 So, we will continue our discussion on reactive distillation. Yesterday we looked at several commercially important systems in which reactive distillation can be used. It is an upcoming field I told you yesterday, the various advantages of using reactive distillation such as enhancement in conversion, enhancement in selectivity, increase in catalyst life, then temperature control and so many. And we looked at various examples like acidification reaction for manufacture of methyl acetate, then etherification for manufacture of MTBE, then acetylization reaction, then isobutene to diisobutene for selectivity engineering. And so many commercially important systems where actually reactive distillation is applied is practiced and some places where reactive distillation finds the potentially important applications. So, today we are going to learn how to develop a process or how to get into the process development studies related to reactive distillation. For example, somebody comes to me and tells me that you have a reaction and I want to know whether I can use reactive distillation in the process. The answer is not so simple because I told you yesterday we cannot apply reactive distillation to all the systems because there are many constraints like the most important one is the volatilities should be right, the temperature of reaction should match the temperature of distillation. And after doing all this even if reactive distillation is feasible whether it is economically viable or not is again a big question. So, feasibility is one issue at the same time whether it is economically beneficial or not is another issue. So, these two things are one has to really see to get the reactive distillation commercialized for a given process. This particular lecture tells you about different activities, different steps you have to go through while developing a reactive distillation process. And we have considered a very common reaction again the acidification reaction as an example. We have done some work on this reaction in our laboratory and we have gone through all the steps for reactive distillation process development studies and I will tell you about that in this particular lecture. Reaction is acetic acid with n-butanol giving butyl acetate and water. Reversible reaction as you know leachatellis principle if you remove one of the products or both the products during the course of the reaction, reaction shifts in forward direction and you get enhanced conversion. And if you look at a process I can use stoichiometric ratio of acetic acid and butanol. So, I do not have any unconverted reactants if the reaction is 100 percent. So, recycle is avoided right and there will be reduction in capital cost as well because I am doing both reaction separation in a single piece of equipment right. So, reactive distillation would be beneficial for such reactions and potentially important system. Now, if I want to design a reactive distillation column, if I want to see how reactive distillation will give you better results then I have to do so many things. So, that is what we are going to see in this particular lecture. Now, look at the system the proposed reactive distillation flow sheet, it comes out of your experience. Initially you are going to start with some configuration in your mind. What I will say is like acetic acid and butanol two reactants they go to a pre-reactor right. Pre-reactor you do as much conversion as possible in the pre-reactor at only the reason what is the reason, why do you have pre-reactor and then the reactive distillation column why can't we do both or rather the entire reaction in just one distillation column sizing at the same time see the yeah the load. I want to reduce the load on reactive distillation column because most of the times in the column I use reactive packings and it would be most of the times expensive. So, I want to reduce the load on this particular piece of equipment. So, as much reaction as possible you do it in a simple reactor it can be a slurry reactor, it can be a fixed bed reactor and then take the equilibrium mixture to the reactive distillation column. So, that is the theme I have in my mind before I start my process development studies and what is expected based on knowledge that I have about the system the vapour liquid equilibrium of the system is that butyl acetate being less volatile or least volatile component in the system will come out from the bottom which is a product of course boiling point at atmospheric pressure is 127 and from the top you have an azeotrope, azeotrope heterogeneous see that you can enter there heterogeneous this azeotrope is a ternary azeotrope a minimum boiling ternary azeotrope of butyl acetate and water. So, once you have this vapours getting condensed I can take the water out which is another product right. Now, the beauty of this LLE or liquid equilibrium is that the vapours when they condense the aqueous layer is almost pure water, aqueous layer is almost pure water. So, with that much knowledge I just say I have a possible configuration like this which can work and give me almost 100 percent conversion. So, I do not need anything in the plant right no further separation purification and all that right I have pure products coming out. So, this is my entire process right now I want to design this it is not so simple because you have both reaction and distillation taking place in a single piece of equipment on the distillation trays or can be packing right. So, you have both things happening here now what are the complexities so let us go ahead. So, in order to do the process development studies we have to follow various steps and I have given a flow chart here the organization of all the work elements as far as process development study of react to distillation is concerned. What do we need to do first experiments for reaction kinetics and regression see now so far we have been talking about distillation and distillation what is important is vapour liquid equilibrium right. But here along with vapour liquid equilibrium the reaction kinetics will play a very important role. So, whatever knowledge we have about distillation I can just extend it to react to distillation it all depends on how the reaction is playing its role whether it is very fast reaction or it is very slow reaction. So, the kinetics is important right and once you do the kinetics regression that means I need to get the kinetic parameters it is like vapour liquid equilibrium I do x vapour liquid equilibrium experiments and estimate a parameters like binary interaction parameters for liquid phase activity coefficient and all similarly in the kinetics if you remember the kinetic equation R is equal to k C A C B minus k dash C C C D. So, I need to get this k and k dash right so I do experiments in laboratory find the data the data is typically in the forms conversion versus time or conversion versus residence time if you are using a plug flow reactor in laboratory right if you are using a batch reactor it would be conversion versus time. So, this data is used to get the kinetic equation right. So, this is one important activity another important activity you know very well you learn so much about distillation. So, in this case we are generating vapour liquid equilibrium data and of course, the many possibilities typically for liquid phase non-ideality the apparatus that we showed you the modified atmosphere still is used if you are really looking at simple systems atmospheric pressure or vacuum rather right. So, vapour liquid equilibrium data generation and again regression or parameter estimation where you determine non-ideality or other the binary interaction parameters for model ok. Now, if I am talking about any quack model then corresponding parameters will be required if it is Wilson then so that is right. So, this is a quaternary system. So, we have to generate vapour liquid equilibrium data for all the binaries we have to generate the vapour liquid equilibrium data for all the binaries and then so it is not just butyl acetate and water right because in my reactive distillation column all the components would be there right. So, you get the binary interaction parameters here and then try and predict the behavior of the quaternary system that is the way we do it for normal distillation. So, these are the two important pillars I would say on which the entire reactive distillation process development study is based on ok. What do we do next? Next is the conceptual design. So, we have been talking about conceptual design for last 3-4 days for enhanced or complex distillation systems whether distillation is feasible or not. The same answer you will get here whether reactive distillation is feasible or not by taking data from kinetics, by taking information from vapour liquid equilibrium experiments. At this particular stage I do some analysis before going for further rigorous experiments and simulation just to know whether reactive distillation is worth going ahead or not. Sometimes you get answer no right, just stop your exercise there right, but most of times if you have some gut feelings, experience and based on the boiling points or velodilities if you decide to go for reactive distillation the answer is always yes, but then it gives you some important information like what should be the reflux ratio, what should be the residence time. Now, I am talking about residence time this is a important consideration in reactive distillation nowhere in all other lectures on distillation we talked about residence time right, we did not talk about hold up and other things to that extent, but here it is very important in the reactive distillation column the residence time plays a very important role. Suppose you have a solid catalyst the catalyst loading it is a very important parameter. So, that is going to play an important role that is why you get some idea at this particular stage what should be the order of magnitude for the reflux ratio, number of stages or the Damkohler number what is the Damkohler number it is it decides the extent of reaction. So, it is proportional to residence time right, so we get some inputs from this particular exercise and then you can do the steady state simulations it can be dynamic or steady state simulation depending on what you are interested in and you can simultaneously perform column experiments. So, you looked at the setups in the laboratory right typically 3 to 4 meter height column to each diameter you get catalytic packing from companies like Schulzer, Koch engineering and all you use those packing for reaction of your interest right perform experiments in laboratory do rigorous simulation the same time taking some inputs from the conceptual design exercise. And then see whether your simulation predictions are matching with experimental data that you have generated. This is a very important step because now the model takes inputs from both kinetics and vapor liquid equilibrium and if you go wrong in one of these then your predictions would go wrong right. So, this step is very important and if there is no match then you have to go back and see where you are going wrong right because once you have a model which is experimentally validated then you can go ahead and do optimization control, scale up and so many other things and you can think of doing commercialization right. But this step is very important simulation and column experiments in fact that is what we have been talking about throughout this program like you have on one side you have conceptual design then you have simulations and you have laboratory experiments all these three things rather should go hand in hand or other should be given equally importance while designing a complex distillation system. If it is very simple ideal systems so much is known about it sorry is not much of a problem. But for system like this especially when I am talking about reactive distillation not much experience it is a novel process right that feelings would not help because there are so many things happening reaction distillation mixing right. And once you have this model ready then you are ready to play with the model and get the optimal solution or optimal configuration can design a control system and you can even go back and fine tune your parameters if required. So at this stage you are ready for commercialization ready for scale up and these are the activities one can perform in laboratory you can see one block here hardware selection. This is again very important in the case of reactive distillation. So in the last lecture you looked at column internals for distillation systems. But in this case it is not just distillation but you have to look at the column internals which are good for both distillation and reaction right. So you have additional constraint on the column internals I think I have some slides on this and we will talk about it when we proceed. So first activity is to get kinetic model for the reaction and of course this is for butyl acetate system this is for butyl acetate system this is a kinetic rate equation. So all I want to tell here is that you have a very complex model here. So as I said kinetics means what normally if you go through any textbook of kinetics like Levenspiel or Fogler most of the time the problem solved use the kinetics R is equal to K C A C B and all that. But in this case the kinetics you can see there is no K C A C B type factors here like of course I can see K that is forward rate constant. But look at what you have inside this bracket all these A's are activities and not concentrations. So normally the rate equation is expressed in terms of activities because all these components are reactive and most of the time the system is non-ideal. Since the system is non-ideal instead of concentration is always better to work in terms of activity and then you get a equation which is valid over the entire composition space. So this is the equation that I have given for butyl acetate and these are the parameters that we have estimated and of course there are many possible equations which can be used and looking at the residual for your regression analysis you select one of them. So there are many possible equations depending on the residual or rather the success of the parameter estimation will be decided based on the residual. But these are the models for solid catalyst. Now one thing I did not tell you that this reaction of butyl acetate synthesis is performed in the presence of ion exchange resin catalyst which is a solid catalyst. Sometimes you may have homogeneous catalyst less ptsa sulfuric acid and all that. In that case this model may not be so complicated. When you are using solid catalyst the reaction takes place on the surface of the catalyst and in that case you have to use adsorption based equations like LA radial or LHSW at Langmuir, Hinshelwood, Huygens-Watson model. So one important point here again probably we talked about it in desolation as well that we should look at a composition space. What does it mean? It means that when you do kinetic studies in laboratory, when you do kinetic studies in laboratory you do it in a batch reactor. So what will I do? I will take acetic acid, I will take butanol, I will do the reaction, I will follow the course with respect to time. I will get a data and generate the kinetics based on that. Now most of the times this kinetics is available for that particular composition space. Now in this case you have batch kinetic runs with pure reactor. This is a normal thing that we do in laboratory. So here in this particular zone of the composition space, look at this composition space. This is a 3D space. I have 4 components. I have 4 components. This is pure acetic acid, n butanol, water and n butanol acetate. When you do batch experiments on kinetics you are in this zone. But then if you do reactor desolation, it is quite likely that you will cross this zone, you will cross this zone and may go here, may go here. That is what we have shown here. Of course these profiles we have plotted based on the experiments that we did later in reactor desolation column. So all I want to say here is that your composition profile in reactor desolation can travel from one corner to another corner. Depending on how the vapor liquid equilibrium interacts with reaction. So you need to generate the data over the entire composition space to get the kinetic parameters because if you are somewhere here and if your kinetic data generated based on or other kinetic parameters are estimated based on the data generated here, they may not be valid. They may not be exactly correct. So it is always better to generate the data in this zone. So you may start with some butanol acetate present in the batch reactor. I am talking about generating kinetic data in laboratory. So instead of starting with pure reactants, you may start with some n butanol acetate present and generate batch reactor data. Another important thing is in the boundary you work with small amount of catalyst and most of the time if you see the kinetic equation, you have this m cat here. What does it mean? The rate is directly proportional to amount of catalyst. But what happens sometimes if you are working with dilute systems, the rate will increase. It is increasing linearly with catalyst loading. This is catalyst loading and this is initial rate. But after certain loading, you see it goes away from the linear graph. And in reactor distillation, the catalyst loading is very high because you are packing column with catalytic bagging. The catalyst amount per unit volume is much higher in reactor distillation. So you are in this particular zone and that is where you have the kinetics given by this plot. And if you have the data generated here and if you use a direct relationship which is given by this linear plot, you will have the error, this much error which is significant. You should be careful about all these aspects when you do scale up. And this is typical or this is a typicality of reactor distillation, such types of behavior normally observed in reactor distillation where there are many other possibilities like vapour liquid equilibrium interacting with reactor kinetics and so on. Now, vapour liquid equilibrium. The one part is over that is kinetics. We are talking about vapour liquid equilibrium. And of course, we have talked so much about it in last 4, 5 days on vapour liquid equilibrium. And this is our apparatus. We have shown this apparatus to you already. We have this equilibrium chamber and we generated data for different binaries and these are the existing correlations available for the same system that is the cotton system of butanol, butalacetate, acetic acid and water. And this is the one that we have come up with. And see, look at this particular system. We have so many azeotropes. So many azeotropes there. We have almost 5 azeotropes in this system, one with water butalacetate, water butanol, butanol, butalacetate, acetic acid butanol and there is a ternary azeotrop. And this is the model that we have used in our simulations later or conceptual design rather. And of course, vapour phase, there is a dimerization of acetic acid. Acetic acid is a tendency to dimerize in the vapour phase that needs to be considered and there are separate models for that. I am trying to tell you the importance of vapour liquid equilibrium. And case like this, where I am dealing with the reacting system. I am dealing with the reacting system. There is so much complexity. This system is highly non-ideal and we should be very careful about it. Just do not go by volatilities. Do not go by boiling points. So we have generated vapour liquid equilibrium data. We have generated kinetics. Now next step is to do the conceptual design. So in last 3, 4 days, we have looked at conceptual design of distillation systems. Now this slide is for conceptual design of reactive distillation systems. Of course, I am not going to tell details about it. I am just going to tell you that there are some tools available which can determine the feasibility in reactive distillation. Like residue curl map, residue curl map, it is used to determine the feasibility for azeotropic systems. Something like that we have here which is co-current cascades with chemical reaction. It is a model. Like residue curl map, this particular model has corresponding equations. We solve them and get some results which can be used to identify or work out the feasibility of reactive distillation. I am not going to talk much about this. As I said, like this is again an important tool or important step of conceptual design that tells you whether I will get pure butylacid at the bottom, whether I will get pure water at the top or not. Because initially, I just started with an assumption. The first configuration that I showed you, I said that I have acetic acid butenol going to the column and from bottom I will get butylacid, from top I will get water. That was my gut feeling. Whether that happens or not, it will happen in reality or not. How can I prove it? It is always good to do it on paper first before I actually build a column or do some regular simulation. This particular exercise tells you. This is a plot of temperature versus Damkohler number in dimensionless form. Damkohler number is the extent of reaction. If I increase the catalyst loading, the Damkohler number increases. If I increase the residence time, the Damkohler number increases. It tells you that I can get the possible bottom compositions. If I solve those equations, I get possible bottom compositions. This is one possible composition that is n butenol acetic acid as your drop or I can get n butenol as a possible bottom composition or I can get n butyl acetate as a possible bottom composition. What do I want? I want n butyl acetate as the bottom product. This particular exercise tells me that this is possible. This is possible. This is possible or potential bottom composition. Top composition is given by this particular line but this is again azeotrop. Once the azeotrop is broken or rather this is a heterogeneous azeotrop, you have the aqueous phase which is almost pure water. This particular exercise tells me that it is possible to remove butyl acetate as a bottom product in pure form and I develop confidence. Of course, how to do this? I am not going to spend much time on this. There are tools available that much I can say. So, I got the answer that I can go ahead. It is feasible. I can perform reactive distillation. Next step is to do experiments in laboratory and simulation. Experiments in laboratory and simultaneously simulation and validate my simulation model so that this simulation model can be used further to do optimization or even design control system. So, we have this laboratory scale reactive distillation column. You all have seen that. It is a multipurpose column, hybrid column. You can see the middle portion, the red one where we have packed the column with catalyst or catalytic packing, typically catapact S which is a soldier packing quite popular. I have given it here. It is a 3 meter tall column. Temperature sensors are situated at various locations. You have facility to withdraw samples. So, I can find out a column composition profile. I can find out a column temperature profile. This is the picture of this particular setup. And for butyl acetate system, conversion is of the order of 98 percent. Laboratory scale column, then concentration of butyl acetate in the bottom stream is greater than 97 percent. So, it shows that it is possible to get almost close to 100 percent conversion. So, these are the experimental results. But it is not very easy to conduct these experiments. Let me tell you that once we start these experiments, it takes at least 10 to 12 hours to get a steady state. The students, they normally work in shifts, even in academic institute. And 10 to 12 is the minimum time required to get a steady state. If everything goes well, no power shutdown and all that. Once you get a steady state, you have to remove samples because there is no control on this column. We purposely do not have any control. Control step will come later. So, we just want to get a steady state results. Typically, what we do is, we charge the re-boiler with the reaction mixture, butyl acetate, water, acetic acid and butanol. We keep the column initially under total reflux as I think we demonstrated that to you before for isotropic distillation as well. And then once the temperature profile is established in the column, the column is relatively hot, then we start feed of butanol, sorry, the equilibrium mixture of butanol, butyl acetate, water and acetic acid. And then we wait till the temperature becomes constant everywhere. The composition becomes constant everywhere, both top feed, overall material balance is satisfied. F is equal to D plus B and component balance is also satisfied. That means, if the reactant is getting consumed, that much product should form. So, everything is established. Then we report the results. So, these are the steady state results. So, what do I generate? I generate a data of temperature profile in the column, composition profile in the column. How much is the conversion? What is the top flow rate, bottom flow rate? And of course, the reballer duty and everything is known. Now, all this is connected to the computer. So, it is always easy to rather identify the steady state. Once the temperatures they become constant, I know I am very close to the steady state. So, I start removing the sample rather and I check them with respect to time. That is the way we conduct the experiment on reactant distillation. So, we have a data available now for the given system. And then of course, I have very certain parameters. Say, reballer duty, feed flow rate, removal ratio in the feed, sometimes the column height, that is a bit difficult. We had to play with the column setup, but that also we do sometimes. So, we have generated data. The next step would be to do simulations. And of course, we have spent so much time on simulation. Simulation of reactive distillation for that matter is not different from simulation of distillation. It is only one difference that you incorporate reaction. When you write the species balance on a given system or rather for on a given tray, then in the species balance, whatever coming is going out plus reacted. So, reaction term will appear in those equations. And of course, like depending on whether reaction is the equimolar or non-equimolar, your flow rates would change and all that. And then you can write a model and simulation software or a simulator, you should be able to take care of that. So, whatever simulation models were developed for normal distillation, when I say normal means not reactive or non-reactive distillation, the same model can be easily extended to reactive distillation without much of a difficulty. The solution method is same. You still use that naphthalene sand old algorithm inside out, right, Newton-Rassam based approach. But there are different possibilities. You can consider the stage as an equilibrium stage. That means the streams leaving a particular stage, they are in equilibrium phase equilibrium. Or sometimes as I said before, you do not provide proper contact, then the streams will not be in equilibrium. So, you have to consider efficiency. Now, when I am talking about equilibrium, I am not talking about reaction equilibrium. When I am saying equilibrium, that means it is only the phase equilibrium. Reaction equilibrium will be achieved if the reaction is very, very fast. But most of the times, the reactions are relatively slow. And that is why you can easily or you can assume comfortably the streams to be in phase equilibrium, but not in reaction equilibrium. But of course, if you have very large catalyst amount of loading, the reaction will be intrinsically very fast. In that case, sometimes you can go to reaction equilibrium as well. But there are models available. There are models available and the solvers now, you can actually solve right programs. Of course, in our lab, we have written our own programs because commercial software like Aspen and all, they have their own limitations in terms of kinetics that we can give and all. So, we have written some programs in our lab. But in the worst case, like of course, you can use Aspen, of course, the solvers are quite robust. Getting convergence is ensured here. But then of course, you have to have user defined kinetics if the kinetics is complex. You can use Aspen. There are some other software in which we can do programming and all, like write equations on our own. Of course, Aspen is very user friendly. You have to just give input and it will come out with the composition of outgoing streams and then the composition profile. So, you are doing simulation here. So, here we are doing actual experiment. Here you are doing virtual experiment. And then you have results and you have to match them. In the match, that means my model is good. My simulator or simulation model that I have used is experimentally validated. So, that is what we see here. So, your comparison of steady state experimental and simulation profiles. I am giving example of butyl acetate system again here. So, I am going through the butyl acetate system. As in when we change the system or if I want to tell you more about other aspects of reactive distillation through other reaction, I will mention that. So, these are different column composition profiles, where you have acetic acid, butanol, butyl acetate, water and temperature. So, this is the experimental data that we have generated in laboratory. Actual points and continuous lines are the model predictions. Not bad, good. So, that means we have confidence in our simulation model. Now, once I have model as I said before, now I can play with the model. Again, I have so many things I can do. I can do, look at dynamics. That means how the things change with respect to time. There is some fluctuation in the feed. How will it get reflected in my results? How will column respond to changes? So, this is what we have done. Rather, we have shown everything here, not just dynamics, but effect of different parameters also. If I change one parameter, effect of Damkohler number, that is catalyst loading, effect of number of reactive stages, effect of feed location, effect of mole ratio in the feed, effect of mole ratio butanol to acetic acid in the feed. I will come to this plot later because it is very important. The effect of flow rate and effect of revolver duty and see how it changes your performance. And that gives me some insight into column behavior and it tells me like what should be the parameter range I should operate the column at. So, that I will get the best possible performance because whatever I have observed here may not be the best performance. This is just for experimental validation. And what I am going to do next is a parametric study where I determine the optimal parameters or the performance which the best possible performance and the parameters required for that. I will come to this plot. What I have shown here is very as an important aspect of reactive distillation that we have realized through our research. You have mole ratio butanol to acetic acid in the feed versus DBE formation. What is DBE? And that is in PPM. DBE is dibutyl ether. Now, I did mention about this particular compound before. This is dibutyl ether which is a side product. And this ether is very high boiling and it comes along with butyl acetate in the bottom. And it spoils the specs. Nobody will buy butyl acetate with DBE present above certain limitation or permissible limit. So, this is an important aspect. If I do the reaction of butanol and acetic acid in a normal reactor, say CSTR or plug flow reactor, fixed bed reactor, I would not see DBE formation. I would not see dibutyl ether formation. In this case, reactive distillation gives me dibutyl ether formation. That means not good. Though it is helping me to enhance the conversion, it is helping me to reduce the capital cost, it is not good because it is giving me unwanted product. Let us try to understand why we get it. And these are the points which we have determined experimentally and we tried to correlate it with model predictions. So, we had a separate kinetics for DBE formation which we included in the simulation model and we compared the data. So, why this dibutyl ether formation is taking place? It is a very important aspect not just for this reaction or this application. It is true for other applications also and we should be very, very careful while looking at reactive distillation as a possible candidate for this process. I am not saying that it is always good. The many disadvantages also if you do not take proper care in process development and design. Why it is forming? Because in the reactive distillation column, in the reactive zone, if you look at the concentrations, this is where your reactive zone would be. And you have large amount of butanol present there and not much acetic acid. See, in this case, you have butanol present there and you do not have much acetic acid very small amount. In a normal reactor, what will happen? Since your stoichiometric ratio of acetic acid and butanol, if one mole of acetic acid gets consumed, one mole of butanol gets consumed. So, both will go together. So, butanol will always find acetic acid to react with. But in this case, because at some places you have more butanol and less acetic acid, butanol sometimes does not find acetic acid to react with and make butanol acetate. It reacts with itself because there is large amount of catalyst there. Temperature is high. So, you get di-dutal ether. Of course, the quantity is less, the PPM quantity, but still it is significant as far as the product specifications are concerned. This is very important aspect and why do you get this and not in a normal reactor? Because in this case, the profiles are not determined only by reaction, but by distillation as well, but by vapour liquid equilibrium as well. So, that is a very important point that profiles in the column will get adjusted according to both reaction and distillation, which is much different from the profile in a normal plug flow reactor, which is determined only by the reaction stoichiometry. And because of that, there is a possibility of formation of side products and which is happening in this particular case and which goes against reactive distillation. But then of course, we looked at butanol to acid ratio. That is why we say that you use more butanol, sorry, you use less butanol, you use more acid in the feed, slightly excess. In that case, your ratio would be less and you will get less amount of DBE. So, you should be very careful side products. But now since I am, see this is, as I said, this goes against reactive distillation, the side product formation because the composition profiles, they take a very peculiar pattern and because of that, you have formation of side products. But I can use this advantageously for other reactions to increase the selectivity towards the desired product. In this case, what has happened is the side reaction, side reaction got enhanced. So, I can play with the profiles in such a way that the main reaction gets enhanced. So, whenever you have selectivity problems, as I was telling you yesterday, A going to B going to C and I want to increase the selectivity towards B. In that case, I can play with composition profiles and increase the selectivity. I will come to that later. Before that, of course, so much discussion on column internals, but about reactive distillation again, we have to be very careful and we have several types of column internals or column packings available. Now, the very important point here is, now I want to use catalyst in the column. This typical catalyst particle size is, suppose it is an exchange resin. What is the particle size? Less than 1 millimeter. Can I just put an exchange resin in the distillation column? I cannot do that. What will happen? Tremendous pressure drop, flooding and so many things. The column performance will collapse. So, how do I use these packings? How do I use this catalyst in the column? The various ways, like your tea bags, you put the catalyst in some envelopes and you have a collection of these envelopes. So, vapor finds space through these envelopes and there is good voidage. So, there is no pressure drop. And I showed you the packing, catapact S, where you had this envelopes in the packing itself. And then many other ways, where you can code the catalytic material on the existing distillation packing. But of course, there are limitations. We cannot go to very high catalyst loading. You do not get surface area, that much surface area, which is required for reaction. Many pictures shown here. Now, I have dynamics. I have a simulation model. I can use the same simulation model, write it in time variant form. So, I look at a dynamics. How the things will change with respect to time? Suppose, I have fluctuation in the field. This is again for butyl acetate. Suppose, I have fluctuation in the field. How the column will respond to that? So, for example, I am not going to talk about all the figures. Look at this particular figure, where I am changing my volumetric flow rate at this particular time. There is a pulse here. There is a pulse here. How the column will respond to that? The mole fraction of butyl acetate in the bottom. You can see this. It is not exactly, actually these two are at the same time. These two are at the same time. So, the column will respond to it. But when it comes back, the column also goes back to its original stage, original steady state. Similarly here, in the negative direction, I give the pulse. Again, the column goes. These are all stage compositions. So, this is the column. Now, in this case, there is not much harm. If there is some fluctuation, column will respond to it but the fluctuation goes away. The column will come back to its original steady state. But then in some cases, it is very important. In some cases, it does not happen that suppose you have fluctuation in your field composition or field flow rate, column is running smooth but some fluctuation occurs in a field composition or flow rate or pressure or any state variable for that matter. It is possible that column will change its or column will respond to it but if the fluctuation goes away, column does not come back to its original steady state. Column will achieve some other steady state. That means, you are getting conversion of 98 percent and there is some small fluctuation in your field composition. I am just giving an example. Small fluctuation in your field composition, you are getting 98 percent conversion, smooth, it is running well, small fluctuation. You do not notice also. Say for few seconds or few minutes and then you realize all of a sudden, the composition has changed in the column, the temperature profile has changed in the column and the bottom composition is changed and your conversion comes down from 98 percent to 20 percent just because of small fluctuation. Such possibilities exist in reactive distillation. Why? Because a very complex process, highly non-linear process is something called as multiple steady states. You must have heard about multiple steady states. A CSTR is a very common example in reaction engineering. CSTR, exothermic reaction, multiple steady states. So, here also it is possible that you get multiple steady states. That means, for the same input, you can have two different outputs and one corresponds to very high conversion and the other would be at low conversion and just small fluctuation will take you from high conversion to low conversion. You should be aware of all this before we design. Because sometimes, if you just go by steady state design, say I have a perfectly designed column, you commission it to work well, but then I am not identified its sensitivity towards the inputs. And I may not know that it is likely to give me multiple steady states. And while during operation, if something goes wrong, for few seconds, few minutes and columns start behaving erratically and such possibilities exist in reactive distillation. You be very careful. Operation and control is also very important. So, non-linear dynamics, this multiplicity would come because of all this, non-linearity basically. If you see the equations, then non-linear equations and non-linearity exist in enthalpy balance, vapor liquid equilibrium, maybe there in reaction kinetics as well. See, non-linearity, look at the equations, it is non-linear. It is not like y is equal to mx plus c. So, because of that, you may see steady state multiplicity or you may see oscillations. The column will oscillate, sustained oscillations. If you just analyze the top composition or temperature with respect to time, you can see oscillations. So, I am going to show you that. Look at this. This is another example. This is not butylocitate. This is TAME. I was telling you about it yesterday. Tertiary Amyl methyl ethyl column. Reliance Jamnagar refinery has this plan. TAME plan. Commercial process, of course, I do not know much about how they operated and all that. But then I am telling you, it is a commercially important product. So, you can see the fluctuation would result. See, there are some fluctuations. The feed concentration is changed and column which are otherwise running smoothly. In this particular case, for this particular design, you start seeing the oscillations. Such possibilities exist in reactor distillation. Of course, you can see the details here. There is a reference. If you want more details, I will provide you. The aim of this particular slide is to show you that there is a possibility of seeing even oscillations and not just multiple steady states. And this further analysis here, I will just skip this slide. So, I will just summarize. The process development studies of reactive distillation would involve generation of kinetic data. It should be very careful. It should have versatile kinetic model. When I say versatile means, it should be valid over the entire composition space in every corner of the composition space, the reason is the composition profile in the reactor distillation column may travel from any corner of the composition space to any corner depending on the vapor liquid equilibrium. So, it should be very careful while generating the kinetic data. Thermodynamics, of course, I did not mention about it. Otherwise, it is well known how to generate kinetic data. It is all established. We had so many lectures on thermodynamics. But for reactive system, it is very challenging. Suppose the system is reacting. You can imagine the modified atmosphere still. How do I generate a data for two components A and B which are reacting? I cannot generate just vapor liquid equilibrium data. The moment they come in contact with each other, they will react also. So, getting that steady state and equilibrium is a problem. In our case, of course, acetic acid and butanol we could generate because they do not react without catalyst. So, in opmancyl, I do not have catalyst. I can generate a binary interaction parameter. But sometimes, we have studied some other reactions also where the catalyst is not required. The self is a catalyst. Self-catalyzed reaction. In that case, it is very tricky generating the vapor liquid equilibrium data. The homogeneous system, in the opmancyl itself, you have the reaction taking place. Column hardware, I told you the main problem here, catalytic packing. Now, an exchange resin packing, we can put them in envelopes like tea bags and all. Whereas, suppose you have zeolite catalysts. Can you imagine zeolite? Small particles crystalline. Typical size is micron. If you increase the particle size, then you have inter-particle diffusion resistance and all coming in picture. Small crystals, I cannot put them in tea bags. I cannot put them in those wire mesh envelopes. They will come out of it, micron size. I cannot have mesh. You can hold micron size particles, crystals. How do I do then? What is the solution? They are trying to deposit these zeolite particles on metallic surfaces. Now, imagine like we talked about structured packing in the last lecture. Structured packing. In fact, deposit zeolite particles on the structured packing. I can effectively use that. But again, like the efficiency of this deposition like because of some, because this is always some flow taking place, the vapor going up and liquid coming down. So, you have some problems of catalyst going, leaving the support. Conceptual design. Now, I talked about equilibrium reactions. But then you have multiple reaction. There is no tool available. There is no tool available. I am talking about selectivity. Suppose, I have two reactions taking place simultaneously. I am interested in one of the products. In that case, conceptual design methods are not well developed. There is tremendous scope in this direction for research of course. And the last aspect of process development that is non-linear dynamics and control is still wide open. As I told you, now people have just identified that such possibilities exist like multiple steady states, oscillations, and all that. But then, what is the cost behind it? If such thing happens, what should be the control strategy? How do I design the control system? No such work is reported in literature. Of course, whatever is available at its primitive stage, so much work required to be done to understand this particular aspect of reactive distillation. So, as I said before, like other distillation systems like, azeotropic distillation, extractive distillation or a normal distillation with non-adial systems including tangent pinches and all that. So, it is all well-known, well-established and we have a book talking about the theoretical aspects. How do thermodynamics interacts with mass transfer? But here, there are so many issues, still need attention and there is so much research going on and there is so much work and that is the main reason. Of course, you need to take efforts on laboratory scale experiments. The distillation as I said, they say that the scalar pressures are very high. There is so much information available. I just need to know the vapor liquid equilibrium. Of course, we had a special lecture for that but then it is not very crucial aspect, of course it is crucial but then as compared to reactive distillation, the design of hardware, it is well-known for that matter and that way the scale of distillation is not so difficult compared to reactive distillation and so much work required to be done as far as process development is concerned on laboratory level.