Kinetics of the Alkaline Hydrolysis
of Nitrocellulose
Christos Christodoulatos, Tsan-Liang Su, Agamemnon Koutsospyros
ABSTRACT: Cellulose nitrate (nitrocellulose) is an explosive solid
substance used in large quantities in various formulations of rocket and
gun propellants. Safe destruction of nitrocellulose can be achieved by
alkaline hydrolysis, which converts it to biodegradable products that can
then be treated by conventional biological processes. The kinetics of the
alkaline hydrolysis of munitions-grade nitrocellulose in sodium
hydroxide solutions were investigated in completely mixed batch
reactors. Experiments were conducted using solutions of alkaline
strength ranging from 0.1 to 15% by mass and temperatures in the range
of 30 to 90 °C. Regression analysis of the kinetic data revealed that
alkaline hydrolysis of nitrocellulose is of the order 1.0 and 1.5 with
respect to nitrocellulose and hydroxide concentration, respectively. The
activation energy of the hydrolysis reaction was found to be 100.9
kJ/mol with a preexponential Arrhenius constant of 4.73 3 1013. Nitrite
and nitrate, in a 3:1 ratio, were the primary nitrogen species present in
the posthydrolysis solution. The kinetic information is pertinent to the
development and optimization of nitrocellulose chemical– biological
treatment systems. Water Environ. Res., 73, 185 (2001).
KEYWORDS: nitrocellulose, kinetics, alkaline hydrolysis, treatment,
energetic materials, propellants, activation energy.
Introduction
Properties of Nitrocellulose. Cellulose nitrate, more commonly
called nitrocellulose (NC), is a fibrous solid and, like all nitrate
esters, in its dry form is a powerful explosive capable of detonating
if subjected to shock, abrasion, or sparks. Because NC is a non-
toxic substance, its public health significance and its classification
as a hazardous waste are solely because of its explosive properties
(AWWA, 1990). High-nitrogen-content NC, a principal energetic
ingredient of propellants, has been classified as reactive (40 CFR
§ 261.21 [U.S. GPO, 1999a]) and is listed under the Resource
Conservation and Recovery Act (RCRA, U.S. GPO, 1999b). The
physical and chemical properties of NC depend on the manufac-
turing process, materials used for its preparation, degree of nitra-
tion, purity, and water content. Preparation of NC involves the
reaction of cellulose with mixtures of varying proportions of nitric
and sulfuric acids. The degree of nitration depends on the source
of cellulose and acidic medium used during production. Fully
nitrated NC is produced by substitution of each of the three alcohol
hydroxyl groups per anhydroglucose monomer with an equal num-
ber of nitrate groups corresponding to a maximum nitrogen content
of approximately 14.15%. Typically, low-nitrogen-content NC
(11.5 to 12.5%) finds application in a wide variety of industries
such as pharmaceuticals, varnishes, ink bases, adhesives, cinema-
tography films, artificial leather, and printing, whereas high-nitro-
gen-content NC (12.5 to 13.5%) is primarily produced for use in
energetic materials such as rocket and gun propellants.
Fate and Transport Issues. Because NC is a widely used
industrial feedstock, its release in the environment can occur via
March/April 2001
numerous routes during military, industrial, and commercial ac-
tivities. Current demilitarization programs in the U.S. generate
large quantities of energetic materials that have to be disposed in
accordance with state and federal laws. From the environmental
standpoint, nitration radically alters the chemical nature, and con-
sequently the fate and transport behavior of cellulose. The envi-
ronmental mobility of NC is severely restricted because this sub-
stance is practically insoluble in water. However, water is used as
a dispersion and transport medium during NC preparation pro-
cesses such as stabilization, primary washing, boiling, pulping,
blending, and final purification. These processes generate aqueous
waste streams rich in NC fines that cannot be discharged to the
environment without prior treatment. Nitrocellulose mobility is
enhanced in the presence of carbonyl compounds, ketones, and
esters because NC is soluble in these compounds virtually in all
proportions. Nitrocellulose is reportedly insoluble in hydrocarbons
and their halogenated derivatives (Quinchon and Tranchant, 1989).
Another important property of NC, with significant environmental
consequences, is its resistance to microbial degradation. Although
cellulose, the most abundant natural polymer on earth, is a rela-
tively stable compound subject to microbial attack by a variety of
cellulolytic microorganisms, nitrocellulose and several other man-
made nitrated esters constitute a real xenobiotic challenge to
naturally occurring microorganisms (Brodman and Devine, 1981).
Treatment Methods. The choice of treatment technology for
NC-containing wastes and process streams is limited by three
properties of the compound, namely water solubility, explosivity,
and biodegradability. Its extremely low water solubility coupled
with its resistance to biodegradation makes NC almost impossible
to treat by biological methods alone. Furthermore, its explosive
properties pose real safety issues for landfill disposal of NC-
containing waste matrices, and air and water pollution concerns
have led to the banning of open burning and open detonation
practiced in the past by military installations around the country.
Supercritical water oxidation, another method to destroy explo-
sives and rocket fuel components, is associated with high costs
because of the high energy requirements (Harradine et al., 1993).
Combined chemical and biological treatment processes seem to be
more effective for the treatment of recalcitrant energetic com-
pounds. A three-step chemical– biological treatment scheme has
shown promising results for the destruction of most nitrate esters
including NC (Wendt and Kaplan, 1976). The scheme essentially
involves hydrolysis of the nitrate esters leading to denitration and
simultaneous destruction of their backbones. The resulting product
stream, consisting of a variety of biodegradable organics, nitrites,
nitrates, and other minor constituents, can be subsequently neu-
tralized and treated by a biological denitrification process leading
to innocuous end products. It should be noted, however, that high
185

Christodoulatos et al.
concentrations of nitrite in the posthydrolysis product stream may
interfere with the denitrification process by imparting toxicity to
denitrifying microbial species (Beccari et al., 1983). The develop-
ment of an efficient, inexpensive, and environmentally sound
technology for the pretreatment of energetic compounds is critical
to their final transformation and disposal. Chemical alteration of
the polymeric structure of NC and conversion to products amena-
ble to microbial degradation presents a plausible treatment scheme
for the treatment of NC-containing wastes (Christodoulatos et al.,
1996; Quivey et al., 1994; and Wendt and Kaplan, 1976).
Hydrolysis of Nitrocellulose. Nitrocellulose, like all nitrate
esters, can be readily hydrolyzed under acidic or alkaline condi-
tions; however, alkaline hydrolysis seems to lead to a more com-
plete destruction of the NC backbone (Spontarelli et al., 1993, and
Wendt and Kaplan, 1976). In literature cited by Kenyon and Gray
(1936) and Urbanski (1965), it has been established that unlike
most esters, the reaction of NC with alkaline agents is not a simple
saponification reaction regenerating cellulose and forming alkali
nitrate. Instead, it involves a profound decomposition of the NC
backbone to yield various biodegradable organics and inorganic
substances including nitrates; nitrites; ammonia; cyanide; nitrogen
oxides; carbon dioxide; malic, oxalic, glycolic, trioxyglutaric, di-
oxybutyric, malonic, tartonic, and other unidentified complex ac-
ids; sugar; modified cellulose and its nitrates; and partially deni-
trated cellulose nitrate. Edge et al. (1990) studied the degradation
of cinematographic films and proposed NC decomposition and
nitrite–nitrate pathways caused by alkaline conditions. The prob-
able mechanism suggested involves three types of decomposition
reaction, namely (a) hydrolysis of nitrate ester (2ONO2) groups,
(b) destruction of linkages between particular glucose units, and
(c) intermolecular oxidation of particular anhydroglucose rings.
Although these studies explain some of the steps involved in the
decomposition of NC in alkaline environments, the exact mecha-
nism of the hydrolysis reaction is still undetermined. Unresolved
issues include the mechanism of nucleophilic substitution (SN1
versus SN2), single- versus multiple-step reaction, reaction order,
nitrogen balance, and relationship between the various nitrogenous
species in the products. These issues are further complicated by the
polymeric nature of NC compounds.
Kinetic studies dealing with alkaline hydrolysis of NC are also
scarce. Garg et al. (1991), using limited data from treatability
studies of NC sludge with ammonium hydroxide at two tempera-
tures (73 and 90 °C), and assuming second-order kinetics, com-
puted an activation energy of 40.7 6 0.3 kJ/mol (9.72 6 0.065
kcal/mol). The authors provide no information regarding the de-
gree of nitration of NC. Quivey et al. (1994) presented a kinetic
study using munitions-grade NC and sodium hydroxide (NaOH) at
four temperatures ranging from 25 to 65 °C. Based on NC time
profiles and initial hydroxide concentrations, they formulated a
pseudo-first-order kinetic expression and calculated the activation
energy of denitration and solubilization to be 96.1 kJ/mol and 85.2
kJ/mol, respectively. Alkali concentration profiles, and nitrate–
nitrite data were not reported in this study.
The present study focuses exclusively on the kinetics of the
alkaline hydrolysis of NC using NaOH. Extensive experimental
data are obtained within the range of operating conditions expected
in field applications of NC conversion by alkaline hydrolysis.
Primary factors investigated include alkali strength, alkali-to-NC
ratio, and reaction temperature. The objective of this study is to
formulate a generalized kinetic expression to describe the process
186
15547531, 2001, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.2175/106143001X138840 by <Shibboleth>-member@cranfield.ac.uk, Wiley Online Library on [05/04/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
and examine the mechanisms and extent of nitrogen conversion to
nitrate and nitrite.
Materials and Methods
A series of batch alkaline hydrolysis experiments were con-
ducted at four temperatures: 30, 50, 70, and 90 °C. At each
temperature, various alkali-to-NC ratios were studied by varying
the concentration of the base and keeping the solids content
constant in all trials. The batch reactor used in the study was
equipped with a mixer and a precision temperature controller (Parr
Instrument Company, Moline, Illinois). The reactor had a working
reaction volume of 100 mL and was sealed during operation.
Munitions-grade NC was used in all assays throughout the
study. The NC was provided by Aqualon, a division of Hercules
Incorporated (Wilmington, Deleware); its nitrogen content was
approximately 12.2 to 13.5%. Nitrocellulose was received with a
moisture content of approximately 20%. It was dried at 105 °C for
1 hour, and was stored in a desiccator overnight before use.
Alkaline conditions were established in the digestion reactor
using NaOH. Solutions of various initial NaOH strengths were
used to evaluate the degradation rate of NC at each temperature.
These solutions were preheated to the desired temperature in a
water bath before their introduction to the reactor. In each batch
experiment, approximately 1 g of NC was placed in the reactor
with 100 mL of NaOH solution of the desired strength. The initial
concentration of NC in all batch experiments (161 batches) was
1% by weight. The concentrations of nitrate, nitrite, and residual
nitrocellulose were monitored during the course of the reaction.
Nitrocellulose concentration was determined by measuring total
suspended solids (TSS). To minimize sampling errors caused by
the difficulty of obtaining homogeneous samples, the entire reactor
content was used for TSS analysis. This required separate exper-
imental runs with identical initial conditions to obtain nitrocellu-
lose concentration profiles for each initial NaOH concentration.
For instance, at 70 °C and an initial NaOH concentration of 4%,
seven batch experiments were performed with identical initial
conditions to cover a total reaction time of 35 minutes with 5
minute sampling time intervals. When the reaction time ended for
each sampling period, the reactor was transferred to an ice bath to
quickly cool down the reaction medium to room temperature. The
purpose of quenching the reactor and its contents, which took less
than 1 minute, was to considerably slow the reaction during
filtration. The reactor contents were filtered through a crucible
equipped with Whatman 934-AH glass microfiber filter (Whatman
Inc., Clifton, New Jersey) to determine TSS as NC residual. The
filtrate was collected and nitrate and nitrite concentrations were
determined by direct injection of filtrate using high performance
liquid chromatography on a Varian LC Workstation (Varian, Sugar
Land, Texas) equipped with a diode array detector and a Durasep
A-1 4.6 mm 3 100 mm, 5mm (Alltech Associates Inc., Deerfield,
Illinois) chromatographic column. The concentration of NaOH in
the filtrate, which was monitored during the 70 °C runs, was
determined by titration using a Corning Ion Analyzer 250 (Corning
Science Products, Corning, New York) equipped with a Corning
general-purpose combination electrode and 0.012 N of hydrochlo-
ric acid solution as a titrant.
Results and Discussion
The experimental results of NC decomposition presented herein,
include NC time– concentration profiles, nitrite–nitrate production
data for all temperatures and NaOH concentrations examined, and
Water Environment Research, Volume 73, Number 2

Figure 1—Concentration profiles during alkaline digestion of NC at various NaOH concentrations: (a) 30 °C, (b) 50 °C,
(c) 70 °C, and (d) at 90 °C.
NaOH concentration–time profiles at 70 °C. The 70 °C data set
was used to fit a general rate expression that requires concentration
histories of both the NC and hydroxide. Based on the data obtained
at the four temperatures investigated, the activation energy for the
alkaline hydrolysis of NC is calculated and a generalized kinetic
expression is derived that can be used for the design and scale-up
of NC treatment systems by alkaline digestion. In addition, the
production of nitrite and nitrate was monitored for all experimental
runs to determine the maximum level expected during NC pro-
cessing.
Nitrocellulose Decomposition. The residual concentrations of
NC, at the four temperatures examined, are presented in Figure 1
0
for various initial sodium hydroxide concentrations, C NaOH . From
the concentration–time profiles it is clear that the digestion rate
depends on temperature and alkali strength. For a given reaction
temperature, the NC digestion rate is highly dependant on the
strength of the alkaline solution. The greater the hydroxide con-
centration, the greater the decomposition rate. For instance, start-
0
ing with 5% C NaOH at 30 °C (see Figure 1a) requires more than
900 minutes to achieve 95% reduction of NC solids; however, at
0 0
a 12.5% C NaOH the reaction time is reduced to approximately 300
March/April 2001
15547531, 2001, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.2175/106143001X138840 by <Shibboleth>-member@cranfield.ac.uk, Wiley Online Library on [05/04/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Christodoulatos et al.
minutes for the same degree of digestion, a threefold increase in
the overall digestion rate. The temperature effect on the rate and
extent of decomposition is also illustrated in Figure 1. For exam-
0
ple, at 50 °C and 1% C NaOH (Figure 1b) the total NC digestion is
less than 1% in 35 minutes reaction time; however, when the
reaction temperature is increased to 90 °C (Figure 1d), NC diges-
tion is more than 95% for the same reaction period. The effects of
temperature and alkali strength on the digestion rate of NC are
further investigated and quantified in the kinetic modeling section
below, where a general kinetic law is formulated and presented.
To obtain data to fit a general rate expression, eight experimen-
0
tal runs were performed at 70 °C with C NaOH ranging from 0.1 to
5% by mass (Figure 1c). The reaction time in these runs varies
from 0 to 35 minutes in 5-minute intervals. The hydroxide con-
centration history is shown in Figure 2. At low initial alkali
0
strengths (C NaOH , 1 g/L), the hydroxide concentration remains
0
virtually unchanged. For C NaOH well above the stoichiometric
requirements, where high NC destruction is attained, a moderate
hydroxide concentration reduction is observed. The greatest de-
crease in hydroxide concentration was approximately 10% for the
runs with 4 and 5% C NaOH . The 10% decrease in hydroxide was
187

Christodoulatos et al.
Figure 2—Hydroxide profiles at various initial NaOH con-
centrations during alkaline hydrolosis of nitrocellulose
at 70 °C.
accompanied by complete digestion of NC. Evidently, shorter
reaction times are required for NC alkaline decomposition as the
initial alkali concentration is increased from 0.1 to 5%. Almost
complete digestion is observed at initial NaOH levels above 3% by
mass. The degree of decomposition is also verified by nitrogen
balance of the released nitrates and nitrites, which accounted for
approximately 92% of the nitrogen originally present in munitions
grade NC.
Kinetics of Alkaline Digestion of Nitrocellulose. A rigorous
nonlinear regression analysis is applied to the experimental NC
digestion data to obtain the Arrhenius constants of the kinetic law.
First, the 70 °C data, which contain also hydroxide concentrations
for the corresponding NC digestion points, are fitted to a general
kinetic model. This model, combined with nonlinear analysis of
data at this temperature, supports the pseudo-first-order approach
followed for estimating the Arrhenius constants and eliminates the
need to determine the hydroxide concentration at every time in-
terval investigated.
Nonlinear regression with two independent variables was ap-
plied to the 70 °C data set, using the statistical package RS/1
(BBN, 1989), to fit the general kinetic model
dCNC
a b
5 2kC NC C NaOH (1)
dt
Where
CNC 5 nitrocellulose concentration, g/L;
CNaOH 5 NaOH concentration, g/L;
k 5 generalized rate constant, (g/L)1-a-b/min21; and
a, b 5 reaction order coefficients for NC and NaOH, respec-
tively.
The nonlinear least-squares estimate of the rate constant k of
equation 1 is 8.72 3 1024 6 1.7 3 1024 (g/L)1-a-b/min (for the
188
15547531, 2001, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.2175/106143001X138840 by <Shibboleth>-member@cranfield.ac.uk, Wiley Online Library on [05/04/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
reqression coefficient R, R2 5 0.995 for multiple curves). The
values of the exponents of the concentration terms are a 5 0.90 6
0.06 and b 5 1.48 6 0.05. Substitution of these values in equation
1 gives the following expression:
dCNC
5 28.72 3 1024C NC
0.90 1.48
C NaOH (2)
dt
The values of the estimated exponents of equation 2 suggest that
the reaction order, for all practical purposes, can be assumed to be
1 and 1.5 with respect to NC and NaOH, respectively. Based on
this finding and working with NaOH levels greater than those
required stoichiometricaly, one can assume that alkaline decom-
position of NC is adequately described by a pseudo-first-order
kinetic expression, thus eliminating the need to monitor changes in
hydroxide concentration. The pseudo-first-order reaction rate con-
stant, k9, can be defined as
b
k9 5 kC NaOH (3)
Combination of equations 1 and 3 results in the pseudo-first-order
kinetic expression
dCNC
5 2k9CNC (4)
dt
0
Equation 4 can be integrated between limits t 5 0, CNC 5 C NC ,
and t 5 t, CNC 5 CNC to yield
0
C NC 5 C NC 3 exp~2k9t! (5)
0
where C NC 5 initial nitrocellulose concentration, g/L.
Nitrocellulose time– concentration profiles for all runs and all
temperatures studied are graphically depicted in Figure 1. A visual
examination of these plots further supports the notion that NC
disappearance follows first-order behavior with respect to residual
NC concentration. First-order curves, fitted to the experimental
data exhibit excellent correlation as evidenced by the multiple
values of the R2 5 0.99. The first-order kinetic constants for all
runs are shown in Table 1. Thus, the assumption of pseudo-first-
order kinetics with respect to NC concentration is strongly sup-
ported by nonlinear regression analysis performed on time– con-
centration data for all runs and all temperatures studied.
The values of the pseudo-first-order rate constant k9 can be used
to statistically determine the value of the generalized rate constant
k. Nonlinear regression analysis of the values of k9 shown on Table
1 versus NaOH concentration was performed for each temperature.
The resulting values of k and the exponent b along with their
respective statistics (multiple correlation coefficient and standard
error) are given in Table 2. As shown, the generalized rate constant
increases by approximately 1 order of magnitude for every 20 °C
temperature increase. A slight variation of the NaOH concentration
exponent b should also be noted (1.37 to 1.61). When using this
model, an average value of approximately 1.5 is recommended.
This value is also close to the one computed using equation 1.
Accepting this value for the NaOH exponent, the pseudo-first-
order constant equation becomes
0
k9 5 k~C NaOH !1.5 (6)
Based on this equation, the value of the generalized constant k can
then be recomputed to obtain more accurate estimates. The ad-
justed values of the generalized constant, kad, are presented in
Table 2. This approach results in an improved statistical determi-
Water Environment Research, Volume 73, Number 2
 15547531, 2001, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.2175/106143001X138840 by <Shibboleth>-member@cranfield.ac.uk, Wiley Online Library on [05/04/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Christodoulatos et al.

Table 1—Pseudo-first-order reaction rate constants for the alkaline decomposition of NC obtained from concentration
profiles through nonlinear regression.

Pseudo-first-order constant k*, min21

30 °C 50 °C 70 °C 90 °C
NaOH, %
by weight k* R2 k* R2 k* R2 k* R2

0.1 0.001 217 0.99

(0.000 232)
0.2 0.001 429 0.99
(0.000 605)
0.5 0.005 957 0.99
(0.000 615)
1.0 0.001 66 0.99 0.020 512 0.99 0.106 632 0.99
(0.000 162) (0.001 785) (0.008 634)
1.5 0.233 243 0.99
(0.004 105)
2.0 0.012 223 0.99 0.061 624 0.99 0.370 917 0.99
(0.001 394) (0.005 718) (0.009 414)
2.5 0.508 749 0.99
(0.003 133)
3.0 0.014 411 0.99 0.098 879 0.99
(0.000 570) (0.010 465)
4.0 0.024 189 0.99 0.165 534 0.99
(0.001 572) (0.004 048)
5.0 0.002 41 0.99 0.028 372 0.99 0.242 162 0.99
(0.000 118)a (0.002 108) (0.005 536)
7.5 3.44 3 1023 0.99 0.054 537 0.99
(0.000 235) (0.002 847)
10.0 4.85 3 1023 0.99 0.080 392 0.99
(0.000 385) (0.002 62)
12.5 7.68 3 1023 0.99
(0.000 605)
15.0 9.73 3 1023 0.99
(0.000 605)

a
Numbers in parenthesis are standard errors.

nation of the generalized constant as evidenced by the significantly Where

smaller values of the standard error.
R 5 gas law constant, 5 0.008 310 4 kJ/deg mol;
These values of kadj are subsequently used to compute the
T 5 temperature, °C; and
activation energy (Ea, kJ/mol) from the Arrhenius equation:
A 5 Arrhenius constant.

k adj 5 A exp 2 S D Ea
RT
(7) A plot of ln kadj versus 1/T presented in Figure 3 exhibits a high
correlation (R2 5 0.998). The value of the activation energy is
100.9 kJ/mol with a standard error of 3.1 kJ/mol and the value of
Table 2—Nitrocellulose hydrolysis rate constants k and the Arrhenius constant A is 4.73 3 1013. Based on these values, the
b, and kadj for b fixed at 1.5. final form of the Arrhenius equation for alkaline decomposition of
nitrocellulose is
Temperature, Adjusted,
°C

30
k(g/L)2b/min

1.86 3 10 24
b

1.46 (0.129)
kadj(g/L)21.5/min

1.68 3 10 24
S
k adj 5 4.73 3 1013 exp 2
T
D
12141.4
min21 (8)

(6.1 3 1025)a (5.00 3 1026) The activation energy value obtained is valid over the temperature
50 3.39 3 1023 1.37 (0.063) 2.60 3 1023 range studied (30 to 90 °C). It is slightly greater than the values
(4.52 3 1024) (6.60 3 1025)
determined by Quivey et al. (1994) for alkaline denitration (96.1
70 1.80 3 1022 1.61 (0.056) 2.11 3 1022
(1.49 3 1023) (3.31 3 1024)
kJ/mol) and solubilization (85.2 kJ/mol) of NC and substantially
90 1.20 3 1021 1.60 (0.084) 1.29 3 1021 greater than the value of 40.7 kJ/mol reported by Garg et al.
(7.99 3 1023) (2.52 3 1023) (1991). The values of the activation energy for alkaline hydrolysis
of other nitrated esters, also constituents of high explosives, are in
a
Numbers in parenthesis represent standard errors. the same range. More specifically, the reported activation energy

March/April 2001 189

 15547531, 2001, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.2175/106143001X138840 by <Shibboleth>-member@cranfield.ac.uk, Wiley Online Library on [05/04/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Christodoulatos et al.

Figure 3—Linearized Arrhenious plot for the determina-
tion of the activation energy of the alkaline hydrolosis of
NC.
of hexahydro-1,3,5-trinitro-1,3,5-triazine (RDX) and octahydro-
1,3,5,7-tetranitro-1,3,5,7-tetrazocine (HMX) are 99.9 6 1.9 and
111.9 6 0.8 kJ/mol, respectively (Heilmann et al., 1996).
Based on the above analysis, the generalized kinetic equation for
alkaline decomposition of NC using NaOH solutions takes the
following form:
tions that would have allowed control of nitrite production are not
known. A possible scheme for the release of nitrate and nitrite by
a partially nitrated NC monomer was suggested by Edge et al.
(1990). This mechanism, when properly modified and applied to
fully nitrated NC, as illustrated in Figure 5, can explain the
production of nitrite and nitrate during alkaline hydrolysis but
leaves the question of nitrite control unresolved. One possible
degradation pathway is simple ester hydrolysis producing nitrate
and the other is the formation of carbonyl groups at the substitution
site, resulting in the formation of nitrite and an oxidized cellulose
sensitive to cleavage at the b-anomeric linkage in alkaline envi-
ronments. The nitrate and nitrite findings suggest that the nitrite-
producing mechanism is favored, probably because of the lower
bond energy required to cleave the oxygen–nitrogen linkages (163
kJ/mol) compared with the energy associated with oxygen– carbon
bonds (360 kJ/mol).
Destruction of the NC fibers seems to proceed by direct nucleo-
philic attack of the hydroxyls at the nitrated sites on the surface of
the NC solids exposed to solution. The solution penetrates the top
layers of the NC surface, causing a swelling and weakening of the
hydrogen bonds holding the polymers together (Urbanski, 1965).
Surface swelling of the fibers allows the hydroxyl to diffuse into
the top layers of the particle and react with the ester groups and the
cellulose rings, resulting in destruction of the polymeric structure.
As the top surface layers are removed, new ones are exposed to
hydroxyl action and the process is repeated until the entire NC
solid matrix is hydrolyzed. According to the kinetic law obtained
in this study, diffusion of hydroxyl groups into the solid matrix
does not seem to control the overall reaction rate. Furthermore, the
independence of the nitrite-to-nitrate ratio from the degree of
digestion suggests that hydrolysis of nitrocellulose is a single-step
process where nitrates and nitrites are formed directly from the
cleavage of the parent molecules. Wendt and Kaplan (1976) also

0
C NC 5 C NC exp FS 0
24.73 3 1013~C NaOH S
!1.5 exp 2
12141.4
T
DD G
t

(9)
Nitrate and Nitrite Production. The production of nitrate and
nitrite was monitored during the NC decomposition experiments
because the concentration of nitrite is critical to denitrifying or-
ganisms if subsequent biological nitrogen removal is to be under-
taken. A plot of nitrite-nitrogen versus nitrate-nitrogen for all
temperatures and NaOH concentration tested is shown in Figure 4.
Linear regression performed on the entire data set resulted in a
slope of 2.90 6 0.07, indicating that production of nitrite is 3 times
greater than that of nitrate. Thus, the ratio of nitrite to nitrate is
approximately 3 and seems to be independent of temperature,
alkaline strength, degree of digestion, and reaction time. Linear
regression of the nitrite–nitrate concentration on the amount of NC
reacted indicates that each gram of munitions-grade NC that de-
composes releases approximately 92.22 6 1.85 mg of nitrite and
32.04 6 1.00 mg of nitrate. These two values combined account
for 92% of the nitrogen originally present in munitions-grade NC.
The expected concentration of nitrite formed after complete diges-
tion of NC in 1% solid mixtures is therefore 922 mg/L. This value
is well beyond the nitrite levels that can be tolerated by denitrifiers
because nitrite inhibition becomes important for the production of Figure 4 —Regression analysis of nitrite and nitrate for
the respective reductase enzymes responsible for denitrification all times and temperatures investigated (161 data
(Beccari et al., 1983). The exact mechanism and reaction condi- points).

190 Water Environment Research, Volume 73, Number 2


Figure 5—Possible nitrate- and nitrite-producing mech-
anisms during the alkaline hydrolysis of NC.
demonstrated that decomposition of NC in alkaline environments
proceeds in a single step by determining the nitrogen, carbon, and
oxygen contained in the residual (undigested) NC during the
course of the reaction.
The kinetic model developed and presented herein combined
with nitrate–nitrite data can be used for process design of treatment
systems involving biological denitrification preceded by alkaline
hydrolysis of NC solids. The constraints imposed on the system by
energy, cost, performance, and biological process requirements
such as nitrite inhibition can be satisfied by use of a process-
optimization scheme that will address all of these issues.
Conclusions
Analysis of the kinetic data obtained from batch alkaline hy-
drolysis experiments indicates that the reaction rate is on the order
of 2.5 overall, and first order with respect to NC solids concen-
tration. The activation energy is 100.9 kJ/mol and the Arrhenius
preexponential constant is 7.43 3 1023. These values are close to
those observed for alkaline hydrolysis of other nitrated esters.
Approximately 92% of the NC nitrogen is converted to nitrite
and nitrate. The nitrite-to-nitrate ratio is 3:1 and is independent of
alkali strength, temperature, and degree of digestion. Nitrite con-
centration in the posthydrolysis liquor is 922 mg/L per gram of NC
digested. This is greater than inhibitory levels for denitrifying
bacterial species and needs to be addressed in the design of
denitrification units for the treatment of posthydrolysis NC
streams.
Acknowledgments
Credits. The work described in this publication was performed
at the James C. Nicoll Jr. Environmental Laboratory at the Stevens
Institute of Technology, Hoboken, New Jersey, and funded in part
by the U.S. Army Construction Engineering Research Laboratory,
Champaign, Illinois, under contracts No. DACA88-91-D-0011
0004 and DACA88-91-D-0011 0005. The authors thank George P.
Korfiatis for his support and advice throughout this project.
Authors. Christos Christodoulatos is Associate Professor and
Director of Laboratory Operations with the Center for Environ-
mental Engineering at the Department of Civil, Environmental and
March/April 2001
15547531, 2001, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.2175/106143001X138840 by <Shibboleth>-member@cranfield.ac.uk, Wiley Online Library on [05/04/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Christodoulatos et al.
Ocean Engineering, Stevens Institute of Technology. Tsan-Liang
Su is Assistant Research Professor at the Center of Environmental
Engineering. Agamemnon Koutsospyros is Associate Professor
and Chair at the Department of Civil and Environmental Engineer-
ing, University of New Haven, Connecticut. Correspondence
should be addressed to Christos Christodoulatos, Center for Envi-
ronmental Engineering, Stevens Institute of Technology, Hobo-
ken, NJ 07030; email: christod@stevens-tech.edu.
Submitted for publication March 20, 2000; accepted for publi-
cation September 26, 2000.
The deadline to submit Discussions of this paper is July 15,
2001.
References
American Water Works Association (1990) Water Quality and Treatment.
McGraw-Hill Inc., New York.
BBN Software Products Corporation (1989) Release 4 User’s Guide.
Cambridge, Mass.
Beccari, M.; Passino, R.; Ramadori, R.; and Tandoi, V. (1983) Kinetics of
Dissimilatory Nitrate and Nitrite Reduction in Suspended Growth
Culture. J. Water Pollut. Control Fed., 55, 58.
Brodman, B.W. and Devine, M.P. (1981) Microbial Attack of Nitrocellu-
lose. J. Appl. Poly. Sci., 26, 997.
Christodoulatos, C.; Su, T.L.; and Koutsospyros, A.D. (1996) Denitrifica-
tion Kinetics of Hydrolyzed Solid Propellant Waste Streams. Proc.
3rd Int. Conf. Protection Restoration Environ., Chania, Gr.
Edge, M.; Allen, N.S.; Hayes, M.; Riley, P.N.K.; Horie, C.V.; and Luc-
Gardette, J. (1990) Mechanisms of Deterioration in Cellulose Nitrate
Base Archival Cinematograph Film. Eur. Polymer. J., 26, 623.
Garg, R.; Grasso, D.; and Hoag, G. (1991) Treatment of Explosives
Contaminated Lagoon Sludge. Hazard. Waste Hazard. Mater., 8, 319.
Harradine, D.M.; Buelow, S.J.; and Doll’Orco, P. (1993) Oxidation Chem-
istry of Energetic Materials in Supercritical Water. Hazard. Waste
Hazard. Mater., 10, 233.
Heilmann, H.M.; Wiesmann, U.; and Stenstrom M.K. (1996) Kinetics of
the Alkaline Hydrolysis of High Explosives RDX and HMX in
Aqueous Solution and Adsorbed Activated Carbon. Environ. Sci.
Technol., 30, 1485.
Kenyon, W.O., and Gray, H.LeB. (1936) The Alkaline Decomposition of
Cellulose Nitrate. I. Quantitative Studies. Am. Chem. Soc. J., 8, 1422.
Quinchon, J., and Tranchant, J. (1989) Nitrocelluloses: The Materials and
Their Applications in Propellants, Explosives and Other Industries.
Wiley & Sons, West Sussex, U.K.
Quivey, D.M.; Aleman, J.E.; and Kim B.J. (1994) Alkaline Hydrolysis/
Biodegradation of Nitrocellulose. 49th Purdue Ind. Waste Conf.
Proc., Lewis Publishers, Chelsea, Mich.
Spontarelli, T.; Buntain, G.A.; Sanchez, J.A.; and Babcock, R.W. (1993)
Destruction of Waste Energetic Materials Using Base Hydrolysis.
Proc. 12th Int. Incineration Conf., Knoxville, Tenn.
Urbanski, T. (1965) Chemistry and Technology of Explosives. Pergamon
Press, London, 2.
U.S. Government Printing Office (1999a) Code of Federal Regulations—
Chapter 40, Subchapter I, Part 261, Characteristics of Ignitability.
Washington, D.C.
U.S. Government Printing Office (1999b) United States Code—Title 42,
Chapter 82, Subchapter 1, § 6901. Washington, D.C.
Wendt, T.M., and Kaplan, A.M. (1976) A Chemical-Biological Treatment
Process for Cellulose Nitrate Disposal. J. Water Pollut. Control Fed.,
48, 660.
191