This is an open access article published under an ACS AuthorChoice License, which permits

copying and redistribution of the article or any adaptations for non-commercial purposes.

Article

Cite This: Ind. Eng. Chem. Res. 2018, 57, 10180−10186 pubs.acs.org/IECR

Process Intensification in Nitric Acid Plants by Catalytic Oxidation of

Nitric Oxide
Carlos A. Grande,*,† Kari Anne Andreassen,† Jasmina H. Cavka,† David Waller,‡ Odd-Arne Lorentsen,‡
Halvor Øien,‡ Hans-Jörg Zander,§ Stephen Poulston,∥ Sonia García,∥ and Deena Modeshia∥
†
SINTEF AS, P.O. Box 124 Blindern, Oslo N0314, Norway
‡
Yara International ASA, Yara Technology Centre, P.O. Box 1130, Porsgrunn 3905, Norway
§
LINDE AG, Engineering Division, Dr.-Carl von Linde Straβe 6-14, Pullach 82049, Germany
∥
Johnson Matthey Technology Centre, Blount’s Court, Sonning Common RG4 9NH, United Kingdom
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

ABSTRACT: We have evaluated the kinetics of the catalytic oxidation of

NO to NO2 using a Pt/alumina catalyst, under conditions relevant to
industrial nitric acid production: NO and steam contents up to 5% and 20%,
Downloaded via 190.24.134.41 on February 4, 2019 at 22:12:50 (UTC).

respectively, with temperatures from 250 to 350 °C, and pressures up to 4.7
bar. The objective is to replace the current homogeneous oxidation process,
which requires cooling of the process gas and a long residence time, with a
more intensive heterogeneous oxidation process, allowing the heat of reaction
(114 kJ/mol) to be recovered. This may give a 10% improvement in overall
heat recovery and, additionally, lead to reduced capital expenditure (CAPEX)
and footprint of new build plants. With world production of nitric acid of 60
million tonnes per annum, the transformation from the homogeneous
oxidation of NO to a heterogeneous oxidation can lead to significant
environmental benefits and cost reduction.

1. INTRODUCTION NO2 is subsequently absorbed in water to form nitric acid and

nitric oxide. The tail gas of the absorption column is treated in
Reduction of anthropogenic greenhouse gas emissions is one
a DeNOx unit before being discharged.
of the key challenges that must be addressed by mankind.
All of the reactions in the Ostwald process are exothermic,
Among the many technologies that should be implemented,
and currently only a small fraction of the energy spent in
improving the energy efficiency of existing industrial processes making ammonia is recovered in its transformation to nitric
is one of the necessary steps to reduce the energy intensity of acid. A summary of the reactions taking place in the process
the manufacture of commodities. together with the heat of reaction is
Nitric acid is a key industrial chemical for the production of
fertilizers. In 2013, annual production capacity was ca. 78 4NH3(g) + 5O2 (g) → 4NO(g) + 6H 2O(g)
million tonnes (Mt) and actual manufacturing ca. 58.5 Mt.1
Production growth of 2.3% is anticipated, considering an (ΔH = −905.2 kJ/mol) (1)
annual production of 65.5 Mt in 2018.
The current production route of nitric acid is known as “the 2NO(g) + O2 (g) ↔ 2NO2 (g) (ΔH = −114 kJ/mol)
Ostwald process” and has been used for nearly a century.2 A (2)
simplified scheme of a modern plant is shown in Figure 1.3 In
this process, ammonia is combusted (oxidized) in air, to nitric 3NO2 (g) + H 2O(l) → 2HNO3(aq) + NO(g)
oxide (NO). This highly exothermic reaction is carried out (ΔH = −117 kJ/mol) (3)
over a highly selective platinum−rhodium catalyst. The
temperature ranges between 800 and 930 °C. The pressure, The oxidation of NO in the state-of-the-art plants is a
on the other hand, varies from ambient to 15 bar, depending homogeneous process that occurs without the use of a
on the technology. Typically, 9.5−11.5% of ammonia in air is heterogeneous catalyst. Because there is a molar reduction,
used as the feedstock. The exact concentration depends on the the reaction rate of NO2 conversion (reaction 2) increases at
plant conditions. NO then reacts with oxygen to form nitrogen higher pressures. The homogeneous reaction rate is third order
dioxide (NO2). The oxidation of NO to NO2 in nitric acid
plants is currently carried out by a combination of two Received: April 5, 2018
methods: gas cooling shifts the equilibrium toward NO2 Revised: July 18, 2018
formation; and giving the sufficient residence time to allow Accepted: July 20, 2018
the homogeneous oxidation reaction to complete this process. Published: July 20, 2018

© 2018 American Chemical Society 10180 DOI: 10.1021/acs.iecr.8b01483

Ind. Eng. Chem. Res. 2018, 57, 10180−10186
Industrial & Engineering Chemistry Research
Figure 1. Simplified scheme of the Ostwald process for nitric acid production.
relative to pressure. Moreover, this is an unusual reaction as
the observed rate is higher at lower temperatures.2−4 The
temperature reduction that is required to shift the equilibrium
toward NO 2 is achieved through a network of heat
exchangers.5 There, a fraction of energy (heat) is recovered.
Furthermore, heat loss through the pipe walls, where heat
energy is not recovered, also occurs. In a large, high
throughput plant, the tube between the ammonia burner
superheater and the absorption column may have a volume of
several hundred cubic meters. This column contains a
relatively high pressure of highly corrosive gas. For these
reasons, the conventional method of NO oxidation constitutes
a significant capital expenditure (CAPEX) for a new plant. In
addition, this pipework occupies a substantial part of the plant
footprint.
In existing plants with significant energy integration
measures, at temperatures between 300−260 °C the degree
of NO to NO2 oxidation is around 20%. However, according
to equilibrium calculations, it could be up to around 90%. This
means that there is a large unexplored potential of conversion
of gas that is not exploited with existing technology; the heat
transfer rate is much faster than the conversion rate. The bulk
NO conversion is made at lower temperatures where the
energy recovery is less effective and with lower efficiency
producing low-pressure steam. An increase of rate of NO
oxidation at temperatures around 300 °C could result in an
additional recovery of the heat of reaction at temperatures
where high-pressure steam can be produced. It is estimated
that the overall plant heat recovery can be increased by around
10% if significant conversion of NO to NO2 can take place at
temperatures around 300 °C.
The recovery of the energy of oxidation of NO at
intermediate temperatures (∼300 °C) is the main motivation
of this work: using a catalyst, a faster conversion of NO to NO2
10181
Article
can be achieved. The conditions under which the catalytic
oxidation of NO to NO2 would occur are challenging: the
mixture is highly corrosive, and that imposes a large problem
even for analysis of the results. The gas composition contains
ca. 10% NOx (NO + NO2) and 15% water vapor. Kinetics of
NO oxidation to NO2 at such high concentrations have not
been reported in the literature in any catalyst.
Because of the well-known properties of NO conversion to
NO2 in platinum at lower concentrations and with low Pt
concentration,6−12 we have selected this metal as an active site
for the catalyst. A catalyst based on uniformly dispersed
platinum deposited on alumina has been used to study the
reaction rate. The lack of literature on NO oxidation close to
the work operating conditions leads to a range of variables
being evaluated: molar fractions of NO, NO2, and water,
pressure, and temperature (yNO, yO2, yH2O, T, P). For this
reason, the kinetic expression was determined with a total of
825 experimental points obtained at different operating
conditions.
2. EXPERIMENTAL SECTION
2.1. Catalyst Preparation and Characterization. The
catalyst was prepared by impregnation of a Pt precursor,
aiming at 2 wt % Pt, on γ-Al2O3 SCFa 140 from Sasol and
subsequent reduction. Granule dimension of catalyst was
determined to be ca. 40 μm using a Mastersizer-3000 laser
particle size analyzer. The particle size of the platinum was
derived from CO metal area measurements. The sample was
reduced under H2/N2 at 300 °C for 10 min. The catalyst was
then cooled under N2 to 35 °C. Subsequently, pulses of CO
were introduced. A Pt:CO stoichiometry of 1:1 was assumed,
and the Pt atomic area was assumed to be 0.0800 nm2. The Pt
particle size calculated from the CO adsorption measurements
was 8−10 nm.13
DOI: 10.1021/acs.iecr.8b01483
Ind. Eng. Chem. Res. 2018, 57, 10180−10186
Industrial & Engineering Chemistry Research Article

The samples were examined in a JEM 2800 scanning A closer TEM image shown in Figure 3 indicates the
transmission electron microscope (STEM) using the following presence of well-defined Pt nanoparticles on the catalyst. A
instrumental conditions: voltage (kV) 200; C2 aperture (μm)
70 (Z-contrast) imaging in scanning mode using an off-axis
annular detector. The SE signal was acquired simultaneously
with the other STEM images providing topological informa-
tion on the sample.
The STEM images showed the distribution of the metal in
the support; in Figure 2a, the bright areas indicate the location
of Pt on the surface of the alumina, while in Figure 2b, the
particles inside the support are shown. These images illustrate
a good dispersion and metal distribution of the metallic
particles.

Figure 3. Transmission electron microscopy showing the Pt

nanoparticles in the Al2O3 catalyst used in the kinetic measurements.

Figure 4. Particle size distribution of Pt nanoparticles obtained from

TEM images in the prepared Al2O3 catalyst.

particle size analysis by TEM, shown in Figure 4, indicates that

most particles were found between 1.20 and 4.40 nm. The
discrepancy between these values and those obtained by CO
metal area suggests the presence of some large metal particles
that were not observed in the TEM areas studied.
2.2. Experimental Setup for NO Oxidation. The
reaction rate data were collected in a differential bed reactor
using 52 mg of catalyst. The unit and a detailed view of the
reactor are shown in Figure 5. The reactor has a feed section
where four mass flow controllers (Bronkhörst, The Nether-
lands) introduced air, nitrogen, and NO (20% diluted in
Figure 2. Scanning transmission electron microscopy of the prepared nitrogen) into the system. All gases were acquired from Yara
catalyst showing the Pt on the (a) surface and (b) subsurface of (Norway). A CEM (controlled evaporation and mixing) unit
Al2O3. (Bronkhörst, The Netherlands) was used to introduce water
10182 DOI: 10.1021/acs.iecr.8b01483
Ind. Eng. Chem. Res. 2018, 57, 10180−10186

Industrial & Engineering Chemistry Research
Figure 5. Experimental unit to measure NO oxidation kinetics where
heated lines are highlighted in red. MFC, mass flow controller; CEM,
controlled evaporation and mixing; T, K thermocouple; BPR, back-
pressure regulator. Zoom of the differential reactor with dimensions.
into the system. All lines after the CEM were heated with
electrical resistance elements to avoid condensation, which
would result in flow and pressure pulsation as well as variations
in molar fractions supplied to the reactor. Because the reaction
also proceeds in the gas phase, it is very important to limit the
contact time of NO and O2 before the catalyst. For the same
reason, the length of the exit stream of the reactor has to be
reduced as much as possible between the reactor and the
analyzer (to limit unconverted NO and O2 interaction). Our
initial reactor design presented a high dead volume, and some
NO conversion was observed in the absence of a catalyst
(homogeneous catalysis). In the final design, no measurable
conversion of NO to NO2 was detected without the presence
of a catalyst. After the reactor, pressure is controlled with a
backpressure regulator (Equilibar, U.S.), followed by an online
IR analyzer. Calibration of the analyzer system for NO, NO2,
N2O, and water was carried out. Finally, all gases passed
through a knockout drum filled with hydrogen peroxide to
avoid high emissions of corrosive and dangerous NOx gases to
the exhaust system. Because of the hazardous nature of the
gases used in this process, the unit is equipped with gas alarms
and was thoroughly tested for possible leaks.
Using this system, the reaction rate was evaluated for NO
concentrations from 0.25% to 5%, O2 concentrations from 1%
to 8%, and water content from 0% to 20%. The flow rate in all
of the experiments was set at 1.698 SLPM (standard liters per
minute). Experiments were carried out at three different
pressures, 1, 2.7, and 4.7 bar, and at four temperatures covering
the range between 250 and 350 °C.
3. ASSUMPTIONS
The data were analyzed assuming that the reactor worked like
a plug flow reactor. The change of the mole fraction xi in the
reactor tube along the dimensionless coordinate z was
calculated according to the following equations:
Ri = ∑ vij·rj
j (4)
10183
ÄÅ ÉÑ
Article
ÅÅ ÑÑ
ÅÅ ÑÑ
ÅÅ R i ÑÑ
·ÅÅÅÅ 2 ∑ kÑ
ÑÑ
Wi
WHSV ÅÅÅ ∑k Mk ÑÑ
ÑÑ
dxi 1 Mi
ÅÅ ÑÑ
= w − · R
ÅÇ ÑÖ
dz
k (∑ )wk
k Mk
k
(5)
1 dxi
xi ,outlet = ∫0 dz
dz
(6)
where rj are the reaction rates, with stoichiometric coefficients
vij and species formation rates Ri. The dimensionless length
coordinate is represented by z. Mole fractions are denoted by
xi, mass fractions by wi, and molecular weights by Mi. The
weight hourly space velocity (WHSV) is the gas mass flow
divided by the catalyst mass. Equations 4−6 are the standard
1D plug flow reactor equations expressed in convenient
variables for the purpose of parameter fitting.
The reaction rate expression is based on catalyst mass, that
is, moles of conversion per catalyst mass and time. This
equation describes the change of mole fractions due to
chemical conversion and/or due to dilution caused by the
volume change of a reaction.
The difference between calculated and the measured mole
fractions was used to define a weighted error square sum,
which was the objective function to a numerical minimization.
The minimum that was found was then defined as the optimal
parameter set. Simulations are stopped when the difference
between successive optimization steps is smaller than 10−10. A
statistical test was used to determine the significance of the
identified parameters. The proposed algebraic structure of the
kinetic equations obtained as result was modified until the fit
quality was under the predefined acceptable difference and all
parameters were significant. For kinetic parameter fits, a 20%
error margin is often expected. This work was supported by an
already existing program code termed “Siamod”.14
There was only one modeled reaction in the system (NO
oxidation). N2O was excluded from the reaction system,
because its inlet concentration was not available in all
experiments and it was only present in few ppm, probably
impurities in the feed gas.
NO oxidation is limited by equilibrium. The equilibrium
constant was calculated from thermodynamic data, standard
enthalpy and entropy of formation, and the cp(T) profiles
by15,16
a a a
log10 K [bar] = a0 + 1 + 22 + 33
T T T (7)
where the coefficients are a0 = −7.955316 × 100, a1 = 6.074796
× 103, a2 = 1.890112 × 104, and a3 = −6.267597 × 106.
4. RESULTS AND DISCUSSION
The typical results of experiments are shown in Figure 6. It
should be noted that the aim of the work is to determine
kinetics so complete or even very high conversion of NO to
NO2 is not desirable. An additional advantage of using a
differential bed is that the generation of heat of reaction is
controlled, and for this reason the system was analyzed as
isothermal. Temperature measurements close to the catalyst
layer confirm this assumption.
Equilibration of NO2 concentration was remarkably quick,
and each of these points was determined after 5 min of stable
operation. As shown in Figure 6, the typical run consisted of
setting the temperature, the amount of water, and NO.
Subsequently, the pressure of the system was modified. The
DOI: 10.1021/acs.iecr.8b01483
Ind. Eng. Chem. Res. 2018, 57, 10180−10186
Industrial & Engineering Chemistry Research Article

Figure 6. Reaction kinetics of NO oxidation to NO2 at (a) 250 °C and (b) 350 °C at 4.7 bar and with 5% NO for different amounts of oxygen and
water.

ij 94.67 kJ yz
jjj− mol z
zz
jj z
R ·T zzz
amount of oxygen then was modified. Experiments at different

jj
mol
r1 = 1.01 × 1017
k {
temperatures with constant amounts of water were initially × exp
made, and water content was the last variable to be evaluated. kg·s·bar 3
NO was varied from 1% to 5%, while the range of oxygen used 2
pNO
2
was between 1% and 8%. Four levels of water, 0%, 2%, 10%, pNO ·pO − 2
2 K
and 20%, were investigated. All experiments were carried out at ×
1.0, 2.7, and 4.7 bar and at 250, 275, 300, and 350 °C. The (1 + 3839 bar −1·pNO2 + 0.43 bar −1·pH O )2 (8)
2

total number of test conditions amounted to 825. All

experiments we made with the same catalyst. Every day that
new runs were performed, the same reference test was
performed. No deactivation has been observed for over 20
days of gas−catalyst contact. The estimated error in the
measurements was ±10% and in most of the experiments
within ±5%. Because no deactivation was found, our kinetic
rate expression does not have deactivation terms caused, for
example, by passivation of Pt by oxygen as reported by other
researchers.17
The following reaction rate (including equilibrium and
adsorption term) was found to fit well the experimental data:
10184
A parity plot of measured versus simulated mole fractions is
shown in Figure 7. The quality of the fitted reaction rate
achieves approximately 20% accuracy; due to the experimental
error of the experimental data, it was considered that such
accuracy is acceptable for first catalyst evaluation. The number
of the adapted parameters is small, and all of them are
statistically significant. The pre-exponential factor is very large.
However, this is multiplied by low numbers (Arrhenius
exponent, the high adsorption factor, and the high reaction
order) at small partial pressures and close to equilibrium. It is
also important to note that the term related to water
DOI: 10.1021/acs.iecr.8b01483
Ind. Eng. Chem. Res. 2018, 57, 10180−10186
Industrial & Engineering Chemistry Research
Figure 7. Parity plot of the different species (a) NO, (b) NO2, and
(c) H2O in all of the simulations. Dotted lines are deviations of 5%,
10%, and 20%.
adsorption has only a small contribution under the conditions
tested.
The results obtained in this work indicated that there is
potential to develop a catalytic step in the nitric oxide
oxidation to recover extra energy in the production of nitric
acid. The impact of this process can also have a significant
effect on the final footprint of the nitric acid plant due to large
reduction of piping now existing to provide time for conversion
of NO. This work is only the first step in the process of catalyst
evaluation, given that the velocities used in an industrial setting
are significantly larger than those in this study raising
diffusional limitation issues. Furthermore, although gains in
production of heat can be achieved, pressure drop increase
could become an important bottleneck for the catalytic
process. For this reason, a good reactor design is required
including a very efficient catalyst distribution.
5. CONCLUSIONS
This work brings up a novel topic of research in production of
nitric acid, by transforming the homogeneous noncatalytic
nitric oxide oxidation into a faster catalytic step. This
conversion may bring advantages both in reduction of
footprint and in increased energy recovery. With this
transformation, the energy recovery of production of nitric
acid is expected to increase up to 10%. Considering that nitric
acid is a major commodity for fertilizers, a 10% increase in
energy is highly beneficial for the environment.
A platinum catalyst on alumina was investigated for this
purpose. The reaction was tested for NO concentrations as
high as 5% with water content up to 20%, resulting in an
extremely corrosive gas exiting the reactor. The reaction rate
10185
Article
obtained indicated that the conversion was rapid with an
activation energy of 94 kJ/mol.
This work is the first step in demonstrating that a catalytic
conversion of NO to NO2 at higher concentrations is possible
and efficient. A compromise between catalyst performance and
pressure drop needs to be achieved for the catalytic process to
accomplish the desired extra energy recovery.
■ AUTHOR INFORMATION
Corresponding Author
*Phone: +47 93207532. Fax: +47 22067350. E-mail: carlos.
grande@sintef.no.
ORCID
Carlos A. Grande: 0000-0002-9558-5413
Notes
The authors declare no competing financial interest.
■ ACKNOWLEDGMENTS
The project leading to this application has received funding
from the European Union’s Horizon 2020 research and
innovation programme under grant agreement no. 680414.
The project belongs to the SPIRE programme, and
information can be found in www.printcr3dit.eu. We want to
acknowledge Martha Bricero de Gutiérrez from Johnson
Matthey for the TEM pictures.
■ REFERENCES
(1) Gubler, R.; He, X.; Suresh, B.; Yamaguchi, Y. Nitric Acid;
Chemical Economics Handbook; IHS Chemical: London, 2014.
(2) Sorgent, H. A.; Sachsel, G. F. Nitric Acid Manufacture. Theory
and Practice. Ind. Eng. Chem. 1960, 52, 101−104.
(3) Perez-Ramirez, J.; Kapteijn, F.; Schöffel, K.; Moulijn, J. A.
Formation and Control of N2O in Nitric Acid Production. Where do
we stand today? Appl. Catal., B 2003, 44, 117−151.
(4) Connor, H. The Manufacture of Nitric Acid. Platinum Met. Rev.
1967, 11, 2−9.
(5) Ivanis, G. R.; Lazarevic, M.; Radovic, I. R.; Kijevcanin, M. L.
Energy Integration of Nitric Acid Production using Pinch Method-
ology. Hem. Ind. 2015, 69, 261−268.
(6) Olsson, L.; Fridell, E. The influence of Pt oxide formation and Pt
dispersion on the reactions NO2 ↔ NO + 1/2 O2 over Pt/Al2O3 and
Pt/BaO/Al2O3. J. Catal. 2002, 210, 340−353.
(7) Mulla, S. S.; Chen, N.; Cumaranatunge, L.; Blau, G. E.;
Zemlyanov, D. Y.; Delgass, W. N.; Epling, W. S.; Ribeiro, F. H.
Reaction of NO and O2 to NO2 on Pt: kinetics and catalyst
deactivation. J. Catal. 2006, 241, 389−399.
(8) Schmitz, P. J.; Kudla, R. J.; Drews, A. R.; Chen, A. E.; Lowe-Ma,
C. K.; McCabe, R. W.; Schneider, W. F.; Goralski, C. T., Jr. NO
oxidation over supported Pt: impact of precursor, support, loading,
and processing conditions evaluated via high throughput exper-
imentation. Appl. Catal., B 2006, 67, 246−256.
(9) Weiss, B. M.; Iglesia, E. NO oxidation catalysis on Pt clusters:
elementary steps, structural requirements, and synergistic effects of
NO2 adsorption sites. J. Phys. Chem. C 2009, 113, 13331−13340.
(10) Li, L.; Qu, L.; Cheng, J.; Li, J.; Hao, Z. Oxidation of nitric oxide
to nitrogen dioxide over Ru Catalysts. Appl. Catal., B 2009, 88, 224−
231.
(11) Bray, J. M.; Schneider, W. F. First-principles analysis of
structure sensitivity in NO oxidation on Pt. ACS Catal. 2015, 5,
1087−1099.
(12) Hong, Z.; Wang, Z.; Li, X. Catalytic oxidation of nitric oxide
(NO) over different catalysts: an overview. Catal. Sci. Technol. 2017,
7, 3340−3352.
DOI: 10.1021/acs.iecr.8b01483
Ind. Eng. Chem. Res. 2018, 57, 10180−10186
Industrial & Engineering Chemistry Research Article

(13) Foger, K. Dispersed metal catalysts. In Catalysis, Science and

Technology; Anderson, J. R., Boudart, M., Eds.; Springer: Heidelberg,
1984; Vol. 6.
(14) Zander, H.-J.; Dittmeyer, R.; Wagenhuber, J. Dynamic
modelling of chemical reaction systems with neural networks and
hybrid models. Chem. Eng. Technol. 1999, 21, 571−574.
(15) Poling, B. E.; Prausnitz, J. M.; O’Connell, J. P. The Properties of
Gases and Liquids, 5th ed.; McGraw-Hill: New York, 2001.
(16) Perry, R. H.; Green, D. W. Perry’s Chemical Engineers’
Handbook, 7th ed.; McGraw-Hill: New York, 1997.
(17) Bhatia, D.; McCabe, R. W.; Harold, M. P.; Balakotaiah, V.
Experimental and kinetic study of NO oxidation on model Pt
catalysts. J. Catal. 2009, 266, 106−119.

10186 DOI: 10.1021/acs.iecr.8b01483

Ind. Eng. Chem. Res. 2018, 57, 10180−10186