 °C: (9) 0.00 g NaNO
3
(0.225 mol total nitrate), ([) 3.05 g of NaNO
3
(0.261 mol total of
nitrate), (2) 6.05 g of NaNO
3
(0.297 mol total of nitrate), (b) 12.05 g
3
of NaNO (0.367 mol total of nitrate).
d(N )/dt ) k [Cl-],
2
obsd
maintained at 0.20 M while nitrate (as NaNO3) was added
(Figure 5). The additional NO3 modestly enhanced the
ammonium nitrate decomposition rates; an empirical data fit
suggested an approximate x3 dependence. Note that at higher
nitrate concentrations evidence for rate acceleration from acid
autocatalysis can be observed early in the reaction profile.
kobsd ) 3.2((0.4) × 10 mol of N s M-1
-
5
-1
-
2
d(N O)/dt ) k [Cl-],
2
obsd
kobsd ) 1.0((0.2) × 10 mol of N O s M-1
-
4
-1
2
Influence of Additional Nitrate. A number of experiments
were performed with both the H and Cl concentrations
Decomposition Mechanisms of Aqueous Ammonium Ni-
trate Solutions. The decomposition of pure and chloride-
+
-

NH4NO3 in Superheated Aqueous Solutions
J. Am. Chem. Soc., Vol. 119, No. 41, 1997 9741
contaminated ammonium nitrate melts has been studied in some
Scheme 1
detail;1
9-21
however, far less is known about decomposition
behavior in aqueous solutions. Ammonium nitrate decomposi-
22
tion is inhibited by small quantities of water added to the melt.
This effect has been attributed to the ionization of nitric acid
+
22
and its subsequent inability to generate NO2 , which is
believed to be a key intermediate in the decomposition
protonated samples. This inverse kinetic isotope effect is
consistent with an equilibrium isotope effect favoring undisso-
pathway.1
9-21
It is incorrect to infer that water can completely
+
block NO2 formation; however, as the nitric acid dissociation
equilibrium shifts toward the undissociated form as temperature
27
ciated nitric acid. The O-D bond is stronger than the O-H
+
bond, and this drives the equilibria of eq 4 toward higher NO2
23
and pressure are increased. Above 150 °C the ionization
constant is less than one.23 Therefore, appreciable amounts of
undissociated HNO3 exist in aqueous solution under the
conditions of our kinetics experiments.
concentrations, thereby accelerating decomposition.
+
-
The synergistic catalytic behavior of H and Cl in dilute
aqueous ammonium nitrate solutions shows that the generation
+
of NO2 alone is insufficient for catalysis. Previous mechanistic
In their classic studies of organic nitration reactions, Ingold
et al. were the first to show conclusively the intermediacy of
H2NO3 during the generation of the nitryl cation.24,25 They
identified two limiting cases for these reactions. When anhy-
drous nitric acid was used as a nitrating agent, the reaction was
best expressed as eq 3.
studies of anhydrous ammonium nitrate melts have shown that
+
NO2 can abstract an electron from chloride to initiate radical
+
28
chain mechanisms. In aqueous solution, however, chloride
+
may stabilize NO2 as NO2Cl. Nitryl chloride has been
postulated recently as a reactive intermediate in the aqueous,
29,30
ionic redox chemistry of other nitrogen species.
The first-
+
order kinetic dependence on chloride at 0.2 M [H ] for both
N2O and N2 formation implicates eq 5 as a common, rate-
determining step in the mechanism.
+
3
-
HNO + HNO h H NO + NO3
h
3
3
2
NO2+ + H O + NO (3)
-
2
3
NO2+ + Cl f NO Cl
-
(5)
2
When stronger acids were used, a significant increase in the
+
rate of nitration was observed, and the formation of NO2 was
attributed to eq 4.
In contrast, experimental rate laws determined for the production
of N2O and N2 from the chlorine-radical-catalyzed decomposi-
tion of NH4NO3 melts express a square-root dependence on
chloride concentrations and yield highly variable N2O/N2
HNO + H h H NO h NO + + H O
+
+
(4)
3
2
3
2
2
1
7
product distributions.
Scheme 1 depicts the formation of the nitryl chloride
The exact KA for nitric acid under the current reaction conditions
23
24
26
can be estimated to lie within the range 1 > KA > 0.1. Under
the dilute aqueous conditions of these experiments, eq 4 is a
better representation of the equilibria responsible for generating
intermediate. Ingold and Bunton have reported the equi-
librium k1/k-1 to be fast for organic nitration steps. They have
also reported that k-2 is more rapid than k2; the aqueous
conditions of the current experiments may shift the second
equilibrium even further to the left. Given that the first two
equilibria are fast, the expression for the rate determining step
+
NO2 . In support of this, Ingold noted that the reaction
2
4
identified in eq 3 was inhibited by the addition of nitrate;
however, excess nitrate (as NaNO3) slightly accelerated de-
composition rates in the current study. Equation 4 mirrors the
+
-
is d(products)/dt ) k3[NO2 ][Cl ]. If one assumes steady state
19-21
+
initiation step previously reported for molten NH4NO3 salts.
behavior for [NO2 ], the rate expression becomes:
Further support for the validity of the equilibria proposed in
eq 4 was obtained when NH4NO3 was decomposed in 70%
-
+
-
d(products)/dt ) (k k [Cl ][H NO ])/(k + k [Cl ])
2
3
2
3
-2
3
1
8
16
18
H2 O. Both N2 O and N2 O were generated in a 4:3 ratio.
18
The quantity of N2 O observed represented 75-80% of that
Inverting both sides yields
rate)-1 ) k /(k k [Cl ][H NO ]) + 1/(k [H NO ])
anticipated from total oxygen scrambling. Control reactions
18
18
showed that no O exchange occurred between N2O and H2 O.
The dehydration/rehydration equilibrium of eq 4 must occur
repeatedly to generate the observed level of exchange. This
agrees with previous kinetic studies, which found that this near
pre-equilibrium proceeded more quickly then subsequent nitra-
tion events and causes 18O incorporation into NO3 .
-
+
+
3
(
-
2
2
3
2
3
2
2
+
-
As [HNO3] can be approximated as [H ][NO3 ]/KA(HNO ), this
rate expression is consistent with the second-order acid depen-
dence (Figure 3) and the first-order chloride dependence (Figure
3
- 26
-
1
4
). Figure 6 was constructed by plotting (rate) as a function
It was also observed that ND4NO3 in 0.2 M DCl/D2O
decomposed at rates approximately double those of analogous
-
-1
of [Cl ] . It should be possible to obtain a direct measure of
the ratio k-2/k3 by dividing the slope of Figure 6 by its intercept
value. A small, positive intercept was anticipated, and the
observed small negative value (-0.15 ( 0.18, 95% confidence
level) is not statistically significant. The observed dependency
upon nitrate is lower than predicted; however, it is also important
to recognize that the high salt concentration involved, as well
(
19) Keenan, A. G.; Notz, K.; Franco, N. B. J. Am. Chem. Soc. 1969,
1, 3168-3171.
20) Rosser, W. A.; Inami, S. H.; Wise, H. J. Phys. Chem. 1963, 67,
753-1757.
21) Colvin, C. I.; Keenan, A. G.; Hunt, J. B. J. Chem. Phys. 1963, 38,
033-3035.
22) Rubtsov, Y. I.; Strizhevskii, I. I.; Kazakov, A. I.; Moshkovich, E.
B.; Andrienko, L. P. Zh. Prikl. Khim. 1989, 62, 2417-2422.
9
1
3
(
(
(
(27) Lowry, T. H.; Richardson, K. S. Mechanism and Theory in Organic
Chemistry 3rd ed.; Harper and Row: New York, 1987; pp 241-244.
(28) Keenan, A. G.; Dimitriades, B. J. Chem. Phys. 1962, 37, 1583-
1586.
(
23) Marshall, W. L.; Slusher, R. J. Inorg. Nucl. Chem. 1975, 37, 1191-
1
202.
(
24) Hughes, E. D.; Ingold, C. K.; Reed, R. I. J. Chem. Soc. 1950, 2400-
2
473.
(29) Johnson, D. W.; Margerum, D. W. Inorg. Chem. 1991, 30, 4845-
(
25) Olah, G. A.; Malhotra, R.; Narang, S. C. Nitration: Methods and
Mechanisms; VCH Publishers Inc.: New York, 1989.
26) Bunton, C. A.; Halevl, E. A. J. Chem. Soc. 1952, 4917-4924.
4851.
(30) Margerum, D. W.; Schurter, L. M.; Hobson, J.; Moore, E. E.
EnViron. Sci. Technol. 1994, 28, 331-337.
(

9742 J. Am. Chem. Soc., Vol. 119, No. 41, 1997
MacNeil
Scheme 2
+
+
+
-
NO Cl + NH f {O NsNH } + H + Cl
(6)
(7)
2
4
2
3
+
+
{
O NsNH } f N O + H O
2
3
2
3
31
Nitryl chloride is also known to react with ammonia. Both
3
2
31
the nitrite and the chloramine produced in eq 8 react further
to yield dinitrogen.
-
+
3
NO Cl + 3NH f 3NH Cl + 3NO + 3H
(8)
(9)
2
3
2
2
+
-
+
3
NH Cl f N + NH + 3Cl + 2H
2 2 4
NO2- + 3NH4 f 3N + 6H O
+
(10)
3
2
2
Figure 6. Plotting the inverse of the decomposition rate as a function
of inverse chloride concentration. Rate data were taken from the study
of chloride dependence. The initial acid concentration was held invariant
at 0.191 M.
Nitrous oxide prepared from the decomposition of 15NH4NO3
contained the label exclusively (>98%) on the terminal nitrogen
1
5
15
( NdNdO), while N2O prepared from NH4 NO3 was labeled
15
only at the central nitrogen (Nd NdO). Neither NdNdO nor
1
5
15
Nd NdO were observed in detectable quantities. This is
consistent with the mechanism of eqs 6 and 7. The dinitrogen
1
5
15
15
gas produced from NH4NO3 contained both Nt N and
Nt N in a 1:3 ((5%) ratio, while only Nt N was recovered
from the decomposition of NH4 NO3. Background N2 in the
mass spectrometer prevented reliable quantification of the
unlabeled dinitrogen. The 1:3 ratio of Nt N and Nt N
prepared from NH4NO3 is inconsistent with the radical
mechanism; however, the stoichiometries of eqs 9 and 10 in
the current mechanism satisfy this isotope distribution.
1
5
15
1
5
1
5
15
15
1
5
2
1
Scheme 2 summarizes the reaction pathways proposed for
the decomposition of aqueous ammonium nitrate solutions. Both
acid and chloride are required to generate NO2Cl, which
subsequently reacts along parallel pathways, producing N2O and
N2. The data presented in Table 1 show that the nitrous oxide/
dinitrogen product ratio is directly proportional to the acidity
of the reaction, although no such correlation was found in the
chloride dependency studies. The linear response of the N2O/
N2 ratio to the initial acidity of the reaction (Figure 7) is a direct
consequence of the common intermediacy of NO2Cl. The rates
Figure 7. Plot of the final ratio of nitrous oxide to dinitrogen evolved
during ammonium nitrate decomposition, as a function of initial acid
concentration. Reactions were carried out at 180 °C in the presence of
-
0
.20 M Cl .
as the reduced dielectric constant of H2O at 180 °C, will
introduce activity coefficient corrections of unknown magnitude.
In spite of these limitations, the experimental data fit the kinetic
model reasonably well.
of formation of N2O and N2 can be expressed as d(N2O)/dt )
+
kN O[NO2Cl][NH4 ] and d(N2)/dt ) kN [NO2Cl][NH3]. Since
2
2
+
+
+ -1
KA(NH4 ) ) [H ] [NH3] [NH4 ] , the ratio of the rate expres-
sions is as follows:
Several lines of evidence point toward NO2Cl as a common
intermediate in the generation of both N2O and N2. The
(31) Jander, J.; Engelhardt, U. DeVelopments in Inorganic Nitrogen
Chemistry; Elsevier Scientific Publishing Company: New York, 1973; Vol.
+
+
Coulombic barrier inhibiting direct attack of NO2 on NH4 is
avoided by initial formation of nitryl chloride, which can
generate nitrous oxide through the mechanism of eqs 6 and 7.
2
, p 228.
32) Greenwood, N. N.; Earnshaw, A. Chemistry of the Elements, 1st
ed.; Pergamon Press: New York, 1984; pp 1542.
(

NH4NO3 in Superheated Aqueous Solutions
J. Am. Chem. Soc., Vol. 119, No. 41, 1997 9743
+
k
[NO Cl][NH ]
d(N O)/dt
N O
2
4
2
2
)
)
d(N )/dt
k [NO Cl] [NH ]
N 2 3
2
2
+
k
[NO Cl][NH ]
N O
2
4
2
+
+ -1
k [NO Cl]K
A(NH4+) 4
[NH ][H ]
N
2
2
This simplifies to yield the expression
k
d(N O)/dt
N O
2
2
+
)
[H ]
d(N )/dt
N2 A(NH4+
k K
2
)
The excellent linear correlation depicted in Figure 7 provides
strong evidence for this mechanistic proposal. Further experi-
mental support for the intermediacy of free NH3 in the N2
evolution pathway was found in the product distribution of the
ND4NO3 decomposition experiments outlined previously. The
greater strength of N-D bonds would be anticipated to favor
ND4 . Therefore, free ND3 concentrations in solution would
be lowered and N2 formation would be inhibited. The N2O/N2
ratio from ND4NO3 decomposition rose to over 4.0:1 as
compared to the 3.5:1 ratio observed in analogous NH4NO3
decompositions.
Figure 8. The effect of increased ammonium nitrate loading on the
rate of decomposition. All reactions were performed at 180 °C with
+
-
0
.20 M H and 0.20 M Cl .
+
27
The chloride-catalyzed decomposition of ammonium nitrate
at 100 °C in strongly acidic solutions (1-8 M HNO3) has been
3
3
examined. This was proposed to occur through a mechanism
based on the initial formation of HOCl. An empirical fit of
the kinetics data showed a third-order dependence on nitrate
concentration, in contrast to the much weaker dependence
observed under the current conditions. It is significant to note
that at 100 °C the generation of NO2+ via eq 4 would be a less
viable reaction pathway, because of the large dissociation
constant for nitric acid at the lower temperature.
Figure 9. Reaction profiles for the decomposition of 70% ammonium
nitrate solutions contaminated with different levels of chloride (as NH -
4
Cl) at 180 °C.
Influence of Increased Ammonium Nitrate Loading. It
is interesting to consider the reaction conditions involved in
fertilizer manufacturing, where the ammonium nitrate loading
of superheated aqueous solutions may be 70-95%. A series
of reactions were performed at elevated ammonium nitrate
concentrations. Figure 8 depicts the increase in reaction rate
as the ammonium nitrate loading is increased from 20 to 50%
by weight. Above 50% (w/w) NH4NO3, decomposition is
exquisitely sensitive to chloride. At 60% NH4NO3, solutions
heated to 180 °C were decomposed by the presence of chloride
even in the absence of added acid. This may reflect the
dehydrating nature of concentrated salt solutions, which would
+
favor NO2 formation in eq 4 as well as product catalysis by
+
H . Figure 9 depicts the reaction profiles for three experiments
performed on 70% ammonium nitrate at 180 °C. The decom-
position was very rapid at the highest chloride loading used,
and more severe conditions were not tested to alleviate the risk
of accidental explosion.
Figure 10. Reaction profile as a function of temperature; 70%
ammonium nitrate solutions initially free of acid were contaminated
with 900 ppm NH
00 °C.
4
Cl and heated to temperatures ranging from 170 to
2
The rate of ammonium nitrate decomposition was also studied
as a function of temperature. Figure 10 shows the data obtained
from decomposition experiments conducted on 70% ammonium
nitrate solutions contaminated with 900 ppm chloride. No initial
acidity was introduced. At temperatures below 170 °C, the rate
of decomposition was slow. An Eyring plot was used to
estimate the apparent activation energy under these conditions
at 180 kJ mol . Temperatures above 180 °C also increased
the proportion of nitrous oxide evolved. Decomposition of 20%
ammonium nitrate at 220 °C generated a headspace that was
greater than 95% N2O.
Ammonium Nitrate Manufacture as a Potential Source
of Atmospheric Nitrous Oxide. Ammonium nitrate is prepared
by mixing ammonia and nitric acid, resulting in a hot aqueous
solution approximately 60% (w/w) in ammonium nitrate. To
produce solid NH4NO3, the remaining water must be driven off.
One of the two common drying methods is known as prilling;
in this procedure the solution is sprayed into the top of a prilling
tower and encounters a countercurrent of hot air as it falls.
Temperatures in the prilling tower can surpass 180 °C, and a
residence time of 15 s is typical. The data suggest the presence
of elevated chloride concentrations (above approximately 500
ppm) would be sufficient to initiate decomposition of am-
-
1
(33) Kelmers, A. D.; Maya, L.; Browning, D. N.; Davis, W., Jr. J. Inorg.
Nucl. Chem. 1979, 41, 1583-1588.

9744 J. Am. Chem. Soc., Vol. 119, No. 41, 1997
MacNeil
monium nitrate and generate nitrous oxide under these condi-
tions. The dangers of chloride-contaminated ammonium nitrate
solids are well-known within the industry, and Cl standards
bated the problem by redissolving accumulated chloride salts
and mixing them with the dissolved ammonium nitrate. It is
unclear whether seawater was used directly to fight the blaze
-
3
5
have been adopted; the currently accepted chloride limit for
manufacture of ammonium nitrate in Europe is 200 ppm. The
rate measurements show that when chloride-limited operating
conditions are maintained, it is unlikely that ammonium nitrate
decomposition could produce globally significant quantities of
nitrous oxide. Higher levels of chloride contamination will
generate nitrous oxide, and monitoring effluent emissions from
neutralizers and prilling towers may be informative.
that preceded the explosion.
During a scheduled plant shutdown in December 1994,
64 000 lbs of NH4NO3 detonated at a fertilizer plant in Port
1
Neal, Iowa, killing four people and injuring 18 others. An EPA
report concluded that the ammonium nitrate had been manu-
36
factured from nitric acid contaminated with chlorides. Before
the explosion, an aqueous solution of 85-90% ammonium
nitrate, with a pH estimated to be less than one, was sprayed
Relevance to the Commercial Manufacture of NH4NO3.
Ammonium nitrate is a commodity chemical manufactured
primarily as a agricultural fertilizer and for producing explosives.
Although often regarded as safe, incidents still occur during its
production and handling. The largest accidental explosion in
the U.S. occurred in Texas City on April 16, 1947, when a fire
aboard the S. S. Grandcamp caused the detonation of 2280 tons
of solid ammonium nitrate fertilizer. Over 600 people were
killed, and the Associated Press war correspondent likened the
36
with steam above 220 °C at 200 psig for 9.5 h. The aqueous
decomposition reactions outlined in this paper involve similar
species and conditions, and present a plausible mechanism for
the buildup of pressure in the titanium reactor and the ensuing
explosion. Although “dry” ammonium nitrate decomposes more
rapidly than concentrated aqueous solutions at similar levels of
chloride contamination, the current data indicate that at tem-
peratures exceeding 170 °C aqueous decomposition may proceed
rapidly.
34
destruction to the atomic-bombing of Nagasaki. The exact
cause of the Texas City explosion was never determined. Some
fertilizer bags were known to have ruptured, spilling granulated
ammonium nitrate onto the floor of the hold, where seawater
residue may have provided chloride. The water used in an
attempt to extinguish the Grandcamp fire could have exacer-
Acknowledgment. The authors thank Prof. Charles L. Perrin
for helpful suggestions. Funding was provided by the National
Science Foundation, grant number CHE-9632311.
JA971618K
(
34) Gray, H. B.; Simon, J. D.; Trogler, W. C. BraVing the Elements;
University Science Books: Sausalito, CA, 1995; p 418.
35) Natl. Fire Protection Assoc. Quarterly 1947, 41, 25-57.
(36) Thomas, M. J.; Cummings, A.; Gomez, M. Chemical Accident
Investigation Report: Terra Industries, Inc. Nitrogen Fertilizer Facility, Port
Neal, Iowa, U.S.E.P.A., Region 7, 1996.
(