Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
From nano-scale studies of working sulfuric acid catalysts to improved industrial-scale sulfuric acid production Kurt Christensen, Filippo Cavalca, Pablo Beato, Jens Hyldtoft and Stig Helveg Haldor Topsoe A/S Nymøllevej 55, DK-2800 Kgs.Lyngby, Denmark Prepared for 39th International Phosphate Fertilizer & Sulfuric Acid Technology Conference Clearwater Beach, Florida June 5th & 6th, 2015 1 TABLE OF CONTENTS INTRODUCTION ...................................................................................................3 BACKGROUND ...................................................................................................... 3 DYNAMICS OF THE ACTIVE PHASE............................................................... 5 CHEMISTRY AND COLORS ............................................................................. 10 CONCLUSIONS .................................................................................................... 14 REFERENCES....................................................................................................... 15 2 INTRODUCTION The main purpose of a catalyst is to increase the rate of chemical reactions without being consumed. This is also the main function of the V2O5-based sulfuric acid catalyst, which has been used for more than 50 years for SO2 oxidation in sulfuric acid production: SO2 + ½ O2 ⇌ SO3 + heat (1) In general, however, a catalyst is also a dynamic system for which the detailed chemical composition and nano-scale structure depend on the composition, temperature and pressure of the surrounding gas. As a consequence, the catalyst adopts different conformations in the individual beds of a sulfuric acid converter, even if the same catalyst type were loaded in all beds. One effect of the dynamic nature of the catalyst is that after a change of feed gas composition, plant load or bed temperatures, the catalyst will adjust to the new local conditions on a time scale from minutes to a couple of days depending primarily on the temperature. Another implication of the dynamic behavior is the different colors of the operating catalyst that may also sometimes be observed after a shutdown depending on the shutdown procedure. As a catalyst producer, Topsoe continuously aims to improve the fundamental understanding of catalysis in the industrial environments and to apply this knowledge to rationally develop better catalysts for improved efficiency of industrial sulfuric acid production. In the current paper, we present results from novel advanced in situ techniques recently introduced in our laboratories and combine the findings with bench-scale investigations and large-scale observations from industrial plants to obtain a more complete picture of the working catalyst. BACKGROUND Commercial sulfuric acid catalysts are based on V2O5 mixed with alkali-metal sulfates on an inactive porous silica support. The carrier is usually made from diatomaceous earth, and the catalyst typically contains 2-5 wt% vanadium and 2-5 moles of alkali metal promoter per mole of vanadium. The alkali promoters are typically potassium, sodium and cesium, and the catalysts are available as cylindrical pellets, cylindrical rings or finned rings resembling a Daisy as shown in Figure 1. Figure 1. Sulfuric acid catalyst from Topsoe in the Daisy shape. 3 Upon start-up, the catalyst takes up sulfur oxides from the gas environment to form molten alkali pyrosulfates: M2SO4 + SO3(g) ⇌ M2S2O7 M2S2O7 + SO3(g) ⇌ M2S3O10 (M=Na, K, Cs) (2) In this type of supported liquid phase (SLP) catalysts, the oxidation of SO2 takes place as a homogeneous reaction in a liquid film covering the internal surface of the support material [1, 2]. The promoting role of the alkali metals is commonly attributed to their ability to form these relatively low-temperature-melting pyrosulfates, which dissolve the vanadium oxides. The amount of sulfur oxides bound as alkali metal pyrosulfates depends on the actual reaction conditions. Under circumstances for which the catalyst is blown with hot air for a sufficiently long time, up to 10% of the catalyst weight is desorbed as SO2/SO3 leaving the alkali metals in the form of sulfates, M2SO4. The catalytic activity of the melt was demonstrated conclusively in 1948 by Topsøe and Nielsen [3], who obtained a high degree of conversion to SO3 by bubbling SO2 and oxygen through a column packed with raschig rings and containing molten potassium pyrosulfate in which 14% wt% V2O5 was dissolved. Since then, numerous investigations on the coordination chemistry of the catalytic melt and the reaction mechanism have been published, but in spite of this the reaction mechanism is not understood in detail yet. For many years, the mechanism was assumed to include reduction of vanadium to V(IV) and re-oxidation to V(V) by oxygen as proposed by among others Mars and Maessen [4]. However, in 1968-1973 Boreskov and coworkers [5, 6] found from transient kinetic studies a number of facts contradicting previously suggested redox mechanisms, for example that the rate of SO2 oxidation is much higher than the rate of V reduction. Numerous subsequent studies have confirmed that an associative cycle including only V(V) complexes is the governing mechanism [7] as illustrated in Figure 2. V SO3 -4 (V O)2O(SO4)4 4 -4 SO2 IV -2 2V O(SO4)2 O2 2VIVOSO4 (s) 1 (VVO)2O(SO4)4SO3 3 -2 2SO4 SO3 SO2 (VVO)2O(SO4)4O2-4 2 (VVO)2O(SO4)4O-4 SO2 SO3 Figure 2. The catalytic cycle of a sulfuric acid catalyst. Adapted from [7]. 4 However, V(IV) compounds still play an important role for the activity of the catalyst, because an equilibrium exists between V(V) and V(IV) compounds in the melt. The degree of reduction to inactive V(IV) increases at low temperature and high SO2 partial pressure. Furthermore, at temperatures below 500°C, part of the V(IV) compounds precipitate and cause depletion of V(V) in the melt, and this partial solidification eventually causes the activity to drop to practically zero at minimum operating temperatures of 300-350°C. The reduction to V(IV) shown in Figure 2 is not a deactivation but rather a dynamic and reversible response of the catalyst to the gas environment. That is, increasing temperature or lowering SO2 partial pressure results in a recovery of the catalyst activity. Understanding the complex composition of the catalytic melt is a scientific challenge that requires advanced characterization techniques in itself. However, the dynamic response of the catalyst to the gas environment is an additional severe complication, as ex situ studies of the cold catalyst in air, inert gas or vacuum may lead to erroneous conclusions. Hence, new advanced in situ techniques are needed to make further progress in the understanding of the dynamics and mechanism of the sulfuric acid catalyst. We have met this challenge by introducing novel experimental methods. In the following sections, examples and learnings from our current studies at Topsoe are presented. DYNAMICS OF THE ACTIVE PHASE The catalytic melt of a sulfuric acid catalyst typically accounts for 20-40% of the catalyst mass. The melt is considered mobile under operating conditions, although the dynamic behavior is difficult to address in detail at room temperature, where the active phase appears as solid lumps of oxides and sulfates of vanadium and alkali metals on the internal surface of the carrier as shown in Figure 3. The cylindrical structures seen in the figure are diatom skeletal fractions, which originate from the diatomaceous earth typically used as carrier. Figure 3. Scanning Electron Micrograph (SEM) of the inner surface of a cleaved sulfuric acid catalyst. 5 When heated and put on stream in SO2/O2/SO3 gas, the sulfates take up SO3 and form pyrosulfates according to (2). This transformation causes the active phase to melt and redistribute on the carrier. Due to capillary forces, however, the liquid stays in the pores of the carrier. This redistribution is illustrated in Figure 4. In this lab experiment, a pelletized porous silica carrier was partially impregnated by submerging one end of the pellets for 30 seconds into an impregnation liquor containing appropriate salts of V, K and Na. After calcination in air at 400°C for 15 min, visual inspection and scanning electron microscope (SEM) energy-dispersive X-ray spectrometry (EDX) show how one end is fully impregnated while V, K and Na are absent in the other end, cf. Figure 4a. Subsequently, the pellets were put on stream in a glass reactor in a feed gas with 10% SO2 and 10% O2 at 500°C for three days. After this treatment, the same analyses demonstrate a uniform color and distribution of V, K and Na over the pellet, revealing a marked migration of the active phase to the unimpregnated end of the pellets, as shown in Figure 4b. Importantly, this redistribution is associated with a 50% increase of catalytic activity from its initial value until the test termination. Figure 4. Redistribution of the active phase in a sulfuric acid catalyst. Photo and SEM EDX line scan of the sample loaded in glass reactors a) before test and b) after three days on stream. In situ Transmission Electron Microscopy (TEM) The lab experiment described above clearly shows the mobility of the catalytic melt in the sulfuric acid catalyst, but in order to directly unlock the state of the melt in the working catalyst and its dynamic behavior in the heterogeneous pore system, direct observations at nano-scale resolution are needed. We have therefore recently introduced transmission electron microscopy (TEM) as a novel means for directly monitoring catalyst samples reacting with an SO2/O2 gas mixture at temperatures 6 up to 600°C and at nanometer resolution. These time-resolved observations provide unprecedented new insight into e.g. a simulated converter startup. As the industrial catalyst is quite heterogeneous across the micro- and nano-scales (cf. Figure 3), a more well-defined model system is used for the initial studies. The model catalyst was prepared by impregnation of 100 nm monodispersed silica spheres with a mixture of vanadium oxide and alkali pyrosulfate to a composition of 4.2 wt% V content and molar ratios of K/V=3.5 and Cs/V=0.4. The as-prepared sample was investigated in a Philips CM300 transmission electron microscope in vacuum (at about 10-9 bar) using a TIETZ F-114 charged-coupled device camera for TEM imaging and a Gatan Imaging Filter (GIF) for electron energy loss spectroscopy (EELS) and energy-filtered TEM (EFTEM). Figure 5 shows combined TEM and V L2,3 EFTEM images revealing that the active phase of the fresh catalyst is present in more concentrated lumps of material (marked by red circles). The combined use of these techniques demonstrates that the vanadia phase can be mapped out at nanometer resolution in the as-prepared catalyst. Figure 5. A cluster of silica spheres with the vanadia phase resolved by TEM (left) and V L2,3 ETEM (right). The red circles indicate regions of relative higher vanadium concentration. Whereas a TEM normally operates at ambient temperature and very low pressure in the order of 10-9 bar, the CM300 transmission electron microscope is equipped with a differentially pumped vacuum system that allows solid samples to be exposed to reactive gas environments up to 1-10 mbar in pressure and up to 600°C while still maintaining atomic resolution [8], cf. Figure 6. Furthermore, as a unique capability, this particular microscope is entirely dedicated to studies involving sulfur-containing gases, in relation to e.g. hydrodesulfurization and sulfuric acid catalysts, in order to avoid that cross-contamination and sulfur-corrosion of the microscope’s interior parts interfere with other experiments [9]. 7 In the present study, crushed catalyst powder was loaded on a micro-electromechanical-system (MEMS) heater and placed in the microscope. After an initial inspection in vacuum, the samples were exposed to a gas mixture of 50% SO2 and 50% O2 at a total pressure of 10 mbar. Although this total pressure is low compared to that in industrial reactors operating at 1.1-1.4 bar, it is much closer to industrial values than a typical TEM operating pressure, and a partial pressure of SO2 of 5 mbar is still sufficient to keep the catalyst in its relevant pyrosulfate form. Furthermore, the effects of the electron beam on the observations have been investigated to establish experimental protocols in which the electron microscopy observations are negligibly affected by the electron beam. Figure 6. A photo of Topsoe’s in situ CM300 transmission electron microscope dedicated to S-rich experiments. The reactive gas environments are confined to the sample region by means of a differential pumping system (sketch adapted from [8]). The gas composition in the experiment is monitored by the ion current from a mass-spectrometer attached to the second differential pumping stage (graph). 8 Figure 7 shows results from a heating of the as-prepared catalyst sample in the gas environment. Up to 350°C, the TEM images show that the nano-scale structure of the catalyst resembles the as-prepared sample as illustrated in Figure 7a. Upon heating to higher temperature, the images reveal significant dynamic changes in the catalyst structure. For instance Figure 7b shows that a redistribution of mass has occurred as the sample is heated to 600°C, corresponding to the high temperature found in the lower part of the first bed of typical sulfuric acid converters. Subsequent cooling to 350°C and a repeated heating to 600°C results in even more pronounced changes (Fig. 7c). Several TEM pictures taken during the experiment are shown consecutively as a movie providing the first-ever vivid impression of liquid phase restribution in the catalyst. Such dynamic observations open up possibilities to extract information on wetting properties and dynamics of melt distribution. Moreover, the observations can also be coupled with chemical composition and catalytic activity from e.g. operando Raman (see next section). Figure 7. In situ TEM of dynamic changes of the model catalyst during exposure to 10 mbar of 50% SO2 and 50% O2 (a) at 350°C (b) at 600°C and (c) after cooling to 350°C and reheating to 600°C. The SiO2 spheres have a diameter of ca. 100 nm. Liquid loading and dust protection One learning from the dynamic studies of the active phase is that activation of the catalyst distributes the melt on the carrier in order to reduce the surface free energy of the combined carrier-melt system at the prevailing temperature and gas composition. At steady-state, the distribution depends on the surface properties of the silica support, e.g. the wettability, surface tension and carrier morphology. However, in the context of catalyst design another related and very important parameter of an SLP catalyst is the liquid loading defined as the fraction of the pore volume of the support filled with melt. This is investigated in Figure 8, where we have varied the melt load on a catalyst in our lab and measured activity in 10% SO2, 10% O2 and 500°C. The activity of the catalyst increases proportionally to the liquid loading (which is proportional to the vanadium content) at low loadings but decreases at high loading, which is explained by gas phase pore diffusion restriction and liquid diffusion restriction in the deeper melt pools. As a consequence, an optimum melt loading exists for a given melt composition and 9 operating condition. Another implication is that higher vanadium content is not always better for activity. The melt normally stays in the pores of the porous catalyst in the attempt to reduce total surface energy. However, if the catalyst is covered in dust, especially fine dust with high surface area, part of the melt will gradually migrate to the smaller pores of the dust causing depletion of vanadium and alkali metals. Depending on the initial melt loading, the consequence for activity may be modest in the beginning but at high dust build-up and repeated screening the conversion will eventually drop (see Figure 8) and call for catalyst replacement. Figure 8. Activity vs. liquid loading of a sulfuric acid catalyst prepared in lab. Another consequence of the molten active phase is the stickiness of the sulfuric acid catalyst under operating conditions. As a result, dust present in feed gases in sulfuric acid plants typically deposits and accumulates within the top 10-15 cm of catalyst in the first bed of the SO2 converter. Depending on the dust concentration, particle size, void fraction and catalyst stickiness, this leads to a pressure drop build-up over time which eventually may cause a costly shutdown of the plant necessary for screening of the catalyst. A simple way to prolong the time between screenings is to install a 10-15-cm top layer of VK38 25-mm Daisy, which distributes the dust further down the bed due to its larger size and high void fraction [10]. The top layer should be changed after a few catalyst screenings as it will lose some of its stickiness when the melt is depleted by the fine dust. CHEMISTRY AND COLORS Sulfuric acid catalysts are normally characterized by yellow, gold or orange colors indicative of vanadium in +5 oxidation state V(V). Catalyst exhibiting a green, pale green or pale blue color are considered to contain vanadium in the +4 oxidation state V(IV). As the oxidation state couples to the reaction environment, it will change and adapt to the surroundings. For instance, if the catalyst is exposed to 10 moisture, the appearance and colors also change. When a catalyst is put on stream, the catalyst composition adjusts to the gas composition and temperature, for example with respect to the V(V)/V(IV) equilibrium illustrated in Figure 2. As a consequence, the catalyst color also changes, although this is not visible in a steel converter. Such changes will also happen during shutdown and depend on the applied procedure. All in all, the color of a catalyst is not a valid method for assessing how it will perform or have performed in a converter. Operando Raman In order to study the chemical transformations taking place in the operating sulfuric acid catalyst, we are using an advanced operando Raman setup enabling both the chemical composition to be studied by Raman spectroscopy and the sample color to be observed by visual inspection [11]. Laser Raman spectroscopy provides a vibrational spectrum, which contains information on the chemical bonding in the sample. Operando Raman spectroscopy is a technique that is able to monitor the structural changes in a sample on a molecular level during the reaction. For operando Raman studies it is particularly important to avoid laser-induced heating, which, in the worst case, can lead to thermal degradation of the sample. To alleviate this problem, we have developed a novel technique that allows the sample to be studied in a small fluidized bed reactor for securing uniform and well-defined temperature, cf. Figure 9. Figure 9. Detailed sketch of the Linkam CCR1000 reactor used for in situ Raman (a) and schematic drawing of the sample holder with flow indications by arrows to illustrate the fluidization principle (b) [11]. In order to simulate a converter startup, the Raman reactor was loaded with a 150300 µm sieve fraction of crushed VK38, and the fluidized bed of catalyst heated first to 380°C in air at 1 atm. Under this condition, the catalyst is yellow/orange, 11 and the Raman spectrum shown in Figure 10a resembles some characteristic features of the dimeric V(V) complex, which are shown as the active species in Figure 2. In particular, the two bands at ∼860 and 770 cm−1, which are assigned to V-O-S and V-O-V bridging bonds, have been claimed to be the signature of the active catalyst [12]. The lack of V(V)-vanadyl bands (typically at ~1040 cm-1) indicates the extended polymeric nature of the catalyst under this condition. After switching the reactor feed gas to 10% SO2 and 10% O2 in nitrogen at 380°C in Figure 10b, the catalyst color changes to greenish, and the Raman spectrum reveals structural transformation from a polymeric vanadate to the well-defined molecular structure of the (VO)3(SO4)54- ion implying a reduction of V(V) to V(IV). At this point, the conversion of SO2 measured by infrared (IR) SO2 analysis of the effluent is very low. As the temperature is raised to 480°C, a significant increase in conversion is observed, and the color changes from greenish to ocher, as seen in Figure 10c. The corresponding Raman spectrum shows a band centered at 1043 cm−1 characteristic for V(V)-vanadyl bonds and a broad and structured band between 940 and 985 cm−1 together with two bands at 665 and 605 cm−1 assigned to the vibrational modes of the SO4 ligands in different vanadium oxosulfate complexes. From a detailed analysis of the spectra, it is possible to conclude that the active phase of the industrial VK38 catalyst at 480°C corresponds to a mixture of predominantly monomeric and to a minor extent dimeric vanadium(V) oxosulfate species, which fits well with the proposed structures in Figure 2. Figure 10. Operando Raman spectra while switching from synthetic air (a) to SO2/O2 (b) at 380°C and in SO2/O2 at 480°C (c). The photographs to the left of the spectra show the catalyst in the reactor under the respective reaction conditions [11]. 12 Colors in industry The studies presented above show that the composition and color of the operating catalyst change with the temperature and gas composition. Catalyst containing vanadium primarily in oxidation state +5 is yellow/orange, but if it is loaded in a converter and operating at low temperature or high partial pressure of SO2, the color will change to greenish/bluish because these conditions favor reversible reduction of vanadium to oxidation state +4. Also the catalyst may change visual appearance during a converter shutdown depending on the purging and cooling procedure or during storage. When purged with air while the catalyst is still hot, vanadium will be oxidized to V(V) and attain a yellow color even if it was green when operating in the converter in e.g. the top of the first bed. As an example, Figure 11 shows how a greenish catalyst sample from industry can change color to yellow when heated in an air flow in a glass tube for 24 hours at 500°C. Figure 11. Colors alone cannot be used for assessment of the catalyst performance. Here reduced vanadium is reoxidized by heating of a greenish sample in air. Catalyst design for optimal performance The reaction rate for a catalyst may be limited either by external mass and heat transfer to the pellet surface, by internal mass transfer in the porous support or by the intrinsic reaction rate itself. For sulfuric acid catalysts operating at a low temperature, the main limitation is the intrinsic reaction rate, which depends on the chemistry in the catalytic melt. In particular, as the molecular turnovers happens on vanadia species in the melt, accessibility to the bulk of the melt by the gaseous reactants is important to optimize activity and it is therefore crucial to control the distribution of the catalytic melt on the support and gas solubility and transport through the melt. The present in situ techniques provide valuable insight into the mechanisms governing the intrinsic catalyst performance, and the learnings are integrated in Topsoe’s continuous developments of improved industrial catalysts. One example is the significant changes of the melt properties, upon replacement of part of the potassium and sodium sulfates by cesium sulfates. The introduction and 13 optimization of cesium-promoted catalysts, e.g. VK59 and VK69, for lowtemperature operation had a significant impact on possible inlet temperatures and SO2 emissions in industry [13]. The fundamental studies also revealed how the activity of normal commercial sulfuric acid catalysts is seriously surpressed by deep pools of catalytic melt restricting gas transport in the active phase. These limitations have been overcome in the new high-activity VK-701 catalyst based on our novel LEAP5™ technology [14]. The improvements include a very high V(V) content and were accomplished by changing the intrinsic morphology and surface properties of the carrier and optimizing the active phase for high SO3 concentration. CONCLUSIONS Catalysts are dynamic systems that adapt to the local environment in which they operate. As a consequence, in situ studies at relevant temperatures, pressure and gas composition are necessary to get a true picture of the working catalysts. This is particularly true for sulfuric acid catalysts for which the active phase is a liquid making up about one third of the catalyst mass, i.e. a so-called supported liquid phase (SLP) catalyst. In order to look directly at the state of the working catalyst, we have recently introduced new advanced in situ techniques including highresolution transmission electron microscopy and Raman spectroscopy to directly resolve the dynamic state of catalyst samples interacting with an SO2/O2/SO3 gas mixture at temperatures up to 600°C. These techniques provide unprecedented insight into sulfuric acid catalysis. In situ high resolution TEM studies combined with lab-scale results show how the active phase takes up SO3 and becomes mobile. From time-resolved series of TEM images, the rate of the dynamic changes and the final wetting characteristics and melt distribution can be derived. The operando Raman observations demonstrate that color changes occurring during a simulated converter startup are directly related to measured catalyst chemistry and observed SO2 conversion. The results indicate that the vanadium changes from a dimeric V(V) complex in hot air to primarily V(IV) compounds in SO2/O2 gas at 380°C. When the temperature is increased to 480°C, vanadium is reoxidized, and the spectral information obtained supports that the active phase of the industrial VK38 catalyst at 480°C corresponds to an equilibrium between predominantly monomeric and to a minor extent dimeric vanadium(V) oxosulfate species. The studies presented here show that the melt distribution and chemistry interact with the gas environment and that its physical and dynamical state is just as important as the gross composition of a sulfuric acid catalyst for understanding and controlling activity. Moreover, the detailed in situ observations also help explaining macro-scale phenomena in the industrial converter such as catalyst colors, oxidation states, melt transformation, melt extraction by dust, pressure drop build-up and catalytic activity. 14 REFERENCES 1. Frazer, J.H. and Kirkpatrick, W.J. (1940). J. Am. Chem. Soc., 62, 1659. 2. Villadsen, J. and Livbjerg, H. (1978). Catal. Rev.-Sci. Eng., 17 (2), 203272. 3. Topsøe, H.F.A. and Nielsen, A. (1948). The Action of Vanadium Catalysts in the Sulfur Trioxide Synthesis. Trans. Dan. Acad. Techn. Sci. 1, 3-24 4. Mars, P. and Maessen, J.G.H. (1968). J. Catalysis 10, 1-12. 5. G.K. Boreskov, R.A. Buyanov, A.A. Ivanov (1967). Kinet. Katal., 8-1, 153 (English translation). 6. G.K. Boreskov, G.M. Polyakova, A.A. Ivanov, V.M. Mastikhin (1973). Dokl. Akab. Nauk SSSR, 210 – 3, 626 (English translation). 7. Lapina, O.B., Bal’zhinimaev, B.S., Boghosian, S., Eriksen, K.M. and Fehrmann, R. (1999). Catalysis Today, 51, 469-479. 8. P.L. Hansen, S. Helveg, and A.K. Datye (2006). Adv. Catal. 50, 77-94. 9. L.P. Hansen, E. Johnson, M. Brorson, S. Helveg (2014). J. Phys. Chem. C. 118 (39), 22768–22773. 10. Christensen, K. and Thellefsen, M. (2007). New Improved Dust Protection Catalyst. Paper presented at the Sulphur 2007 Conference, Montreal. 11. Pablo Beato, Eva Schachtl, Kata Barbera, Francesca Bonino and Silvia Bordiga (2012). Catal. Today. 205 (2013), 128–133. 12. Boghosian, S., Chrissanthopoulos, A., and Fehrmann, R. (2002). J. Phys. Chem. B, 106 (1), 49-56. 13. Jensen-Holm, H. and Hansen, L. (1997). Demonstrated Performance Improvements in Sulphuric Acid Plants Using VK69 Catalyst. Sulphur ’97, pp 193-206, British Sulphur Publishing, London. 14. Christensen, K. (2010). Meeting future SO2 emission challenges with Topsøe’s new VK-701 LEAP5™ sulfuric acid catalyst. Paper presented at the Sulphur 2010 Conference, Prague. 15