Catalytic membrane reactor for the oxidation of NHg to NO. and method for the oxidation of NH3 to NO.
The present invention relates to a catalytic membrane reactor for NH3 oxidation in nitric acid manufacture, the membrane reactor comprising a membrane comprising a muliticomponent oxide having a perovskite or perovskite-related structure, and a method for the oxidation of NH3 to NO.
Ammonia is oxidized over Pt-Rh alloy gauzes to form NO as the first step in the industrial production of nitric acid, a process that has been practically unchanged for over 80 years. The reaction typically yields 94 - 96 % of NO and 4 - 6 % of by-products (N2O and N2) at 1073 - 1223 K. Major drawbacks associated with the platinum-based catalysts are: (i) high production cost, (ii) metal loss in the form of volatile oxides, necessitating efficient metal recovery (Pd catchment) and refining systems, and (iii) the production of N2O, an environmentally harmful gas. Nitric acid production is the largest single source of N2O in the chemical industry (125 Mton CO2 - eq. per year), and the development and implementation of abatement technology for this gas is being required. The above aspects have stimulated research for replacing noble metals by oxide catalysts for NH3 oxidation. Oxides may offer the advantage of lower investment and simpler manufacture and a reduced N2O emission. It is known in the art to use spinels or perovskites in the reaction, preferably containing Co, but also Fe, Mn, Bi, or Cr (see review paper V.A. Sadykov, L.A. Isupova, I .A. Zolotarskii, L.N. Bobrova, A.S. Noskov, V.N. Parmon, E.A. Brushtein, T.V. Telyatnikova, V.I. Chernyshev, V.V. Lunin, Applied Catalysis A. Geneneral 2000, 204, 59). Laboratory, pilot, and industrial tests have typically been carried out in fixed-bed reactors with oxides in the form of particles, pellets or monoliths. Several key aspects have prevented the industrial implementation of oxides: (i) a relatively low NO selectivity (< 90 %), (ii) the rapid deactivation under relevant reaction conditions, and (iii) the lower optimal operation temperature compared to noble metal catalysts, causing difficulties with the steam balance in a revamped plant.
US 5306411 relates to a solid, gas-impervious, electron-conductive, oxygen ion- conductive, single-phase membrane for use in an electrochemical reactor, said
membrane being formed from a perbvskite represented by the formula AsA'tBuB'vB"wOx wherein A represents a lanthanide, Y, or mixture thereof; A' represents an alkaline earth metal or mixture thereof; B represents Fe; B' represents Cr1 Ti, or mixture thereof; and B" represents Mn, Co, V, Ni1 Cu, or mixture thereof and s, t, u, v, w, and x each represent a number such that s/t equals from about 0.01 to about 100; u equals from about 0.01 to about 1 ; v equals from about 0.01 to 1 ; w equals from zero to about 1 ; x equals a number that satisfies the valences of the A, A', B, B' and B" in the formula; and 0.9<(s+t)/(u+v+w)<1.1.
WO 03/037490 relates to a solid multicomponent mixed proton and electron conducting membrane for use in a reactor, where the membrane comprises a mixed metal oxide having a structure represented by the formula A1-xA'x(B1-yB'y)wθ3-ci> wherein A is a lanthanide element or a mixture thereof, A' represents an alkaline earth metal or a mixture thereof, B is chromium, manganese, or iron, B' is titanium, aluminium, zirconium, or hafnium, and x, y, w, and d each represent a number such that 0 <x <1 , 0 <y <1 , 0.9 <w <1.1 , and d equals a number that renders the compound charge neutral and is not less than zero and not greater than about 0.6.
The main object of the present invention was to arrive at a membrane reactor and a method for the oxidation of NH3, where oxygen would be separated from nitrogen by the use of a membrane and ammonia would be oxidized on the surface of the same membrane by the separated oxygen.
Another object was to arrive at a more cost effective method for NH3 oxidation by using a membrane reactor.
A further object was to reduce the emission of N2O.
These and other objects of the invention were obtained by the product and method as described below. The invention is further characterized by the patent claims.
To solve the object of the invention, the inventors started a work to find suitable membranes. It was found that with a lanthanum ferrite-based perovskite membrane for ammonia oxidation, it could be possible to attain NO selectivities up to 98 % and no N2O formation during the oxidation. The strategy was to integrate in a single reactor the separately reported properties of the above perovskites as oxygen conductors and
catalysts for NH3 oxidation. Accordingly, the applied configuration, combines (i) separation of O2 from air in the feed side by transport through oxygen vacancies in the mixed conducting membrane and (ii) reaction of oxygen species with ammonia on the membrane surface at the permeate side to form NO. This achievement gave rise to a radically intensified process for nitric acid manufacture, since large amounts of inert N2 (2/3 of the total flow in today's plants) are excluded.
Important features of perovskite membranes in ammonia oxidation include oxygen flux, catalytic performance (activity and selectivity), and chemical stability in reducing and oxidizing atmosphere at high temperature. These interrelated parameters could be tailored by tuning the degree of calcium and strontium substitution in lanthanum ferrite- based perovskites (Lax(Sr1Ca)xFeO3-2). A higher substitution level (x') increases the number of oxygen vacancies (z) and thus the oxygen flux, but at the expense of lower chemical stability. Besides, La-substitution by alkaline-earth cations in the perovskite structure is also known to influence the catalytic properties of these mixed oxides. This can be caused by (i) variation in the charge and/or coordination of 3d cations, (ii) change of the surface chemical composition, and (iii) development of micro-heterogeneities at the catalyst surface.
The present invention will in its widest scope comprise a catalytic membrane reactor for the oxidation of NH3 to NO, where the membrane reactor comprises a mixed electronic and ionic conducting membrane which is capable of transporting oxygen, comprising a multi-component oxide having a perovskite or perovskite-related structure represented by the formula
AxAVBO3*
wherein A represents a lanthanide element or a mixture thereof, A' represents an alkaline earth metal or a mixture thereof, B represents a transition metal or a mixture thereof, x and x' each represent numbers such that 0 <x <1.1 , 0 <x' <1.1 and 0.9 < (x+x1) <1.1 , and z represents a number rendering the compound charge neutral.
Particularly suitable compositions according to the present invention are represented by said general formula wherein the element A of the enumerated formula substantially represents La. Even more suitable compositions according to the present invention are represented by said general formula wherein the element A of the enumerated formula
substantially represents La and the element A' represents Sr or Ca or a mixture thereof. Even more suitable compositions according to the present invention are represented by said general formula wherein the element A of the enumerated formula substantially represents La, the element A' represents Sr or Ca or a mixture thereof, and the element B substantially represents Fe. Preferably, x and x' each represent numbers such that 0.5 <x <0.98, 0.05 <x' <0.5 and 0.97 <(x+x') <1.03.
The invention further relates to a method for the oxidation of NH3 to NO where an oxygen containing gas is fed to one side of a mixed electronic and ionic conducting membrane which is capable of transporting oxygen, and an NH3 containing gas is fed to the other side of said membrane, oxygen is dissolved into the membrane from the oxygen containing gas and transported through the membrane to the surface contacting the NH3 containing gas, NH3 adsorbs on the surface of the membrane and reacts with oxygen dissolved in the membrane to a reaction product mixture mainly consisting of NO and H2O, wherein said membrane comprises a multi-component oxide having a perovskite or perovskite-related structure represented by the formula
AxAVBO3*
wherein A represents a lanthanide element or a mixture thereof, A' represents an alkaline earth metal or a mixture thereof, B represents a transition metal or a mixture thereof, x and x' each represent numbers such that 0 <x <1.1, 0 <x' <1.1 and 0.9 < (x+x1) <1.1 , and z represents a number rendering the compound charge neutral.
Particularly suitable compositions according to the present invention are represented by said general formula wherein the element A of the enumerated formula substantially represents La. Even more suitable compositions according to the present invention are represented by said general formula wherein the element A of the enumerated formula substantially represents La and the element A' represents Sr or Ca or a mixture thereof. Even more suitable compositions according to the present invention are represented by said general formula wherein the element A of the enumerated formula substantially represents La, the element A' represents Sr or Ca or a mixture thereof, and the element B substantially represents Fe. Preferably, x and x' each represent numbers such that 0.5 <x <0.98, 0.05 <x' <0.5 and 0.97 <(x+x') <1.03. The temperature during oxidation is 900 - 1500 K1 preferably 970 - 1350 K.
The invention is further described and explained in the following examples and figures.
Figure 1 shows the principle of separation of O2 from air in the feed side of a mixed conducting membrane by transport through oxygen vacancies in said membrane and reaction of oxygen species with ammonia on the membrane surface at the permeate side to form NO.
Figure 2 shows quartz reactor used in the membrane tests.
Figure 3 shows NO selectivity and O2 flux vs. inlet ammonia flow at different temperatures for a Lao.8Sro.2Fe03-2 membrane.
Figure 4 shows Maximum NO selectivity (solid symbols) and corresponding oxygen flux (open symbols) at 1200 K vs. the substitution level in {*,o) Ca and (♦.<>) Sr-substituted lanthanum ferrite perovskites.
Example 1
Lanthanum ferrite-based perovskites (LaxAVFeO3-2, with A' = Ca, Sr, x = 0.8 - 0.9 and x' = 0.1 - 0.2) were prepared by a conventional wet complexing route using citric acid.
Standardised nitrate solutions (1 M) of the metals were mixed with citric acid in a molar ratio of acid to total cations = 3. The resulting solution was slowly heated to 433 K
(40 K h"1) to ensure complete complexation and left overnight at this temperature to dry.
After further heating to 773 K for 5 h in flowing air for the removal of organic matter and final calcination at 1173 K for 10 h, all powders were single - phase perovskites as evidenced by XRD. The particle size obtained by laser scattering was in the range of 0.1
- 0.5 μm. Dense membrane disks (> 95 % of theoretical density) were made by uniaxial pressing and sintering at 1573 K. After final grinding and polishing, membrane disks with a diameter of 10 mm (ca. 80 mm2) and a thickness of ca. 0.9 mm were obtained.
Example 2
A membrane disk of composition Lao.8Sr0.2Fe03-z was prepared as described in Example
1 , and mounted in a quartz microreactor, as shown in Fig. 2, which was heated to 1333 K for sealing it using gold rings. The oxidation of ammonia was investigated at
1000 - 1333 K by feeding an equimolar mixture of O2:Ar to the feed side and mixture of
NH3:He to the permeate side. The inlet ammonia flow was varied in the range of 0.05 - 4.5 NmI NH3 min"1, keeping a total flow of 130 NmI min"1. Product gases were analyzed on-line using a mass spectrometer and a gas chromatograph. The absence of leakages was verified by the absence of Ar in the permeate side of the membrane. The NO selectivity was obtained from the concentrations at the reactor outlet according to S(NO) = C(NO)/(C(NO)+2C(N2O)+2C(N2)). The oxygen flux J(O2) was determined from the measured concentrations of all the O-containing species. Stable membrane performance over the testing period (10 days) was verified by periodically repeating measurements at selected temperature/flow conditions.
Fig. 3 illustrates the dependence of the NO selectivity and O2 flux on the inlet NH3 flow at different temperatures. In the temperature range investigated, NO selectivities in the range of 90 - 100 % could be obtained by adjusting the inlet NH3 flow. No N2O was formed (< 10 ppm) in the experiment, N2 being the only N-containing by-product. The conversion of ammonia (80 - 95 %) increased with temperature and decreased at high ammonia flow rates. The oxygen flux is expected to increase with temperature at a fixed oxygen potential gradient. Within our experimental conditions, however, the oxygen flux was almost exclusively controlled by the ammonia flow rate and thus independent of temperature. Consequently, at constant ammonia flow rate the oxygen potential gradient decreased with increasing temperature at a fixed oxygen flux. This feature is caused by the steepness of the oxygen partial pressure vs. the O2/NH3 ratio around stoichiometric conditions corresponding to complete conversion to NO and H2O. The NO selectivity, however, revealed a dependence on temperature, and moreover exhibited a maximum with the inlet ammonia flow. As expected, the NO selectivity decreased in favour of N2 under conditions of excess ammonia due to a favourable recombination of adsorbed NHx fragments in close proximity. The additionally observed feature of decreasing selectivity at low ammonia flow rates, giving rise to the maximum, was attributed to partial oxidation on the hot reactor walls of NH3 bypassing the membrane with the excess oxygen recombining and escaping the membrane surface. This undesired reaction was enhanced with temperature, which explains the progressively decreased maximum NO selectivity when going from 1000 K (98 %) to 1333 K (92 %). Hence, higher NO selectivities than those displayed in Fig. 3 would have been obtained if wall effects could be excluded. The shift of the NO selectivity maximum to lower ammonia molar flow with decreasing temperature reflects the reduced oxygen flux (at constant oxygen potential) at lower temperatures.
Example 3
Two membrane disks of composition Lao.9Cao.i FeO3-2 and one of composition Lao.85Cao.i 5FeO3-2 were prepared as described in Example 1 , and experiments were carried out as described in Example 2 for each of the membranes. The general observations regarding dependence of NO selectivity and oxygen flux on temperature and ammonia flow rate were as described in Example 2. Fig. 4 shows the maximum NO selectivity and corresponding oxygen flux as functions of the substitution level of Ca and Sr in the lanthanum ferrite perovskites at 1200 K for the three membranes of the current Example and the membrane of Example 2. As expected, the oxygen flux corresponding to maximum NO selectivity was proportional to the substitution level. The maximum NO selectivity (95 - 96 %) did not depend on the type and degree of substitution within this range, indicating a similar catalytic performance of these membrane compositions.
The above results demonstrate the potential of the present invention, leading to NO selectivities up to 98 %, comparable to the state-of-the-art Pt-Rh alloys. In addition, formation of undesirable N2O is totally suppressed. The implementation of an ammonia oxidation process based on oxygen-conducting membranes would constitute a major step change in nitric acid production (a top -10 product in the bulk industry), strongly impacting the fertilizer industry. Apart from the superior NO selectivity, in situ separation of O2 from air by application of membranes enables extremely compact and intensified production units, since N2 represents ca. 70 % of the total flow in a current plant. The total flow reduction poses the potential for a drastic size reduction of key units of plants, including the absorption tower and the tail-gas train, as well as the obvious intensification in piping. Moreover, there are energy savings in compression of the NOx gas before the absorption step, which requires high pressure. In summary, a more efficient and sustainable process may be obtained, which is especially attractive for on- purpose nitric acid production, i.e. decentralized from large existing plants.