Article

pubs.acs.org/IECR

Kinetic Model Development for Steam Pyrolysis of

Dimethylformamide in a Tubular Reactor
Vaishali B. Thaore and Vilas G. Gaikar*
Department of Chemical Engineering, Institute of Chemical Technology, Nathalal Parikh Marg, Matunga, Mumbai-19, India
*
S Supporting Information

ABSTRACT: This work presents the development and experimental validation of a mathematical model for the steam pyrolysis
process for the thermal degradation of CHON amides. Experimental data on the steam pyrolysis of dimethylformamide (DMF)
in a nonisothermal tubular stainless steel reactor for different residence times over the temperature range of 973−1273 K were
used to develop a reaction network involving 15 components and 10 reactions to model the process. The General Process
Modeling System (gPROMS) was used for kinetic parameter estimation and simulation. The model simulations results were
found to agree well with experiments carried out over a wide range of operating conditions.

■ INTRODUCTION The main objective of the present work was to develop a

suitable kinetic model for the vapor-phase steam pyrolysis of
The recovery and purification of plutonium (Pu) from uranium
amides. Another objective was to determine the kinetic
(U) and of 233U from thorium (Th) in spent nuclear fuel
parameters of the set of reactions involved in the steam
reprocessing is accomplished by processes that employ pyrolysis of amides by fitting the experimental data in the
organophosphorous compounds as extractants.1 Tri-n-butyl reaction network model. The experimental data and analysis are
phosphate (TBP), dissolved in a diluent such as dodecane, is expected to be useful in the steam pyrolysis of amide waste in
the most common extractant in the nuclear industry, despite its fuel reprocessing in the nuclear industry. The basic approach
several limitations. In recent years, several linear- and branched- was to understand the reaction mechanism of the steam
chain tetraalkylamides have been developed as alternatives to pyrolysis of DMF and then to estimate the reaction rate
TBP. Among them, N,N,N′,N′-dimethyl dibutyl tetradecyl constants for the generation and consumption of different
malonamide (DMDBTDMA), 2,2′-(methylimino)bis(N,N-dio- species by solving the model equations, allowing for the
ctylacetamide) (MIDOA), and N,N,N′,N′-tetraoctyl diglycola- determination of important species in the reaction network.
mide (TODGA) are promising extractants for the separation of
minor actinides and lanthanides from spent nuclear fuel.2−7 REACTION MECHANISM: STEAM PYROLYSIS OF
However, after a few cycles of the solvent extraction process, a ■ DMF
significant degradation of the amides is expected because of
their radiolytic and hydrolytic instabilities. This secondary low- Understanding a reaction mechanism is an important aspect of
designing and operating a reactor. It is quite difficult to develop
level radioactive waste is of concern for its safe disposal.
an actual reaction mechanism for thermal cracking reactions
Amides, unlike phosphates, are completely incinerable, as they
because short-lived and unstable free radicals are usually
are composed of C, H, O, and N elements and do not leave generated but might not be identified in the final product
behind any solid radioactive residue. mixture. By using a quasi-steady-state approximation, however,
Steam pyrolysis is a well-known process in petroleum it is possible to express the formation and disappearance of free
refining, used to crack long-chain hydrocarbons and heavy oil radicals in the material and energy balance equations.12,13
into low-molecular-weight components.8,9 In a recent work, we Previous experimental studies showed the presence of
reported experimental data on steam pyrolysis as a method of dimethylamine (DMA), H2, CH4, CO, CO2, N2, and O2 in
continuous disposal of amidic solvents. An overall pseudo-first- the final products from the steam pyrolysis of DMF.10,11 Based
order reaction was considered to represent the kinetics of the on these products, a reaction network model is proposed here
steam pyrolysis of dimethylformamide (DMF), as a representa- that includes a total of 10 main reactions involving 15
tive amide, in the temperature range of 973−1273 K.10,11 components (Table 1). No reports on a detailed reaction
However, the degradation of an amide is too complex to be mechanism for the steam pyrolysis of amides are available in
represented by a simple first-order reaction. We have thus the literature.
developed a reaction network model to describe the steam Based on the products identified in the final mixture, a set of
pyrolysis of amides, as reported in this article. We expect that plausible reactions between different species was proposed.
such an approach will help in the identification of the species
formed and destroyed during these reactions in the short Received: January 28, 2013
reaction time of a few seconds. The model is a rigorous Revised: June 5, 2013
simulation of the reactions under different conditions, studied Accepted: July 9, 2013
in a laboratory-scale continuous tubular reactor.11 Published: July 9, 2013

© 2013 American Chemical Society 10601 dx.doi.org/10.1021/ie4003238 | Ind. Eng. Chem. Res. 2013, 52, 10601−10608
Industrial & Engineering Chemistry Research Article

Table 1. Proposed Reactions for the Steam Pyrolysis of DMF and Heats of Reaction in kJ mol−1 at 298 K
reaction scheme heat of reaction at 298 K, ΔHa (kJ mol−1) reaction no.
k1 31.7 R1
(CH3)2 NCOH + H 2O → (CH3)2 NH + HCOOH
k2 262.8 R2
HCOOH → HCHO + 0.5O2
k3 4.5 R3
HCHO → CO + H 2
k4 401.65 R4
(CH3)2 NH → (CH3)2 N• + H•
5 k 306.8 R5
(H3C)2 − N• → H3C − N:̈ + H3C•
6 k 30.52 R6
2H3C − N:̈ → 2H3C• + N2
k7 −438 R7
CH3• + H• → CH4
k8 −41.2 R8
CO + H 2O ⇔ CO2 + H 2
k9 205.9 R9
CH4 + H 2O → CO + 3H 2
k10 −283 R10
2CO + O2 → 2CO2
a
Heats of reaction estimated using heats of formation of reactants and products.17

Although it still might not be possible to take every reaction analysis of the experimental data by mathematical modeling.
into account, an attempt was made here to incorporate all The reactor was modeled as a tubular reactor with a constant
major reactions. cross-sectional area (Ac) and a constant wall temperature. The
The preferential breaking point of the hydrolytic attack is steady-state model equation for each species is given by
considered to be the C−N bond of the amide.2 The hydrolysis 10
of dimethylformamide leads to the formation of formic acid and dFj
dimethylamine (reaction R1, Table 1). Previous studies of
= ∑ ri
dz i=1 (1)
formic acid decomposition reported dehydration as the main
reaction that forms CO and H2O.14,15 The combination of where Fj is the molar flux of component j (j = 1−15); is ∑10
i=1ri
reactions R2 and R3 (Table 1) gives the total decomposition of the corresponding net rate of formation and utilization of
formic acid to oxygen, carbon monoxide, and hydrogen. species j, summed over all reactions i; and z is the length of the
Dimethylamine is considered to undergo further decom- tubular reactor (Supporting Information). A total of 15
position through a free-radical mechanism. A highly reactive components based on the chemical analysis of the products
(CH3)2N• radical is formed by the removal of the hydrogen was considered: DMF, H2O, DMA, formic acid, carbon
from the N−H bond in dimethylamine. This bond cleavage monoxide, formaldehyde, hydrogen, oxygen, nitrogen, methane,
process could occur through hydrogen abstraction or homolytic carbon dioxide, (CH3)2N•, CH3N•, hydrogen radical, and
dissociation. The (CH3)2N• radical decomposes to methyl methyl radical. The quasi-steady-state approximation was
radical and CH3N• radical. The set of reactions R4−R7 (Table applied to each free radical so as to generate an algebraic
1) represents the decomposition process finally resulting in the equation for each one. This set of equations was solved to
formation of CH4. Carbon monoxide and hydrogen were express the free-radical concentrations in terms of the
observed from the decomposition of formic acid, whereas measurable concentrations of other species.
nitrogen and methane are considered to originate from the The pressure drop was modeled as16
decomposition of dimethylamine.
dP 4ρm v 2f
Water−gas shift reactions R8 and R9 (Table 1) are − =
important reactions for hydrogen formation. The steam dz 2d i (2)
reforming of methane is the major reaction considered for
methane conversion, which is a highly endothermic reaction. where P is the reactor pressure, di is the inner diameter of the
tubular reactor, ρm is the density of the gas mixture, and v is the
REACTOR MODELING AND SIMULATION velocity. The friction factor ( f) for a straight tube is given as a
■ CONDITIONS function of hydrodynamics, as represented by the Reynolds
number (Re), by the equation
The steam pyrolysis of DMF was thus considered in this work
to involve a set of reactions that operates in a nonisothermal f = 16/Re (3)
manner. The major equations of this model form a set of with
differential equations consisting of a mass balance equation for
each species and an energy balance equation along the length of d ivρm
the tubular reactor. Operation under steady-state conditions Re =
μm (4)
reduces the continuity, energy, and pressure drop equations to
ordinary differential equations. To obtain the kinetic parame- where μm is the viscosity of the reaction mixture and ρm is the
ters for the reaction network of the steam pyrolysis of DMF, a density of the reaction mixture.
nonisothermal plug-flow model was considered for the reactor. The total molar flux (NT) is the sum of all molar fluxes in the
Ideal gas-phase behavior was considered throughout the reactor at any point, which, in turn, determines the volumetric
10602 dx.doi.org/10.1021/ie4003238 | Ind. Eng. Chem. Res. 2013, 52, 10601−10608
Industrial & Engineering Chemistry Research Article

Table 2. Estimated Kinetic Parameters for the Steam Pyrolysis of DMF

kinetic parametersa (m3 mol−1 s−1 or s−1) rate constanta (m3 mol−1 s−1 or s−1)
−3 −1
reaction rate (mol m s ) k = k0 exp(−E/RT) 775 K 850 K 875 K
R1 = k1C DMF 240 exp(−39106/RT) 0.55 0.951 1.110
R 2 = k 2C HCOOH 4.99 × 1014 exp(−185550/RT) 1.55 × 102 1975 4.17 × 103
R3 = k 3C HCHO 1.2 × 1015 exp(−180308/RT) 8.71 × 102 10373 2.13 × 104
R 4 = k4C DMA 1.12 × 1017 exp(−267067/RT) 0.11 4.35 12.75
21 3 4
R 5 = k5C(CH3)2 N• 7.15 × 10 exp(−275035/RT) 2.07 × 10 8.95 × 10 2.7 × 105

R 6 = k6CCH3N•2 9.3 × 108 exp(−29476/RT)a 9.59 × 106a 1.44 × 107a 1.62 × 107a

R 7 = k 7CCH3•C H• 2.9 × 1012 (1/T)0.4 a 2.02 × 1011 a 1.97 × 1011 a 1.93 × 1011 a
a 2a a
⎛ CCO2C H2 ⎞ 1.0 × 106 exp(−53035/RT) 2.66 × 10 551 7.82 × 102a
R 8 = k 8⎜CCOC H2O − ⎟ [1.0 × 106 exp(−52960/RT)a,b]
⎝ Kw ⎠
CCO2C H2 Kw = 491 exp(−59614/RT) 0.047 0.106 0.135
Kw = Kw = 520 exp(−60110/RT)b]
CCOC H2O
R 9 = k 9CCH4C H2O 3.1 × 105 exp(−125148/RT)a 1.1 × 10−3 a 6.4 × 10−3 a 1.01 × 10−2 a
[3.1 × 105 exp(−124710/RT)a,b]
R10 = k10CCOCO2 5.31 × 108 exp(−122919/RT)a 2.75a 14.82a 24.37a
[8.83 × 1011 exp(−99768/RT)a,c]
a
Second-order rate constant, m3 mol−1 s−1. bCorella and Sanz.20 cKim et al.23

flow rate and velocity of the reaction mixture at any point in the conductivity of the reaction fluid as applicable under the given
reactor hydrodynamic conditions
N 0.33
h id i ⎛ d⎞
∑ Fj(z) = NT(z) = 1.86⎜Re × Pr i ⎟
j=1 (5) ki ⎝ L⎠ (10)

NT(z)RT (z) The reaction rate constant was considered as a function of

v (z ) = temperature according to the Arrhenius law
P(z) (6)
⎛ −E ⎞
The concentration of each component j can thus be k = k 0 exp⎜ ⎟
⎝ RT ⎠ (11)
expressed as
Fj
Cj = PARAMETER ESTIMATION
v (7) ■
Considering the conversion versus time data obtained from the
The residence time (τ) at length z was then estimated from the previously reported kinetic experiments at different temper-
dimensions of the reactor and velocity profile under given set of atures and residence times,10,11 a total of 256 data points were
operating conditions. collected to generate the set of experimental reaction rate data.
An energy balance on the reactor leads to the equation A total of 64 data points each was used at all four temperatures
10
∑ riΔHi − Ua(T − Tout) for the kinetic parameter estimation.
dT
= i=1 15
The tubular reactor flow model, described in the previous
dz ∑ j = 1 FjCp section, was used to simulate experimental data under various
j (8)
conditions. gEST, the parameter estimation tool of the General
where ΔHi is the heat of reaction for the ith reaction, U is the Process Modeling System (gPROMS) modeling package was
overall heat-transfer coefficient, and a is the heat-transfer area used to obtain the solution of the dynamic equations of the
per unit volume. Tout is the heater temperature, outside the reaction network model, to estimate the model parameters.18
reactor wall. The differential-algebraic equations (DAEs) were integrated by
The overall heat-transfer coefficient, U, is constructed from employing DASOLV, a differential-algebraic solver included in
the individual heat-transfer coefficients and the resistance of the the gPROMS library. The parameter estimation optimization
tube wall solver uses the maximum-likelihood approach to estimate all
1 model parameters simultaneously. The initial guess parameters
U= were iteratively adjusted using a comparison to the results of
1
hi
+
Ai
2π k w L ( )+R
ln
do
di fi
A
+ R fo A i +
o
Ai 1
Ao ho (9)
the experiments. The kinetic parameters estimated from the
proposed model are summarized and compared with the
The outside convective heat-transfer coefficient (ho) was literature values, wherever available, in Table 2. To reduce the
large with respect to the inside heat-transfer coefficient (hi). degree of nonlinearity, a reparameterization of the model was
Therefore, the last term in the denominator was neglected. hi used for the Arrhenius equation with a reference temperature.
was calculated in terms of the Nusselt number and thermal The Arrhenius law was replaced by the equation
10603 dx.doi.org/10.1021/ie4003238 | Ind. Eng. Chem. Res. 2013, 52, 10601−10608
Industrial & Engineering Chemistry Research Article

Figure 1. Reaction pathway for the steam pyrolysis of dimethylformamide.

⎡ −E ⎛ 1 1 ⎞⎤ ■ SIMULATION RESULTS AND DISCUSSION

k = k Tref ⎢ ⎜ − ⎟⎥
⎢⎣ R ⎝ T Tref ⎠⎥⎦ (12)
On the basis of the reported product distribution,10,11 a
schematic reaction pathway of a steam pyrolysis process for the
Two statistical tests were performed to check the reliability of thermal degradation of dimethylformamide is shown in Figure
the parameter estimation. The t-test established the statistical 1. A detailed steady-state model described by the set of ordinary
differential equations and algebraic equations was simulated to
significance of each parameter. Table S.3 of the Supporting
predict the behavior of the overall DMF steam pyrolytic
Information shows the 95% t-value of each estimated process. The simulation considered the effects of process
parameter. The χ2 test concerning the goodness of the model conditions including the feed flow rates of DMF and steam
fit was also satisfied with a weighted residual value less than the entering the reactor and the temperature of the product
χ2 value. mixture. The temperature was varied in the range of 973 to
Another measure indicating whether parameters are 1273 K, whereas the feed flow rates were varied between 0.2
optimally defined is the confidence ellipsoid, which takes into and 3 cm3 min−1. The molar ratio of steam to DMF in the feed
account the correlation between the estimates of a kinetic was kept fixed at 4.3 for all runs.
constant at a reference temperature and the activation energy Because the reactant mixture was heated in a preheater to
for each reaction. This statistical significance test is considered 673 K, the inlet feed temperature was fixed at 673 K for the
in Figures S.1−S.11 of the Supporting Information and was simulation. Because of the presence of steam in large excess, the
based on whether the elliptically shaped area representing the reactor wall remained free of any carbon deposition, and the
joint confidence region enclosed the optimal point at its center measured pressure drop was negligible.10 This indicates that the
neighborhood. The joint confidence region illustrates that the pressure drop due to the fluid flow in the tubular reactor was
parameters were known with the same precision and that the balanced by the pressure increase due to the considerable
increase in the number of moles as the reaction proceeded
estimated point was optimal for both of them.
along the length of the reactor.
Another measure indicating whether an estimation is
Figures 2−4 show that a good agreement exists between the
performed successfully is the percentage residuals, which can theoretical results based on the proposed kinetic model and the
be examined using a corresponding diagram that illustrates the experimental values, within the experimental error. The
distribution of the observed error with time. Figure S.12 conversion of DMF increased with increasing temperature
(Supporting Information) shows a horizontal band of residuals, and residence time, as shown in Figure 2. This resulted in a
which indicates that long-term effects did not influence the corresponding increase in the product species formation. It was
data. Also, the observed error between the experimental and observed that the DMF conversion increased from 20% at
predicted data was independent and randomly distributed. higher flow rates to nearly complete conversion (94.5%) at
10604 dx.doi.org/10.1021/ie4003238 | Ind. Eng. Chem. Res. 2013, 52, 10601−10608
Industrial & Engineering Chemistry Research Article

Figure 3. Effects of residence time on contents (mol %) of (A) (◊)

Figure 2. Effect of residence time on the percentage conversion of H2, (○) N2, (×) CO2, (+) dimethylamine, and (▲) unconverted
DMF at various reactor wall temperatures: 973, 1073, 1173, and 1273 DMF and (B) (*) O2, (□) CO, and (Δ) CH4 at a reactor wall
K. temperature of 973 K.

lower flow rates. The DMF conversion increased from 82% at a percentages showed a maximum with increasing time,
reactor surface temperature of 973 K to 94.5% at a reactor indicating that their conversions decreased in subsequent
surface temperature of 1273 K at the lower flow rate of 0.2 cm3 reactions. The mole percentage of DMA decreased with
min−1. The hydrolysis of dimethylformamide to dimethylamine increasing time as the further secondary reactions readily
represented by reaction R1 (Table 1) is a first-order reaction consumed it. The nitrogen mole percentage increased with
with respect to the dimethylformamide concentration. The rate increasing temperature and residence time.
constant value for reaction R1 increased from 0.55 to 1.11 s−1 The formation of compounds such as formaldehyde and
as the temperature increased from 775 to 875 K (Table 2). The formic acid was not observed experimentally.10 Based on these
activation energy estimated for the reaction is thus 39 kJ mol−1. observations, a possible route for the decomposition of formic
This is lower than the activation energy of 267 kJ mol−1 for acid and formaldehyde was suggested by reactions R2 and R3
reaction R4 (Table 1). The estimated activation energy for (Table 1). The rate of formaldehyde decomposition was higher
breaking a N−H bond is higher than the activation energy for than the rate of reduction of formic acid to formaldehyde, and
breaking a C−N bond. Dimethylamine is thus more stable than thus, formaldehyde was consumed as soon as it formed. The
DMF at temperatures up to 800 K. Above 800 K, the rate of rate constant for formaldehyde decomposition thus had a
DMA degradation increased more than the rate of DMF higher value, 2.13 × 104 s−1, at 875 K than that for formic acid
degradation. reduction to formaldehyde at the same temperature (4.17 × 103
The reaction network of DMF steam pyrolysis was further s−1). To reduce the discrepancy between the calculated and
used to compare the experimental and predicted mole experimental product distributions, additional reactions to
percentages of products at various temperatures and inlet account for the decomposition of higher compounds were
flow rates. Figures 3 and 4 show the experimental validation of included. The radical reactions were used as a guide for further
the model at two reactor surface temperatures, 973 and 1073 K, model development. The major cause of formation of nitrogen
respectively, for different residence times. The data at the two and methane is given by the set of reactions R5−R7 (Table 1).
other temperatures investigated also showed similar agreement These reactions were very fast as compared to the
between the simulated values and the experimental data. decomposition reactions of formic acid (reaction R2) and
Depending on the stabilities of the different species and the formaldehyde (reaction R3). The rate constants at 875 K were
secondary reactions, the mole percentages of various species found to be 2.7 × 105 s−1 for reaction R5 and 1.62 × 107 m3
increased, decreased, or showed a maximum with increasing mol−1 s−1 for reaction R6. Hydrogen radicals generated from
residence time. Thus, the hydrogen mole percentage increased reaction R4 and methyl radicals generated from reactions R5
with residence time, whereas the CO and CH4 mole and R6 combined to produce methane. Generally, the
10605 dx.doi.org/10.1021/ie4003238 | Ind. Eng. Chem. Res. 2013, 52, 10601−10608
Industrial & Engineering Chemistry Research Article

the hydrolysis of DMF to DMA in reaction R1 and then the

further decomposition of DMA by secondary reaction R4.
Using the estimated activation energies and frequency factors
from Table 2, reaction rates at any temperature within the
experimentally tested range can be obtained. Figures 5 and 6

Figure 4. Effects of residence time on contents (mol %) of (A) (◊)

H2, (○) N2, (×) CO2, (+) dimethylamine, and (▲) unconverted
DMF and (B) (*) O2, (□) CO, and (Δ) CH4 at a reactor wall
temperature of 1073 K.

activation energy for the recombination of free radicals must be

zero.19 The second-order rate constant was, therefore, ex- Figure 5. Effects of residence time on contents (mol %) of (A) (◊)
H2, (○) N2, (×) CO2, (+) dimethylamine, and (▲) unconverted
pressed in a different manner as a function of temperature, as DMF and (B) (*) O2, (□) CO, and (Δ) CH4 at a reactor wall
shown in Table 2. temperature of 1173 K.
In the presence of steam, the water−gas shift reaction is an
important reaction for hydrogen formation (reactions R8 and
R9, Table 1). The water−gas shift reaction is an equilibrium are the results of model predictions at reactor wall temperatures
exothermic reaction. Steam reforming of methane is another 1173 and 1273 K for various residence times. The simulation
major reaction considered for methane conversion. This is the and experimental results exhibit the same trends, but they also
highly endothermic reaction. Figures 3 and 4 show the effects show some minor discrepancies. Only the measured mole
of residence time on the hydrogen, methane, and carbon percentage of CO at higher residence times was greater than
dioxide mole percentages at the exit of the reactor at different the predicted value. Under the experimental conditions, the
outer surface temperatures of the tubular reactor. Considering Reynolds number was in the range of 100−200. For this
the product distribution of the steam pyrolysis of amides, for reactor, the length to diameter ratio was large (L/di = 35),
which the mole percentages of hydrogen and nitrogen increase which means that variations in the radial directions could be
with increasing temperature whereas the methane mole neglected.21 However, because of the laminar flow conditions,
percentage decreases, the changes in the mole percentages of the distribution of residence times was probably not uniform.
H2, CO, and CO2 are considered to be mainly due to the mildly This would affect the product distribution, which would lead to
exothermic water−gas shift reaction. The decreased mole differences between the predicted and experimental results.
percentage of CH4 is considered to be mainly due to further Thermal degradation is carried out in a tubular reactor,
reforming of methane to CO and H2. The kinetic parameters because of the endothermicity, as heat is supplied to the reactor
for these reactions are given in Table 2, along with their through the reactor wall. In the present work, the surface-to-
literature values.20 The results show a slight deviation from the volume ratio employed was 4 cm−1 (4/di). As a result of the
estimated parameters in the current work. When the feed flow increased surface-to-volume ratio, the temperature difference
rate was decreased, the CH4 and CO conversions increased between the reactor wall and the reaction mixture passing
because of the sufficient residence time in the reactor. A through was not significant.22 This resulted in the effective
comparison of the rate constants of the water−gas shift reaction thermal degradation of DMF in the short reaction time of a few
and methane reforming shows that the water−gas shift reaction seconds. This model was developed for the degradation of
is much faster than reforming of methane. simplest amide. However, the structures of amides used as
The estimated kinetic parameters values showed that the extractants in the nuclear industry are more complex, and one
rate-controlling steps of the DMF steam pyrolysis process were could expect a larger number of intermediates and correspond-
10606 dx.doi.org/10.1021/ie4003238 | Ind. Eng. Chem. Res. 2013, 52, 10601−10608
Industrial & Engineering Chemistry Research Article

■ AUTHOR INFORMATION
Corresponding Author
*E-mail: [email protected]. Tel.: +91-22-33612013.
Fax: +91-22-33611020.
Notes
The authors declare no competing financial interest.

■ ACKNOWLEDGMENTS
V.B.T. thanks the University Grant Commission (India) for
financial assistance.

■ NOMENCLATURE
a = surface area per unit volume, m−1
Ac = cross-sectional area of the reactor, m2
C = concentration, mol m−3
Cpg = specific heat capacity of the reaction mixture, J kg−1
K−1
Cpj = specific heat capacity of jth component in J mol−1 K−1
di = inside diameter of the reactor, m
do = outside diameter of the reactor m
dz = differential length of the reactor, m
E = activation energy, J mol−1
f = fanning friction factor
Fj = molar flux of component j, mol m−2 s−1
hi = inside heat-transfer coefficient, W m−2 K−1
Figure 6. Effects of residence time on contents (mol %) of (A) (◊) k = reaction rate constant, s−1 or m3 mol−1 s−1
H2, (○) N2, (×) CO2, (+) dimethylamine, and (▲) unconverted ki = thermal conductivity of fluid flowing inside the reactor,
DMF and (B) (*) O2, (□) CO, and (Δ) CH4 at a reactor wall W m−1 K−1
temperature of 1273 K. kw = thermal conductivity of the stainless steel reactor wall,
W m−1 K−1
L = length of the reactor, m
M = molecular mass, kg kmol−1
ing products. However, the approach should remain the same NT = total molar flux, mol m−2 s−1
and can help in developing proper models for other amides as Nu = Nusselt number
well. P = pressure drop in the reactor, N m−2
Pr = Prandlt number
■ CONCLUSIONS r = rate of reaction per unit volume, mol m−3 s−1
The effects of different operating conditions temperature and R = universal gas constant, 8.314 J mol−1 K−1
residence time on the noncatalytic steam pyrolysis of DMF Re = Reynolds number
were investigated by mathematical modeling of the reaction T = temperature of the reactor, K
network. An increase in pyrolysis temperature was found to Tw = reactor wall temperature, K
lead to higher DMF conversion. Degradation products from the U = overall heat-transfer coefficient, W m−2 K−1
steam pyrolysis of DMF are dimethylamine (DMA), H2, CH4, v = velocity of fluid flowing through the reactor, m s−1
CO, CO2, N2, and O2. The model successfully predicted the Zc = critical compressibility factor
variations in the residence time behavior at different temper- Greek Letters
atures. Each reaction considered in the model was found to
affect the performance of the tubular reactor. This study ΔCp = specific heat of reaction, J mol−1 K−1
provides a more detailed description of the amide pyrolysis ΔH = heat of reaction, kJ mol−1
process in the presence of steam. Schematic reaction pathways μ = component viscosity, N s m−2
of the steam pyrolysis of DMF were then proposed, although ρ = density of fluid flowing through the reactor, kg m−3
many secondary decompositions occurred, making the
mechanism complex. This study has provided a great deal of ■ REFERENCES
information about the thermal decomposition of amides at high (1) Manchanda, V. K.; Pathak, P. N. Amides and diamides as
temperatures. The same methodology could also be used to promising extractants in the back end of the nuclear fuel cycle: An
study the steam pyrolysis process for the thermal degradation overview. Sep. Purif. Technol. 2004, 35, 85.
of other amides used as nuclear extractants. (2) Gasparini, G. M.; Grossi, G. Long chain disubstituted aliphatic
amides as extracting agents in industrial application of solvent
ASSOCIATED CONTENT extraction. Solvent Extr. Ion Exch. 1986, 4, 1233.
■
*
S Supporting Information
(3) Ansari, S. A.; Pathak, P. N.; Manachanda, K. V.; Hussain, M.;
Prasad, A. K.; Parmar, V. S. N,N,N′,N′-Tetraoctyl diglycolamide
Additional information as noted in text. This material is (TODGA): A promising extractant for actinide partitioning from high-
available free of charge via the Internet at http://pubs.acs.org. level waste (HLW). Solvent Extr. Ion Exch. 2005, 23, 463.

10607 dx.doi.org/10.1021/ie4003238 | Ind. Eng. Chem. Res. 2013, 52, 10601−10608

Industrial & Engineering Chemistry Research Article

(4) Tian, Q.; Hughes, M. A. Synthesis and characterization of

diamide extractants for the extraction of neodymium. Hydrometallurgy
1994, 36, 79.
(5) Guoxin, S.; Min, L.; Yu, C.; Meilong, Y.; Shaohong, Y. Synthesis
of N,N′-dimethyl-N,N′-dioctyl-3-oxadiglycolamide and its extraction
properties for lanthanides. Solvent Extr. Ion Exch. 2010, 28, 482.
(6) Gujar, R. B.; Ansari, S. A.; Bhattacharyya, A.; Mohapatra, P. K.;
Kanekar, A. S.; Pathak, P. N.; Manchanda, V. K. Studies on the
radiolytic stability of N,N,N′,N′-tetra-2-ethylhexyl diglycolamide in n-
dodecane solution containing different phase modifiers. J. Radioanal.
Nucl. Chem. 2011, 288, 621.
(7) Sugo, Y.; Sasaki, Y.; Tachimori, S. Studies on hydrolysis and
radiolysis of N,N,N′,N′-tetraoctyl-3-oxapentane-1,5-diamide. Radio-
chim. Acta 2002, 90, 161.
(8) Crynes, B. L.; Albright, L. F. Pyrolysis of propane in tubular flow
reactors: Kinetics and surface effects. Ind. Eng. Chem. Process Des. Dev.
1969, 8, 25.
(9) Basu, B.; Kunzru, D. Catalytic pyrolysis of naphtha. Ind. Eng.
Chem. Res. 1992, 31, 146.
(10) Dicholkar, D. D.; Gaikar, V. G.; Kumar, S. Studies on steam
pyrolysis of amides as a waste solvent management method. Energy
Proc. 2011, 7, 534.
(11) Dicholkar, D. D.; Gaikar, V. G.; Kumar, S.; Natarajan, R.
Modeling and optimizing of steam pyrolysis of dimethyl formamide by
using response surface methodology coupled with Box−Behnken
design. J. Anal. Appl. Pyrolysis 2012, 96, 6.
(12) Fogler, H. S. Elements of Chemical Reaction Engineering, 4th ed.;
Pearson Education: New York, 2006.
(13) Nauman, E. B. Chemical Reactor Design, Optimization and
Scaleup; McGraw-Hill: New York, 2002.
(14) Satio, K.; Kakumoto, T.; Kuroda, H.; Torii, S.; Imamura, A.
Thermal unimolecular decomposition of formic acid. J. Chem. Phys.
1984, 80, 4989.
(15) Jianli, Y.; Savage, P, E. Decomposition of formic acid under
hydrothermal conditions. Ind. Eng. Chem. Res. 1998, 37, 2.
(16) Huntrods, R. S.; Nighswander, J. A.; Mehrotra, A. K.; Behie, L.
A. Modeling of coke formation in gas quenchers of industrial ethane
cracking furnaces. Chem. Eng. Res. Des. 1990, 67, 632.
(17) Reid, R. C.; Prausnitz, J. M.; Poling, B. E. The Properties of Gases
and Liquids, 4th ed.; McGraw-Hill: New York, 1987.
(18) Ziogou, C.; Voutetakis, S.; Papadopoulou, S.; Georgiadis, M. C.
Modeling, simulation and experimental validation of a PEM fuel cell
system. Comput. Chem. Eng. 2011, 35, 1886.
(19) Tsang, W. Rate constants for the decomposition and formation
of simple alkanes over extended temperature and pressure ranges.
Combust. Flame 1989, 78, 71.
(20) Corella, J.; Sanz, A. Modeling circulating fluidized bed biomass
gasifiers. A pseudo-rigorous model for stationary state. Fuel Process.
Technol. 2005, 86, 1021.
(21) Hill, C. G. An Introduction to Chemical Engineering Kinetics and
Reactor Design; Wiley: New York. 1977.
(22) Butt, J. B. Reaction Kinetics and Reactor Design, 2nd ed.; CRC
Press: New York, 2000.
(23) Kim, Y. J.; Lee, J. M.; Kim, S. D. Modeling of coal gasification in
an internally circulating fludized bed reactor with draught tube. Fuel
2000, 79, 69.

10608 dx.doi.org/10.1021/ie4003238 | Ind. Eng. Chem. Res. 2013, 52, 10601−10608