Characterisation of NH3 adsorption on a V- SCR catalyst

(1)

Characterisation of NH3 adsorption on a V- SCR catalyst

Elin Sjöberg

Chemical Engineering, master's level 2019

Luleå University of Technology

Department of Engineering Sciences and Mathematics

(2)

Acknowledgements

This master’s thesis was made as the final part of my master’s degree in chemical engineering at Luleå University of Technology. The project was carried out in the department Emission Solutions under Research & Development at Scania CV AB in Södertälje.

I want to thank my supervisor Björn Westerberg for answering my questions and sharing his knowledge with me, Andres Suarez, for always helping me with my questions and guidance during my entire time here at Scania. I would also like to express my appreciation for Jonas Hedlund, my examiner at LTU, for reading my report and giving me valuable feedback to improve my work.

Södertälje, June 2019 Elin Sjöberg

Elin Sjöberg

Thesis in Sustainable Process Engineering, Master 2019

Luleå University of Technology

Department of Civil, Environmental and Natural Resources Engineering

(3)

Abstract

All around the world the environmental focus is growing rapidly. The aim of this master’s thesis has been to work on developing experimental methodologies to achieve qualitatively, reliable and optimal data used on modelling. The selected study case is the analysis of ammonia adsorption on a V-SCR catalyst, which is an important step for future SCR-performance.

Over the last years the mechanism of the V-SCR catalyst and the nature of its active sites have been widely discussed in order to be understood. However, so far, no clear conclusions have been established about the catalytic mechanism, particularly on atomic level, that demonstrates the complexity of the catalytic chemistry within the SCR process. The kinetics of the adsorption of ammonia on a commercial V-SCR catalyst needs further investigation and much more knowledge is needed to understand the chemistry and the behavior behind it. In order to obtain this, new methods where high-quality data is achieved for the adsorption phenomenon, need to be developed.

The results from this work has demonstrated the used method to be a reliable way to ensure reproductively data at each temperature used for ammonia adsorption performed over a V-SCR catalyst.

Further work with focus on the in-depth of the kinetics in the adsorption process is needed.

It is necessary to enhance the understanding of the behaviour of the adsorption on the active sites of the V-SCR catalyst surface as well as performing tests with lower ammonia concentrations for an even better understanding of the processes.

(4)

Sammanfattning

Över hela världen växer fokus på miljön snabbt. Insikten om behovet av att minska föroreningar och avgaser kräver nya lösningar inom många områden. Syftet med examensarbetet har varit att utveckla experimentella metoder för att säkerställa att data som används vid modellering erhåller en bra kvalitet, är pålitlig och optimal. Det valda studiefallet har varit att analysera ammoniakadsorptionen på en V-SCR-katalysator, vilket är ett viktigt steg för SCR-prestandan.

Under de senaste åren har mekanismen för V-SCR-katalysatorn och naturen av dess aktiva säten diskuterats i stor utsträckning för att bättre kunna förstås. Hittills har ingen klar slutsats etablerats om den katalytiska mekanismen, speciellt på atomnivå, vilket visar på komplexiteten i den katalytiska kemin inom SCR-processen. Kinetiken för adsorption av ammoniak på en kommersiell V-SCR-katalysator behöver undersökas ytterligare och mer kunskap behövs för att förstå kemin och beteendet bakom. För att uppnå detta måste nya metoder utvecklas där högkvalitativa data uppnås från adsorptionen av ammoniaken på en V-SCR katalysator.

Resultaten från detta arbete har visat att den valda metoden är ett tillförlitligt sätt att säkerställa reproduktivt data vid varje temperatur som används för den studerade adsorptionen av ammoniak på en V-SCR katalysator.

Ytterligare arbete krävs med större fokus på djupet av kinetiken inom adsorptionsprocessen.

Detta för att öka förståelsen för beteendet hos adsorptionen på de aktiva sätena hos ytan på V- SCR katalysatorn samt utförande av test med lägre koncentration av ammoniak för en ännu

bättre förståelse.

(5)

Abbreviations

ASC Ammonia slip catalyst deNOx Denitration

DPF Diesel particulate filter

FTIR Fourier transform infrared spectroscopy

HC Hydrocarbon

NSC Ammonia storage capacity PM Particulate matter

SCAT-rig Synthetic gas catalytic activity testing rig SCR Selective catalytic reduction

TPD Temperature programmed desorption V-SCR Vanadia-based SCR catalyst

(6)

Symbols

Ai Pre-exponential factor in (s^-1) 𝐸_" Activation energy in J/mol

𝐸_"^# Activation energy at zero coverage in J/mol ΔG Gibbs free energy

∆Hads Heat of adsorption ΔHR Enthalpy of the reaction ΔS Entropy, (J/K^-1)

T Temperature

X Conversion

Z1 Transition state of the gas-phase reaction Z2 Transition state of the surface reaction α Coverage dependence in the reaction ϴi Coverage of species i

(7)

Introduction

How a V-SCR catalytic system actually is working is still widely discussed. For better understanding the problem needs to be broken down into several steps (Tsakoumis, York, Chen

& Rønning, 2015). There are currently relatively few guidelines for evaluating models of data from ammonia adsorption over a V-SCR catalyst and even among experts there is an argument as to what makes a good data model.

New legislation in Europe regarding diesel engines are constantly being developed with the aim to impose stricter emission limits (Koebel, Elsener & Kleemann, 2000). Diesel engines are considered to be one of the largest contributors of exhaust emissions from vehicles and stationary engines (Resitog˘lu, Altinisik & Keskin, 2014). The standards for emissions from diesel vehicles and trucks such as oxides of nitrogen (NOx) particulate matter (PM) carbon monoxide (CO) and hydrocarbons (HC) are constantly becoming stricter (Guan, Zhan, Lin &

Huang, 2014). Diesel engines are widely used due to the exceptional mechanical efficiency.

However, the drawbacks with the engine is the high emissions of NOx and particles which needs to be handled in order to meet the emission targets.

NOx is one of the major pollutants emitted from both mobile and stationary sources and scientists agree that these emissions must be reduced, or even removed, because of its harmful effect on the human health. One of the major issues of the diesel engine is the excess of oxygen in the exhaust, making it unfeasible to use a three-way catalyst. With a large excess of oxygen, the most selective reducing agents for NOx are compounds containing nitrogen, such as cyanuric acid, urea and ammonia (Koebel, Elsener & Marti, 2007). Selective catalytic reduction (SCR) by ammonia is one of the most efficient technologies used today for removing NOx in the exhaust from diesel engines (Iwasaki & Shinjoh, 2010). Currently the vanadium-based V- SCR catalyst and the zeolite-based catalysts used in the exhaust of the diesel engine to meet the emission targets in Europe (Guan et al., 2014).

Even if numerous experimental and theoretical works have been carried out during recent years in order to understand the structural aspects of the catalyst and the nature of its active sites, the models can be improved further (Anstrom, Topsøe & Dumesic, 2002). This may lead to the development of new advanced technologies used in diesel engines in the future (Nova &

Tronconi, 2014; Guan et al., 2014).

(11)

About the project

This is the final part of my master’s degree of chemical engineering at Luleå University of Technology. The project was carried out in the department Emission Solutions under Research & Development at Scania CV AB in Södertälje.

Objective

The aim of this thesis has been to work on developing experimental methodologies to achieve qualitatively, reliable adsorption data used for modelling. The selected study case is the analysis of ammonia adsorption on a V-SCR catalyst. This data is needed for future optimization of the SCR-performance. During this project, different experimental plans were developed, with different data processing and assessment techniques that will be implemented in future projects.

Focus has been on the adsorption phenomena occurring on the surface of the V-SCR catalyst.

Problem formulation

When developing a new model, it is important to have a fundamental understanding of the chemical and physical processes which give rise to the final performance. By focusing only on the adsorption phenomena and eliminating the transport phenomena, the reaction kinetics can be evaluated. In this work the approach will be restricted to the adsorption of ammonia over a V-SCR catalyst. The project will strive to answer some questions listed below:

- How should the method be developed?

- How should quality data be obtained?

- How should the condition, (temperature, ammonia concentrations, time, space velocities) for the tests be evaluated?

- How should the temperature be defined?

- How should the catalyst be prepared?

- How should regeneration of the catalyst be evaluated?

(12)

Literature review

As a first part of the project, a literature study was done where the most important topics concerning this work are described in this section.

Chapter 1 – Models

3.1.1 The need of new models

The word model has a colloquial meaning referring to how some procedure should be carried out or how something should look like. A model represents the reality where some degree of approximation is involved. Depending on complexity and scope, models differ from each other (Carson & Cobelli, 2014; Gerlee & Lundh 2016). To achieve high quality data from a method, a reliable model needs to be developed where addressing the complexity of the catalytic system in the models are also essential. Complexity proves itself through elements covering any physical system through the nature of their connectivity effects (Carson & Cobelli, 2014). For further understanding of the kinetics regarding ammonia adsorption, current models need to be improved and new flexible and robust models must be developed. These models are required in order to manage even stricter emission limits as well as the high concern for the transients like cold start, urea dosing dynamics etc (Tsakoumis et al., 2015).

3.1.2 Model framework

Across levels within a physical system, complexity is exhibited at each of the hierarchy levels.

When developing a mathematical model, the experimental data assist in achieving an understanding of the physical and chemical processes. This can, in other words, be referred to as modelling the system. Figure 1 illustrates a methodological framework for how data can be modelled where a key point for successful modelling is the precision in the acquired data (Tsakoumis et al., 2015; Carson & Cobelli, 2014).

Figure 1. A methodological framework for modelling the data (Carson & Cobelli, 2014).

(13)

The critical part when developing new models is the experimentation part. This part needs to be free from artefacts and configured to synchronize the response from the active compound and the product concentration response with the surface species (Tsakoumis et al., 2015). By focusing on one isolated part where the design of the tests is carefully made, elimination of transport phenomena can be done.

3.1.3 Construction of a kinetic model

When creating a model there will always be a risk that the model loses its usefulness because it becomes too complicated. A model is used to assist in obtaining knowledge and to give a better understanding of the studied phenomenon. The aim of the model is to make valid predictions of the phenomenon to enhance further understanding of the entire process. To reach this knowledge both the comprehensibility and the predictive power is essential to take into consideration when constructing a model (Gerlee & Lundh, 2016).

Different kind of techniques can be used when creating a new model for a meaningful investigation of the kinetic reaction. The kinetic model of Langmuir-Hinshelwood is a model having a successful way of controlling two important parameters. The data needs to be free from the phenomena of transport and the temperature must have been measured with a very high accuracy. Assumptions must be done since the kinetic data obtained is not enough for drawing conclusions of the reaction mechanism. The assumptions should be thermodynamically consistent close to the reality. There are many criteria that needs to be taken into account when testing the validity of the data for the kinetic evaluation. Examples are;

pressure drop, dead volumes, wall effects as well as flow characteristics (Tsakoumis et al., 2015; Berger, Kapteijn, Moulijn, Marin, De-Wilde, Olea, Chen, Holmen, Lietti, Tronconi &

Schuurman, 2008).

The complexity of a model has a high influence on the predictive power and the comprehensibility, as illustrated in Figure 2. A simple model is often easy to understand but has a low predictive power while a complex model has a high predictive power but is difficult to understand. This results in an increase in the predictive power and a decrease in the comprehensibility with increasing complexity of the model where the highest knowledge is reached at an intermediate complexity.

Figure 2. A model’s complexity influenced by the predictive power and the comprehensibility (Gerlee & Lundh, 2016).

(14)

This can be applied to classify different types of models based on their characteristics,

Figure 3. Field (I) include models with both low comprehensibility and low predictive power which is often reflected by a model in the beginning of its development. Most of the current models have a very high predictive power but with a low comprehensibility, (IV). This type of models often fits the data, but the phenomenon is still not fully understood and in order to reach a higher comprehensibility is it essential to strive for more knowledge of the phenomenon. The attempt when developing and creating a model is to reach field (III) where both high predictive power and comprehensibility are achieved (Gerlee & Lundh, 2016).

Figure 3. Model based on knowledge of the predictive power (x-axis) and comprehensibility (y-axis) (Gerlee & Lundh, 2016).

Chapter 2 – The diesel engine

3.2.1 Diesel engine

The diesel engine was developed in 1893 by Rudolf Diesel. The durability of the diesel engine, its fuel economy and low requirements for maintenance has made it very attractive on the market (Heck, Farraut & Gulati, 2009). A diesel engine is preferably operated under lean conditions to improve the efficiency and to reduce the emission of carbon dioxide (Si, Weng, Wu, Yang & Wang, 2010). The diesel engine has an inherently high thermal efficiency, due to the high compression ratio and lean fuel operation (Zheng, Reader & Hawley, 2004). The fuel efficiency for the diesel engine is 30-50 % higher with comparable power output, resulting in a 30-50 % lower emission of CO2 compared to a gasoline engine. CO2 is one of the greenhouse gases which is believed to contribute to the global warming, signifying that the diesel engine is more environmentally friendly compared to the gasoline engine regarding the emissions of CO2

(Heck et al., 2009). However, the drawbacks with the diesel engines is the high emission of NOx where an exhaust gas aftertreatment is required on diesel vehicles to reach the emission targets (Hsieh & Wang, 2011).

(15)

3.2.2 Formation of NO

x

In diesel exhaust, the most dominant problematic compound is nitrogen oxides, NOx. The definition of NOx generally includes the compounds NO and NO2 (Hoekman & Robbins, 2012).

The formation of NOx in diesel engine has been widely studied and there are in general three processes which describes the formation of it; thermal, prompt and fuel NOx formation (Sun, Caton & Jacobs, 2010).

Thermal NOx is formed at temperatures above 1500 °C and increases with increasing temperatures. Formation of prompt NOx involves fragment of hydrocarbons from fuel combustion. These hydrocarbons, mainly CH and CH2, react with nitrogen in the chamber and result in species containing C-N and will, in contact with air, form NOx (Moser, Williams, Haas

& McCormick, 2009). Formation of these species is only possible under fuel rich conditions when there is an excess of hydrocarbon fragments (Hoekman & Robbins, 2012). The process forming fuel NOx occurs when fuel species containing nitrogen are oxidized to NOx in the diesel engine. However, this process is often negligible due to the low nitrogen level in both diesel fuel and biodiesel (Hoekman & Robbins, 2012).

3.2.3 Exhaust gases from diesel engine

When operating a diesel engine, chemical energy in the fuel is converted into mechanical power. If an ideal combustion process in a diesel engine was achieved, only CO2 and water vapor should be formed. In reality, other compounds are also formed during the combustion process. 1 % of the emissions from diesel engines contains pollutants of HC, CO and NOx

which are having a negative effect on both the environment and on human health. An illustration of a typical composition of these pollutants are showed in

Figure 4. These pollutants can be formed in several ways when operating the engine; through incomplete combustion of the fuel, during reactions with mixed components under high temperatures and pressure and during combustion of non-hydrocarbon components of diesel fuel, including sulphur and engine lubricating oil (Majewski & Addy, 2019).

(16)

Figure 4. Typical composition of the exhaust from a diesel engine (Resitog˘lu et al., 2014).

3.2.4 Reducing agent in SCR system

Ammonia has proven to be one of the most effective reducing agents in a SCR system in order to convert NOx in the exhaust gases into water and nitrogen. Ammonia cannot be used directly in the vehicles for safety reasons and for its toxicity (Guan et al., 2014). Therefore, AdBlue which is a 32.5 % urea aqueous solution, is often used instead (Hsieh & Wang, 2011). Urea- SCR has been investigated in detail for over 10 years and is today a well-established technique for reducing NOx in heavy duty vehicles (Koebel et al., 2000). This solution is injected upstream the SCR catalyst where urea is converted to ammonia by a three-step reaction. These reactions include evaporation, decomposition and hydrolyzation and are described in reaction (R.1-R.3).

R.1 describes the evaporation of urea solution and adsorption and R.2 shows the decomposition of urea. The third reaction involves the HNCO (Hsieh & Wang, 2011).

𝑁𝐻_'− 𝐶𝑂 − 𝑁𝐻_'(𝑎. 𝑞. ) → 𝑁𝐻_'− 𝐶𝑂 − 𝑁𝐻_'^∗ (R.1)

𝑁𝐻_'− 𝐶𝑂 − 𝑁𝐻_'^∗ → 𝑁𝐻₃+ 𝐻𝑁𝐶𝑂 (R.2)

𝐻𝑁𝐶𝑂 + 𝐻_'𝑂 → 𝑁𝐻₃+ 𝐶𝑂_' (R.3)

The evaporation process is mainly affected by the temperature and the droplet size which is dependent on the operation conditions and the design of the injector. Through the decomposition process, equimolar amounts of ammonia and isocyanic acid is generated from the urea (Hsieh & Wang, 2011). This process is temperature dependent and slow allowing urea to react with the isocyanic acid forming biuret and cyanuric acid (Strots, Santhanam, Adelman, Griffin & Derybowski, 2010). The reaction starts at a temperature of 200 °C and the maximum reaction rate is reached at a temperature of 350 °C. With a temperature below 200 °C in the exhaust gas, by-products are formed by the decomposition reaction of urea which are highly unwanted. This can be avoided by only injecting the urea solution when the exhaust gas is

(17)

higher than 200 ˚C. From the decomposition process of urea, the hydrolyzation process can convert isocyanic acid to equimolar amounts of CO2 and NH3 (Hsieh & Wang, 2011).

3.2.5 European standards for emissions

With a growing concern for the environment, the regulations for emissions become increasingly more stringent and as mentioned above new robust and flexible models need to be developed to meet the upcoming targets.

In the European Union, (EU) the legislation for the limited amount of the most dangerous emissions; CO, HC, NOx, NH3, PM and PN are clearly described. During the last years, stricter legislations concerning the emissions have been developed. Since 1992 six emission standards for heavy duty vehicles have been developed, Euro 1 to Euro VI, where Euro VI is the emission standard currently used. In Table 1, the current emission limits for Euro V and Euro VI for diesel heavy-duty vehicles are described (Williams & Minjares, 2016).

Table 1. Emission standards for heavy-duty diesel vehicle. a Steady-state testing; b Transient testing; c For Euro V for Natural Gas only, for Euro VI, NG and LPG; d Total HC for diesel engines, non-methane HC for others (Williams & Minjares, 2016).

Euro V Heavy-Duty Euro VI Heavy-Duty

Euro V SS^a Euro V T^b Euro VI SS^a Euro VI T^b

Emission limits (g/km)

CO 1.5 4.0 1.5 4.0

HC 0.46 0.55 0.13 0.16^d

CH4c 1.1 0.5

NOx 2.0 2.0 0.4 0.46

PM 0.02 0.03 0.01 0.01

PN(#/km) 8.0 x 10ⁿ 6.0 x 10ⁿ

Smoke (1/m) 0.5

Ammonia (ppm)¹² 0.01 0.01

Fuel Sulfur Limit (ppm)

10 10 10 10

Test Cycle ESC & ELR ETC WHSC WHTC

3.2.6 Aftertreatment system for the exhaust gas

In order to reach the emission standards for diesel engines, an exhaust gas aftertreatment system is a must (Williams & Minjares, 2016). Figure 5, illustrates the setup of the Scania Euro VI aftertreatment system (Scania, 2019). After the engine a diesel oxidation catalyst DOC, is placed to oxidize the emissions of HC and CO. The DOC is followed by a diesel particulate filter DPF, which is used to remove particles by filtration by the porous walls (Nova &

(18)

Tronconi, 2014). After the DPF, two parallel SCR catalysts are used followed by an ammonia slip catalyst ASC, after each SCR catalyst to remove any excess of ammonia. By using this construction in a diesel car, is it possible to reduce a large amount of the four major pollutants mentioned earlier and reach the current emission target (Scania, 2019).

Figure 5. Scania Euro VI aftertreatment system (Scania, 2019).

Chapter 3 – Catalysts

3.3.1 Catalysts

Catalysts have been known for over 2000 years and are defined as a substance which accelerates the rate of a reaction but emerges from the process unchanged. The function of a catalyst is to provide another molecular route for the reaction (Scott, 2016). Catalysts are today used in the aftertreatment of exhaust gases in diesel cars, converting the toxic gases by a chemical reaction thus minimizing the amount of toxic gas polluting the environment (Roy, Hegde & Madras, 2009).

There are two types of existing catalytic processes, homogeneous and heterogeneous processes.

The homogeneous catalysis process is when the catalyst is in the same phase as at least one of the reactants. In a heterogeneous process there are more than one phase involved and normally the catalyst is in a solid form and the reactants and products are in liquid or gaseous form. In this work, heterogeneous processes have been evaluated. To make the catalyst as effective as possible, the surface area of the catalyst should be as large as possible in order to maximize the amount of gas coming in contact with the catalytic material (Scott, 2016).

3.3.2 Catalytic monolith

Monoliths as catalysts are widely used in processes where the major considerations are the pressure drop (Sharma, Birgersson & Vynnycky, 2015; Scott, 2016). Catalytic monoliths have uni-body structures containing small parallel channels working as adiabatic reactors, limiting

(19)

the control of the temperature (Cybulski & Moulijn, 2006). The size of the channels which the gaseous reactant transfers through is the most important physical characteristic when monoliths are used as a catalyst support (Heck, Gulati & Farrauto, 2001). When using monoliths, the purpose is to enhance the contact area between the exhaust gas and the coating where the active material is located and at the same time maintain a low pressure drop over the system (Nova &

Tronconi, 2014). This is achieved by division of the catalyst monolith into channels, each surrounded by a porous washcoat, illustrated in Figure 6 (Heck et al., 2001).

Figure 6. Structure of a typical catalyst monolith (Tomašić & Jović, 2006).

3.3.3 Model of catalyst monolith

A typical monolith channel is illustrated in Figure 7. The modelled monolithic catalyst consists of small channels containing two phases which are the gas phase and the washcoat phase. The active catalytic material is located as mentioned above in the walls, forming the washcoat phase where diffusion between the channels occurs due to the porous walls (Aberg, Widd, Abildskov

& Kjøbsted.Huusom, 2016). There are two gradients within the catalyst, one along the catalyst and one through the active phase around the catalyst where the gradient along the catalyst decreases with an increasing space velocity. In Figure 7, a single channel in bulk, stagnant film and washcoat phase is shown where molecules from the gas phase is able to migrate through the stagnant film to the washcoat, where they react when in contact with the active sites (Aberg et al., 2016).

(20)

Figure 7. Top picture: Illustrate a single monolith channels physical quantity. Bottom picture: Illustrates a channel in bulk phase, stagnant film, and the washcoat (Aberg et al., 2016).

3.3.4 Steps included in a heterogenous catalytic reaction

The catalytical system includes a sequence of seven fundamental steps describing the kinetic behavior of the process occurring inside the monolithic catalyst (Tsakoumis et al., 2015), illustrated in Figure 8, (Scott, 2016). The steps involve diffusion where reactant molecules diffuse through the pore network of the catalyst to the surface of the washcoat walls shown in step 1 and 2. Step 3 describes the adsorption of the reactant onto the active sites located on the surface of the washcoat where step 4 involves the reaction on the active sites. Step 5 illustrates the desorption of the intermediates (Tsakoumis et al., 2015; Scott, 2016) and step 6 and 7 describes the diffusion of the product molecules through the pore mouth to the bulk fluid. This multistep process involves a complex kinetic information and includes transport phenomena.

By focusing on the adsorption step occurring on the active sites on the surface and limit the mass transfer resistance by increasing the space velocity, reaction kinetics can be estimated (Tsakoumis et al., 2015; Chen, Yang, Wanga, Ring & Dabros, 2008).

(21)

Figure 8. Illustration of the steps in a heterogeneous catalytic reaction (Scott, 2016).

3.3.4.1 Active sites

The active sites in catalysts are described as a group of atoms on a solid surface which take part in the rate-limiting step of the catalytic reaction. As mentioned above the active sites are distributed in the washcoat of the monolithic catalyst where the importance of finding the active site in the catalyst is to be able to prepare it in larger abundance to improve the catalyst (Nova

& Tronconi, 2014). There are two types of active sites in a catalyst which contributes to the acidity of the catalyst where adsorption of ammonia can occur. These are called Brønsted and Lewis acid sites where Brønsted acid sites works as a proton donor and Lewis acid sites works as an electron pair acceptor (Anstrom et al., 2002).

Chapter 4 – Selective catalytic reduction

3.4.1 Selective catalytic reduction, SCR

Selective catalytic reduction, SCR, has been used since the 1970s and is nowadays mainly used to reduce NOx formation in diesel engines. The SCR process has been improved, particularly for high-duty vehicles. During the 1990s, V-SCR was introduced to the mobile applications, especially for trucks with heavy-duty diesel engines (Nova & Tronconi 2014). The amount of stored ammonia on the SCR catalyst surface should be as high as possible to maximize the efficiency. Excess amount of ammonia, however, may result in undesired amounts of ammonia in form of ammonia slip (Resitog˘lu et al., 2014). V-SCR was introduced early into the market of diesel vehicle because of its high efficiency of NH3-SCR reaction and high resistance for SO2 poisoning (Shan, Liu, He, Shi & Zhang, 2012). The working temperature for the V-SCR catalyst is around 300-400 ˚C and the flue gas temperature is as low as 100-200 ˚C. This necessitates a heating process to heat the SCR system above 300 ˚C. A lot of energy is consumed during the heating prosses and therefore a wider temperature window and activity at lower temperatures of the catalysts are necessary to develop (Xu, Chen, Guo & Xie, 2018).

(22)

3.4.2 SCR reaction modelling

There are three main reactions occurring when operating an SCR catalyst. These are, the

“standard” reaction, the “fast” reaction, and the “slow” reaction, described in (R.4-R.6) (Colombo, Nova & Tronconi, 2010).

Standard SCR reaction:

4𝑁𝐻₃+ 4𝑁𝑂 + 𝑂_' → 4𝑁_'+ 6𝐻_'𝑂 (R.4)

Fast SCR reaction:

2𝑁𝐻₃+ 𝑁𝑂 + 𝑁𝑂_' → 2𝑁_' + 3𝐻_'𝑂 (R.5)

Slow SCR reaction:

8𝑁𝐻₃+ 6𝑁𝑂_' → 7𝑁_'+ 12𝐻_'𝑂 (R.6)

The standard reaction R.4, is based on the reaction producing water and nitrogen by reacting NH3, O2 and NO. The “fast” SCR reaction R.5 consists of an equimolar amount of NO and NO2

and is more active at a lower temperature compared to the standard reaction. The “slow” SCR reaction R.6, occurs when the ratio of NO2:NO is larger than 1:1 and NO2 react with NH3 to form N2 and H2O. This reaction is slower compared to the standard SCR reaction over a V- SCR catalyst (Devadas, Kröcher, Elsener, Wokaun, Söger, Pfeifer, Demel & Mussmann, 2006).

3.4.3 NH

3

and NO oxidation

Beside the standards reactions, there are side reactions occurring during the SCR reaction where products other than N2 are formed which results in unwanted products (Busca, Lietti, Ramis &

Berti, 1998). This implies when ammonia is converted in other ways than the “standard reaction” and where the SCR catalyst acts as an oxidation catalyst. Due to the excess of oxygen in the SCR reaction these reactions occur at higher temperatures (Xu et al., 2018). The unwanted side reactions are described below:

4𝑁𝐻₃+ 5𝑂_' → 4𝑁𝑂 + 6𝐻_'𝑂 (R.7)

4𝑁𝐻₃+ 7𝑂_' → 4𝑁𝑂_'+ 6𝐻_'𝑂 (R.8)

2𝑁𝐻₃+ 8𝑁𝑂 → 5𝑁_'𝑂 + 3𝐻_'𝑂 (R.9)

6𝑁𝐻₃+ 8𝑁𝑂_'+ 3𝑂_' → 4𝑁_'𝑂 + 6𝐻_'𝑂 (R.10)

4𝑁𝐻₃+ 4𝑁𝑂 + 3𝑂_' → 4𝑁_'𝑂 + 6𝐻_'𝑂 (R.11)

These reactions are unwanted since they reduce the formation of N2 and instead form N2O or NO and NO2 and decreases the selectivity of the catalyst (Xu et al., 2018). Due to these unwanted side reactions new catalysts need to be established with high activity at a lower temperature range and in a wider operation window (Xu et al., 2018).

3.4.4 Materials of SCR catalyst

There are three common types of SCR catalysts which are currently used in heavy-duty diesel cars; V-SCR, Cu-zeolites and Fe-zeolites (Nova & Tronconi, 2014).

(23)

3.4.4.1 V-SCR

The vanadium-based catalysts washcoat is primarily composed of oxides of vanadium, tungsten, and titanium (V2O5/WO3–TiO2) (Liu, Ottinger & Cremeens, 2015). Although several studies have been published concerning the characterization of the V-SCR catalyst, there is not a well-established intrinsic understanding of the V-SCR catalyst. However, the best catalyst in all cases has been found to consist of just less than a full monolayer of vanadium and tungsten oxides over the anastase-titania support (Busca et al., 1998).

The V-SCR catalyst consist of TiO2 used as support material due to its large surface area and because of the uniformly dispersion of both vanadium and tungsten oxides on the surface (He, Ford, Zhu, Liu, Tumuluri, Wu & Wachs, 2016). Vanadia V2O5 stands for the active catalyst component of the catalyst and WO3 is added as a promotor to make the catalyst more active and stable (Nova & Tronconi, 2014). WO3 act as a promoter due to its Brønsted acidity which enhances the ammonia adsorption on the catalyst during the SCR reaction (He et al., 2016). As mentioned above the V-SCR is commonly used for Euro IV, V and VI because of its efficiency to remove NOx from the diesel exhaust with ammonia as reducing agent (Xi, Ottinger & Liu, 2014). The SCR material is either wash coated onto an inert monolith or alternatively the entire monolith walls comprise the active catalyst material (Nova & Tronconi, 2014).

3.4.4.2 Reaction mechanism of ammonia over a V-SCR catalyst

The reaction mechanism over a V-SCR catalyst was, during the 1980s and 1990s, intensively investigated (Forzatti, 2001). Selective catalytic reduction of NOx has been applied and extensively used in the industry for over 50 years (Lai & Wachs, 2018). Even if SCR has been widely investigated there are still many important details about the catalysts that are not yet fully understood, especially on the atomic level (Arnarson, Falsig, Rasmussen, Lauritsen &

Moses, 2017) and there is no general agreement concerning the nature of the active sites and mechanism of the SCR process (Yin, Han & Miyamoto, 2000). However, today there is a general agreement that the SCR process is a redox process where NH3 and NO participate in the reduction step and while oxygen participate in the reoxidation part (Arnarson et al., 2017).

When using ammonia as a reducing agent over a V-SCR catalyst there are several other mechanisms built on intermediate active sites disturbing the process. For example, active species on the catalyst surface or interactions of the reactants with active sites. These mechanisms need further investigation in order to understand the surface chemistry behind the system (Nova & Tronconi, 2014).

The first proposed mechanism for the standard SCR reaction R.1, over a V-SCR catalyst was the Eley-Rideal mechanism where NO in gaseous form reacts with adsorbed NH4+ to form H2O and N2 (Xu et al., 2018). The cycle includes the final oxygen species containing two nearby vanadium species V=O, on the redox sites and the hydroxyl group V-OH, on the Brønsted acidic site as illustrated in Figure 9. The reaction involves adsorption on the Brønsted acidic sites and the reaction is then followed by activation of the adsorbed ammonia at the redox sites, as demonstrated in R.12 and R.13 (Nova & Tronconi, 2014).

(24)

Figure 9. Suggest catalytic cycle for a standard SCR reaction over a V-SCR catalyst. Acidic site is associated with V⁵⁺-OH and the redox site is associated with V⁵⁺=O (Xu et al., 2018).

𝑉^>?− 𝑂𝐻 + 𝑁𝐻₃ ↔ 𝑉^>?− 𝑂𝑁𝐻_A (R.12)

𝑉^>?− 𝑂𝑁𝐻_A + 𝑉^>? = 𝑂 ↔ 𝑉^>?− 𝑂𝑁𝐻_A^?− 𝑂 − 𝑉^A? (R.13)

Many studies have considered the reaction mechanisms for the SCR process of NOx reduction and the results have shown that the reaction rate is highly related to the temperature and the dispersion of the vanadyl species (Xu et al., 2018).

3.4.4.3 Polymeric/monomeric vanadyl species

Polymeric vanadyl species have shown to have a much higher activity compared to the monomeric vanadyl species in NH3-SCR processes. A suggested reaction mechanism is shown in Figure 10. The coupling effect of the structure of the polymeric species shortens the pathway for the regeneration of the redox sites as well as it shortens the overall barrier of the reaction in the cycle. Therefore, the polymeric species of vanadyl therefore determines the activity of the NH3-SCR with a V-SCR catalyst, especially at low temperatures (He, Lian, Yu, Yang, Liu, Shi, Yan, Shan & He, 2018).

(25)

Figure 10. Suggest reaction mechanism for the standard NH3-SCR reaction, where (A) represent the reactions over a monomeric vanadia/TiO2 surface and (B) over a dimeric vanadia/TiO2 surface (He et al., 2018).

Chapter 5 – The adsorption phenomena

3.5.1 Adsorption

At least one reactant must attach to the surface for a catalytic reaction to occur, this phenomenon is called adsorption. In general, there are two kinds of adsorption; physical and chemical adsorption (Rouquerol, Rouquerol, Llewellyn, Maurin & Sing, 2013). Energy is released during reaction in both types of adsorption processes meaning that the processes are exothermic (Scott, 2016). Adsorption takes place by an attraction between the catalytic surface of the adsorbent of a solid and a vapor or a solution of the adsorbate. When an atom binds to the surface, heat is involved in a process called heat of adsorption ∆Hads (Li & Somorjai, 2010).

Physical adsorption is described as similar to condensation (Scott, 2016). Physical adsorption arises with intermolecular forces (van der Waals) interaction between the adsorbent and the adsorbed molecule. Physical adsorption can exist in multiple layers which result in a low heat of adsorption and in low binding energy (Rouquerol et al., 2013).

Chemical adsorption form bonds between the adsorbent and the adsorbate where only one layer is possible between the adsorbent and adsorbate on the surface. In chemical adsorption higher heat of adsorption is obtained together with higher binding strength. The force between the molecule and the surface is significantly stronger and more reactive compared to physical adsorption (Scott, 2016; Ross, 2012; Rouquerol et al., 2013).

3.5.2 Energetic aspects of adsorption on the surface

A molecule must undergo activated adsorption to be able to enter a reactive state on the surface.

For effective catalysis, the strength of adsorption of the molecule is significant where the molecule should adsorb not too strong but neither too weak on the surface of the catalyst. The energy aspects and chemisorption play an important role to understand heterogeneous catalysis.

Figure 11 is an illustration of the comparison between a heterogeneously catalyzed reaction and

(26)

an uncatalyzed reaction in gas-phase (Ross, 2012). The activation energy for the uncatalyzed reaction is described as Ea,0. The true activation energy of the reaction is represented as Ea,1 and the noticeable activation energy of the reaction is referred to as Ea,2.

Figure 11. Comparison between an uncatalyzed reaction with a catalyzed reaction AG – PG (Ross, 2012).

Three elementary steps with corresponding energies; Ea,0, Ea,1, Ea,2 are occurring on the surface of the catalyst as shown in Figure 11. Chemisorption of gases on metals has been investigated widely where the structure of the metal has a large impact on the reactivity of the metal surface towards different types of gases. Titania has shown to have a high reactivity towards other metals and are able to strongly adsorb different gases (Ross, 2012).

3.5.3 Adsorption Isotherms

Adsorption isotherms describe the amount of adsorbed gas on a catalyst surface at different partial pressures at the same temperatures, where the adsorption increases with decreasing temperature, as illustrated in Figure 12 (Scott, 2016). The exothermic adsorption process embraces strong binding forces between the atom on the catalyst surface and the adsorbed gas molecule. When the gas phase molecule adsorbs on the surface of the catalyst, the degree of freedom of the molecules decreases. This results in a negative entropy (∆S) and Gibbs free energy ∆𝐺, (E.1) should be negative for a feasible adsorption process. Due to small variation in the reaction entropy, the enthalpy ΔH, of the adsorption will mainly depend on the strength of the chemical bonding between the catalyst and the gas molecule (Ross, 2012).

∆𝐺 = ∆𝐻 − 𝑇∆𝑆 (E.1)

(27)

Figure 12. Isotherm of adsorption showing that the degree of adsorption increases with decreasing temperature (Scott, 2016).

3.5.3.1 Langmuir adsorption Isotherms

To produce meaningful investigations of the reaction kinetic involving the adsorption phenomena occurring on the active sites on surface of the catalyst there are different kinds of techniques which can be of a great support when constructing a model (Tsakoumis et al., 2015).

To quantify the number of adsorbed molecules on the surface of the catalyst, Langmuir adsorption model is one of the most common to use where two essential parameters are controlled. When creating this model, the dataset needs to be free from the transport phenomena while the temperature has to be measured with high accuracy. The kinetic data itself is insufficient to draw any conclusion about the reaction mechanism from and some realistic assumptions need to be made to stay close to the reality (Scott, 2016; Tsakoumis et al., 2015).

The surface of the catalyst consists of a certain number of active sites and are proportional to the area of the surface where one molecule is able to adsorb on each active site. For Langmuir Isotherms there are some assumptions which need to be considered; The adsorbed molecule on the active site remains at that site until the desorption of it from the surface. When maximum of adsorption occurs, only a monolayer can be formed on the surface of the catalyst, which implicates that molecules already adsorbed do not deposit on each other (Ross, 2012). For all adsorption isotherms, the adsorption and desorption rate need to be in equilibrium, meaning that the rate of adsorption and desorption is equal (Bartholomew & Farrauto, 2006).

3.5.4 Surface reactions

Several studies have shown that when using ammonia as the reducing agent for NOx over a V- SCR catalyst, the ammonia adsorbs as a NH3* species on the Lewis acid sites of titanium-, vanadium- and tungsten oxides. Ammonia adsorbs as NH4+species on the Brønsted acid sites on the oxides of vanadium and tungsten (Lai & Wachs, 2018). The adsorption on the Brønsted acid sites are stronger and thereby favourable (Yin, Han & Miyamoto, 2000). The V-SCR catalyst has shown to consist of several active sites where both WO3 and V2O5 consists of both

(28)

Lewis and Brønsted acid sites. TiO2-support only contain Lewis acid sites. There are therefore five possible active sites within the catalyst where adsorption of ammonia may occur as illustrated in Figure 13 (Zhu, Kun.Lai, Tumuluri, Wu & Wachs, 2017). The ammonia species will first adsorb on the most active sites on the surface and thereafter on the weaker active sites (Lai & Wachs, 2018). Studies have shown that both Lewis and Brønsted acid sites takes part in the SCR reaction while their relative occupation depends on the coverage of the surface vanadia and tungsta sites, the temperature, and the moisture (Zhu et al., 2017).

Figure 13. Standard SCR reaction over a V-SCR catalyst with the different possibly active sites present on the surface (Zhu et al., 2017).

3.5.5 Desorption

Desorption is the opposite from adsorption, it is the phenomena where a molecule is released from the surface. During desorption the bond between the adsorbed molecule and the active site on the surface of the catalyst is broken (Li & Somorjai, 2010). Desorption occurs at higher temperatures where it is more challenging for the molecules to adsorb on the surface resulting in lower adsorbed concentration. Therefore, high temperature results in desorption of the molecules from the surface.

3.5.6 Ammonia adsorption characterisation

The adsorption process of ammonia can be described as a two-way reaction, (R.14):

𝑁𝐻₃ ↔ 𝑁𝐻₃^∗ (R.14)

𝑁𝐻₃^∗ represents the adsorbed ammonia on an active site of the surface of the SCR catalyst. The amount of ammonia that is able to be adsorbed or desorbed by the SCR catalyst have from many studies been shown to have a high dependency on the temperature (Hsieh & Wang, 2011).

(29)

3.5.7 Ammonia storage capacity

As mentioned above the amount of stored ammonia on the SCR catalyst should be maximized for achieving as high efficiency as possible. The reaction of the SCR catalyst starts when the ammonia adsorbs on the active sites of the catalyst where the temperature has shown to have a large impact on how much ammonia the catalyst is able to store. At lower temperatures, more ammonia is able to adsorb on the surface compared with higher temperature where almost no ammonia is able to be stored. Another parameter affecting the catalyst capacity is the aging of the catalyst where the reactivity of the surface is reduced with time (Guan et al, 2014).

3.5.8 Effect of H

2

O on the SCR reaction

The role of water in the SCR reaction has not been widely investigated but studies have shown that water has a large impact on the SCR reaction where water inhibit the reaction by inhibiting the active sites. This has shown to occur at lower temperatures. At higher temperature, the inhibition effect decreases (Arnarson et al., 2017; Lai & Wachs, 2018). In industrial combustion processes there are between 10-30 % water present in the flue gas. When water interacts with the Lewis and Brønsted active sites the sites modifies resulting in a change in the distribution (Lai & Wachs, 2018). Recent studies have shown that water hydrolyses on the surface of the vanadium sites and transforms the Brønsted acid sites into Lewis acid sites. Competitive adsorption for the active sites between the water and the ammonia has also been observed (Arnarson et al., 2017; Lai & Wachs, 2018).

3.5.9 Effect of GHSV on ammonia adsorption

Conversion of NOx is strongly influenced by the temperature, but other parameters have shown to have an effect on the conversion as well. Gas hourly space velocity (GHSV) is one of the parameters which affect the NOx conversion. The GHSV vary the residence time of the reacting components in chemical reactors and then changes their conversion. By increasing the gas velocity, the residence time decreases. The differences in the conversion are not the same for all different temperatures with different gas velocities and the conversion should therefore be set by the gas velocity in combination with temperature controls (Nahavandi, 2015).

3.5.10 Temperature programmed desorption

The heat of adsorption is frequently measured by desorption experiments. By heating the surface fast, the adsorbed molecules are removed. This technique is called temperature programmed desorption, TPD (Li & Somorjai, 2010).

(30)

Method

Instructions for how the experiment, limitations and the calculations were developed and executed are described in this chapter. All tests have been performed in a SCAT-rig at Scania, SCAT- Synthetic gas catalytic activity testing. The gas outlet has been analysed with a FTIR instrument, Fourier transform infrared spectroscopy. In order to achieve a better understanding of the adsorption mechanism on the V-SCR catalyst and to choose right conditions for the experimental method, where most information about the adsorption will be achieved, some screening tests with an unspecified V-SCR catalyst were performed initially.

Limitations

The limits with these kinds of experiments will be to understand the setup and the design of the equipment in order to ensure high quality data. The operating limits of the sensors, control valves and uncertainties are essential to understand when operating the rig. Other limitations when pre-treating the catalyst in an oxidized environment are to choose the right temperatures in order to avoid oxidation reactions with the ammonia. When using a monolithic catalyst, it is difficult to keep the temperature constant and to control it during the entire experiment due to the axial temperature gradient in the monolith which also needs to be taken into consideration.

SCAT-rig

All tests have been performed in the reactor of the SCAT-rig, where the gas outlet from the chamber have been analysed by a FTIR instrument. The function of the SCAT-rig is to expose the catalyst to simulated exhaust gases. To obtain a method where high-quality data can be ensured, two parameters are essential to understand; the setup and the system of the SCAT-rig.

The setup includes the limitations of the rig, its operation ranges where results can be obtained, the uncertainty as well as the size of the data accessibility for the rig to handle. When analysing the system of the rig, parameters which should be taken into account are; the ammonia adsorption, compounds included, properties of the rig and feasibly interactions within the rig.

4.2.1 Reactor design

The monolithic catalyst used for the experiment is placed in a sampler holder which in turn is placed in a quartz tube. Before the catalyst is placed in the sampler holder, a wool thread is wrapped around it to ensure all gas will go through the channels of the catalyst monolith and not slip around it. The temperature is measured by four thermocouples where two are placed inside the catalyst, at the beginning and the end of it and the other two thermocouples are placed around the sample holder. The sampler holder is then heated up to the desired temperature in a heating coil surrounding the sample holder, while the two thermocouples around it measures the temperature. Mass flow controllers are used to feed the gas mixture in to the reactor and to control the flow of the gas. The gas composition out from the chamber is measured by an AVL SESAM i60 FT FTIR instrument.

(31)

To be able to ensure high quality data an optimal parametric stochastic filter ARX filter was applied to the dataset. The ARX filter was used to account for the dispersion effects in the data material to ensure more reliable values from the amount ammonia adsorbed on the active sites.

A parameter correlation with multiplied experiments had been estimated before based on this and was thereby already existing. To ensure high quality data, the flow of the gas and possible instrumental delays in the SCAT-rig had to be taken into consideration. A filter was created taking these two parameters into account increasing the possibility to reach an even more accurate result concerning the adsorption occurring on the V-SCR surface. The design of the reactor in the SCAT-rig where the tests has been performed is illustrated in Figure 14.

Figure 14. Illustration of the reactor design in the SCAT-rig.

(32)

4.3 Catalyst

Two fresh V-SCR catalysts with different loadings have been used for the project where the catalytic active material has been coated on the inert monolith. Catalyst 80, containing the higher loading should have more active material around the channels in the monolith. This leads to more available sites where adsorption of ammonia can occur. Catalyst 81, used for the experimental part of the project can be seen in Figure 15, and the dimensions of the catalysts are presented in Table 2.

Table 2. Characteristics for the two catalysts used in the experiment.

Sample loading (g/l) weight (g) length, (mm) diameter (mm) density (g/cm³)

Catalyst 81 289 21 76 24 0.6

Catalyst 80 371 24 77 25 0.6

Figure 15. Catalyst 81 containing loading of 288,8 g/l.

(33)

4.4 Experimental procedure

How the experimental plan for the project was executed is described below in Figure 16.

Figure 16. Description of the experimental procedure.

In the beginning of the project, a scope was decided within which the tests should be performed.

This was created with help of literature studies from previous performed tests concerning ammonia adsorption. Different state of the catalyst has been used as a parameter where the catalyst either had been pre-treated in a reduced or in an oxidized environment. The water and ammonia concentration and at which temperatures the tests should be focused on was essential to determinate.

4.4.1 Screening tests

As an initial part of the experimental procedure, several screening tests were performed on an already aged V-SCR catalyst. This was executed to find out more about the ammonia adsorption phenomena occurring on the active sites of the catalyst surface and in what way the method should be created. The screening tests were executed with different pre-treatment methods of the catalyst, where the catalyst had either been pre-treated in a reduced or in an oxidized environment. The ammonia adsorption was performed in a wide concentration span at different

Define scope Screening tests Steady state tests

Catalyst state Experimental

region

Temperatures Required

resolution

Equilibrium parameters

% H2O Obtain

information

Limit the mass transfer resistance

NH3 conc

Adsorption Isotherms

(34)

temperatures. By evaluating the screening tests more information about the behaviour of the adsorption at extreme conditions could be obtained. With this knowledge the actual catalyst tests could be constructed.

4.4.2 Test performed with steady state

Steady state tests have been performed to limit the influence of the mass transfer resistance (see Figure 8). At steady state, the reaction rates of all consecutive elementary steps are approximately equal to each other and equal to the slowest step, the limiting reaction step. By performing the tests with a high space velocity, the mass transfer resistance is decreased. By considering the thermodynamic equilibrium as constant at each step, where steady state is obtained, the thermodynamic parameters can be evaluated based on Gibbs free energy, (E.1).

4.4.3 Experimental procedure

Two fresh V-SCR catalysts with different loadings have been used in the experimental part, see under Catalyst for exact description. Tests with catalyst pre-treated in a reduced or oxidized environment have been performed. Both dry tests where only ammonia is able to adsorb on the active sites and tests including water have been executed. The space velocity has been held constant at 85 000 h^-1 for all tests and the temperatures used have been within the range of 100-450 ˚C. The commercial software Matlab has been used for creating the recipe for the experimental procedure in the SCAT-rig and for analysing the results by applying an ARX filter and pre-processing code from the data material to achieve a high-quality data. How the experimental procedure was executed is illustrated in Figure 17 and are further described below.

Figure 17. The setup of the experimental procedure.

4.4.4 Process sequence

The fresh V-SCR catalyst is initially placed in the reactor tube. The experiment has been carried out with common-response gases where different steps in the inlet concentration of the reactant are used where the response is analysed. The gases are pre-mixed together in a single stream and then enriched with water vapor. For these tests a gas mixture of ammonia and nitrogen have been used in order to reach an ammonia concentration of 3,6 %. Nitrogen has been used to stabilize the sample before initiating the test. Ammonia is added to the stream when the flow has become constant and is thereafter fed to the catalyst monolith in the reactor through a heated line. Different ammonia concentrations between 75-400 ppm are fed to the catalyst where steady state is allowed to be reached at each outlet level before the concentration is changed.

The temperature is kept constant during each test in the sequence, where the inlet gas to the reactor is preheated to the given temperature before the hot stream of gas is directed through

Pretreatment Adsorption Desorption Regeneration De-greening

(35)

the catalyst monolith with a constant flow. The temperature of the gas is measured by thermocouples both before and after the catalyst monolith. The outlet gas composition is cooled and analysed with the FTIR instrument. Between each test within the sequence, each sample has been purged at a high temperature to desorb all ammonia left on the surface of the catalyst before increasing the temperature.

4.4.5 De-greening

Fresh catalysts exhibit an activity loss in the primary short period of their operation. The aim for the de-greening is to obtain steady state activity in the fresh catalyst by exposing it to the process conditions. This procedure is only done once when the catalyst is new. During this period the catalyst was cleaned from any residues remaining from the production rendering the catalyst re-structured and active. This procedure was done by oxidizing the catalyst at 450 ˚C for a total of four hours in the presence of gas containing NH3, NO, H2O and O2.

4.4.6 Pre-treatment

The catalyst has been pre-treated either in a reduced or in an oxidized environment before the activity test.

To oxidize the surface of the of the catalyst it has been held in the presence of 5 % O2 at a temperature of 500 ˚C for 10 minutes. This will lead to a fully oxidized surface where the vanadium on the surface of the catalyst is present in V⁵⁺.

To reduce the catalyst surface, it has first been oxidized in the presence of 5 % O2 for 10 minutes. Afterward the catalyst was flushed with N2 for 5 minutes before introducing 500 ppm of ammonia to the reactor for 20 minutes, allowing the ammonia to reach steady state. All steps have been performed at 500 ˚C.

4.4.7 Activity test

The activity test involves the adsorption and the desorption process of ammonia on the two sites located on the surface of the catalyst.

4.4.7.1 Adsorption

A sequence of different ammonia concentrations at a set temperature is introduced to the reactor where the ammonia is first adsorbed on the most active sites of the catalyst at lower concentrations. The test starts with a low concentration level followed by a number of sequences with increasing concentration where the outlet of ammonia concentration is allowed to reach steady state at each inlet level.

4.4.7.2 Desorption

Same concentration steps as for the adsorption is used for desorption where the ammonia is desorbed from the active sites of the catalyst where the molecule bonded by the weakest bond

Characterisation of NH3 adsorption on a V- SCR catalyst