Ullmann's HNO3.pdf

c© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim10.1002/14356007.a17 293

Nitric Acid, Nitrous Acid, and Nitrogen Oxides 1

Nitric Acid, Nitrous Acid, and Nitrogen Oxides

Nitrates and Nitrites is a separate keyword.

Michael Thiemann, Uhde GmbH, Dortmund, Federal Republic of Germany

Erich Scheibler, Uhde GmbH, Dortmund, Federal Republic of Germany

Karl Wilhelm Wiegand, Uhde GmbH, Dortmund, Federal Republic of Germany

1. Nitric Acid . . . . . . . . . . . . . . . 1

1.1. Introduction . . . . . . . . . . . . . . 1

1.2. Properties . . . . . . . . . . . . . . . . 2

1.3. Industrial Production . . . . . . . . 3

1.3.1. Oxidation of Ammonia . . . . . . . . 4

1.3.2. Oxidation and Absorption of Nitro-

gen Oxides . . . . . . . . . . . . . . . 8

1.3.3. Equipment . . . . . . . . . . . . . . . . 15

1.3.3.1. Filters and Mixers . . . . . . . . . . . 15

1.3.3.2. Burners and Waste-Heat Boilers . . 16

1.3.3.3. Compressors and Turbines . . . . . . 18

1.3.3.4. Heat Exchangers and Columns . . . 21

1.3.3.5. Construction Materials . . . . . . . . 22

1.3.4. Processes . . . . . . . . . . . . . . . . 24

1.3.4.1. Weak Acid Processes . . . . . . . . . 24

1.3.4.2. Concentrated Acid Processes . . . . 30

1.4. Environmental Protection . . . . . 35

1.4.1. Wastewater . . . . . . . . . . . . . . . 35

1.4.2. Stack Gas . . . . . . . . . . . . . . . . 35

1.4.2.1. Emission Limits . . . . . . . . . . . . 35

1.4.2.2. Analysis . . . . . . . . . . . . . . . . . 36

1.4.2.3. Control of NOx Emissions . . . . . 36

1.5. Storage and Transportation . . . . 40

1.6. Uses and Economic Aspects . . . . 40

2. Nitrous Acid . . . . . . . . . . . . . . 40

3. Nitrogen Oxides . . . . . . . . . . . . 41

3.1. Dinitrogen Monoxide . . . . . . . . 42

3.2. Nitrogen Monoxide . . . . . . . . . 43

3.3. Nitrogen Dioxide and Dinitrogen

Tetroxide . . . . . . . . . . . . . . . . 44

3.4. Dinitrogen Trioxide . . . . . . . . . 45

3.5. Dinitrogen Pentoxide . . . . . . . . 45

4. Toxicology and Occupational

Health . . . . . . . . . . . . . . . . . . 46

5. References . . . . . . . . . . . . . . . 47

1. Nitric Acid

1.1. Introduction

Nitric acid is a strong acid that occurs in natureonly in the form of nitrate salts. When large-scale production of nitric acid began, sodium ni-trate (soda saltpeter, Chile saltpeter) was used asthe feedstock. At the beginning of the 20th cen-tury the reserves of Chile saltpeter were thoughtto be nearing exhaustion, so processes were de-veloped for replacing nitrogen from natural ni-trates with atmospheric nitrogen. Three tech-niques were used industrially:

1) Production of nitrogen monoxide by re-acting atmospheric nitrogen and oxygen at> 2000 ◦C (direct processes)

2) Production of ammonia by hydrolysis of cal-cium cyanamide under pressure

3) Production of ammonia from nitrogen andhydrogen

The direct combustion of air in an electricarc, developed by Birkeland and Eyde, wasabandoned because of its poor energy efficiency.Later direct processes, such as thermal nitrogenmonoxide synthesis with fossil fuels or nitro-gen monoxide synthesis in nuclear reactors, didnot gain widespread acceptance. Ammonia pro-duction from calcium cyanamide was of onlytransitory value. Ammonia produced from nitro-gen and hydrogen by the Haber – Bosch process(→Ammonia) is, however, still used as feed-stock for nitric acid production.

The crucial step in nitric acid production, thecatalytic combustion of ammonia, was devel-oped by Ostwald around the turn of the cen-tury. The most important design parameters fora nitric acid plant were determined first in labo-ratory tests and later in a pilot plant. The first pro-duction facility employing the Ostwald processcame on stream in 1906 at Gerthe in Germany[1–3].

2 Nitric Acid, Nitrous Acid, and Nitrogen Oxides

Most of the nitric acid produced is used toform inorganic fertilizers (→Phosphate Fertil-izers); it is mostly neutralized with ammonia toform ammonium nitrate (→Ammonium Com-pounds, Chap. 1.2.1.).

1.2. Properties

Nitric acid was known to the ancient Egyptiansbecause of its special ability to separate goldand silver. Many well-known alchemists in theMiddle Ages experimented with the acid. In themiddle of the 17th century, Glauber reportedits preparation from saltpeter and sulfuric acid.

Physical Properties. Nitric acid [7697-37-

2], Mr 63.013, is miscible with water in allproportions. At a concentration of 69.2 wt %, itforms a maximum-boiling azeotrope with water.The azeotropic mixture boils at 121.8 ◦C. Pureanhydrous nitric acid boils at 83 – 87 ◦C; the rea-son a range of boiling points are cited in the lit-erature is that the acid decomposes on heating[4]:

4 HNO3 −→ 2 H2O + 4 NO2 + O2

The nitrogen dioxide formed on decompositioncolors the acid yellow or even red at higher con-centrations. Because the vapors can also absorbmoisture, the term “red fuming nitric acid” isused. In the pure anhydrous state, nitric acid isa colorless liquid.

The most important physical properties ofpure nitric acid follow:

fp −41.59 ◦C

bp 82.6± 0.2 ◦C

Density, liquid

at 0 ◦C 1549.2 kg/m3

at 20 ◦C 1512.8 kg/m3

at 40 ◦C 1476.4 kg/m3

Refractive index n24D 1.3970

Dynamic viscosity

at 0 ◦C 1.092 mPa · s

at 25 ◦C 0.746 mPa · s

at 40 ◦C 0.617 mPa · s

Surface tension

at 0 ◦C 0.04356 N/m

at 20 ◦C 0.04115 N/m

at 40 ◦C 0.03776 N/m

Thermal conductivity (20 ◦C) 0.343 W m−1 K−1

Standard enthalpy of formation

Liquid 2.7474 J/g

Gas 2.1258 J/g

Heat of vaporization (20 ◦C) 626.3 J/g

Specific heat

at 0 ◦C 1.7601 J g−1 K−1

at 20 ◦C 1.7481 J g−1 K−1

The decomposition of nitric acid makes itsphysical properties difficult to determine athigher temperature. Up to ca. 50 ◦C, conven-tional methods of measurement are possible; be-yond this, indirect thermodynamic calculationsor special short-time measuring methods mustbe employed. Figure 1 illustrates the vapor pres-sure curve of pure nitric acid. Table 1 lists impor-tant physical properties of aqueous nitric acid. Acomprehensive review of the physical and chem-ical properties of nitric acid is given in [6].

Figure 1. Vapor pressure curve for pure nitric acid

Chemical Properties. Concentrated nitricacid, with nitrogen in the + 5 oxidation state,acts as a strong oxidizing agent. The reaction

NO−3 + 4 H+ NO + 2 H2O

goes to the right for all substances with oxi-dation potentials more negative than + 0.93 V[7]. For example, copper (+ 0.337 V) and silver(+ 0.799 V) are dissolved by nitric acid, whereasgold (+ 1.498 V) and platinum (+ 1.2 V) are re-sistant. In practice, 50 % nitric acid (aqua fortis)


Table 1. Physical properties of aqueous nitric acid as a function of composition [5]

HNO3 concentration, wt % Density (20 ◦C), g/cm3 mp, ◦C bp, ◦C Partial pressure (20 ◦C), kPa

HNO3 H2O

0 0.99823 0 100.0 2.23

10 1.0543 −7 101.2 2.26

20 1.1150 −17 103.4 2.02

30 1.1800 −36 107.0 1.76

40 1.2463 −30 112.0 1.44

50 1.3100 −20 116.4 0.03 1.05

60 1.3667 −22 120.4 0.12 0.65

70 1.4134 −41 121.6 0.39 0.35

80 1.4521 −39 116.6 1.4 0.12

90 1.4826 −60 102.0 3.6 0.03

100 1.5129 −42 86.0 6.0 0

is used for separating gold from silver. Somebase metals, such as aluminum and chromium,are attacked by nitric acid only at their surfacesbecause a thin oxide layer is formed, which pro-tects (i.e., passivates) the core against further ox-idation. As a result of this passivation, alloyedsteel equipment can be used in nitric acid tech-nology.

Highly diluted nitric acid is almost com-pletely dissociated

HNO3+ H2O H3O+ + NO−3

and does not attack copper and more noble met-als. Due to its acid nature, however, it reacts withbase metals, liberating hydrogen and forming ni-trates.

A mixture (volume ratio 3 : 1) of concen-trated nitric acid and concentrated hydrochloricacid (aqua regia) also dissolves gold.

1.3. Industrial Production

The industrial production of nitric acid by theOstwald process is described in this section. Theprocess involves three chemical steps (Fig. 2):

1) Catalytic oxidation of ammonia with atmo-spheric oxygen to yield nitrogen monoxide:

4 NH3+ 5 O2 −→ 4 NO + 6 H2O (1)

2) Oxidation of the nitrogen monoxide productto nitrogen dioxide or dinitrogen tetroxide:

2 NO + O2 −→ 2 NO2 N2O4 (2)

3) Absorption of the nitrogen oxides to yieldnitric acid:

3 NO2+ H2O −→ 2 HNO3+ NO (3)

Figure 2. Ostwald process

The way in which these three steps are imple-mented characterizes the various nitric acid pro-cesses. In monopressure (single-pressure) pro-

cesses, ammonia combustion and NOx absorp-tion take place at the same working pressure.These include medium-pressure (230 – 600 kPa)and high-pressure (700 – 1100 kPa) processes.Very few plants currently employ low pressure(100 – 220 kPa) for both combustion and absorp-tion.

In dual-pressure (split-pressure) processes,the absorption pressure is higher than the com-bustion pressure. Modern dual-pressure plantsfeature combustion at 400 – 600 kPa and absorp-


tion at 900 – 1400 kPa. Some older plants stillemploy atmospheric combustion and medium-pressure absorption.

Intensive work on ammonia combustion aswell as the oxidation and absorption of nitro-gen oxides has been well documented since theearly 1920s. Up to now the combustion of am-monia has been considered as a heterogeneouscatalytic surface reaction with undesirable sidereactions. The oxidation of nitrogen monoxideis a rare example of a homogeneous third-ordergas-phase reaction. The absorption of NOx is aheterogeneous reaction between fluid (i.e., gasand liquid) phases.

1.3.1. Oxidation of Ammonia

The oxidation of ammonia to nitrogen monoxideis described by the equation

4 NH3+ 5 O2 −→ 4 NO + 6 H2O ∆H =−904 kJ/mol (4)

Over a suitable catalyst, 93 – 98 % of the feedammonia is converted to nitrogen monoxide.The remainder participates in undesirable sidereactions leading to dinitrogen monoxide (ni-trous oxide, laughing gas)

4 NH3+ 4 O2 −→ 2 N2O + 6 H2O

∆H =−1105 kJ/mol (5)

and to nitrogen

4 NH3+ 3 O2 −→ 2 N2+ 6 H2O ∆H =−1268 kJ/mol(6)

Other undesirable reactions are the decomposi-tion of the nitrogen monoxide product

2 NO −→ N2+ O2 ∆H =−90 kJ/mol (7)

and its reaction with ammonia

4 NH3+ 6 NO −→ 5 N2+ 6 H2O

∆H =−1808 kJ/mol (8)

Reaction Mechanism. Early theories of themechanism for the oxidation of ammonia dif-fered as to which intermediate product wasformed. The classical theories were proposed byAndrussov (nitroxyl theory, 1926), Raschig

(imide theory, 1927), and Bodenstein (hy-droxylamine theory, 1935) [3]. According tothe hydroxylamine theory, ammonia reacts with

atomic oxygen adsorbed on the catalyst to yieldhydroxylamine. This in turn reacts with molec-ular oxygen to form nitrous acid, which thendissociates to nitrogen monoxide and hydroxylions:

O2 2 O (adsorb.)

O (adsorb.) + NH3 −→ NH2OH

NH2OH + O2 −→ HNO2+ H2O

HNO2 −→ NO + OH−

2 OH− −→ H2O + O

In oxidation experiments under high vacuum,Bodenstein and coworkers detected both hy-droxylamine and nitrous acid as intermediateproducts on the platinum catalyst. Later studieswith a mass spectrometer showed that no suchintermediate products are found above 400 ◦C[8]. In these tests, both molecular and atomicoxygen was found on the catalyst surface. A di-rect reaction between oxygen and ammonia wastherefore postulated according to the followingequations:

NH3+ O2 −→ NO + H2O + H

NH3+ O −→ NO + H2+ H

In fast reactions such as the combustion ofammonia (the reaction time for Eq. 4 is ca.10−11 s), however, the possible intermediatesmost likely cannot be detected because they arevery short-lived and present in very low concen-trations. Nowak [9] has compared the numer-ous theories of ammonia combustion. Pignet

and Schmidt present a classical Langmuir –Hinshelwood treatment of the process [10]. Thepossibility that the combustion of ammonia maybe considered as a homogeneous catalytic reac-tion is now being discussed.

At low temperature (200 – 400 ◦C), the rateof ammonia oxidation is limited by the reac-tion kinetics. The byproducts nitrogen, and es-pecially dinitrogen monoxide, are formed pref-erentially. Between 400 and 600 ◦C, the reactionrate becomes limited by mass transfer, which isthen dominant above 600 ◦C. Control by masstransfer relates to the diffusion rate of ammo-nia through the gas film to the catalyst surface.The product in this reaction regime is nitrogenmonoxide.


The catalyst surface experiences a very lowpartial pressure of ammonia and a diminishedpartial pressure of oxygen due to mass-trans-fer limitations. The catalytic side reaction bet-ween ammonia and nitrogen monoxide is accel-erated by an increase in the partial pressures ofboth species on the catalyst surface. If catalystpoisoning occurs,the number of active centersdecreases and the partial pressure, particularlythat of ammonia, increases at the remaining ac-tive centers. The result is increased formation ofundesirable byproducts. In practice, this shift isoffset by raising the temperature, i.e., the reac-tion is again displaced into the region limited bygas-film diffusion.

Figure 3. Photograph of platinum – rhodium gauze (De-gussa, FRG) taken with a scanning electron microscope (en-largement 100 : 1)A) Initial stage; B) Highly activated stage

Precious-metal catalysts are usually em-ployed in the form of closely stacked, fine-meshgauzes. At the beginning of a campaign theyhave a smooth surface and therefore only slightlylimit mass transfer in the gas phase. This gives aninitially lower yield of nitrogen monoxide. Af-ter a short time in service, however, the surfacearea of the catalyst increases greatly becauseof microstructural changes and interactions withvolatilizing constituents of the catalyst (Fig. 3).Catalyst growth increases the limitation of masstransfer in the gas phase and thus the yield ofnitrogen monoxide.

Reaction Engineering. From an engineer-ing standpoint, the combustion of ammonia isone of the most efficient catalytic reactions(maximum possible conversion 98 %). Accord-ing to the stoichiometry of Equation (4), theammonia – air reaction mixture should contain14.38 % ammonia. A lower ammonia – air ra-tio is, however, employed for a variety of rea-sons, the most important being that conversiondecreases with too high a ratio; ammonia andair also form an explosive mixture. Because themixing of ammonia and air in industrial prac-tice is incomplete, locally higher ratios may oc-cur. A ratio that includes a safety margin be-low the explosion limit is therefore necessary.The explosion limit declines with pressure sothat high-pressure burners can operate with upto 11 % ammonia, whereas 13.5 % ammonia ispossible in low-pressure systems. The explosionlimit also depends on the flow velocity (Table 2,see also [3]) and the water content of the reactionmixture [11].

Table 2. Explosion limits of ammonia – air mixtures at

atmospheric pressure

Flow velocity, Lower explosion Upper explosion

m/s limit, % NH3 limit, % NH3

0 15.5 27.5

3 28 38

5 – 8 30 40

12 32 37

14 none none

According to Le Chatelier’s principle, the in-crease in volume in Equation (4) implies thatconversion declines as pressure rises (Fig. 4, seenext page).

If the gas velocities are too high or the numberof catalyst gauzes is insufficient, conversion to


nitrogen monoxide decreases because of the slip(leakage) of ammonia which reacts with it ac-cording to Equation (8). If the gas velocities aretoo low or too many gauzes are used, decompo-sition of nitrogen monoxide according to Equa-tion (7) is promoted. Table 3 gives typical gasvelocities.

Figure 4. Conversion of ammonia to nitrogen monoxide ona platinum gauze as a funtion of temperature [5]a) 100 kPa; b) 400 kPa

Table 3. Typical design data for ammonia burners

Pressure, Number Gas Reaction Catalyst

MPa of gauzes velocity, temperature, loss, g/t

m/s ◦C HNO3

0.1 – 0.2 3 – 5 0.4 – 1.0 840 – 850 0.05 – 0.10

0.3 – 0.7 6 – 10 1 – 3 880 – 900 0.15 – 0.20

0.8 – 1.2 20 – 50 2 – 4 900 – 950 0.25 – 0.50

Figure 5. Losses of precious metals in the combustion ofammonia to nitrogen monoxide as a function of temperatureand catalyst composition [5]a) Pt; b) Pt – Rh 98/2; c) Pt – Rh 90/10

A high reaction temperature promotes am-monia combustion (Fig. 4), but high tempera-ture decreases the conversion. Maximum tem-peratures up to 950 ◦C have been realized, butcatalyst losses, chiefly due to vaporization, in-

crease greatly (Fig. 5). The usual way to main-tain the conversion toward the end of a campaignin spite of catalyst deactivation and losses, isto raise the temperature. The ammonia – air ra-tio is directly related to the reaction tempera-ture: a 1 % increase in the proportion of ammo-nia increases the temperature by ca. 68 K. If thetemperature is too low, conversion to nitrogenmonoxide decreases. The reaction temperaturealso depends on the temperature of the inlet gasmixture. The following formula provides an ap-proximate value in ◦C:

treactor=tgasmixture + r ·68where r= NH3/ (NH3+air) (9)

Catalysts [12]. Since the introduction of theindustrial production of nitric acid by the Ost-wald process, several hundred materials havebeen tested as catalysts for ammonia combus-tion. Platinum catalysts have proved most suit-able and are used almost exclusively today.Oxides of non-noble metals can also be em-ployed; these cost less than platinum catalysts,but conversion is lower. Another disadvantage ofnon-noble-metal catalysts is that they are morequickly poisoned [13].

The usual form of catalyst is a fine-meshgauze (standardized at 1024 mesh/cm2 ). Man-ufacturers offer wire diameters of 0.060 and0.076 mm. The Johnson Matthey “TailoredPack” combines gauzes made of different wirediameters [14]. The upstream gauzes have a wirediameter of 0.076 mm and the downstream ones0.060 mm to optimize the catalyst surface area.

High-pressure plants also utilize “fiberpacks” in which a fraction of the fibrous cata-lyst material is held between two gauzes [15].The packs have the advantage of reduced plat-inum loss and thus longer run times. “ Fiberpacks” have not yet been used in medium-pres-sure ammonia combustion because fewer gauzesare used in medium-pressure than in high-pres-sure processes (see Table 3), so the lower pres-sure drop results in poorer distribution.

To reduce catalyst consumption, particularlyin gauzes at the bottom of the reactor, ceramicgrids coated with platinum – rhodium have alsobeen developed [16]. Platinum is usually alloyedwith 5 – 10 % rhodium to improve its strengthand to reduce catalyst loss. During combustion,


the metal surface becomes enriched in rhodium,thus improving catalytic activity [17]. Becauserhodium is more expensive than platinum, a rho-dium content of 5 – 10 % has proved optimal.Figure 6 shows the efficiency of conversion tonitrogen monoxide and the platinum loss as afunction of the rhodium content of the alloy [18].

Figure 6. Conversion of ammonia to nitrogen monoxide onplatinum gauze (a) and platinum losses (b) as a function ofcatalyst composition

If a low reaction temperature (< ca. 800 ◦C)is unavoidable, a pure platinum catalyst shouldbe employed, because otherwise rhodium(III)oxide would accumulate at the catalyst surfaceand decrease catalytic activity [18]. Palladium isalso used in catalyst alloys. Laboratory studieson the catalytic activity of 90 % Pt – 10 %Rh and90 %Pt – 5 %Rh – 5 %Pd gauzes reveal no greatdifferences in nitrogen monoxide yield [19]. Theternary catalyst is of economic interest becausepalladium costs much less than platinum or rho-dium.

Platinum gauzes are pretreated by the manu-facturer (Degussa and Heraeus, FRG; JohnsonMatthey and Engelhard, UK) so that they donot need to be activated after installation in theburner. The reaction is ignited with a hydro-gen flame until gauzes glow red. After the re-action has been initiated, crystal growth rapidlyincreases the surface area. The highest conver-sion on the catalyst is obtained after a few daysuse. Specially pretreated platinum gauzes thatreach maximum activity after just a few hoursare also on the market [20]. Start-up problemswith platinum gauzes are discussed in [21]. Theprolonged decline in activity after the conversionmaximum is due to catalyst poisoning and plat-

inum loss. Catalyst poisons include iron oxide(rust) and dust in the process air, which is es-pecially concentrated in the vicinity of fertilizerplants. Any material deposited on the catalystsurface acts as a poison. A multistage filter isused to prevent these substances from reachingthe catalyst (see Section 1.3.3.1).

Platinum loss during operation is caused byvaporization and abrasion. Vaporization losspredominates, but if heavy vaporization lossweakens the structure then abrasion can substan-tially lower burner efficiency. Platinum loss dueto vaporization is thought to involve the forma-tion of short-lived platinum dioxide:

Pt + O2 −→ PtO2(g)

Mechanical losses are accelerated by impuritiesin the reaction mixture, which can lead to theformation of whiskers at the grain boundaries.The whiskers are more easily broken off by theflowing gas.

Recovery of Precious Metals. Precious-metal losses from the catalyst have a significanteffect on the operating costs of a nitric acid plant.The approximate losses and resulting costs canbe determined from the ammonia combustiontemperature and the loading. Losses increasewith temperature and loading; however, smallerlosses may be measured at higher loadings be-cause a more uniform incoming flow causes lessmovement of the gauzes relative to one anotherand thus reduces mechanical losses. Table 3 liststypical loss rates. Platinum recovery systems areinstalled in most nitric acid plants to improveprocess economics.

Mechanical Filters. Mechanical filters aremade of glass, mineral wool, or ceramic or as-bestos fibers. They are normally installed wherethe gas temperature is < 400 ◦C. Their most im-portant drawback is their large pressure drop,which can reach 25 kPa between filter changes.

Mechanical filters recover 10 – 20 % plat-inum, but recovery rates up to 50 % have beenreported [22]. These filters have the advantageof low capital costs and the disadvantage of highoperating costs resulting from the large pres-sure drop across the filter. For economic reasons,they are only employed in mono-high-pressureplants.


In low-pressure burners, marble chips 3 –5 mm in length were used immediately down-stream of the catalyst gauzes [23]. At high re-action temperature, the marble is converted tocalcium oxide, whose high absorption capacityfor platinum and rhodium permits up to 94 % re-covery. During reactor shutdown, however, thecalcium oxide must be protected against mois-ture.

Recovery Gauzes. A technically elegant ap-proach to platinum recovery is to install, justdownstream of the catalyst gauzes, a materialthat is stable at high temperature and that canabsorb platinum oxide vapor and form an al-loy with it. The first recovery gauzes in produc-tion plants (supplied by Degussa in 1968) con-sisted of a gold – palladium alloy (20/80). Nowa-days most such gauzes are gold-free but have alow content of nickel (ca. 5 %) to enhance theirstrength. The gauzes have 100 – 1350 mesh/cm2

and are made of wire with a diameter of 0.2 –0.06 mm. Several recovery gauzes with differ-ent specifications are usually installed. Loss ofplatinum from the catalyst is greater than that ofrhodium; for example, the calculated loss com-position for a 90/10 Pt – Rh gauze is 95/5. A “re-covery factor” can be determined for each recov-ery gauze, indicating how much of the availableplatinum is retained by the gauze. This recov-ery factor depends on the specifications of thegauze, the operating pressure, and the loading.Figure 7 shows the dependence of the recoveryefficiency on wire diameter and mesh fineness[24]. A method for selecting the optimum plat-inum recovery gauze system for a particular ni-tric acid plant is described in [25].

The captured platinum diffuses into the pal-ladium gauze and forms an alloy. The topmostgauze is often designed to reach saturation in onecampaign, usually 180 d. Determining the eco-nomic optimum for a recovery system (the mostimportant criterion is the weight of palladium re-quired) necessitates taking into account the factthat palladium is lost. This loss corresponds toabout one-third of platinum recovery.

“Getter” systems can recover more than 80 %of the platinum lost from the catalyst. The mech-anism of rhodium recovery has not yet been fullyelucidated. Up to 30 % of rhodium from the cat-alyst is captured by recovery systems.

Figure 7. Performance of a palladium-based recovery gauzeas a function of mesh size and wire diameter

1.3.2. Oxidation and Absorption of Nitrogen

Oxides

After the combustion of ammonia, the nitrogenmonoxide product gas is cooled en route to ab-sorption and, if necessary, compressed. As a re-sult, part of the nitrogen monoxide is oxidized tonitrogen dioxide or dinitrogen tetroxide (Eq. 2),which is converted to nitric acid by absorptionin water (Eq. 3).

Absorption is usually performed in platecolumns where the nitrous gases are convertedto nitric acid in the liquid phase on the plates. Ni-trogen monoxide is constantly reformed in thisstep and prevents complete absorption of the in-let gases. Oxidation of nitrogen monoxide oc-curs mainly in the gas phases between the plates,but also in the gas phase of the bubble layer onthem.

Gas Phase. The gas phase is the site of nitro-gen monoxide oxidation

2 NO + O2 −→ 2 NO2 ∆H =−1127 kJ/mol (10)


as well as nitrogen dioxide dimerization

2 NO2 N2O4 ∆H =−58.08 kJ/mol (11)

and dinitrogen trioxide formation

NO + NO2 N2O3 ∆H =−40 kJ/mol (12)

Reaction of the oxidized nitrogen oxides withwater vapor leads to formation of nitrous andnitric acids. These reactions are of secondaryimportance from an engineering standpoint.

To describe the kinetics of nitrogen monoxideoxidation (Eq. 10), a third-order rate equation isused [26]:

r=kp

RTp2NOpO2 (13)

where

r = reaction rate, kmol m−3 s−1

kp = reaction rate constant, atm−2 s−1

R = universal gas constant, m3 atm kmol−1

K−1

T = temperature, Kp = partial pressure, atm

The rate constant is defined by

logkp=652.1

T− 1.0366 (14)

This reaction is unusual because it goes to theright faster at low temperature than at high tem-perature, i.e., the reaction rate has a negativetemperature coefficient.

Because equilibrium in the nitrogen dioxidedimerization system is reached very quickly, anequilibrium formula

r=kp

RT

(

p2NO2 −

pN2O4

Kp

)

(15)

can be assumed where Kp is the equilibrium con-stant. The dimerization rate is virtually indepen-dent of temperature [27], so the rate constant at25 ◦C

kp=5.7× 105atm−1s−1 (16)

can be used at all temperatures of technical in-terest.

Hoftyzer and Kwanten [28] give the fol-lowing equation for the NO2 – N2O4 equilib-rium constant:

Kp=pN2O4

p2NO2

= 0.698× 10−9exp

(

6866

T

)

(17)

An equilibrium formula can also be written fordinitrogen trioxide formation:

r=kp

RT

(

pNOpNO2 −pN2O3

Kp

)

(18)

According to [27], the equilibrium constant canbe determined as

Kp==pN2O3

pNOpNO2

= 65.3× 10−9exp

(

4740

T

)

(19)

Liquid Phase. The large number of reac-tive components in the gas phase (NO, NO2,N2O3, N2O4) means that the reaction model forabsorption is complicated. Figure 8 (see nextpage) shows a sophisticated model devised byHoftyzer and Kwanten [28].

The main route for the formation of nitric acidin this model involves two steps in the liquidphase. First, dissolved dinitrogen tetroxide re-acts with water, yielding nitric and nitrous acids:

N2O4+ H2O −→ HNO3+ HNO2

∆H =−87.0 kJ/mol (20)

Nitrous acid then dissociates to nitric acid, water,and nitrogen monoxide, the latter being trans-ported across the interface into the bulk gas:

3 HNO2 −→ HNO3+ H2O + 2 NO

∆H =−15.3 kJ/mol (21)

According to Andrew and Hanson [29], a first-order equation can be written for the rate of dini-trogen tetroxide hydrolysis:

r=kcN2O4(22)

where cN2O4is the concentration of dinitrogen

tetroxide in kmol/m3. The rate constant is givenby

logk= − 4139

T+ 16.3415 (23)

The kinetics of nitrous acid dissociation werestudied by Abel and Schmid [30], with the fol-lowing result:

r=kc4HNO2

p2NO

=k

H2NO

c4HNO2

c2NO

(24)


Figure 8. Nonstoichiometric model of absorption of nitrogen oxides in water [28]

where HNO is the Henry coefficient for nitrogenmonoxide in m3atm/kmol and c ist the concen-tration in kmol/m3. The rate constant is definedby

logk= − 6200

T+ 20.1979 (25)

Equations (23) and (25) were derived by Hoff-

mann and Emig [31] and are based on experi-mental data from [32].

Mass Transfer. For more than 50 years sci-entists have studied the transport of nitrogen ox-ides in water and in dilute and concentrated ni-tric acid. A variety of mechanisms have beenproposed depending on the gas composition andthe acid concentration. These are partly in agree-ment with the absorption model of Figure 8.

In the NOx absorption region important foracid formation, dinitrogen tetroxide transport isconsidered to be the rate-limiting step. The quan-tity of dinitrogen tetroxide transported from thebulk gas to the interface depends chiefly on theNO2 – N2O4 equilibrium [28,33]:

JN2O4=

kgNO2

2RT

(

poNO2 − pi

NO2

)

+kgN2O4

RT

(

poN2O4

− piN2O4

)

(26)

where

J = absorption rate, kmol m−2 s−1

kg = gas-side mass-transfer coefficient, m/spo = partial pressure in the bulk gas, atmpi = partial pressure at the interface, atm

At higher NO2 /N2O4 concentrations in the reac-tion gas, primarily dinitrogen tetroxide crossesthe interface and reacts in the fast first-order re-action (Eq. 20) to form nitric and nitrous acids.Nitrous acid dissociates (Eq. 21), and the result-ing nitrogen monoxide is transported back intothe gas space.

The following equation can be written for theabsorption rate of dinitrogen tetroxide [34]:

JN2O4=HN2O4

pN2O4

√

kDN2O4(27)

where

H = Henry coefficient, m3 atm/kmolk = rate constant, s−1

D = diffusion constant, m2 /s

Numerous measurements in laboratory ab-sorbers support this mechanism. Interpretationsof measurements in absorption columns havegiven the following correlations for the mass-transfer coefficients [35]:

1) Bubble-cap trays


2) Sieve trays

where

H = Henry coefficient, m3kPa/kmolT = temperature, KW = mass fraction

The absorption of NO2 – N2O4 in concen-trated nitric acid can be treated as pure phys-ical absorption [36]. Deviations from the de-scribed chemisorption also occur at low gas-phase NO2 – N2O4 concentrations [29].

Absorption Equilibrium. Equilibrium con-siderations allow determination of the maximum

possible acid concentration for a given compo-sition of the inlet gas. For example, suppose theoverall reaction is

3 NO2 (g) + H2O (l) 2HNO3(l) + NO (g)

∆H =− 72.8 kJ/mol (30)

and the equilibrium constant is defined as

Kp=p2HNO3pNO

p3NO2

pH2O

(31)

with the partial pressures of nitrogen monox-ide ( pNO) and nitrogen dioxide ( pNO2

) in thegas phase and the vapor pressures of nitric acid( pHNO3

) and water ( pH2O) over the dilute acid.In industrial absorption the equilibrium constantdepends only on temperature. The equilibriumconstant is conventionally split into two terms:

Kp=K1 ·K2

whereK1=pNO

p3NO2

and K2=p2HNO3

pH2O

(32)

K2 depends on the acid concentration and thetemperature. Gases ( N2O4, N2O3, HNO2 ) dis-solved in the dilute acid affect the activities of

Figure 9. Vapor pressures of nitric acid and water as a func-tion of acid strength and temperature [28]1mm Hg = 133.2 mPa

Figure 10. Value of K1 from Equation (32) as a function ofacid concentration and temperature [28]


nitric acid and water. Up to a nitric acid concen-tration of 65 wt % and an inlet nitrogen oxidepartial pressure of 100 kPa, however, the vaporpressures from the binary system HNO3 – H2Ocan be used to determine K2 (Fig. 9).

Figure 10 shows how K1 depends on theacid concentration and the temperature. This pa-rameter is particularly interesting from an en-gineering standpoint because it allows the acidstrength to be determined if the gas concentra-tion is known.

If the equilibrium is based on the overall re-action

3/2 N2O4(g) + H2O (l) 2 HNO3(l) + NO3 (g)

∆H =− 75 kJ/mol (33)

the equilibrium constant is

Kp=K1 ·K2

whereK1=pNO

p3/2

N2O4

and K2=p2HNO3

pH2O

(34)

Again Figure 9 gives the K2 values. In thisformulation, K1 is a function of the acid con-centration and does not depend on temperature(Fig. 11). The curve of Figure 11 can be fittedby an empirical equation [37]:

logK1=7.412 − 20.29WHNO3 + 32.47W 2HNO3

−30.87W 3HNO3 (35)

where W denotes the mass fraction.

Figure 11. Value of K1 from Equation (34) as a function ofacid concentration

Another way of describing the absorptionequilibrium is to start with the hypothetical com-ponent NO2 denoted as “chemical NO2”:

p ˜NO2=pNO2 + 2pN2O4

(36)

The nitrogen dioxides, considered as completelydissociated, are assumed to react according to

This hypothetical reaction leads to the equilib-rium relation

Kp=K1 ·K2

whereK1=pNO

p3˜NO2

and K2=p2HNO3

pH2O

(38)

The equilibrium constant Kp can be estimatedwith the following equation [38]:

logKp= − 7.35+2.64

T(39)

Then K1 can be determined from Kp, and the K2

values obtained from Figure 9. The parameterK1 determines the equilibrium curves for con-stant temperature and constant acid concentra-tion when pNO2

is plotted against pNO (Fig. 12).The K1 values are necessary to calculate the ab-sorption of nitrous gases according to Toniolo

[39]. The application of the Toniolo diagram isdescribed in [40].

Modeling of the Absorption Tower. The lit-erature offers many methods for calculating theabsorption of nitrous gases; they can be classi-fied according to the number of reactions, thereaction kinetics, or the type of mathematicalmethod:

1) Number of Reactions. The simplest calcula-tion model uses Equation (30) to describe theevents on an absorption plate and a kineticformula to describe nitrogen monoxide ox-idation between two oxidation plates [41].Extended models [42–45] also take into ac-count the dimerization of nitrogen dioxide todinitrogen tetroxide. In this way it is possi-ble to describe nitric acid formation not justvia nitrogen dioxide but also via dinitrogentetroxide [46–49].Models that include the formation and dis-sociation of nitrous acid in the liquid phasegive the best description of known chemicalevents [31,50]. A model that permits nitrousacid formation in the gas phase has also beenproposed [51].


Figure 12. Toniolo diagram for an absorption tower [39]A→B absorption process; B→C oxidation process

2) Reaction Kinetics. Most design computa-tions use equilibrium equations for the ba-sic reactions. These static methods have thedrawback of requiring the introduction of ef-ficiency factors [51] that take into accountlimitation of transport of the reactants acrossthe interface. Dynamic methods do not needempirical efficiency factors because they de-scribe mass transfer in terms of physical andchemical laws [28,31,50,52].

3) Mathematical Methods. In view of the largenumber of nonlinear equations to be solved inboth equilibrium and dynamic models, plate-to-plate calculations are generally describedin the literature. The relaxation method [31]and a special stage-to-stage calculation [52],however, have also been used for the designof absorption towers.

By way of example, an equilibrium modeland a transport model are described in more de-tail.Equilibrium Model.

The equilibrium model is based on Equations(11) and (33). If a gas of known composition isfed to the plate and an acid of known strengthleaves the plate, the calculation begins at the bot-tom of the tower. If the gas throughput is takento be constant (because of the high content ofinerts), the following mass balance is obtainedfor the n-th plate:

3(

yn,NO − yn−1,NO

)

=(

yn−1,NO2+2yn−1,N2O4

)

−(

yn,NO2+ 2yn,N2O4

)

(40)

where y is the mole fraction in the vapor phase.Equilibrium Equations (34) for Kp

1 and (17)for Kp2 give the equation

ay3n,NO2

+ by2n,NO2

+ yn,NO2+ d=0

with a=3 (Kp1

√p) (Kp2p)1.5

b=2 (Kp2p)

d= −(

3yn−1,NO + yn−1,NO2+ 2yn−1,N2O4

)

(41)

This formula allows the compositions of the gasleaving the n-th plate and the acid arriving at then-th plate to be determined.

Before the (n + 1)-th plate is calculated, theoxidation of nitrogen monoxide must be deter-mined. The flow behavior of the nitrous gasesin an absorption tower is not known exactly, sothe oxidation space can be described as either atubular or a stirred-tank reactor. Actual condi-tions probably lie somewhere between the two[28], but trends can be identified on the basis ofplate type. The stirred-tank model is preferredfor bubble-cap trays [47], but flow-through con-ditions can be expected in the case of sieveplates. Because virtually all absorption towerstoday are fitted with sieve plates, only these arediscussed.

When the reaction is carried out isothermally,integration of the rate equation

−dpNO

dt=kp2

NOpO2 (42)

gives a solution

θ =2

k (2pO2 − pNO)· (43)

{

x

pNO(1− x)− 1

2pO2 − pNO

ln

[

2pO2 − pNOx

2pO2(1− x)

]}


where

θ = residence time, sk = rate constant, atm−2 s−1

p = partial pressure, atmt = time, sx = fraction of oxidized nitrogen monoxide

in which x must be determined by iteration fromthe residence time θ and the inlet partial pres-sures of nitrogen monoxide ( pNO) and oxygen( pO2

) [47]. The partial pressures of the othercomponents leaving the reactor can be calcu-lated from the fraction x of oxidized nitrogenmonoxide.

If an excess of oxygen is assumed to bepresent, the rate equation can be simplified [53]as

−dpNO

dt=kp2

NO (44)

so that the partial pressure of nitrogen monoxideat the reactor outlet ( p′NO) can be determined:

1

p′NO

=1

pNO

+ (2pO2 − pNO)kθ

2(45)

Because the oxidation reaction is highly exother-mic, the tubular reactor can be designed forisothermal operation only when conversion isvery low (e.g., when the tail gas contains littleNO). Otherwise, adiabatic reaction conditionsmust be assumed. To describe the reaction un-der adiabatic conditions the oxidation space isdivided into n layers, in each of which the re-action is isothermal. The temperature increasefrom one layer to the next is given by [33]

∆T =1/2CNO,0 (−∆HR)

%0Cp

GNO,0 −GNO

GNO,0(46)

where

CNO, 0 = initial concentration of NO, kmol/m3

Cp = molar heat capacity, kJ kmol−1 K−1

GNO = molar gas flow rate, kmol/sGNO, 0 = initial gas flow rate, kmol/s∆HR = reaction enthalpy, kJ/kmol%0 = initial density, kg/m3

Transport Model. A transport model for the ab-sorption tower is based on a series of units,each containing a bubble- column reactor and anadiabatic plug flow reactor (Fig. 13) [52]. Thebubble-column reactor, modeled as two ideal

stirred-tank reactors, simulates the liquid withgas bubbles on a tray (absorption space). Theadiabatic plug flow reactor simulates oxidationbetween plates (oxidation space).

Figure 13. Hydrodynamic model of absorption towerg = molar gas flow rate of a component; G = molar flow rateof gas; L = molar flow rate of liquid; p = pressure; T = tem-perature; Tcw = temperature of cooling water; V = volume;x = mole fraction in liquid phase; y = mole fraction in gasphase

The balances are carried out with the equa-tions

Gnyn,i − Gn−1yn−1,i − kg,iaVpn

RTn

(

y∗n,i − yn,i

)

= Vg

Mg∑

j=1

νg,ij rg,j (47)

i = N2, O2, NO, NO2, N2O4 andk∑

i=1

yn,i= 1

for the components in the gas phase (subscriptg) where

a = interfacial area per unit volume, m2 /m3

G = molar gas flow rate, kmol/skg = gas-side mass-transfer coefficientMg = number of reactions in the gas phasen = plate numberr = reaction rate, kmol m−3 s−1

V = volume of bubble layerVg = volume of gas phase in bubble layeryi = mole fraction of i-th component in liq-

uid phase


Figure 14. Equipment needed in a nitric acid plant

y ∗i = mole fraction of i-th component in liq-uid phase at the phase boundary

ν = stoichiometric coefficient

Analogously for the liquid phase (subscript l)

Lnxn,i − Ln+1xn+1,i + kl,iaV cn(

x∗n,i − xn,i

)

±Ln xn,i=Vl

Ml∑

j=1

νl,ij rl,j (48)

i = NO, N2O4, HNO2, HNO3, H2O andk∑

i=1

xn,i = 1

where

c = molar concentration, kmol/m3

L = molar liquid flow rate, kmol/sx = mole fraction in liquid phaseL = side streamM l = number of reactions in the liquid phase

Thermodynamic equilibrium is assumed at theinterface,

y∗n,i=Hn,ix∗

n,i (49)

where H is the Henry coefficient and the formula

kg,ipn

RTn

(

yn,i − y∗n,i

)

=kl,icn(

x∗n,i − xn,i

)

i= NO, N2O4 (50)

is used for dynamic mass transfer betweenphases. A separate heat balance is performed for

each phase, so that the temperature may differ inboth phases of a single stage. The mathematicaltreatment of this model is described in [52].

1.3.3. Equipment

Figure 14 summarizes the equipment needed toimplement the three chemical steps involved innitric acid production (Fig. 2). In addition to pro-ducing nitric acid, a nitric acid plant also gener-ates steam or mechanical energy from the chem-ical energy liberated.

1.3.3.1. Filters and Mixers

To obtain the highest possible conversion of am-monia on the catalyst and a long catalyst life,the starting materials (air and ammonia) mustbe purified. The air must be very clean to pre-vent catalyst poisoning ( page 6).

Air Filter. A multistage filter is normallyused on the air intake to a nitric acid plant andshould remove 99.9 % of all particles larger than0.5 µm.

Typical filter media are plastic and glassfibers. Filter frames should be made of stain-less steel. Air filters must be replaced regularlybecause filter bags can tear or filter mats can be-


come overloaded and cause an excessive pres-sure drop. Filter life depends on the particulateload in the air (typically 0.8 mg/m3 ). In areaswhere sandstorms may occur (maximum load-ing 500 mg/m3 ), the use of a sand separator(a centrifugal collector) as a prefilter is recom-mended.

Ammonia Filters. The liquid ammonia fil-ter removes solid contaminants, especially smallrust particles; 99.9 % of all particles larger than3 µm should be eliminated. Selection of a liq-uid ammonia filter requires consideration of thefact that traces of oils and chlorides are alsopresent in ammonia. Proven filter materials areTeflon and sintered metals. Polypropylene andceramic filter candles (cartridges) are also em-ployed. Magnetic filters have found some usebut have limited capacity.

Filtration of ammonia gas should remove99.9 % of oil and solid particles larger than0.5 µm. The principal filter media are glassfibers, sintered metals, and ceramics. The pres-sure drop is up to 10 kPa, depending on filtertype.

Filters for the Mixed Gas. The mixed-gasfilter provides final cleaning of the ammonia –air mixture and improves the mixing of the airand ammonia; it should remove 99.8 % of allparticles larger than 1.5 µm. Contaminants in thegas mixture not only originate externally (in theprocess air and ammonia) but are also formed bycorrosion inside the system (rust). The gas mix-ture filter should therefore be installed as near aspossible to the burner. Ceramic filter cartridgesare generally used in which silicon dioxide is thedominant constituent. For a filter cartridge witha surface loading of 250 m3 /m2, the maximumpressure drop in the clean state is 10 kPa.

Mixers. Static gas mixers are used for twopurposes in nitric acid plants:

1) To mix ammonia and air in a ratio of ca. 1 : 10for catalytic oxidation (see Section 1.3.1)

2) To mix ammonia and tail gas in a ra-tio of 1 : 100 for the catalytic reduction ofNOx as part of tail-gas treatment (see Sec-tion 1.4.2.3).

From an economic standpoint, homogeneousmixing of the gas streams upstream of the reac-tor is important in both cases. Local ammonia

excesses in the burner are a risk for plant safety(explosion limit) and may also cause overheat-ing of the catalyst gauze. Poor mixing lowers theconversion of ammonia to nitrogen monoxideand increases platinum loss from the catalyst.

The efficiency of a given mixing operation isdescribed by the standard deviation σ of mea-sured samples from the theoretical value x a cer-tain distance downstream from mixing. In nitricacid plants, the deviation should be < 1 %. If thetheoretical ammonia concentration in a mixtureis 10 %, for example, measured values fluctuatebetween 9.9 and 10.1 % [54]. Only static mix-ers are used (e.g., the Uhde tubular mixer or theSulzer three-element mixer).

1.3.3.2. Burners and Waste-Heat Boilers

Figure 15. Reactor for catalytic ammonia oxidation withwaste-heat recovery system (Lentjes)a) Burner head; b) Perforated plate; c) Platinum gauzes andplatinum recovery gauzes; d) Inspection glass; e) Super-heater and evaporator tubes; f) Hydrogen ignition; g) Re-fractory packing; h) Nitrous gas outlet


Figure 16. Water and steam connections of heat exchangers (Lentjes) downstream of catalysisa) Reactor; b) Economizer; c) Steam drum; d) Boiler feedwater pumps; e) Preevaporator; f) Upper superheater; g) Lowersuperheater; h) Main evaporator

The heat of reaction liberated during am-monia combustion is utilized to produce steamand preheat the tail gas. Steam is produced in awaste-heat boiler located immediately below theburner (Fig. 15). The burner consists of a burnerbasket packed with filling material to ensure uni-form distrubution of the downward flowing reac-tion gas. Platinum recovery gauzes ( page 7) arelocated above the filling material. The catalystgauzes ( page 6) are located above the recoverygauzes. The burner basket is clamped betweenthe flanges of the burner head and the waste-heat boiler. To ensure better distribution of thereaction mixture, the burner head also containsa perforated plate or honeycomb grid. Hydro-gen burners rotating above the surface are of-ten used for ignition. The gauze temperature ismeasured, and the space above the gauze can beobserved through inspection glasses. The gauzesglow bright red during catalysis. The throughputper element can produce up to 1200 t of 100 %nitric acid per day.

Waste-heat boilers up to 6 m in diameter areused. Figure 15 shows a waste-heat boiler formedium-pressure burning. The preevaporator islocated directly below the burner basket and pro-

tects the downstream superheater against exces-sive temperature. The superheater is followed bythe main evaporator. Evaporator tubes are alsoinstalled at the wall to cool it and protect it fromexcessive temperature.

Figure 16 shows how heat from the waste-heat boiler is used to raise steam. Boiler feed-water is led through the economizer (b) into asteam drum (c). Water at thermal equilibriumwith the steam is pumped from the drum throughthe evaporators. Steam from the drum goes firstto the lower superheater (g) and is then cooledin the drum so that the desired temperature canbe attained in the upper superheater (f).

Systems with an economizer integrated in thewaste-heat boiler have also proved serviceable(Fig. 17, see next page). Here, as in the waste-heat boiler of Figure 15, the heat-exchangertubes can be arranged spirally in disk form orin a square pack.

A shell-and-tube evaporator with natural cir-culation can also be used (Fig. 18, see next page).The superheater tubes (e) are again arranged asa flat spiral. These waste-heat boilers are partic-ularly suitable for smaller plants or those usinglow steam pressure.


Figure 17. Reactor for catalytic ammonia oxidation with in-tegrated waste-heat recovery system (Steinmuller)a) Burner head; b) Perforated plate; c) Platinum gauzes andplatinum recovery gauzes; d) Inspection glass; e) Super-heater and evaporator tubes; f) Feedwater preheater; g) Ni-trous gas outlet

Figure 18. Reactor for catalytic ammonia oxidation with in-tegrated waste-heat recovery system (Steinmuller)a) Burner head; b) Perforated plate; c) Platinum gauzes;d) Packing; e) Superheater tubes; f) Evaporator; g) Nitrousgas outlet

Stress cracking corrosion is a special threatto economizer tubes. Start-up always leads tothe formation of ammonium salts; nitrous andnitric acids may also condense. To prevent veryrapid corrosive attack, the equipment is there-fore heated before start-up and the feedwater ispreheated to a high temperature.

Waste-heat boilers can be designed for pres-sures up to 10 MPa and temperatures up to550 ◦C in the superheater. The steam is used todrive a turbine or is exported.

1.3.3.3. Compressors and Turbines

Machines on a common shaft are used to deliverand compress the gases and supply the drivepower needed for this purpose. Power sourcesare the tail-gas turbine and a steam turbine orelectric motor. The tail-gas turbine can supply35 – 100 % of the compression energy requiredfor the process, depending on the degree of pre-heating; the remainder comes from the steam tur-bine. The machinery used depends on the nitricacid process and may consist of an air compres-sor, a nitrous gas compressor, a tail-gas turbine,and a steam turbine or an electric motor.

Figure 19 shows the nitric acid processwith atmospheric combustion and medium-pres-sure (400 – 600 kPa) absorption. The nitrousgas compressor (b) sucks the reaction mixturethrough the burner and simultaneously providesthe pressure needed for absorption. This com-pressor does not have interstage cooling and canbe of radial or axial design. The tail gas fromthe absorption column is expanded in a tail-gasturbine (d), usually a single-stage device. Therest of the power required to drive the nitrousgas compressor is supplied by the steam turbineor an electric motor.


Figure 19. Flow sheet of a nitric acid process with atmospheric combustion and medium-pressure absorptiona) Reactor with waste-heat recovery system; b) Nitrous gas compressor; c) Absorption tower; d) Tail-gas turbine; e) Steamturbine; f) Cooling system; g) Heat exchanger network

Figure 20. Flow sheet of a nitric acid process with medium-pressure combustion and medium-pressure absorptiona) Air compressor; b) Reactor with waste-heat recovery system; c) Absorption tower; d) Tail-gas turbine; e) Steam turbine;f) Heat exchanger network

Figure 20 shows a medium-pressure pro-cess (400 – 600 kPa), as preferred for small-or medium- capacity plants (≤ 600 t/d 100 %HNO3 ). The machinery for this process includesa radial or axial air compressor (a) without in-terstage cooling, a single-stage tail-gas turbine(d), and a steam turbine (e) or electric motor.

A dual-pressure process with medium-pres-sure combustion and high-pressure absorptioncalls for both an air compressor and nitrous gascompressor (Fig. 21). The air for combustion isdelivered at 400 – 600 kPa by an uncooled radialor axial air compressor (a). The radial nitrousgas compressor (c) produces the 1.0 – 1.2 MPapressure needed for absorption. The drive power

for these two devices comes from a multistagetail-gas turbine (e) and a steam turbine (f) orelectric motor. These systems are preferred forlarge- capacity plants (≥ 400 t/d 100 % HNO3 ),especially in Europe.

High-pressure plants (Fig. 22) are preferredin the United States because of lower investmentcosts. The machinery includes an air compres-sor with interstage cooling. Radial devices aregenerally used, but the low-pressure section inlarger machines may be axial. The tail-gas tur-bine is usually a multistage reaction machine.

The nitrous gas compressors are virtually al-ways of radial design, but axial machines areavailable. The design and operation of the com-


Figure 21. Flow sheet of a nitric acid process with medium-pressure combustion and high-pressure absorptiona) Air compressor; b) Reactor with waste-heat recovery system; c) Nitrous gas compressor; d) Absorption tower; e) Tail-gasturbine; f) Steam turbine; g) Cooling system; h) Heat exchanger network

Figure 22. Flow sheet of a nitric acid process with high-pressure combustion and high-pressure absorptiona) Air compressor; b) Reactor with waste-heat recovery system; c) Absorption tower; d) Tail-gas turbine; e) Steam turbine;f) Heat exchanger network; g) Interstage cooler

pressors must take particular account of the fol-lowing properties of the gas stream:

1) Nitrites and nitrates may be formed2) The nitrous gas is toxic and highly corrosive3) Chemical reactions take place when nitrous

gas is compressed

When the nitric acid plant is started up orin continuous operation, unreacted ammonia ispresent in the burner exit gas. The level in con-tinuous operation is ca. 30 ppm of ammonia. Asa result, nitrates and nitrites would accumulatein the nitrous gas compressor if they were notremoved regularly. If the temperature exceeds240 ◦C, combustion or explosion is possible. Toavoid an explosion, water or steam is brieflysprayed into the suction of the nitrous gas com-pressor every 8 h [55]. Because nitrous gas is

highly toxic, it must not be allowed to escapefrom the compressor. The machine is thereforeequipped with air-purged labyrinth seals. Chem-ical reactions must be considered in the designof nitrous gas compressors. The endothermicdissociation of dinitrogen tetroxide to nitrogendioxide removes heat from the stream [56] andcan cool the gas by 20 ◦C or more.

Because of their large capacities, modern ni-tric acid plants usually employ axial air com-

pressors. The possibility of surging in the com-pressor, especially during plant start-up, requiresthe installation of appropriate surge protection.Adjustable inlet guide vanes allow operation atvarious plant capacities.

Tail-gas turbines can be single-stage or mul-tistage devices of axial or radial design. For part-


Figure 23. Ammonia evaporatora) Evaporator tubes; b) Sparger; c) Drop separator

load operation, these machines are equippedwith a group of valve-controlled nozzles or ad-justable guide vanes in the inlet.

Steam turbines may be condensing or back-pressure machines with or without an extrac-tion/side stream.

1.3.3.4. Heat Exchangers and Columns

Heat Exchangers. The shell-and-tube de-sign predominates for tail-gas heat exchangers,with hot nitrous gas on the tube side and cool tailgas on the shell side. The pressure drop shouldbe kept as low as possible to secure a favorableenergy balance for the plant as a whole.

Ammonia Evaporator. Various types ofequipment can be used for ammonia evapo-ration. Figure 23 shows a shell-and-tube heatexchanger with mist collector. Cooling wateron the tube side is cooled on evaporation. Inmedium-pressure plants, the heat of vaporiza-tion is also taken from chilled water used ascoolant in the absorption step. The shell-sideheat-transfer coefficients for ammonia can becalculated according to [57].

Gas Cooler – Condenser. In the cooler –condenser, the temperature is lowered below thedew point of the inlet nitrous gas. The nitric acid

condensate has a concentration of, for exam-ple, ca. 40 wt % at medium pressure. Suitablematerials of construction are discussed in Sec-tion 1.3.3.5. The thermal design of a gas cooler –condenser is difficult because the cooling stepinvolves chemical reactions in both the gas andthe condensed liquid phases. In the apparatusshown in Figure 24, nitrous gas passes into theshell side through two inlets; each of the par-tial streams reverses direction four times. Heatis absorbed by cooling water.

Absorption Tower. Absorption tower designis governed by the calculations for the oxida-tion process (i.e., tower height, see also Sec-tion 1.3.2) and thermal design. Because oxida-tion proceeds more slowly as the NOx content inthe tower decreases, the spacing between platesincreases. Figure 25 shows an absorption towerfor a plant with a daily production capacity of1830 t of 100 % nitric acid. Acid formation takesplace chiefly in the bottom third of the tower,whereas NOx is reduced in the upper two-thirds.As a result, most of the heat must be withdrawnin the lower third; this is done with cooling coilson the plates. The coils on the upper plates pri-marily cool the nitrogen monoxide gas againstchilled water, because oxidation proceeds fasterat lower temperature. Most modern absorptiontowers have sieve plates. Compartments builtinto the bottom of the tower separate acid ofmedium strength from weak acid in case of plant


Figure 24. Cooler – condenser with feedwater preheaters attached to nitrous gas inlets

Figure 25. Absorption towera) Nitrous gas inlet; b) Inner compartment; c) Outer com-partment

shutdown. The collected acids are pumped backto the tower at restart so that a steady state isreached more quickly. Before the tail gas leavesthe tower, entrained acid droplets are collectedin the demister.

Bleacher. The NOx gases contained in theacid are stripped out with secondary air in thebleacher. A distribution tray delivers the acidevenly onto the Raschig ring packing. The pack-ing rings are made of stainless steel, materialno. 1.4301 (AISI 304) or 1.4306 (AISI 304L).

1.3.3.5. Construction Materials

Because of the special behavior of nitric acid to-ward metals (see Section 1.2), materials of con-struction for nitric acid plants must be selectedvery carefully. When industrial production be-gan, acid-resistant masonry linings were usedfor plant components that came in contact withthe product. By the 1920s, materials technologyhad advanced to the point that high chromiumcontents could be incorporated into alloy steels.Today, the principal materials of construction innitric acid plants (< 70 % HNO3 ) are austeniticchromium – nickel steels containing 18 % chro-mium and 10 % nickel. The corrosion resistance


Table 4. Chemical analyses (wt %) of standard and special stainless steels [59]

Type of steel C max. Si max. P max. S max. Cr Ni Other

AISI 304L 0.035 0.75 0.040 0.030 18 – 20 8 – 13

321 0.08 0.75 0.040 0.030 17 – 20 9 – 13 Ti

347 0.08 0.75 0.040 0.030 17 – 20 9 – 13 Nb

ISO 1.4306 0.03 1.0 0.045 0.030 17 – 20 10 – 12.5

1.4541 0.10 1.0 0.045 0.030 17 – 19 9 – 11.5 Ti

1.4550 0.10 1.0 0.045 0.030 17 – 20 9 – 11.5 Nb

Sandvik 3R12∗ 0.030 0.60 0.030 0.030 18.5 10.5

2R12∗ 0.020 0.10 0.015 0.010 18.5 11

2RE10∗∗ 0.020 0.30 0.020 0.015 24.5 20.5

∗ ISO material no. 1.4306. ∗∗ISO material no. 1.4335.

of these steels decreases with increasing tem-perature. Normal austenitic steels are generallynot stable above 70 %. The chromium contentdetermines the corrosion resistance; its anticor-rosive action is influenced by the carbon con-tent of steel. The chromium in the alloy formsa carbide (Cr3C2 ) at the grain boundaries; onlysmall amounts of free chromium remain at highcarbon contents, and the steel is therefore lesscorrosion resistant. If excess chromium remainsafter the amount needed for carbide formationis consumed, only this excess is responsible forthe anticorrosion properties.

Figure 26. Anodic polarization curve for stainless steelCr denotes the effect of chromium on the anodic polarizationrate

The effect of chromium on the corrosion rateis shown in the anodic polarization curve of Fig-ure 26. When chromium is present in the alloy,

the passive region Ep (see Section 1.2) is reachedsooner. Furthermore, chromium lowers the cor-rosion rate in the passive state and raises thetranspassive potential Etr. The last point is im-portant because the corrosion potential of stain-less steel in strong (highly concentrated) nitricacid often reaches the limit of the transpassivepotential. In the transpassive range the passivelayer starts to dissolve. A steel with too lit-tle chromium or with elements that lower thetranspassive potential tends to corrode at a highrate [58].

Steels of a given specification behave differ-ently under nitric acid attack because of per-mitted batch-to-batch variations in composition.The only way to be sure that a selected steel issuitable for an apparatus in contact with nitricacid is to perform tests. The best known methodis the Huey test, in which a clean, polished pieceof steel is immersed in boiling 65 % nitric acid atambient pressure [59]. The specimen is exposedto these conditions for five successive 48-h pe-riods, the acid being renewed at the beginningof each test period. The Huey test gives a goodindication of how individual alloy constituentsaffect intergranular corrosion.

To ensure good weldability, the carbon con-tent is limited to 0.03 %; alternatively, titaniumor niobium is added as a stabilizer.

The stainless steels generally used in nitricacid plants are those classified unter materialno. 1.4306 (AISI 304L); in the Federal Republicof Germany, however, the corresponding stabi-lized qualities, material no. 1.4541 (AISI 321)and 1.4550 (AISI 347), are also employed (Ta-ble 4). These steels are usually specified for tow-ers and heat-exchanger shells, and also for manyheat-exchanger tubes. The 1.4335 steel (AISI310L) has a very low corrosion rate and is used


particularly where corrosive attack is severe, forexample, if the temperature is below the dewpoint of nitrous gas, even at higher temperatures[59]. These conditions normally occur at the in-let to the tail-gas preheater, in the dew point re-gion of the cooler – condenser, and at the outletof the feedwater preheater. Corrosion at theselocations is caused mainly by reevaporation ofnitric acid condensate, which brings the acidconcentration up to the azeotropic level (69 %HNO3 ). The acid is then very aggressive, even inthe vapor phase. Corrosion can take place in theboiler feedwater preheater if condensate formsat the outlet. Reheating of the condensate by thehot process gas can lead to critical conditions.Process design parameters should normally beselected so as to avoid condensation; materialno. 1.4306 is then adequate. If, however, con-densation or reboiling occurs and this steel can-not be used, 1.4335 should be employed [60].

In plants producing concentrated nitric acid,aluminum (99.8 %), ferrosilicon, tantalum, andspecial austenitic steels are used. Because tanta-lum is very expensive, it is employed only withboiling concentrated nitric acid. Ferrosilicon canbe used only in castings because of its brittle-ness.

1.3.4. Processes

Figure 14 illustrates the steps necessary for im-plementation of the Ostwald process. Industrialprocesses differ in the sequence and design ofthese steps. This section describes processes forthe production of weak acid [61, pp. 61 – 98] andstrong acid [61, pp. 99 – 130].

1.3.4.1. Weak Acid Processes

The first industrial plants for the production ofweak acid employed atmospheric combustionand low-pressure absorption [3]. This type ofplant is no longer built. It was followed by aprocess employing atmospheric combustion andabsorption at medium pressure. This process hasthe advantage of lower absorption costs becausethe marked improvement in absorption obtainedwith the pressure generated by a nitrous gas com-pressor allows the apparatus to be downsized.

The energy of compression can be partially reco-vered with a tail-gas turbine. To circumvent thedifficulty of operating a compressor for nitrousgas and obtain a further decrease in plant vol-ume, to ammonia was also oxidized at mediumpressure. This type of process is particularly eco-nomical for smaller capacities and is thus de-scribed first.

As plant capacities continued to grow andlower tail-gas levels of NOx were necessary theabsorption pressure was increased still further.Lower energy costs and shorter depreciation pe-riods in the United States have led to a pref-erence for the high-pressure process, whereasmore favorable consumption and production fig-ures in Europe have favored the dual-pressureprocess with medium-pressure combustion andhigh-pressure absorption.

The principle of medium- and high-pressure(monopressure) plants is that air for ammoniaoxidation and NOx absorption is compressed tothe desired pressure. The “primary” air for oxi-dation is then mixed with ammonia and forcedthrough the burner. The “secondary” air for strip-ping dissolved NOx out of the raw acid and ox-idizing the intermediate nitrogen monoxide issupplied upstream of the absorption tower. Thetail gas is heated and then expanded in a tur-bine (on a common shaft with the compressor)to produce mechanical energy. The remainingdrive power is supplied by a steam turbine run-ning on process steam.

Medium-Pressure Process. Figure 27 is aflow sheet of a medium-pressure process (ca.550 kPa). Liquid ammonia is evaporated at ca.700 kPa in the ammonia evaporator (a). Thecold generated can be used in the absorptionstep. Water is removed from liquid ammonia inan ammonia stripper (b). The ammonia – waterblowdown mixture is evaporated batchwise withsteam. Residual aqueous ammonia can be usedin fertilizer production. The ammonia vapor isheated with steam to ca. 90 ◦C in the ammoniapreheater (c) and then filtered (d). If necessary, asmall stream of ammonia is diverted to tail-gastreatment.

All the air needed for the process is suckedin through the air filter (f ) by the air compressor(g). The compressed air is divided into two sub-streams in a tail-gas preheater (h). The secondaryair goes to the bleacher (x) for stripping raw acid.


Figure 27. Simplified flow sheet of a medium-pressure processa) Ammonia evaporator; b) Ammonia stripper; c) Ammonia preheater; d) Ammonia gas filter; e) Ammonia – air mixer; f) Airfilter; g) Air compressor; h) Tail-gas preheater III; i) Reactor; j) Waste-heat boiler; k) Tail-gas preheater I; l) Economizer;m) Tail-gas preheater II; n) Feedwater preheater; o) Cooler – condenser; p) Absorption tower; q) Ammonia – tail-gas mixer;r) BASF catalytic tail-gas reactor; s) Tail-gas expansion turbine; t) Feedwater tank with deaerator; u) Steam drum; v) Steamturbine; w) Steam turbine condenser; x) Bleacher

The secondary airstream exiting the bleacher isladen with NOx and is added to the nitrous gasbefore it enters the absorption tower ( p). Thegreater part of the process air, the primary air,is mixed with the superheated ammonia in theammonia – air mixer (e); it then flows throughthe mixed-gas filter (not illustrated) and, nowcleaned, enters the reactor (i) where ammoniareacts with oxygen in the primary air over theplatinum – rhodium catalyst, yielding nitrogenmonoxide and water. The reaction is exother-mic and proceeds at ca. 890 ◦C. The nitrous gasstream is cooled in the waste-heat boiler ( j),raising steam. Nitrogen monoxide is oxidizedto nitrogen dioxide in the downstream pipingand equipment, heating the nitrous gas stream.Further heat recovery takes place in the tail-gaspreheater (k) and the economizer (l). After ex-iting the economizer, the nitrous gas stream iscooled in a second tail-gas preheater (m). Con-densate can form in the next unit downstream,

the feedwater preheater (n), which is located di-rectly on the inlet of the cooler – condenser (o).Additional water of combustion condenses outin the cooler – condenser before the nitrous gasstream is mixed with the NOx-laden secondaryair and admitted to the absorption tower (p). Theacid condensate from the feedwater preheaterand the cooler – condenser flows directly into thebottom of the gas cooler – condenser. This acidcondensate is then delivered by an acid conden-sate pump to a tray with the appropriate concen-tration in the absorption tower.

In the lower part of the tower, the nitrousgas is oxidized further, thus reaching the neces-sary degree of oxidation before entering the ab-sorption section. The tower contains sieve traysthrough which the gas flows countercurrent tothe acid. The product acid concentration is at-tained on the lowermost absorption tray of thetower. The acid is then pumped to the bleacher(x), where dissolved NOx is stripped, before the


acid is stored in tanks. Chlorides in the absorp-tion tower feed accumulate on trays where theacid concentration is ca. 21 wt %. Because of thedanger of corrosion, these trays must be emp-tied into the product acid tank from time to time.Cooling water flowing through coils on the sievetrays removes heat generated by the oxidationof nitrogen monoxide to nitrogen dioxide andits further conversion to nitric acid. Part of theheat absorbed by the cooling water is utilizedto evaporate ammonia. A demister in the headof the tower collects liquid droplets entrained inthe tail gas; the condensate runs back to the lasttray of the absorption tower.

The tail gas absorbs heat from the secondaryair and the nitrous gas in three tail-gas pre-heaters. In the BASF catalytic tail-gas treatment(r; see also Section 1.4.2.3), ca. 60 vol % of theNOx in the tail gas is selectively reacted withammonia. Hot tail gas containing < 200 ppmNOx goes to the tail-gas expansion turbine (s),in which mechanical energy is produced to drivethe air compressor. Finally, the tail gas is dis-charged to the atmosphere via the stack.

The tail-gas expansion turbine supplies partof the power needed to drive the air compres-sor. The rest is generated by a condensing steamturbine supplied with product steam. The tur-bine may be bled. The steam turbine conden-sate goes to the boiler feedwater preheater (n)where it absorbs heat from the NOx gas and thenflows through the deaerator into the feedwatertank (t). Nondeaerated feedwater from the bat-tery limitalso flows via a preheater into the feed-water tank. The boiler feedwater pump raises thepressure of the deaerated feedwater to the requi-site boiler pressure. Product steam is raised withheat from ammonia combustion in the waste-heat boiler ( j).

Most of the steam generated in the waste-heatboiler drives the condensing steam turbine. Theremainder can be utilized in the ammonia pre-heater and stripper and for deaerating the feed-water; any excess is exported as product steam.

High-Pressure Process. Figure 28 shows ahigh-pressure process running at ca. 1 MPa. Liq-uid ammonia is fed into the evaporator (a), whereit is evaporated at ca. 1.15 MPa against warm wa-ter. The evaporation temperature rises slightlyabove 35 ◦C as the water content builds up inthe evaporator. The ammonia – water blowdown

mixture is evaporated with low-pressure steamin the ammonia stripper (b). The residual am-monia concentration is ca. 2.5 %.

Ammonia exiting the evaporator system isheated to ca. 130 ◦C in the ammonia preheater(c), and contaminants are removed in the ammo-nia gas filter (d).

Air is sucked in through the air filter (f ) andcompressed to ca. 1 MPa in an air compressor(g) with interstage cooling (h). The primary air(ca. 80 % of the total) is heated to ca. 180 ◦Cagainst nitrous gas in the air heater (m) and thenfed to the ammonia – air mixer (e). The mixedgas stream has a temperature of ca. 175 ◦C andcontains 10.7 vol % ammonia. It passes througha filter (not illustrated) into the reactor (i) wherethe ammonia and atmospheric oxygen react overthe platinum – rhodium catalyst at ca. 900 ◦C toform mainly nitrogen monoxide and water.

The reaction gases next pass through thewaste-heat boiler ( j), where they are cooled toca. 400 ◦C. The heat is used to raise high-pres-sure steam and superheat it to 500 ◦C in the pree-vaporator, superheater, and reevaporator of theboiler.

The boiler system is fed with condensate fromthe steam turbine condenser (w) plus demineral-ized water from the battery limit. Both streamsare led to a tank and from there go through thefeedwater preheater (n) to the deaerator mountedon the feedwater tank (t) and operating at a slightgauge pressure. The tank serves as suction ves-sel for the feedwater pumps, which deliver waterthrough the economizer into the boiler drum (u).

The water – steam mixture in the evaporatoris circulated by the boiler circulation pump. Sat-urated steam and water are separated in the boilerdrum (u).

After leaving the waste-heat boiler ( j), thenitrous gas is cooled to ca. 260 ◦C in a tail-gaspreheater (k). After passing through a run of pip-ing for oxidation, nitrous gas gives up more use-ful heat in the economizer (l), and its tempera-ture falls to ca. 210 ◦C. After another oxidationrun, the stream is led into the air heater (m) andcooled to ca. 180 ◦C. The heat recovered is usedto preheat the primary air.

Before the nitrous gas is led into the gascooler – condenser (o), it passes through feed-water preheaters and through the warm-waterheater (n) both mounted on the two inlets of thecooler – condenser. The nitrous gas is cooled to


Figure 28. Simplified flow sheet of a high-pressure processa) Ammonia evaporator; b) Ammonia stripper; c) Ammonia preheater; d) Ammonia gas filter; e) Ammonia – air mixer; f) Airfilter; g) Air compressor; h) Interstage cooler; i) Reactor; j) Waste-heat boiler; k) Tail-gas preheater; l) Economizer; m) Airpreheater; n) Feedwater and warm-water preheaters; o) Cooler – condenser; p) Absorption tower; q) Tail-gas preheater; r) Tail-gas preheater; s) Tail-gas expansion turbine; t) Feedwater tank with deaerator; u) Steam drum; v) Steam turbine; w) Steamturbine condenser; x) Bleacher

ca. 115 ◦C and partly condensed before enteringthe cooler – condenser.

In the water- cooled cooler – condenser (o),the gas is cooled to < 50 ◦C so that water formedduring ammonia oxidation condenses to pro-duce ca. 45 % acid. This product is delivered tothe appropriate tray of the absorption tower (p)by the acid condensate pump. The NOx-ladensecondary air recycled from the bleacher (x) ismixed with the nitrous gas stream before it is fedinto the tower ( p).

In the sieve-tray absorption tower ( p) the ni-trous gas flows countercurrently to the processwater fed at the column head, and nitric acid isformed. Heat from the absorption tower is trans-ferred to the cooling water. Raw acid from theabsorption tower is treated with secondary air inthe bleacher (x), then delivered to battery limitwith the aid of the system pressure.

Tail gas exits the absorption tower at ca. 20 –30 ◦C. It is heated to ca. 80 ◦C in a tail-gas pre-heater (q) against the tail-gas stream from the ex-pansion turbine. The tail gas is further heated to140 ◦C against low-pressure steam (r) and thento ca. 375 ◦C in a tail-gas preheater (k). It is sub-sequently expanded in the tail-gas turbine (s);this step produces ca. 70 % of the energy re-quired to drive the air compressor.

Tail gas exits the turbine at ca. 135 ◦C and isled to a tail-gas preheater (q), where it is cooledto ca. 90 ◦C before being discharged through thestack.

The rest of the power needed to drive theair compressor can be supplied by a pass-outcondensing steam turbine (v). The superheatedsteam is delivered to the steam turbine or ex-ported. Part of the steam supplied to the turbine iswithdrawn as low-pressure steam before admis-sion to the condensing section. This low-pres-


sure steam is used in the plant or exported. Therest of the steam supplied to the turbine passesthrough the condensing section and is condensedin the steam turbine condenser (w).

Dual-Pressure Process. The dual-pressureprocess combines the favorable economics ofmedium-pressure combustion with the effi-ciency of high-pressure absorption. Figure 29 isa flow sheet of such a plant with ammonia com-bustion at 0.5 MPa and absorption at 1.1 MPa.

Liquid ammonia is fed to the ammonia evap-orator (a), which is under a pressure of ca.0.65 MPa. The entire stream then goes to ammo-nia evaporator I, where ca. 80 % of the ammoniais evaporated against cold water at a constanttemperature of ca. 12 ◦C. Residual liquid am-monia is fed to ammonia evaporator II (not illus-trated) and evaporated at varying temperaturesagainst cooling water. The evaporation temper-ature increases from 12 ◦C to ca. 20 ◦C as water

content builds up in the evaporator. Ammoniaevaporator II is designed so that the entire am-monia stream can be evaporated in it, if neces-sary, at a maximum temperature of 20 ◦C. If thewater content or the evaporation temperature inammonia evaporator II is too high, the ammo-nia – water mixture is blown down and fed tothe ammonia stripper (b). Ammonia is strippedto a residual content of ca. 2.0 % as the tempera-ture is raised to 150 ◦C with low-pressure steam.The residue, greatly depleted in ammonia, is dis-charged.

Gaseous ammonia from the evaporator sys-tem is passed through the ammonia gas filter(c) to remove contaminants, then heated to ca.150 ◦C in the ammonia preheater (d).

Air is sucked through the filter (f) and com-pressed to 0.5 MPa in the compressor (g). About86 % of the air is fed as primary air to the am-monia – air mixer (e). The mixed gas stream hasa temperature of ca. 220 ◦C and contains ca.

Figure 29. Simplified flow sheet of a dual-pressure processa) Ammonia evaporator; b) Ammonia stripper; c) Ammonia gas filter; d) Ammonia preheater; e) Ammonia – air mixer; f) Airfilter; g) Air compressor; h) Reactor; i) Waste-heat boiler; j) Tail-gas preheater III; k) Economizer; l) Tail-gas preheater II;m) Feedwater preheater; n) Cooler – condenser I; o) Cooler – condenser II; p) Nitrous gas compressor; q) Tail-gas preheaterI; r) Cooler – condenser III; s) Absorption tower; t) Tail-gas expansion turbine; u) Feedwater tank with deaerator; v) Steamdrum; w) Steam turbine; x) Steam turbine condenser; y) Bleacher


10 vol % ammonia. It is transported through thegas-mixture filter (not illustrated) to the reac-tor (h). Here, ammonia reacts on the platinum –rhodium catalyst with atmospheric oxygen at890 ◦C to produce a 96.5 % yield of nitrogenmonoxide and water.

The reaction gases next pass through thewaste-heat boiler (i) and are cooled to 355 ◦C.The heat is used to raise high-pressure steam inthe preevaporator, superheater, and reevaporatorof the boiler.

The boiler system is fed with condensate fromthe steam turbine condenser (x) plus demineral-ized water from the battery limit. Both streamsare led to a tank and then pass through the feed-water preheater (m) to the deaerator mountedon the feedwater tank (u). The tank serves as asuction vessel for the feedwater pumps, whichdeliver the water through the economizer (k) intothe steam drum (v).

The water – steam mixture in the evaporatoris circulated by the boiler circulation pump. Sat-urated steam and water are separated in the steamdrum (v).

After exiting the waste-heat boiler and pass-ing through a run of piping for oxidation, the ni-trous gas is cooled from 390 to 280 ◦C in tail-gaspreheater III (j). After further oxidation, it entersthe economizer (k), where it is cooled from 310to 200 ◦C. The nitrous gas gives up further use-ful heat to the tail gas in tail-gas preheater II (l)and its temperature falls to ca. 170 ◦C.

The gas passes through the feedwater pre-heater (m), mounted on the cooler – condenserinlet and thereby cools to ca. 100 ◦C. In thewater- cooled cooler – condenser I (n), the gas iscooled to < 50 ◦C so that the water formed dur-ing ammonia oxidation condenses. A ca. 43 %acid results and is pumped to the appropriate trayof the absorption tower (s). The secondary airladen with NOx and recycled from the bleacher(y) is now mixed with the nitrous gas stream. Themixed stream is then led into cooler – condenserII (o), where it is cooled to ca. 45 ◦C againstchilled water. More water condenses and a ca.52 % acid is formed, which is forwarded forabsorption along with acid formed in cooler –condenser I. The gas is passed through a mistcollector (not illustrated) upstream of the nitrousgas compressor ( p).

The nitrous gas is compressed to 1.1 MPa inthe compressor ( p) and has an outlet tempera-

ture of ca. 130 ◦C. It passes to tail-gas preheaterI (q ), where it is cooled to ca. 75 ◦C. Down-stream of the preheater is cooler – condenser III(r), which the nitrous gas leaves at 50 ◦C. Theresulting acid condensate has a concentration ofca. 58 wt % and is mixed with the raw acid fromthe absorption column.

From cooler – condenser III the nitrous gas,now at least 90 % oxidized, enters the absorptiontower (s). The gas first passes through oxidationtrays that receive only a slight flow of liquid, toattain the degree of oxidation needed for equilib-rium with 68 % acid. In the absorption section,gas and liquid move countercurrently. Requiredprocess water is pumped to the tower head. The68 % raw acid is withdrawn from the first ab-sorption tray and then treated with secondary airin the bleacher (y). It is delivered to the batterylimit with the aid of the system pressure.

Approximately 80 % of the heat generated byabsorption is transferred to the cooling water andthe remainder to the chilled water employed inthe last part of the absorption step where verylittle heat of reaction is liberated.

The tail gas contains < 150 – 200 ppm NOx

and exits the absorber at 20 ◦C. It is heated to110, 200, and finally 350 ◦C in the three down-stream tail-gas preheaters (q, l, and j) before be-ing expanded in the tail-gas turbine (t). The tur-bine supplies ca. 67 % of the power required bythe air and the nitrous gas compressors. The ex-panded tail gas has a temperature of ca. 100 ◦Cand is discharged through a stack.

The remaining power required to drive therotating machinery can be obtained from a pass-out condensing steam turbine (w). Superheatedsteam is fed to the turbine or exported. Part of thesteam supplied to the turbine may be withdrawnas low-pressure steam before it enters the con-densing section. This low-pressure steam maybe used in the plant or exported. The remainderof the steam fed to the turbine passes throughthe condensing section and is condensed in thesteam turbine condenser (x).

Comparison of Medium- and High-

Pressure Processes [62]. Two trends can beseen in the worldwide development of weak acidprocesses. First, from the process-engineeringstandpoint, a progression occurs from low-through medium- to high-pressure processes [5].Second, capacities continue to increase; single-


Table 5. Comparison of significant specific consumption figures for nitric acid plants (values are given per tonne of 100 % HNO3, the tail

gas contains < 200 ppm of NOx )

Parameter Monopressure processes Dual-pressure process

Medium pressure High pressure

Operating pressure, MPa 0.55 1.080 0.45/1.1

Ammonia, kg 282∗ 283 279

Electric power, kW · h 8.5 8.0 9.0

Platinum, g 0.14 0.30 0.11

Cooling water, t (∆t = 10 ◦C)∗∗ 120 125 130

Process water, t 0.3 0.3 0.3

Low-pressure heating steam, t 0.1 0.1 0.1

High-pressure excess steam, t (2.5 MPa, 400 ◦C) 0.87 0.78 0.81

∗ Includes 1 kg NH3 for NOx reduction from 600 to < 200 ppm. ∗∗Includes water for steam turbine compressor.

train plants producing up to 2000 t of nitric acidper day are now being built. In Europe, the dual-pressure design is preferred for larger plants,whereas smaller ones employ the monopressuredesign. Where feedstock and energy prices arelow, monopressure operation offers special ad-vantages; low investment costs ensure a quickpayout, particularly in North America. If, on theother hand, feedstock and energy prices are veryhigh (as in Europe), yield and energy efficiencymust be maximized, so higher investment costsare acceptable. Table 5 compares consumptionfigures per unit product for monopressure anddual-pressure processes.

New pollution control regulations and the en-ergy crisis of the mid-1970s have also led to thedevelopment of new processes or the improve-ment of existing ones. To avoid the need for cat-alytic tail-gas treatment in large-tonnage plants(and thus an increase in specific ammonia con-sumption), new facilities employ higher absorp-tion pressures.

1.3.4.2. Concentrated Acid Processes

Industrially produced nitric acid contains 50 –70 wt % HNO3. This is high enough for fertil-izer production, but nitration processes in indus-trial organic chemistry call for concentrated acid(98 – 100 %). Distillation, the simplest way toconcentrate dilute acid, fails because nitric acidand water form an azeotrope (68.4 % HNO3 atatmospheric pressure).

Concentrated nitric acid is manufactured di-rectly or indirectly [63]. In the direct process,the water generated in ammonia combustion iswithdrawn by rapid cooling to give a nitrous gasmixture from which concentrated acid can be

produced directly in one of two ways. First, thecompletely oxidized NOx can be separated inliquid form by absorption in concentrated ni-tric acid and then led to a reactor where it isreacted with oxygen and water (or weak acid)under pressure to yield concentrated acid. Sec-ond, nitrous gases can be reacted with azeotropicacid to form a concentrated acid that can be con-verted easily to concentrated and azeotropic acidby distillation. The latter product is either com-pletely recycled or used in the production of or-dinary weak acid.

In the indirect processes, concentration isbased on extractive distillation and rectificationwith sulfuric acid or magnesium nitrate.

Direct Process. The essential feature of theclassical direct process (Fig. 30) is that liquiddinitrogen tetroxide is produced and reacted un-der pressure with pure oxygen and a certainquantity of dilute nitric acid. A detailed accountof concentrated acid production via azeotropicnitric acid is given in [61, pp. 99 – 130].

Ammonia Oxidation. As in weak acid pro-cesses, feed ammonia is evaporated, super-heated, and filtered (see Section 1.3.4.1). Evap-oration takes place with the aid of water heatedwith process heat. Because traces of water andoil accumulate in the ammonia evaporator thesump must be drained into the steam-heatedstripper from time to time. Air for ammoniacombustion is sucked through a filter (a) andmixed with purified ammonia in the mixer ( b).The ammonia content of the mixture is held inthe range 11.5 –12.3 vol %, depending on thetemperature. The optimal combustion temper-ature is 830 – 850 ◦C. Ammonia oxidation takesplace over conventional platinum – rhodium cat-


Figure 30. Simplified flow sheet of a process for direct production of strong nitric acid with oxygena) Air Filter; b) Ammonia – air mixer; c) Gas mixture filter; d) Reactor; e) Cooler; f) Compressor; g) Cooler; h) Oxidationtower; i) Recirculating acid tank; j) Postoxidizer; k) Cooler; l) Cooler; m) Absorption tower; n) Final absorber; o) Raw acidtank; p) Cooler; q) Precondenser; r) Stirred tank; s) Reactor; t) Bleacher; u) Cooler

alysts in the reactor (d). The heat of reaction isused to raise steam in a waste-heat boiler andan economizer. The process gas exits the econ-omizer at ca. 170 ◦C and is then cooled to 25 ◦Cin a cooler – condenser (e), giving an acid con-taining ca. 2 – 3 wt % nitric acid. The cooler –condenser is specially designed to condense asmuch water as possible without much NOx ab-sorption. Part of the process heat is also trans-ferred to the warm water used for ammonia evap-oration. The weak condensate generated in thecooler – condenser is used as scrub liquor for fi-nal absorption.

Oxidation of Nitrogen Monoxide. Secondaryair is mixed with the cooled process gas to oxi-dize nitrogen monoxide to nitrogen dioxide. Theamount of secondary air must be sufficient toensure that the tail gas has a minimal oxygencontent (2.5 – 3.5 vol %) before it is dischargedto the atmosphere. The concentrated nitric acidventing system is connected to the secondary airintake in such a way that oxygen and nitrogen ox-ides liberated on venting are recycled to the pro-cess. The nitrogen monoxide gas and secondaryair are compressed together to ca. 0.14 MPa (f ).After further cooling (g) to remove the heat ofcompression, the mixture enters the oxidationtower (h), which is usually equipped with sievetrays. Tube coils on the trays remove the heat ofoxidation. To maintain a liquid level on all trays,

acid from the tower bottom is pumped (recircu-lated) to the head.

Final Oxidation of Nitrogen Monoxide. Thenitrous gases leaving the oxidation tower are ap-proximately 95 % oxidized and are led to a finaloxidizer ( postoxidizer) ( j) where they are com-pletely oxidized by desorption of concentratednitric acid. The acid moves countercurrently tothe nitrous gases. Water produced in this reac-tion dilutes the concentrated acid; acid leavingthe postoxidizer contains ca. 75 % HNO3. Flowcontrol of the concentrated nitric acid feed intothe postoxidizer is important because too higha rate lowers the output of concentrated acid,whereas too low a rate displaces final oxidationinto the absorber, with the result that the concen-trated nitric acid entering the absorption tower istoo dilute. The product gas is saturated with ni-tric acid vapor. For the next step, absorption, theprocess gas is cooled to −10 ◦C against brinein a cooler ( l). Virtually all the nitrogen diox-ide is dimerized to dinitrogen tetroxide at thistemperature.

Absorption of Dinitrogen Tetroxide. The ni-trous gases are completely dimerized to dini-trogen tetroxide and, in this condition, are fedinto the absorber (m). The absorber consists offour sections, each packed with ceramic Raschigrings. The dinitrogen tetroxide is absorbed inconcentrated nitric acid that comes from thebleacher and is cooled to ca.−10 ◦C (k, u). The


acid leaving the absorption tower contains ca.25 – 30 wt % dinitrogen tetroxide and is pumpedto the raw acid tank (o).

Final Absorption. The gas exiting the absorp-tion tower is practically free of nitrogen dioxidebut saturated with nitric acid vapor. The vapor isremoved by scrubbing in the final absorber (n),which consists of two sections. The scrub liquor(weak nitric acid from the cooler – condenser) isrecirculated or used to make up the water bal-ance in the stirred tank.

Dinitrogen Tetroxide Production. The ladenconcentrated nitric acid is sent from the raw acidtank (o) to the bleacher (t), which consists ofa stripping section and a reboiler. The acid en-ters the head of the tower and flows downwardcountercurrently to the nitric acid vapor, whichstrips nitrogen dioxide from the concentrated ni-tric acid. The bleached concentrated acid goes tothe reboiler section of the bleacher, where it ispartially evaporated and partially withdrawn asproduct acid. The vapor exiting the bleacher con-tains ca. 95 % nitrogen dioxide and 5 % nitricacid. Most of the vapor condenses in a down-stream precondenser (q). Dinitrogen tetroxide isliquefied in the liquefier and fed to a stirred tank(r).

Production of Concentrated Nitric Acid. Theformation of concentrated nitric acid in the re-actor (s) is described by the following equation:

4 NO2( 2 N2O4) + O2+ 2 H2O 4 HNO3

This requires a N2O4 – H2O molar ratio of 1 : 1and a weight ratio of 5.11 : 1. The reactiontime can be considerably shortened if dinitro-gen tetroxide is present in excess. On the otherhand, an excess of dinitrogen tetroxide lowersthe production rate of the reactor and increasesthe dinitrogen tetroxide recycle rate, resulting inhigher steam consumption for bleaching and agreater cold requirement for dinitrogen tetroxideliquefaction.

The N2O4 – H2O ratio, denoted the R factor,is the most important criterion for monitoring re-actor operation. In practice, it fluctuates between6.5 and 7. A product concentration of > 98 wt %nitric acid requires a higher R factor. The eco-nomically most favorable R factor can be foundonly by empirical means; the ratio can be var-ied to achieve optimal operating conditions. Themakeup acid used to maintain the water balance

of the raw acid mixture is taken from the bottomof the final absorber (n).

Concentrated nitric acid is formed in the re-actor (s) at ca. 5.0 MPa and 60 – 80 ◦C. Highmechanical strength and corrosion resistance areneeded, but these two requirements cannot besatisfied by any one material. The reactor thushas a carbon steel jacket to provide mechanicalstrength and a pure aluminum inner shell to pro-vide corrosion resistance. Unreacted oxygen issupplied to the suction of the nitrous gas blowervia the venting system. The concentrated nitricacid product is led to the raw acid tank (o) andthen to the bleacher. The bleached concentratedacid is cooled to + 30 ◦C (u). Most of the con-centrated nitric acid is returned to the absorptiontower and final oxidation.

Indirect Processes. Two types of indirectprocess are used to produce concentrated nitricacid (i.e., product containing> 97 wt % HNO3 ):

1) Sulfuric acid process2) Magnesium nitrate process

Both concentration techniques are based onextractive distillatfication. Weak acid is firstproduced by a conventional process (Sec-tion 1.3.4.1), then a third component is addedto extract the water so that up to ca. 99 % nitricacid can be distilled from the ternary mixture.

Figure 31. Simplified flow sheet of a process for preconcen-tration of nitric acida) Preconcentrating tower; b) Separator; c) Recirculatingevaporator system; d) Cooler

The starting product is ordinary commercialnitric acid (ca. 55 – 65 % HNO3 ). Weaker acids


can be preconcentrated in a single-stage or mul-tistage process to give ca. 68 wt % nitric acid.Figure 31 illustrates a single-stage apparatus.Preconcentration takes place in a continuouscountercurrent distillation tower (a), which canbe of either the packed or the bubble-tray type.The system includes one or, if appropriate, twopreheaters, a recirculating evaporator system forthe tower bottom, a stripping and rectifying sec-tion for the tower, an overhead condenser, and areflux separator (b).

Sulfuric Acid Process. Sulfuric acid can beused as extracting agent (see also→ 4. Distilla-tion and Rectification). The concentrated nitricacid product is clear and colorless; it contains98 – 99 wt % nitric acid and < 0.05 % nitrogendioxide.

Consumption figures for the production of 1 tof 99 % nitric acid from dilute nitric acid aregiven in Table 6.

Table 6. Consumption figures for the production of 1 t of 99 %

HNO3 from dilute nitric acid by the sulfuric acid process∗

Parameter Starting HNO3

concentration, wt %

55 60 65

Heating steam (1.0 – 1.8 MPa), t 2.0 1.75 1.45

Cooling water, m3 80 60 50

Electric energy, kW · h 17 14 11

Water evaporated, t 0.82 0.66 0.53

∗ Source: Plinke – NASAC plant data.

The nitric acid concentration step now oper-ates with indirect heating, an advance over ear-

lier technologies that reduces steam consump-tion by as much as 50 %.

The materials of construction are highly cor-rosion resistant. Towers are made of borosili-cate glass, enameled steel, and steel – polytetra-fluoroethylene. Heat exchangers are made ofglass, polytetrafluoroethylene, stainless steel,tantalum, titanium, high-purity aluminum, andspecial alloys.

Figure 32 is a flow sheet of a plant makingconcentrated nitric acid by the sulfuric acid pro-cess. The feed nitric acid, concentrated to about68 wt %, is preheated (e) and then fed into thedistillation tower (a). At least 80 wt % sulfuricacid is fed to the head of the tower. The part ofthe tower above the nitric acid inlet, which isirrigated with sulfuric acid, can be regarded asthe rectifying section, with the sulfuric acid alsofunctioning as reflux. The part of the tower be-low the nitric acid inlet is the stripping section.

A circulating evaporator system takes care ofbottom heating. The ca. 70 % sulfuric acid leav-ing the bottom of the tower goes to the concen-trator (c), which operates under vacuum (8 kPa).The overhead vapor product from the tower iscondensed to form 99 % nitric acid and thendeaerated (b). The tail gases, which still con-tain nitric acid vapor, are scrubbed with dilutenitric acid (d).

Magnesium Nitrate Process. In this processa magnesium nitrate solution is used to extractwater from the nitric acid. The resulting dilutemagnesium nitrate solution is restored to the de-

Figure 32. Simplified flow sheet of a process for concentrating nitric acid with sulfuric acid (Plinke – NACSAL process)a) Concentrating tower; b) Condenser – deaerator; c) Sulfuric acid concentrating tower; d) Tail-gas treatment; e) Preheater;f) Cooler; g) Blower; h) Separator


Figure 33. Simplified flow sheet of a process for concentrating nitric acid with magnesium nitrate (Plinke – MAGNAC process)a) Concentrating tower; b) Vacuum evaporator; c) Tail-gas treatment; d) Preheater; e) Cooler; f) Blower; g) Separator

sired working concentration of about 72 % in avacuum concentrator before being returned tothe nitric acid concentration step.

Consumption figures for the production of 1 tof 99 % nitric acid from dilute nitric acid aregiven in Table 7.

Table 7. Consumption figures for the production of 1 t of 99 %

HNO3 from dilute nitric acid by the magnesium nitrate process∗

Parameter Starting HNO3

concentration, wt %

55 60 65

Heating steam (1.0 – 1.8 MPa), t 2.0 1.75 1.45

Cooling water, m3 80 70 60

Electric energy, kW · h 10 9 8

Water evaporated, t 0.82 0.66 0.53

∗ Source: Plinke – MAGNAC plant data.

Figure 33 is the flow sheet of a plant mak-ing concentrated nitric acid by the magnesiumnitrate process. Weak acid is fed to the dewa-tering tower (a). Extractive distillation with aconcentrated (72 wt %) solution of magnesiumnitrate at 140 ◦C gives an overhead product con-taining 99 % nitric acid. A small fraction of thecondensed overhead product is refluxed to thedewatering tower. The bottom of the tower isheated, and the dilute magnesium nitrate solu-tion, containing < 0.1 % nitric acid, is led to thevacuum evaporator (b). Condensed weak acid isprocessed in the recovery tower (c) and recycledif appropriate. The tail gas still contains nitricacid vapor and can be scrubbed with dilute ni-tric acid (c).

Comparison of Indirect Processes. Whichconcentration process is better suited to a giventask must be decided case by case. From aprocess-engineering standpoint, the magnesiumnitrate process has the following advantages:

1) The installation is very compact. No hold-ing tanks are needed, which reduces energylosses.

2) Because the reconcentration step for mag-nesium nitrate is carried out in stainless steelvessels, virtually no critical construction ma-terials are required.

3) Magnesium nitrate can be transported insacks, and is therefore much easier to han-dle than sulfuric acid.

4) The weak nitric acid condensate, which isproduced when the magnesium nitrate so-lution is reconcentrated, is partly reused inNOx absorption. Losses of nitric acid arevery slight.

As in weak acid production, development ofconcentrated acid production has followed dif-ferent paths in the United States and the FederalRepublic of Germany (and, in part, in Europe).In the United States estimated operating costsare significantly lower for concentration withmagnesium nitrate than with sulfuric acid. TheU.S. view is largely that, even after the changefrom autoclaves to continuous direct production,the equipment cost (i.e., investment cost) is toohigh for the latter process to be competitive.


1.4. Environmental Protection

In recent years, more stringent water pollutionregulations have taken effect and legislative pro-visions for air pollution have been amended.New processes have been developed and in-troduced to significantly reduce pollutant emis-sions from nitric acid plants. The problem of ni-trogen oxide pollution is treated in more detailin→ Air.

1.4.1. Wastewater

Wastewater problems can be overcome by ap-propriate design of the nitric acid plant. An espe-cially simple form of wastewater treatment canbe employed when the nitric acid is processeddirectly for the production of mineral fertilizers.

The solution from ammonia stripping con-tains up to 10 % ammonia; it can be neutralizedwith nitric acid and then subjected to absorptionalong with process water. The acid product thencontains a small amount of ammonium nitrate,but this is not a problem when the acid is pro-cessed into mineral fertilizers.

Leaks from pumps, vessels, etc., are pumpedinto a separate acid drain tank, and then pro-cessed directly or indirectly. The apparatus usedfor this purpose is completely separate from thesewage system and thus prevents contaminationof wastewater.

If heat is removed by recooling systems, cool-ing water blowdown is fed into the wastewatersystem to limit thickening of the cooling wa-ter. Fresh water must be supplied to compensatefor the blowdown and evaporation losses. Corro-sion inhibitors, hardness stabilizers, and in somecases, biocides are also carried out of the systemby blowdown.

1.4.2. Stack Gas

Obsolescent nitric acid plants can be recog-nized by their strongly colored yellow to reddishbrown plumes of stack gas. The coloration is dueto nitrogen dioxide. Some stack gases contain upto 3000 ppm NOx; very old plants may surpassthese values [64].

The absorption of nitrous gases with waterforms part of the nitric acid process. However,

the laws of nature prevent complete absorption,and some residual emission cannot be avoided(see Section 1.3.2). These emissions can be min-imized by optimizing process conditions, in-creasing the efficiency of absorption, or usingspecial tail-gas treatment methods.

Crucial parameters for the absorption of ni-trous gases in water are the following [65]:

1) Pressure2) Temperature3) Reaction volume4) The efficiency of the absorption tower5) The partial pressures of nitrogen oxides and

oxygen

Emissions in the Federal Republic of Ger-many in 1986 totaled ca. 3×106 t NOx [66].Transportation, households, and small con-sumers accounted for ca. 66 % of this; powerplants and district heating plants for ca. 25 %;and industry, including furnaces, for ca. 9 %. Ni-trogen oxides from nitric acid production wereresponsible for < 1 % of the total emissions.

1.4.2.1. Emission Limits

Heightened sensitivity to environmental con-cerns, coupled with the need for larger and moreefficient production systems, will lead to fur-ther reductions in NOx emissions. Legislatorsare following this trend and modifying pollutionlimits in accordance with changing technical ca-pabilities.

At the beginning of the 1980s the first gen-eral administrative regulation was issued in theFederal Republic of Germany under the FederalPollution Control Act: the Technische Anleitungzur Reinhaltung der Luft (TA-Luft; Technical In-structions for Air Pollution Control) [67]. Thisregulation and the corresponding VDI guide-lines [68] form the basis for issuing operatingpermits. Regulations in many other countries arebased on these guidelines.

The designation NOx is commonly used tospecify emission and concentration values be-cause stack gases always include mixtures ofnitrogen oxides. For the purpose of standard-ization, the oxides are calculated in terms of ni-trogen dioxide. The acceptable concentrationsfrom TA-Luft 1986 [67] are


1) Mass of substances emitted as mass concen-tration in units of g/m3 or mg/m3, referredto the volume of stack gas at standard con-ditions (0 ◦C, 101.3 kPa) after deducting themoisture content of water vapor, or to thevolume of stack gas at standard conditions(0 ◦C, 101.3 kPA) before deducting the mois-ture content of water vapor; and

2) Mass of substances emitted, referred to time,as mass velocity in units of kg/h, g/h, or mg/h;and

3) Ratio of the mass of substances emitted tothe mass of products generated or processed(emission factors), as mass ratio in units ofkg/t or g/t.

The usual concentration unit ppm (mL/m3 )can be converted to the prescribed mass concen-tration (mg/m3 ) with the following factor [69]:1 ppm = 1.88 mg/m3 (based on monomeric NO2

at 101.3 kPa and 298 K and ideal-gas behavior).The TA-Luft [67] states that the emission

of nitrogen monoxide and nitrogen dioxide inthe stack gas of nitric acid plants must not ex-ceed 0.45 g/m3 as nitrogen dioxide. Further-more, stack gases may be discharged only if col-orless; as a rule, this is the case if the mass con-centration of nitrogen dioxide in the stack gasdoes not exceed the value given by the follow-ing formula:

Massconcentrationof nitrogen dioxide(

mg/m3)

=1200

Internaldiameterof stackorifice (dm)

The requirement of a colorless discharge setsa practical limit of < 200 ppm NOx. Older low-and medium-pressure plants should comply withthese standards by March 1, 1996.

Plants whose emission of nitrogen monox-ide and nitrogen dioxide (as nitrogen dioxide)exceeds a rate of 30 kg/h must be equipped withdevices that continuously measure the mass con-centration of nitrogen oxides. Quantitative re-lationships between emission and ground-levelconcentration are also described, thus settingguidelines for stack design.

If ground-level concentration limits are ex-ceeded, the TA-Luft emission limits are furtherreduced by the relevant authorities. The ground-level concentration limits for gaseous pollutantsin air are as follows [67]: nitrogen dioxide, long-

term exposure (IW1), 0.08 mg/m3; short-termexposure (IW2), 0.2 mg/m3.

1.4.2.2. Analysis

The guidelines for the measurement and mon-itoring of NOx emissions from nitric acid pro-duction are specified in [67,70,71]. Photometry

without auxiliary chemical reaction and chemi-

luminescence [71, c – e] are suitable for the con-tinuous measurement of NOx emissions [65].

Difficulties in gas sample preparation due tocondensation are discussed in [71, c – g]. Ad-ditonal apparatus-related problems occur whenchemiluminescence is used to measure emis-sion of moist stack gases at higher NOx lev-els. A chemiluminescence analyzer with a max-imum range of 0 – 10 000 ppm has been devel-oped [72].

For stack gases containing traces of ammo-nia, precipitation of ammonium salts in the mea-suring cells has prevented continuous measure-ments. Two methods useful for the discontinu-ous measurement of nitrogen oxide emissions[65] are acidimetric titration and the photomet-ric phenoldisulfonic acid technique [71, a, b].

1.4.2.3. Control of NOx Emissions

Four basic approaches are used to reduce tail-gas NOx levels: improved absorption, chemicalscrubbing, adsorption, and catalytic tail-gas re-duction. Recent decades have seen intense re-search and development effort invested in thesemethods. More stringent environmental restric-tions have lent special urgency to developmentin this area. The most important techniques arecovered in this section.

Improved Absorption. Absorption effi-ciency depends chiefly on the absorption pres-sure, the number of stages, and the tempera-ture. The temperature of the gas between stagesis especially important because it governs theprogress of oxidation, which is the limitingquantity for the entire absorption process (seeSection 1.3.2). In an already existing nitric acidplant, the options are to expand the absorptionvolume and/or lower the absorption temperature.In the first approach, large additional volumes


only result in small reductions of tail-gas NOx

levels because the oxidation of nitrogen monox-ide to nitrogen dioxide proceeds very slowlywhen the NOx concentration is low. The draw-back to this method is that added absorptionvolume in stainless steel is very expensive. Theadvantage is that it does not require any newtechnology. The absorption volume is addedin the form of a second tower; new absorptiontower designs with very few stages have beendevised [73].

The use of cold energy in the absorption pro-cess greatly accelerates oxidation of nitrogenmonoxide to nitrogen dioxide. The disadvantageof this method is that the necessary refrigerationequipment and piping demand further invest-ment. Another technique involving the use ofcold to lower the NOx level is to cool the nitrousgas so that more dinitrogen tetroxide than nitro-gen dioxide is formed. The dinitrogen tetroxideis then scrubbed with nitric acid at ca. 0 ◦C; itcan be stripped out, and converted to nitric acidby the reaction

N2O4 + 1/2 O2 + H2O −→ 2 HNO3

in an absorption reactor at ca. 60 – 80 ◦C. Thetechnique is similar to that in concentrated ni-tric acid production. The method is, however,more suitable for investment in a plant being de-signed than for the expansion or upgrading of anexisting plant.

Chemical Scrubbing. A number of patentsand scientific publications deal with options forscrubbing NOx out of tail gas. Problems relatedto the scrub liquor (cost, regeneration, quantity,and environmental impact) are encountered inall methods. These problems are always easierto manage if the nitric acid plant is part of an inte-grated chemical plant, which is usually the case.However, in some instances a nitric acid plantor sections of a fertilizer complex must operateindependently. The following scrub liquors havebeen proposed: aqueous suspension of magne-sium carbonate and magnesium hydroxide [74];solution of vanadium in nitric acid [75]; ammo-nium sulfide and bisulfide [76]; milk of lime[77]; ammonia [78]; hydrogen peroxide [79];and urea [80].

Ammonia scrubbing is used in the UnitedStates. Goodpasture (Texas) has developed a

safe process based on scrubbing the tail gas withammonium nitrate solution; nitrite formation issuppressed by aeration and the presence of freeacid. The tail gas is then led through an am-moniacal scrub liquor, pH 7.5 – 8.5. The scrub-bing product is a 30 – 50 % ammonium nitratesolution. The tail gas has a residual level of ca.200 ppm NOx.

Hydrogen peroxide scrubbing is based on thefollowing overall reactions:

NO + NO2 + 2 H2O2 −→ 2 HNO3+ H2O

2 NO2 + H2O2 −→ 2 HNO3

The reactions are carried on sieve trays or inpacked towers with recirculation of the hydro-gen peroxide solution. The advantage of thisscrubbing process is that the reaction time isvery fast; the disadvantage, that the hydrogenperoxide scrub liquor is expensive.

A solution containing 20 % urea and 10 %free nitric acid has also been suggested for scrub-bing tail gas to remove NOx. The process is car-ried at ca. 50 ◦C. Both nitrogen oxides are se-lectively reduced to nitrogen by urea, which de-composes to yield nitrogen and carbon dioxide.The process, developed by Norsk Hydro in Nor-way, has the advantage that urea is readily avail-able and relatively cheap. The resulting stack gasplume is colorless but heavily laden with watervapor.

Adsorption Processes. The adsorption ofNOx by molecular sieves has long been knownbut has not yet been tested intensively in full-scale nitric acid plants.

The Pura-Siv-N process ( Union Carbide) isclaimed to reduce stack gas levels below 50 ppmnitrogen dioxide [81], but high investment costsand other problems have prevented its adop-tion. A patent granted to Kernforschungsan-lage Julich GmbH has a similar object [82], butagain industrial experience is slight. A “wet”adsorption system has also been developed byCofaz (France). A carbon- containing adsorbentis sprayed with water or dilute nitric acid andbrought in contact with the tail gas. This pro-cess has been installed in three monopressureplants [83].

Catalytic Reduction. Catalytic reductionwas the first method used to reduce NOx emis-


sions; as early as 1924, a patent for such a pro-cess was issued to Fauser [84]. A fuel is added,at a level below the flash point, to the nitro-gen oxides which then react on the catalyst sur-face to form nitrogen and water vapor. Mostfuels (e.g., hydrocarbons and hydrogen), how-ever, react preferentially with the free oxygenthat is always present in tail gases from nitricacid production. As a result, the nitrogen ox-ides are not destroyed until free oxygen has beenconsumed. Under these conditions the reactionswith methane can be described as

CH4 + 2 O2 −→ CO2+ 2 H2O

CH4 + 4 NO −→ CO2+ 2 H2O + 2 N2

CH4 + 2 NO2 −→ CO2+ 2 H2O + N2

Ammonia can also be used as a reducing agent,preferentially reducing the nitrogen oxides:

6 NO + 4 NH3 −→ 5 N2+ 6 H2O

6 NO2+ 8 NH3 −→ 7 N2+ 12 H2O

NO + NO2+ 2 NH3 −→ 2 N2+ 3 H2O

4 NO + O2+ 4 NH3 −→ 4 N2+ 6 H2O

Catalytic reduction processes are accordinglyclassified as nonselective or selective.

Catalysts for nonselective reduction pro-

cesses are usually based on platinum, vanadiumpentoxide, iron oxide, or titanium. The fuel re-quirement is the stoichiometric amount neededto reduce all the oxygen present (free and in ni-trogen oxides) plus a small excess (ca. 0.5 vol %CH4 ).

Unfortunately, the principal byproducts ofthis process are carbon monoxide (≤ 1000 ppm)and hydrogen cyanide. When hydrocarbon fuelsare used, the tail gas must be preheated beforethe reaction on the catalyst proceeds at all. Thepreheat temperature depends directly on the fuelselected:

Natural gas 450 – 480 ◦C

Propane 340 ◦C

Butane 340 ◦C

Naphtha 340 ◦C

Hydrogen 250 ◦C

The use of hydrogen allows the preheat tem-perature to be lowered to 150 – 200 ◦C but is gen-erally too expensive.

If the quantity of fuel is not enough to reduceall the oxygen, nitrogen dioxide is reduced only

to nitrogen monoxide. The stack gas is then de-colorized; in some countries, this is sufficient.Complete NOx removal generally requires pre-heating, multiple fixed-bed catalysts (with cool-ing between beds to prevent overheating), andcareful heat recovery to offset part of the fuelcost.

Most nitric acid plants constructed or mod-ified by Weatherly use nonselective catalyticNOx abatement units with tail-gas exit tempera-tures of 650 – 675 ◦C. Some plants have gas exittemperatures > 815 ◦C, which is higher than themaximum for admission to the tail-gas expan-sion turbines. The tail gas must then be cooled,either against tail gas entering the catalytic unitor with the aid of waste-heat boilers.

In the process marketed by Du Pont the tailgas is dried, preheated, and then heated in twostages. The gas is subsequently divided into twostreams. One substream is heated further, mixedwith methane, and led over the first fixed-bedcatalyst. The second substream is mixed withmethane without further heating and led overthe second fixed-bed catalyst, along with the firstsubstream. Hot tail gases exit the bottom of thereactor and are led through the tail-gas expan-sion turbine, which recovers part of the work ofcompression.

Ammonia is the only economically relevantreducing agent for selective catalytic processes.The consumption of reducing agent in the se-lective treatment of NOx is much smaller thanin nonselective processes. Furthermore, the tail-gas temperature after reduction is significantlylower, allowing the use of simpler, cheaper con-struction materials. The optimal catalyst servicetemperature is 250 – 350 ◦C, but operation at upto 500 ◦C is possible. The costs due to the useof expensive ammonia must, however, also beconsidered. To prevent formation of ammoniumnitrate (a potential explosion hazard) in the ex-pansion turbine or downstream, emissions mustbe monitored for ammonia (10 – 20 ppm).

When BASF, Ludwigshafen [85] began workon catalytic NOx reduction in the early 1960s,existing patents indicated the use of ammoniatogether with noble-metal (Pt, Rh, Ru, Pd) andiron-group (Fe, Co, Ni) catalysts [86,87]. Alter-natives were found: vanadium pentoxide, tung-sten oxide, and molybdenum oxide gave excel-lent results. Vanadium pentoxide on an aluminasupport proved to be most economical.


Selective processes also require oxygen,which oxidizes part of the nitrogen monoxideto nitrogen dioxide to ensure roughly equal lev-els of these two oxides (the best condition forreduction). Because the BASF process operatesat 200 – 350 ◦C, side reactions between ammo-nia and oxygen can be neglected. The reactionis not affected by dinitrogen monoxide, carbondioxide, or water vapor.

For a plant with a tail-gas stream of37 000 m3 /h (STP) operating at a pressure of730 kPa with an inlet NOx concentration of500 –1000 ppm, this process (270 ◦C) gives anexit concentration of 50 – 150 ppm NOx. Thetreated tail gas contains < 20 ppm ammonia(usually ca. 5 ppm), and the nitrogen dioxidelevel is 30 –50 ppm. The stack gas is generallycolorless. Careful temperature monitoring at thetail-gas expansion turbine and measurement ofthe ammonia level can prevent the depositionof ammonium nitrate. To minimize the pressuredrop, the reactor features an annular design withinward flow (Fig. 34).

Figure 34. BASF tail-gas reactor for NOx reductiona) Hole plate; b) Catalyst; c) Wire gauze

Ammonia is uniformly mixed with the tail-gas stream in static mixers. The performance ofthe treatment process is monitored by measur-ing the temperature rise during reduction (ca.10 ◦C/1000 ppm NOx ).

Uhde, Dortmund, as licensee, has retrofitted14 plants with the BASF process. The tailgas, usually containing 500 – 1000 ppm NOx, isheated to about 260 ◦C. Ammonia gas is addedto the tail-gas stream in a mixer and then ledthrough the reactor. The pressure drop is ca.25 kPa. The effluent meets the requirement ofcolorlessness.

The HGW – Didier process, developedjointly by Hamburger Gaswerke (HGW) andDidier, Essen uses a chromium oxide catalystwith a service temperature of 265 – 350 ◦C. Thenitric acid plant planned for DSM Miststoffen(Geleen, Netherlands) is designed for a tail-gasstream of 76 000 m3/h (STP) with a maximumNOx level of 2200 ppm. The exit concentrationwill not exceed 100 ppm NOx. The tail gas ispreheated before entering the reactor. To meetthe strict emission limit of < 100 ppm, the cat-alyst is located in three beds. Ammonia canbe admitted separately over each. The followingdistribution has proved most efficient: 70 % overthe first bed, 20 % over the second, and 10 %over the third catalyst bed.

Other companies also have selective catalyticreduction processes for nitric acid tail gases. Forexample, Mitsubishi Chemical has developed aprocess operating at 400 – 500 ◦C under pressure[88,89]. The “DN-Cat” catalyst is installed bet-ween the heat exchanger and the tail-gas expan-sion turbine in the conventional flow sheet.

The Weatherly selective catalytic process op-erates at 250 – 310 ◦C. The noble-metal catalystsare the same ones employed in nonselective pro-cesses, except that the fuel is replaced by ammo-nia.

The Bergbauforschung – Uhde process [90]is chiefly employed for stack-gas cleanup atpower plants but may later find use in emissionabatement at nitric acid plants. The Bergbau-forschung – Uhde process for simultaneous sul-fur and NOx removal from stack gases is uti-lized at the cold end of the power-plant boiler,i.e., downstream of the air preheater and elec-trostatic filter. It employs the adsorptive andcatalytic properties of “activated coke” at 50 –150 ◦C. The reducing agent is again gaseous am-


monia. The technique combines a selective cat-alytic process for NOx and an adsorptive one forsulfur dioxide. The catalyst is therefore installedin a moving-bed reactor.

1.5. Storage and Transportation

As a rule, nitric acid is stored in stainless steeltanks and transported in stainless steel contain-ers.

The regulations for transport by rail [91,92]and road [93,94] differentiate among > 70 % ni-tric acid, 55 – 70 % nitric acid, and < 55 % ni-tric acid, and also between mixtures with sul-furic acid (“mixed acid”) containing > 30 % ni-tric acid and those containing < 30 % nitric acid.Similar regulations [95] apply to transport byship. Important transport regulations for > 70 %nitric are as follows:

IMDG Code Class 8, D 8261, E – F 8186

UN No. 2032

RID/ADR Class 8, no. 2 a

CFR 49: 172.101, Cor. M

Regulations for < 70 % nitric acid are

IMDG Code Class 8, D 8260, E – F 8185

UN No. 2031

RID/ADR Class 8, no. 2 b

CFR 49: 172.101, Cor. M

Provisions range from details of technicalequipment to labeling and give instructions incase of accidents. Regulations for the road trans-port of nitric acid (≤ 70 % HNO3 emphasizethe associated hazards ( poisoning by inhalationof vapors, danger of chemical burns on contactwith flammable substances, formation of nitrousgases). Protective equipment includes respira-tory protection apparatus, goggles, and clothingaffording complete coverage, as well as an eye-wash bottle with pure water. In case of accident,the fire and police departments are to be notifiedimmediately. Spilled nitric acid must not be ab-sorbed with sawdust or other flammable material(because of the fire hazard); instead, its spreadmust be prevented by the construction of earthbarriers.

1.6. Uses and Economic Aspects

The principal use of nitric acid is as a starting ma-terial in the manufacture of nitrogen fertilizers;

large amounts are reacted with ammonia to yieldammonium nitrate (→Ammonia Compounds,Chap. 1.2.1.). Weak acid (ca. 60 % HNO3 ) ismost suitable for this purpose. Smaller amountsof weak acid are used to digest crude phos-phates. About 75 – 85 % of the nitric acid pro-duced is used in the fertilizer sector. In addition,porous ammonium nitrate prills are still an im-portant component of explosives (→Explosives,Chap. 7.2.).

Nitric acid is used as a nitrating agent inthe preparation of explosives (→Explosives)and organic intermediates such as nitroalka-nes (→Nitro Compounds, Aliphatic) and ni-troaromatics (→Nitro Compounds, Aromatic).It is also used in the production of adipic acid(→Adipic Acid, Chap. 4.1.).

Other applications include use as a chemicalin metallurgy (e.g., as an etchant and picklingagent for stainless steels) and in rocket fuel pro-duction.

Worldwide output of nitric acid in 1987 was26.882×106 t, representing a slight increase ofca. 4 % over 1976. Production in industrializedcountries has stagnated or even declined, but in1976 – 1985 increases in capacity in less indus-trialized countries occurred [96]. Production fig-ures for nitric acid in 1987 follow (106 t):

Europe, total 16.472

United States 6.553

Eastern Europe 4.289

Asia 1.544

Federal Republic of Germany 2.112

Poland 2.136

Belgium 1.470

Spain (1986) 1.249

Italy 1.195

Hungary 1.018

2. Nitrous Acid

Nitrous acid [7782-77-6], HNO2, Mr 47.01, is amoderately strong to weak acid. Its dissociationconstant K in highly dilute solutions at 18 ◦C is4.5×10−4. The acid is stable only in cold diluteaqueous solution. It can be prepared as follows:

Ba ( NO2)2 + H2SO4 −→ 2 HNO2+ BaSO4

On heating in the presence of sand, glass splin-ters, or other sharp-edged objects, or even at lowtemperature, it disproportionates as

Documents

Ullmann's HNO3.pdf