Addressing Unique Catalyst Deactivation Challenges for Converting Biomass-Derived Feedstocks Webinar

This is the text version for the Addressing Unique Catalyst Deactivation Challenges for Converting Biomass-Derived Feedstocks webinar.

>>Moderator: Hello, everyone, and welcome to today's webinar, Addressing Unique Catalyst Deactivation Challenges for Converting Biomass-Derived Feedstocks. Our speaker today is Huamin Wang of Pacific Northwest National Labs. Before we get started, I'll go over a few items so you know how to participate. You're listening in using your computer speaker system by default. If you would prefer to join over the phone, please select "telephone" in the audio pane of your control panel and the dial-in information will be displayed.

We are taking questions via the chat function on the bottom of your control panel. Questions will be answered at the end of the webinar. We will be recording today's webinar. You will receive a follow-up email within a few weeks with a link to view the recording. Thank you so much for attending, and now I will turn it over to Huamin.

>>Huamin: Hey, Hannah. Can you hear me? Hello, Hannah.

>>Moderator: Yes, go ahead and start your presentation now. Thank you.

>>Huamin: Can you hear me now?

>>Moderator: Yes, we can hear you now.

>>Huamin: Sorry, I had some little problem. Okay, great. Thank you. Hello, everyone and welcome to today's ChemCatBio webinar. My name is Huamin Wang and I'm a senior research engineer here in the Pacific Northwest National Laboratory. And today I'm going to share with you how we are addressing those unique biomass conversion challenges including catalyst deactivation challenges for biomass conversion in our ChemCatBio consortium

So, specifically in today's presentation I will show you that ChemCatBio is tackling overarching catalysis challenges for biomass conversion, especially including improving catalyst lifetime as one of those critical goals. This is specifically important for biomass conversion technology because biomass-derived feedstocks bring a lot of new challenges for catalyst lifetime or catalyst stability. And I will specifically present some examples of our efforts on understanding and mitigating catalyst deactivation within ChemCatBio and other related projects, especially on zeolite and oxide catalysts for aqueous phase conversion reactions and also specifically catalysts, especially the metal catalysts for pyrolysis vapor and oil upgrading.

And ChemCatBio, or Chemical Catalysis for Bioenergy, is a national-lab led R&D consortium. Our mission is to accelerate the development of the catalysts and related technology for the commercialization of biomass-derived fuels and chemicals. So, we are leveraging expertise and capabilities across seven national labs for catalysis study, including advanced synthesis characterization and multi-scale catalyst evaluation and assisted with modeling and simulation effort by computational.

So, as you see there, ChemCatBio is working on catalysis because catalysis plays a central role in biomass conversion, and catalysis is challenging; it is the biggest barrier in biomass conversion. When you look at across the variety of different conversion pathways for biomass conversion, including biochemical approach or thermochemical approach or hybrid approach, there is always a catalysis step, including heterogeneous catalysis. And within ChemCatBio we are specifically working on a lot of important catalytic processes within biomass conversion, including catalytic upgrading and biologic intermediates and catalytic fast pyrolysis, I think there's upgrading and—catalytic upgrading of aqueous phase or gaseous phase in the waste streams, and catalytic hydrotreating.

And what makes catalysis R&D in biomass conversion more challenging is that the biomass presents lots of unique properties, as shown here. It has lots of oxygen there—so, and it also has lots of diverse chemical functionalities. Therefore, we have to consider competing reactions and make our catalysts more selective to certain desired reactions. The more important, as shown, highlighted over here is that lots of those kind of biomass-derived feedstocks have a water content and the impurity there causes a lot of problems for catalyst deactivation and catalyst lifetime.

So, we all know that catalyst lifetime is important, as important as its activity and selectivity. For industrially relevant process we have to demonstrate catalyst stability to meet the production performance goal or requirement for this process. However, in some early stage R&D the catalysis or catalyst stability is normally receiving less attention compared to its activity and selectivity. Within ChemCatBio Consortium, we believe that understanding of the catalyst deactivation and developing strategies to extend the catalyst lifetime are critical to the success of the process development for biomass conversion.

And it's especially important for biomass-derived feedstocks because it brings new challenges to this issue. Compared to the fossil feedstocks, most biomass-derived feedstocks are complex, highly functionalized, and unstable. I'm showing here examples of the bio-oil, which is a fast pyrolysis product from biomass; it contains lots of reactive species. It will go through some condensation reactions or polymerization reactions to form high molecular stuff and to plug your catalysts’ pore and eventually plug your reactor.

At the same time, biomass-derived feedstock could contain contaminants, either from the biomass itself or from the process units. Those contaminants can be organic form, like sulfur or nitrogen containing species, or be inorganic like lots of minerals as shown over here.

A set of problems or challenges for biomass-derived feedstock is that it normally contains lots of water and corrosive species like carboxylic acid. Basically, this requires our catalyst to be working on the aqueous space or corrosive conditions.

So, here in the next few slides, I will provide a summary of how these unique properties of biomass-derived feedstock could cause catalyst deactivation on different types of catalysts we normally use for biomass conversion, and how the mitigation approach can be used to address those issues. Those are information based on reported results in literature or based on our research at ChemCatBio. So, we can group those kind of challenging properties of biomass-derived feedstock into heavy species, reactive species, contaminants, and water/corrosiveness. And a different type of catalyst, causes different types of deactivation, for instance, catalyst fouling, poisoning, or degradation. A mitigation approach can normally include catalyst regeneration or catalyst improvement to be more resistant to this deactivation mechanism, or process improvement including removing those troublemaking species from our feedstock.

Let's first look at the heavy species and reactive species issues. For heavy species, it's normally from the conversion of a biomass, which is a high molecular weight polymer into our intermediate product, and it normally has residual high molecular species. Those high molecular species can also be formed by reactive species in the biomass-derived feedstock, especially from those aldehydes and ketones by condensation reaction or polymer reactions that forms those heavy species. So, those heavy species can easily cause catalyst deactivation by fouling, and this fouling is somehow more chemical or just a physical problem. It kind of is working on pretty much all the catalysts we are using over this biomass conversion.

And the mitigation process can be regeneration to burn off those coke or by process improvement to remove those hot, heavy species or reactive species feedstock. There are some good examples, like listed here. For instance, bio-oil hydrogenation can be fouling by heavy species. It's really a problem and can be regenerated by cleaning the catalysts. Or, zeolite catalyst using for catalytic fast pyrolysis can be fouled by those condensation products and coke can be regenerated by oxidation. Also, other examples, one unique one is the HMF furandicarboxylic acid approach. Normally, the catalyst will have fouling by FDCA will cause catalyst degradation and this can be partially addressed by using special solvents for this process; therefore, you can remove FDCA as soon as it was formed.

A second big problem for this biomass conversion regarding catalyst deactivation is contaminants. And we know that as I just listed that a lot of sulfur nitrogen species in certain biomass-derived feedstock and also in organics from biomass or process units. The kind of type of catalyst—like sulfur can cause poisoning of metal catalysts. Organic can cause poisoning of our acid-based catalyst and nitrogen that can poison our Brønsted acid catalysts. So, the poisoning will cause catalyst degradation and mitigation approach can be regeneration or process improvement.

Some examples are including ruthenium catalysts for different approach, different application, and it can be poisoned by sulfur. And this can be either regenerated by removing sulfur from the poisoned catalyst or by removing sulfur from our feedstock. I will give you a more specific example on this sulfur problem.

The last one is water and corrosiveness. I just mentioned that there's a lot of water involved in our biomass conversion process. And in this case it's really going to impact a lot of metal oxide-based type or zeolite type catalysts, causing catalysts to disintegrate or degradation or phase transition. And the way to normally to do it is—to mitigate those deactivation—is either by catalyst improvement to make catalysts more robust towards those water issues or by process improvement, especially to remove water to get relatively clean, pure, and organic feedstock for this chemistry.

Examples in literature include aqueous hydrogenation reactions with nickel catalysts, and by aqueous phase it can form a nickel hydroxide. And this can be partially mitigated by catalyst surface hydrophobization to limit the water near the catalyst surface. Or, for the copper catalyst and the copper leaching problem, you can stabilize these carbon species on the catalyst surface by atomic layer deposition.

So, hopefully I showed you that overall big picture, how those unique properties of the biomass-derived feedstock can impact the catalyst degradation and what's kind of the general approach to mitigate that. So, the following slides are 20 minutes. I will give you some specific examples for those aspects, and those examples, are either from our related biomass conversion project and also primarily from our ChemCatBio consortium. So, I’ll start with, I just mentioned the aqueous phase conversion and water/corrosiveness issues. I’ll show you two examples of catalyst stability improvement for aqueous phase reaction, especially in hot liquid water. I'll give you the first example on "stability of zeolite improvement for aqueous phase dehydration reaction." This is based on research at PNNL funded by our internal LDRD. I'll give you a second example: "enhancing hydrothermal stability of zirconia catalyst for the ketonization reaction of acetic acid in the aqueous phase." And this is work that is funded by our ChemCatBio and it's, as you can see, that is—all that work is reported in the literature, and you can see more information from the paper I cited over here.

So, let's start with zeolite catalyst. So, why are we working on zeolite for this liquid phase reaction? So, I'm showing you this figure that—you know, the dehydration reaction as shown here, alcohol dehydration, is one of the most important and straightforward methods to remove oxygen from our biomass intermediates. However, most of this dehydration reaction can be catalyzed by a Brønsted acid site like proton over zeolite.

The problem for this reaction is that this reaction was very fast on the gas phase over zeolite catalyst. As soon as you put this zeolite catalyst in liquid water and therefore you want to convert those alcohols in liquid water without separation. So, now the problem is that water will interact with those protons and form a hydrated hydronium ion, which will significantly lower the ability of those Brønsted acid sites to do the dehydration catalysis.

So, by using zeolite, I show in this figure that zeolite cannot only enhance or enrich those organics in the pore; therefore, they increase the concentration of the alcohol in the pore, but also can provide the confinement impact to get a much faster turnover rate of those Brønsted acid sites after water hydration. So, as shown in this picture, by using HMFI we can see a hundred times faster turnover rates over alcohol dehydration compared to free liquid acids.

This makes zeolite to be a very attractive catalyst for aqueous phase dehydration reaction. Of course, the immediate question is how stable is this zeolite for this application? And yes, these are problems that zeolite is not stable in hot liquid water, as shown in this table. And just simply process or treat the zeolite, beta zeolite, in liquid water at 160 degrees C, we do see a significant drop of surface area and pore volume. This is because the hydrolysis of the zeolite framework and the partial dissolution of the crystal in liquid water is causing deconstruction of the zeolite structure.

So, the further extensive study shows that this kind of deconstruction is because of a silanol defect site over those zeolites. And I'm showing this table that after we treat the zeolite with silylation, as shown on the left-hand side, after silylation, NMR shows that we actually remove those silanol defect sites from zeolite. But the thing is we actually lose some surface area and pore volume, but they largely maintain stability and as showing in the last line, after water treatment of this silylation of zeolite, we do see the surface area and the pore volume is larger maintained, really showing that it is true that the silanol defect side is the cause of structural instability of zeolite in hot liquid water. And by silylation treatment, we can improve its stability.

However, by just controlling the silylation of silanol defect sites is not sufficient. The concentration of intraporous water also is directly related to the rate of hydrolysis of zeolite framework. I am showing you on the left-hand side that we are changing the Brønsted acid density by changing the silicon/aluminum ratio of zeolite. And you can see that we are increasing the water uptake with an increase in Brønsted sites density. This is both true for a defect beta zeolite or low-defect beta zeolite. We are showing that because of the existence of Brønsted acid sites we do have a lot of water in the pores. This water will cause a problem of the stability of zeolite. And on the right-hand side you can see that this is really true, that only at a lower Brønsted acid density zeolite we can have a much-increased lifetime. And the silylation method or the lower defect density is only important at low concentrations of the Brønsted acid sites or low concentration of water inside. This really tells us that we have to control the water contact within the zeolite pore to further mitigate this deactivation problem.

And another approach is doing external hydrophobization. And in this figure we show that by doing external hydrophobization of the zeolite we can have a much improved lifetime activity. We believe this is because of the limitations of diffusion of liquid water into the pores, and therefore to further decrease the concentration of water inside the zeolite and further enhance its stability.

So, with this research we hope that we have now a toolbox for zeolite stability enhancement in liquid, in hot liquid water. This toolbox includes starting with a lower defect density, especially a silicon defect density, by using a synthesized fluoride media to get a more crystalline structure and therefore less defect site, or by post-synthesis treatment like silylation. Or at the same time we need to lower the water content of zeolite by tuning the silicon/aluminum ratio and generally lowering the density of the Brønsted site density, at the same time, by some external surface modifications, like external hydrophobicity.

I'm showing here that we do have a lot of this kind of existing knowledge on catalyst stability and mitigation from related projects at different national labs. At ChemCatBio we will definitely leverage this knowledge for developing stable catalysts, especially for this zeolite system, and we are going to leverage what has been done over here for upgrading biologically derived intermediates in the liquid phase.

So, now, I'm getting into a more complicated system. And this is another system, working to enhance the stability of zirconium catalysts for direct catalytic upgrading of aqueous phase carboxylic acids. This is a relatively more complex system compared to the alcohol in water because this is a system from the hydrothermal liquefaction. And for
woody biomass, after the hydrothermal liquation or HTL a lot of carbon will end up leading to aqueous fractions. And in these aqueous fractions, the majority of the carbon is in carboxylic acid form. And there is also other species, especially sugar-derived species. And one of the strategies to use this carboxylic acid, carbon in HTL aqueous phase is to do ketonization reactions shown here, to convert those carboxylic acids into ketones. And the ketones, from the ketones, it will be more straightforward for the steam reforming process to produce hydrogen, or by hydrogenation, dehydration we can get other things from that.

So, for this ketonization reaction, the zirconium catalyst is showing very good performance, especially in gas phase. However, as soon as you put the zirconium in this liquid phase and especially in very diluted carboxylic acid, like five percent or below, in water, as shown in the left-hand side we do see a fast deactivation for zirconium in this application. And for the characterization of this zirconium catalyst, I am showing you in the middle of XRD, and it reduces—at the bottom is fresh zirconium, as well as treating zirconium in water or in carboxylic acid, and it induces the phase transition from tetragonal zirconium to monoclinic. And the team actually spent a lot of effort to really mitigate this catalyst deactivation by using lanthanum doping to really not only enhance the activity of this catalyst, as showing on the left-hand side, and also prevent those phase transitions, as showing you on the right-hand side. After using lanthanum doping this catalyst remained intact, the tetragonal zirconium phase in water or in carboxylic acid condition. By that, this catalyst was much more than a thousand hours lifetime for this application, and this we reported in this paper as cited over here.

I think I showed you that we do have a lot of nice studies to understand how we can mitigate the deactivation bring by condensed water reaction. And now I'm looking to a much more complex system here, fast pyrolysis. So, fast pyrolysis is a process that I can treat the biomass directly, in high temperature, and follow the fast quenching that you can get a liquid. And here the bio-oil itself, I will talk more about that. The bio-oil itself does have significant problems regarding its upgrading to produce fuel products. In order to enhance the quality of the bio-oil there are two approaches to do that. One is that you can produce bio-oil and then to do a bio-oil stabilization to increase its quality for hydrotreating. Or you can use catalyst during fast pyrolysis, which can cause catalytic fast pyrolysis to directly react with the fast pyrolysis vapor to produce a stable or high-quality CFP bio-oil.

Today I am going to show you that—how we are going to mitigate catalyst deactivation during those fast pyrolysis bio-oil stabilization step. This is a work done by PNNL and Oak Ridge under ChemCatBio for fast pyrolysis bio-oil quality improvement. And also, I will give you another example that is an ongoing effort at ChemCatBio on catalytic fast pyrolysis. How are we understanding the deactivation mechanism of the catalyst used for this catalytic fast pyrolysis? It's a collaboration effort between PNNL, NREL, and Oak Ridge, and also multiple projects within ChemCatBio.

So, I will just mention that after fast pyrolysis we have the bio-oil, and this bio-oil is really low quality. The quality is not only because it has a lot of oxygen, a lot of water, and low heating value, but the major problem is that the bio-oil is unstable. And it will be causing—it will undergo some condensation reaction to form the high molecular species and encapsulate your hydrotreating catalyst, if you directly feed this bio-oil into the hydrotreater. Therefore, I just mention the bio-oil quality improvement can enable those direct hydrotreating and those improvements are currently including bio-oil stabilization by a low temperature hydrogenation reaction, or pyrolysis vapor catalytic upgrading—we call it catalytic fast pyrolysis. Both processes are using kind of reduced metal catalysts, and with hydrogen and doing a lot of catalytic reactions including hydrogenation and HDO.

Let's start with fast pyrolysis bio-oil stabilization by hydrogenation. So, as I just mentioned, raw bio-oil if you put it into a reactor or hydrotreater at around 400 degrees C, pretty much the reactor will plug within in one day. And our previous work showed that if you can stabilize the bio-oil by doing low-temperature hydrogenation and the first generation of low-temperature stabilization by using sulfide catalyst you can partially stabilize the bio-oil; however, it is not sufficient enough and normally we will see a reactor plugging around 100 hours. This is really showing—asking us to really kind of conduct deep stabilization. And so-called deep stabilization, I'm showing you here, is really doing a low-temperature hydrogenation. But we learned that if we do deep stabilization, the bio-oil will be able to be processed in hydrotreater, more than 60 days without plugging.

So, what is really happening during deep stabilization? So, before that let's look at what's in the bio-oil and what is causing problems. In bio-oil you will have a lot of sugar-derived species, which is levoglucosan or a lot of aldehyde ketones. And at the same time we have phenolics; all those phenolics or oligomers from the lignin. And of course we have a lot of carboxylic acids. So, basically, you see those species over there, that they can undergo a lot of condensation reactions. They either form humins or form those highly molecular species or resins. And those products, those reactions will occur as low as 80 degrees C. And that means that in order to compete in this reaction we have to do a stabilization reaction at low temperature, and the approach we are doing is to low-temperature hydrogenation, and we saturate those aldehyde ketones phenolics into alcohols.

And we have to do this reaction, again, at a very low temperature. And what we find is that the reduced ruthenium catalyst can do this job very well and it can do hydrogenation at low temperature as low as 120 degrees C. Therefore, it can really compete with those condensation reactions and stabilize the bio-oil.

However, in order to do that we have to use a reduced metal, and sulfur poisoning is always a problem for this reduced metal catalyst. I'm showing here that with time on stream we use hydrogen consumption as an indicator for the catalyst activity, and we are showing that the catalyst activity is dropping with time on stream. We find the deactivation rate is directly linked to the sulfur content in our bio-oil. We are comparing two bio-oils. They are the exact same bio-oil but we are managing to vary the sulfur content, and we see that at a lower sulfur content we can have a slower deactivation rate. This further confirms that post-catalyst characterization—we have much more data than that but I will just highlight the most important one here, that is really showing that at a bio-oil with lower sulfur content, that we had much lower sulfur in the spent catalysts, it's really showing that sulfur poisoning is a major deactivation mode for this catalyst.

However, at the same time we do see a coke formation over those spent catalysts, as shown here that it's always 5 to 6 percent of carbon on a spent catalyst, really showing that polymerization—and we call it coke here, but it's not exactly the same coke that a refinery is talking about—that those polymer formations or coke formations also contribute to catalyst degradation as a secondary deactivation mode, over a sulfur poisoned catalyst.

So, in order to manage this sulfur problem, what we do is we use a lower sulfur content in a bio-oil to enable and improve the catalyst stability. As shown here, for bio-oil, if we use a lower cost, a base metal as a guard bed, we can kind of treat this bio-oil to lower this sulfur content below 10 ppm. By doing that, as shown here, the carbonyl content with time on stream, we do see a much-improved carbonyl reduction and we can have much less carbonyl in our product, which is our desired goal. We do see a much stable catalyst over here.

And most recently, together with Feedstock-Conversion Interface Consortium, we also found that we can control sulfur content in bio-oil by controlling the moisture ash content in your biomass feedstock. I'm showing in this figure that normally if you have a lower ash or lower moisture content we will get bio-oil with much lower sulfur content, as shown here: 19 ppm sulfur. And it's definitely showing a much-improved catalyst stability for this hydrogenation over the Ruthenium catalyst.

A second strategy is that we can do catalyst regeneration. So, we are—since we are running our reaction at a relatively low temperature, it makes a lot of sulfur poisoning is a more likely chemisorption of those sulfur-containing organic species, rather than in a high temperature H2S and surface sulfiding. So, with this one we can use some solvent methods to kind of clean or extract those sulfur species and those coke species over our entire surface. So, we developed a two-stage cleaning approach to really—first stage, to remove those polymers from spent catalysts, and the second stage to remove the sulfur species from these catalysts, and then we do re-reduction and then put it back for bio-oil stabilization. So, basically, by doing that as showing in this figure that we will see after the first cycle we will lose 20 percent of the catalyst lifetime. This is just because there are certain sulfurs strongly bonded to our surface that cannot be removed. However, after the first cycle, in the second or third or fourth cycle we do see that it kind of reaches sort of a pseudo steady state, that this catalyst is maintaining its activity after this regeneration. So, based on that we can design a material, multiple reactor system to do this bio-oil hydrogenation, and for certain reactors in service and certain reactors in regeneration. Together with a lower sulfur content, if we control our biomass feedstock very well, we can achieve a year, more than a year lifetime on this, we've seen in catalysts for bio-oil hydrogenation.

However, we have to do some further R&D when looking for this process, because so far this process is demonstrated over oak, hardwood bio-oil, but we still have some questions that need to be answered to use this process for other bio-oils, especially soft wood bio-oils.

So, let's now focus on CFP. CFP, as just mentioned, uses a catalyst during fast pyrolysis. Especially here at ChemCatBio, we are more focusing on this ex situ fixed bed CFP. Basically, after the fast pyrolysis reactor the pyrolysis vapor will be cleaned by a hot gas filter, and then the vapor will go through a fixed bed reactor to do various reactions to upgrade this vapor into a more improved bio-oil for hydrotreating. So, this enables more diverse catalysts and chemistry and it requires—however, you can imagine as a fixed bed it requires a much longer catalyst time. And therefore, understanding the deactivation mechanism of this catalyst is critical. I am showing here that we are using—the NREL team developed this very nice platinum-titanium dioxide catalyst, and it's leveraging its bi-functionalities, including a metallic functionality here provided by platinum and also acid-based functionality provided by TiO2 support.

However, this catalyst also has a deactivation problem and it will deactivate during a single cycle run, primarily by coke formation. Oxidation regeneration catalyst activity will largely recover. However, we do have concerns with its long-term stability and understanding this mechanism will be critical.

So, we received a CFP catalyst after 200 hours run, and we probe its change compared to a fresh catalyst by combining detailed characterization with kinetic study. And I'm showing here that we were specifically focusing on the HDO activity over the platinum active site and also dehydration or condensation reaction over the TiO2 active site. What we learned is that the first active site activity is largely maintained, and we would not find sulfur on this spent catalyst, which is really good news that the majority of the active sites in this catalyst is maintained, and sulfur poisoning, the most challenging issue, is not here.

However, we do see the accumulation of the potassium in this case. And potassium is from the biomass feedstock. And it keeps accumulating over these spent catalysts. And also, we do see that platinum particle size is changing upon the different treatments and different regenerations during this process. And lastly, that we did a kinetic study and acid site titration we find that most of the acid site was largely maintained, but there was some loss just because of a surface area loss. With this preliminary characterization of the spent catalyst, we find that changes of the platinum particle size and the accumulation of potassium could potentially impact this catalyst lifetime even longer term.

And we specifically focused on potassium issues and potassium accumulation is not only for this specific system. It is also widely observed in in situ and ex situ CFP and hydropyrolysis. Specifically, for this process we find that in terms of accumulation, it's more occurring on the leading edge of the catalyst bed, as shown in this figure, that the top of the catalyst bed is seeing more potassium deposition. However, we believe that with time-on-stream this type of deposition will start to impact our catalyst in the middle and the bottom. So, now it's really a question of how is potassium distributed at an atomic scale, and what's the consequence in catalysis?

So, we have some preliminary data with ACSC, and it's showing that the potassium deposition on the leading edge is true by XPS analysis. And further TEM images suggest that this potassium is uniformly distributed on the catalyst surface. That's really important information because if this potassium is going to form a lot of clusters of crystals somewhere, it may be less a problem, but if it's uniformly distributed on the catalyst surface, it may impact the active site and cause a problem. Therefore, we focus that we have catalyst deactivation and mitigation project within ChemCatBio, and it especially can focus on this potassium accumulation on the catalyst surface and understand its consequence regarding its catalyst activity.

So, in this case we controlled—we conducted controlled deposition of potassium with different loading over platinum and titanium catalysts and we conducted extensive characterization and evaluation. The first thing we found is that by doing this we actually can duplicate what we observed on real spent catalyst and we get a uniform distribution of potassium over the catalyst surface.

And characterization really shows that by those potassium depositions we are not influencing surface area or metal density, metal site density or metal surface area of this catalyst but more acid-based site density, as shown here. And we further conducted extensive kinetic studies over this spent catalyst, this potassium-deposited catalyst, and here we are showing that for dehydration reactions what we learned is that by loading potassium on this surface we do see increased dehydration rate. At the same time, we are not only seeing a decrease in dehydration rate, but also we see an increase of activation energy for this reaction. This really tells us that the potassium is not only killing the active site but also killing a different active site or modifying an active site from a strong active site to a weaker active site.

However, for a HDO reaction we find that it will require almost 2,000 ppm of potassium to start to see a negative impact of potassium deposition for HDO reaction around a platinum active site. At the same time, it looks like the activation energy for this reaction at different potassium loading is pretty consistent. What I'm showing is that this is more like deposition or poisoning of active sites rather than modifying active sites.

Based on this information we propose that for potassium impact on this catalyst system and—basically, for a clean surface, you do have a metal active site and Lewis acid site on TiO2. And at low potassium loading, the potassium looks more a bit—the deposits are remotely loose at this site; they only impact a Lewis acidity, and therefore the hydration or acid-based catalyst reactions. At a higher potassium loading, the potassium starts to interact with the interfacial sides, and therefore starts to interact—start to inhibit the HDO reaction, as shown in the bottom.

So, basically, we know that interaction of potassium with active sites really depends on the potassium loading and it leads to different catalytic consequence. So, with this information it will enable us to predict the catalyst performance at different times-on-stream, and therefore different potassium deposition, and also help us to understand what's the catalytic deactivation—or, catalytic regenerations were like and how much potassium we should remove from this system, therefore to recover that activity.

Okay. Hopefully, today I showed you that within ChemCatBio we are tackling overarching catalysis challenges for biomass conversion, and we are specifically including this improving catalyst lifetime as one of our most important goals. This is important just because biomass-derived feedstocks bring new challenges to catalyst lifetime and catalyst stability issues. I hope that I showed you those examples on ChemCatBio's efforts on understanding and mitigating catalyst deactivation, especially on how we are modifying zeolite and oxide catalysts to enable improved stability for aqueous phase reaction, how we are managing sulfur or how are we developing catalyst regeneration protocol to extend ruthenium catalyst lifetime for bio-oil stabilization, and how are we conducting foundational understanding of deactivation mechanism of those palladium or titanium—platinum or titanium for catalytic fast pyrolysis, which could lead to process and catalyst life improvements.

So, this is really a team effort at multiple national labs and multiple projects within ChemCatBio. And the people participating in this research are listed here and I appreciate it. And I especially thank the support of the Bioenergy Technologies Office for those projects and also for the whole ChemCatBio.

And with that, I would like to thank you for your attention and I would like to take any questions if we have enough time. Thank you.

>>Moderator: Thank you, Huamin. We have a couple of questions. The first one is about the slides, getting a copy of the slides. If you look on the control panel, there is a gray bar that says "Handouts," and if you click on that you can actually download the slides right now. But we will also be sending them out after the presentation with the link to the video recording.

We have another question for Huamin: "What is the fouling of catalyst by heavy species?"

>>Huamin: Okay. Thank you for the question. So, fouling the catalyst by heavy species is like, for instance, like fast pyrolysis, that not all the primary structure of a biomass will decompose with fast pyrolysis. You do have those oligomers of sugar or lignin existing in bio-oil. Those really high molecular species do have diffusion problems or do have reactivity problems; therefore, it cannot be converted in our—in special low-temperature hydrogenation reactors. It will just physically deposit on the catalyst surface and cause problems.

This is also true for other processes, for instance, like aqueous phase products from hydrothermal liquefaction. The unconverted sugars and oligomers will do the same thing, where basically just a high molecular species deposit on the catalyst surface to encapsulate the catalysts; therefore, to prevent diffusion of reactants into the catalyst pore and even further plug your reactor. And this is what I'm referring for in the high molecular species in the biomass-derived feedstock for catalyst fouling.

>>Moderator: Thank you. Are there any other questions from the audience? You can type it in the chat panel on the right-hand side in the control panel. There's a bar, a gray bar that says "chat." I'm going to leave it open for just one more minute to see if there are any other questions.

All right. Here's another one: "During ex situ CFP where the flue gases from the char combustor passed over the catalyst bed, have you tried bypassing the catalyst bed when combusting the char? And did you still observe K poisoning?"

>>Huamin: Yeah, my understanding is that for this process currently we are using a hot gas filter before this fixed bed reactor. Therefore, the char from the fast pyrolysis vapor will be largely removed from this—by this hot gas filter. And therefore, definitely there will be coke formation on this catalyst system, and we are conducting regeneration of this catalyst by burning off the coke by the oxidation approach definitely, and we do see the potassium accumulation in this case. So, that means there's potassium still—although char is removed, but still there is potassium in the vapor and depositing on the catalyst surface.

>>Moderator: All right. The next question: "Any research on catalyst for xylose conversion to xylitol via hydrogenation?"

>>Huamin: My understanding, at least within ChemCatBio, there is no such research. But I will definitely double check with that.

>>Moderator: Okay. I'll leave it open for just any other questions that the audience might have. We can also answer questions offline if you'd like to add them in. Oh, here's another one: "Have you tried surface hydrophobization-supported metal catalysts?"

>>Huamin: Yes. Actually, this is one of our ideas of dealing with the bio-oil hydrogenation or similar catalyst systems, that by doing surface hydrophobization you can limit the water concentration near the catalyst surface, and therefore near our metal sites or near our acid-based sites. Therefore, it can somehow protect this catalyst from any problem induced by water. That's definitely a good approach for that. Thank you.

>>Moderator: A question is—a follow-up on that is: "What support?"

>>Huamin: Do you mean this question? Yeah, it will be more like a transition metal oxide or zeolite support, those oxides, definitely. But of course, if you can do modification of carbon, it's also important. But somehow, carbon has some limitations because oxidation regeneration may not be able to be applied to a carbon-supported catalyst.

>>Moderator: Okay. Thank you. All right, we have another question: "Process control is going to be key here, as in feedstock quality and product formation. The feedstock quality is going to vary as much as any current production line that exists in this area. Understanding that will be key in its control. We'll be in touch."

>>Huamin: Thank you for the question. Exactly. I think as I just presented briefly, that we do have—the Feedstock-Conversion Interface Consortium is basically asking this type of question, that what's the changes in feedstock on the process that's going to be impacted on the downstream quality—the quality of downstream—a product—the quality of product and also the downstream process? And the catalyst lifetime or catalyst stability is one factor we're considering there. As I just presented, for instance, ash content or moisture content in feedstock could impact the sulfur content in our product. Therefore, it impacts our bio-oil hydrotreating or hydrogenation catalysts’ lifetime. So, this is one very important area that I think national labs are dealing under BETO. And we do have those consortiums like Feedstock-Conversion Interface that's dealing with these questions. Yeah.

>>Moderator: Thank you. Next question is: "What are your regeneration methods?"

>>Huamin: Can you specify which process?

>>Moderator: That was not specific about—oh, bio-oil.

>>Huamin: Okay. So, in general, if you are talking about bio-oil stabilization and regeneration, it's solvent cleaning. As I mentioned, since the reaction is at relatively low temperature. So, lots of those sulfur species are chemisorption of those organic sulfur or sulfur carbon—sulfur compounds over the catalyst surface. It's not an H2S poison, that catalyst; therefore, we can use some solvent to really remove—weaken the metal-sulfur bond and then remove this organic sulfur to regenerate it. In general, for this low-temperature hydrogenation of bio-oil we are using a solvent to washing methods to regenerate catalyst. The high temperature, I'll just mention that we use the oxidation method to reduce, to remove those coke, and for potassium we are still working on that. But in the literature, you can see for the zeolite catalyst, a mild acid washing can somehow remove those potassium and therefore regenerate the catalyst.

>>Moderator: All right. Thank you. Wait one more minute? Any other questions from the audience? All right. Well, thank you so much. Thank you, everyone. Thank you, Huamin. Thank you for attending today's webinar. You will receive a follow-up email within the next few weeks with a link to view a recording of today's webinar. You can share that. Thank you for joining us today. And have a great rest of your day.

[End of Audio]