 Great. Thank you, Anurag and Veronica for the kind invitation and thank you all for being here to participate in this webinar and please feel encouraged to use the chat window to ask questions because it's always more fun than just being alone in my office and talking to my computer. So as you heard already from Anurag, I'm at the Max Planck Institute for Chemical Energy Conversion and I wanted to just briefly introduce you first to my department. So we're the department of inorganic spectroscopy. I'm fortunate to work with a relatively large team of really wonderful postdocs, PhD students, undergraduates as well as technical staff. This is the last picture we have when we were still allowed to gather in large groups and what you can see from the flags underneath this photograph is we're also a very international group, which I think also brings a lot of diversity and different perspectives. The other nice thing about it though is we aren't just internationally diverse, we're also scientifically diverse. So my group is comprised of biologists, chemists, physicists and engineers and it's very much my feeling that we need this diverse range of perspectives really to tackle some of the most challenging problems in energy research. And so before I get to the topic of today's seminar, I want to give you just a brief overview of the research in my research group. And so fundamentally, what we're really interested in understanding are key reactions in energy research. So how do I best store and release energy from a chemical bond? And how do I do that in the most efficient way possible? So for those of you in the audience who are chemists, you're of course well aware that this requires a catalyst to lower the kinetic barrier for that conversion. And this catalyst could be a heterogeneous catalyst, some sort of surface space catalyst, it could be a homogeneous catalyst that a chemist synthesized in a lab, or it could be a biological catalyst found in an enzyme. But when we study any one of these reactions that's important in terms of chemical energy conversion, be it nitrogen reduction or hydrogen production, water oxidation, CO2 reduction, methane oxidation. What we really want to know is, on an atomic level, how does this conversion occur? And how do some metals best optimize conversion to happen efficiently? Now at our institute, we're particularly interested in earth abundant metals, because they're sustainable. And we feel that's sort of a key to our energy future. But it's also one of the reasons we're very interested in looking at enzymes, because when we look at how biology does it, all those reactions I showed you on the previous slide, they're also used or carried out by enzymes. And I'm showing the metals that all of these enzymes use. So the hydrogenase enzymes for H2 production, nitrogenase enzymes for N2 reduction. The methane monoxygenase is to take methane to methanol, photosystem 2, which does water oxidation and things like carbon monoxide dehydrogenase, which can take CO2 to CO. And much of our focus is really on the act of sites within these enzymes. I want to make it clear that the act of site is where the substrate binds and is converted. But I put to the left, just to make it very clear, these active sites are part of larger protein systems. And so when we really want to understand, we want to know on an atomic level what's happening at this act of site. But also how does the whole protein system integrate to best facilitate these reactions to deliver electrons and protons at the time that we need it. And so there's really a complex system understanding that we need. And the question is, how do we go about achieving this detailed atomic level understanding. And I think there are many different approaches, and I think there's lots of work from groups around the world contributing to these kinds of problems. In my research group, we focus mainly on how X-ray spectroscopy can help us reach these goals. So we use synchrotron facilities, or we did and we could travel to them. Now we're doing some remote beam times. Or we use some of our in-house built instruments where we can do absorption and emission in-house on relatively concentrated samples. And so we do X-ray absorption, as I already said, we do X-ray emission. But we also combine absorption and emission in sort of a two-dimensional experiment to do something called resonant inelastic scattering or resonant X-ray emission spectroscopy. We also use techniques like X-ray magnetic circular dichrosome and nuclear resonance vibrational spectroscopy. But what we want to do is take all these X-ray based techniques and combine them with more standard in-house methods like optical spectroscopy, EPR, Raman, and Moss power in order to really arrive at a more holistic picture. We correlate this to computational studies, and in doing so we start to try to piece out biological reactions and understand what's happening. And so today I'm going to talk about just one of the enzymes that my group works on, but one that's particularly close to my heart and that is nitrogenase. What I'm showing in this picture are the molybdenum dependent nitrogenase. So as I already said, a protein, it is a system. Nothing happens all in one place. And so here you can see just a schematic of some of the complex machinery that's involved in biological nitrogen reduction. Now the way this happens is you require minimally these two proteins. One is called the iron protein. The iron protein is where two molecules of ATP actually bind. They're hydrolyzed to ATP and they provide the energy to basically drive this reaction. And what happens just in a very simplified picture is that electrons come from the iron protein. So from this four iron four sulfur containing iron protein in the upper right of my slide. My apologies, my pointer is not working. And then it's transferred to the protein below it. This is the so-called Mofi protein, which is 250 kilodaltons. It has in it something called the P cluster that you can see labeled there. It has eight irons all bridged by sulfur. And from there, electrons are transferred to Fimo Co, which is the site where nitrogen is converted to ammonia. Now what I want to point out, and we'll get back to this later, is when you look at the complexity of this, you see there's a lot of iron in this protein. There's a lot of sulfur in this protein, not just in the active sites, but also in the cysteines and the thionines. And so this is one of the big challenges is how do you get selective enough to actually see what you need to see and understand how the protein is changing. And this has been a huge challenge in nitrogenase research. But because those drawings on the previous slide were a little bit small, I want to zoom in and let you look a little bit closer at the Mofi protein. So this is this 250 kilodalton protein I told you about it. And inside are these two active sites we talked about. On the right, we have the P cluster. This is this eight iron cluster that's bridged by sulfur. Its job is just to shuttle electrons. Okay, it just does electron transfer, it doesn't do catalysis. On the left you see Fimoco, and this is the site of catalysis where N2 was converted to ammonia. It has seven irons. In blue you see the capping molybdenum, and in black in the middle is a very unusual musics carbide. Now what's interesting is that all of these components actually have genes that express specific proteins that actually are involved in this assembly. So I'm not going to talk about it today, but these days, thanks to some of the really beautiful work of Yilan Hu and Marcus Ruby, we understand a lot about how these clusters are assembled. And it also tells us that nature knows something. It knows that it has to convert the ordinary iron sulfur cluster on the right to something more complex in order to carry out more challenging chemistry. And these are the questions that really fascinate us. What modulations occur in the electronic structure on going from right to left that enable this new activity. So ultimately I told you that we need eight electrons and eight protons to produce two molecules of ammonia. We're going to get back to that later because for those of you who are paying attention, you probably already noticed there's extra electrons and protons there. And those extra electrons and protons are actually thought to be part of the mechanism. Part of the mechanism is the obligate release of H2, and we'll get back to that later. But what we'd like to know in principle is that every step from the ground state, which we call E0, to each addition of electron and proton. How is this active site transform? What is the atomic composition at every state? What's the charge distribution, the oxidation state distribution, and the magnetic coupling. And so what I'd like to talk to you about today is the role that x-ray spectroscopy has played and will hopefully continue to play in us understanding the complex electronic structure of this co-factor. And so at the start what I want to do is just take you back in time a little bit. Let's go back to some of the earliest structures of nitrogenase. So this structure actually for this co-factor on the protein first came out in 1992. I was actually an undergraduate at the time. I'm finishing up my bachelor's studies in chemistry. And I remember this was a really exciting structure because it did something that all of our chemistry textbooks told us didn't occur. The irons in this 1992 crystal structure, if you look at that on the left, they're all three coordinate. And at the time we would have taught people that iron had to be minimally four coordinate. So this crystal structure actually inspired synthetic chemists to produce a low coordinate iron complexes. It opened up a whole new field. Interestingly though, later crystal structures showed there was density in the middle. And in 2002, although there was density in the middle, we didn't know if it was carbon, nitrogen, or oxygen. And I won't read through this whole slide, but I just want to emphasize that there were many spectroscopists and computational chemists trying to understand what this central atom could be, but they couldn't pin it down. And at the time I was a starting assistant professor on that Cornell, and having come from being a beamline scientist and having a synchrotron background, I was wondering if the new developments with these high resolution emission setups are seen in many, many places now. They'll soon be one. I'm at the Balder beamline at Max 4 that's coming pretty far along. Could these spectrometers actually help us solve this puzzle? Now, I want to emphasize this puzzle is particularly challenging for a protein because we have so many light atoms. We have so many carbons, oxygens, nitrogens. I'm not even including all the structural water. And so when we try to talk about what is that central atom, we needed to characterize one out of thousands and thousands of atoms. And so why valence emission? Why did we think this was a good idea? So, just a reminder, I know you've had some other webinars already on X-ray spectroscopy, but X-ray emission spectroscopy, like most X-ray spectroscopy is element selective. And so what we're doing is we're tuning to an energy where the probability of ionizing an iron 1S electron is very high, okay? And we ionize that to the continuum. And then we follow all the processes that occur afterwards. That could be a 2P to 1S emission event that generates the K-alpha lines. It could be a 3P to 1S event that generates the K-beta lines. Or it could be the very unlikely event that's boxed there in red and blown up so you can see it. The idea that an electron in a filled ligand orbital could refill a 1S pore hole on the iron. Now, why does this matter? It matters because the valence ionization energies of carbon, nitrogen, and oxygen are incredibly different. In fact, their 2S ionization energies differ by nearly 8 electron volts or so. And so it becomes much more sensitive because you have an element selective way. It says this light atom has to be coordinated to the iron. And then the energy of these valence ionization features are going to tell me something about its identity. And so this is something we did now, almost a decade ago, was to first demonstrate the utility of valence to core X-ray emission to identify the central carbide. And so what you can see in the left panel in the inset are these valence ligand 2S to metal 1S transitions. You can see where it appears for iron oxide in red at relatively low energy. It goes up by several electron volts to iron nitride, but it goes up even more quite significantly for the Mophie protein. And what this told us empirically, this very high K-beta double prime is what this feature is called, it told us very likely that this was a carbide. And it was the first example of a carbide in biology, actually so far the only, although it's found another nitrogenesis as well. But this observation is only empirical. And so what we wanted to do was to also correlate it to calculations. And to best highlight what's happening in the calculations, we actually subtracted Fimo Co from the P cluster because remember the P cluster was only iron sulfur and Fimo Co had the central carbide. So to emphasize the contribution of that central atom, we did the subtraction and on the right on the top, you see the different spectrum in blue from experiment. Then in the computer, we can calculate whatever we're looking for, we could put Fimo Co with a carbon with a nitrogen or an oxygen. But what you can clearly see is that the blue spectrum where Fimo Co had a carbon in the middle is clearly in the best agreement with our experimental data and neither nitrogen or oxygen fit. So this kind of confirms what we thought just empirically, right. And so, I won't go through all of this because it's more fun to talk about more recent papers, but this is sort of many of the synchrotron studies that followed really focused on trying to figure out the electronic structure of the testing state of this E0 form. So we've done a lot of experiments like molybdenum and iron-herft XAS. We've correlated that with Mossbauer and molybdenum Allegis. And we've also done different kinds of XMCD pictures to arrive at the fact that Fimo Co is best described as an MO Fe7 S9 carbide one minus cluster. Now some people who maybe aren't into details of electronic structure say, well, why do I care like what the exact charge of this cluster is? But remember, we have to put eight electrons in to get this cycle to complete. And if I'm a computational chemist trying to calculate a mechanism, and I have no idea what charge state I'm starting with, this is pretty challenging. And when we started this, the charge state varied over four units of total charge. And so now we feel we've pinned this down, and we at least know where our mechanism is starting. And this was really important for us to reach this point. There's something that I want you to look at in this drawing that you might find a little bit strange, especially for the chemists in the audience. If you look at the molybdenum three, there's something really funny about it. It has an electron configuration that goes down, down, up. So something that we teach doesn't happen or we call it a non-honed electron configuration. And we were very interested in trying to understand why this happens. And based on computer studies, we believe that this is stabilized by the interaction between the iron and the molybdenums. There's an actual bonding interaction that helps stabilize this. But we wanted an experiment that actually showed us this wasn't normal molybdenum three. And so the experiment we did was molybdenum allege XMCD. So what you see in the upper left panel of this slide are molybdenum allege XMCD data. So this is a 2P to 4D transition. And what you can see is FIMOCO in blue looks very similar to what is labeled MOFE3S4. This is a synthetic cubane that in all of our studies looks very much like FIMOCO and it has those same molybdenum iron bonds that I just talked about. And we think it also has this non-honed configuration. And then we looked at a normal molybdenum three, this molybdenum TTCNCl3, which is a completely normal three halves. Now if you look at just the allege's on the top panel, they all look pretty similar. They all have a maximum about the same energy. And that's all consistent with molybdenum three, but it doesn't tell me anything about the configuration of the spin. Is it three up or is it two down and one up like I drew on the previous slide. So to get insight into the spin configuration, we did X-ray magnetic circular dichroism where we used the difference between left and right circularly polarized X-rays. And now you can see the red and black spectrum, which looked quite similar before, now actually look quite different in the bottom panel. The bottom left is the experimental XMCD. And then on the right I'm showing calculations. And what you can see is the calculations clearly show that when molybdenum interacts with iron, in this bonding-like situation you can stabilize this local one-half state. And this was something quite surprising and we found a really nice way to use XMCD spectroscopy to detangle complex couplings. Maybe before I go on to the next part of this, were there any questions so far? I can't actually see the chat screen at all. Okay, I think you can maybe continue. We are receiving some questions, but they always come at the very end of the talk, according to our experience. Okay, that's completely fine. I'll keep going. Okay, I just wanted to, when I always wants to find out that you haven't like disconnected and I'm, you know, talking to myself. Yes, yes. So let's continue. What I've really given you so far is mostly background on how we've really tried to deconvolute the complexity of the electronic structure in the E0 state. And one of the challenges about going to this protein under turnover conditions, as I already tried to emphasize before, is we have a large number of irons and a large number of sulfurs. We actually have 72 sulfurs in this protein. And so one of the things that's been important is to ask the question, can we get greater selectivity? So for instance, can we substitute selectively selenium for sulfur in order to maybe use selenium based spectroscopy as a unique probe? And this is something we did in collaboration with Doug Reese's group at Caltech, with a very talented former PhD student of his, Renee Arias, and together with a postdoc in my group, Justin Hensorn. And this was a real tour de force effort from both Justin and Renee, who I think made a really beautiful story out of this. So let me tell you the background on this. One of the things that Doug and his group had shown previously is that when you add potassium selenium cyanide to nitrogenase and put it under turnover with protons, you basically see that the selenium selectively goes into the 2B position shown there in green in the middle panel. And what's interesting is this selenium substituted, this Se2B, still has an EPR that is similar to the resting state. And importantly, it still shows the same activity. So it can turn over with nitrogen. But when it does that, it can turn over with nitrogen or acetylene. And this produces this variant that's shown on the far upper right, where that one selenium has now distributed roughly 50-50 between the 3A and the 5A positions. So it suggests that there's sort of a dynamics in this cofactor during the course of catalysis. And this resting form is still a 3S equals 3S EPR signal. And so it seems to suggest the electronic structure isn't significantly perturbed by the introduction of selenium. And the activity is not perturbed at all. The other thing that's kind of cool about this then is it allows us to use the selenium as a probe sort of to crawl around this cofactor and to say, what does the cofactor look like from the perspective of selenium 2B when the selenium interacts with iron II and VI versus what does the cofactor look like from the perspective of selenium 3A and 5A. Okay. And then we can also inhibit this cofactor with CO. This generates a different EPR signal and S equals one-half that is characteristic of what's known in the literature of low CO FIMOCO. But what's cool is that the selenium stays in the 3A and 5A positions. So when we compare the cofactor above and below, this allows us to actually see how the electronic structure is perturbed upon CO binding. And because CO only binds to more reduced states, it effectively allows us to characterize a reduced substrate bound intermediate of FIMOCO. So how are we going to do this? The way we did this is to use selenium herft X-ray absorption spectroscopy. And so here, if you looked at fully reduced selenide, which is shown in the bottom left in black, you would see basically you have a relatively featureless edge, and this corresponds to the selenium 1S to selenium 5P transition. In herft, this is basically just a cut of a rick's plane. What you see is spectral sharpening. And so maybe I just bump ahead. This would be the rick's plane that we're actually cutting from. So we're detecting selenium K-alpha. You can see there's no off diagonals in this rick's plane. So this means that taking a herft cut through this plane should give us spectral sharpening without losing any information. Now, the reason we want that spectral sharpening is because selenium is relatively high energy. It's about 12KV. And so we see a lot of coral lifetime broadening in this spectra. Now, the selenium spectra are interesting to us because although in reduced selenide there are no pre-edge features, once selenium is bound to an open shell metal like iron, you start to see pre-edge features. And what I'm showing you on the slide are calculated pre-edges where the selenium is bound to two ferric irons. So two iron threes are shown in red. And then I'm showing you what progressively happens as we reduce those irons. And so as we populate the d-orbitals of iron, the intensity goes down because there's less whole character. The energy of the pre-edge can also be adjusted to the selenium 1S to 5P energy. And this allows us to see how the iron 3D manifold is moving relative to the selenium. And so what's effectively happening is as I increase the effect of charge on the iron, the pre-edge comes down in energy because the effective charge has increased and the selenium 1S to pre-edge transition goes down in energy. So there are two factors, both intensity and energy, that will tell me something about how the iron electronic structure is changing. Now in order to do this in a quantitative and rigorous way, we use model complexes. And so what I'm showing you here are the models that Justin made. And in the middle panel, you have this double cubane with a single selenium. And then Justin can oxidize that by either one or two electrons. And what I'm showing you here are the pre-edges. And so you can see what happens upon oxidation. We go from black to blue to red. Or the pre-edge is increased in intensity as it becomes more oxidized. And we can also then align the edges like we did in our computer experiment. And you can see as it becomes more oxidized, the pre-edge is moving somewhat down in energy. Now these effects aren't huge. The intensity modulations are relatively modest. The changes in energy aren't that dramatic, but maybe that's not surprising. It's a large cluster with many irons. We expect this effect to be delocalized, but it certainly demonstrates the utility of selenium-herft as a spectator probe for the cofactor charge. So what Justin did next many, many times, this is now the distilled version, he looked at what Fimoca looks like from the selenium-herft when the selenium sits at the 2B position in blue, or when the selenium is distributed between the 3A and 5A positions, which is shown in red. And what you can see is that these edges are dramatically different. I don't have to shift for you or anything. What you can see is from the perspective of the 2B, the irons look much more oxidized, but from the perspective of the 3A and 5A, the irons look much more reduced. And this is kind of a surprising finding because it suggests a much more localized electronic structure than we typically think of for iron sulfur clusters. So why might this be happening? Well, here it's interesting to just take a closer look at the crystal structure. And what we see from the crystal structure is that there are electrostatic and hydrogen bonding interactions at the 3A and 5A positions that likely localize the charge density to that iron phase. There's only a single hydrogen bonding interaction in contrast to the 2B position, and so we think that favors locally more localized oxidized site at the 2,6 edge relative to the other iron phase. But this is kind of cool because it suggests these hydrogen bonding and electrostatic interactions may actually be tuning the protein environment so that one site is kind of activated for substrate binding. Okay, so let's take this step further and look what happens when we actually have a CO bound intermediate. Now I already told you to bind CO, we have to reduce it. So this is a reduced substrate bound intermediate. But what's happened to the pre-edge? It's gone up in intensity. So recall that means we're seeing irons that effectively look more oxidized. Now why might this happen? Well, what we think is happening is for CO to bind, you may actually have redox reorganization in the cluster. Because if CO binds on this iron 2,6 phase, then you need to actually have more reduced irons because CO likes to back bond at that phase. So this actually might be giving us a hint as to what the role of the carbon is. The carbon may actually allow more facile redox reorganization as it's needed. Okay. And so this is something that we're pretty excited about. But let's take it one step further in the last part of my talk and talk about really characterizing a native B1 state. Before I do that, I want to acknowledge the whole team that was involved in this work. So, at the CEC, this was the PhD work of a very talented PhD student who's now doing a wrap up post-op with me. I'm very fortunate he's agreed to stay longer at Casey Van Stoppen. Laura DeCounts, who is the group leader who leads our nitrogenase efforts in our biochemistry labs, as well as Ragnar Bjornsson and Albert Thorhelsen on the computational side. And this work was done as a collaborative effort with the group of Brian Hoffman and his group member Roman Davidoff, I'm at Northwestern, as well as the group of Lance Seafields at Utah. Okay. So you might be asking, as my group spent more than a decade trying to understand E0, why didn't we study other intermediates sooner? And I basically put this diagram up to point that this is a very challenging problem to try to isolate intermediates in nitrogenase. And I've shown here only a sub-partial cycle here. And so you already saw before, there are more E states, but this is just a partial E0 to E4 state. And this is what happens that in the absence of nitrogen, we can only populate intermediates E0 through E4. Now this cycle is driven by rapid electron and proton transfer. And remember those electrons are coming from the iron protein. So basically, in the actual biological system, there's a large excess of iron protein to Mophie protein in order to drive this electron transfer rapidly forward. Okay. And so the ratio of different E states that you actually get when you do this experiment in the lab depends on the ratio of Mophie protein to iron protein. Okay. And so we can vary that ratio in order to control the amount of intermediates that we actually get. So if we want to populate the early E intermediates, what we want to do is have very low electron flux. That means we want to have a lot more Mophie protein than iron protein. And what we find is if we go to ratios of about 50 to 1, we can populate a fair amount of E1. So how do we follow this? Before we go to the synchrotron, we monitor this by EPR. So I already alluded to the fact that E0 has a characteristic S equals 3 halves EPR signal. And if I put an electron in that by definition means I'll go to something that is integer spin and at least a normal EPR, it's going to be silent. We think it's probably an S equals 2 state. And as I reduce it, you see the EPR signal that was purple for the 100% E0. It goes down to about 50% of that signal. And we know that if we went too far and started populating E2, we would have a new S equals 3 house signal with a unique set of G values because that doesn't appear. So we've been in low enough Eflux to only populate these two states, and we can use EPR spin quantitation to get those ratios. And the later we'll use that on our XIS to deconvolute pure spectrum. Okay. Now I want to point out there's more than one way to reduce what I showed you on the previous slide is what we call a native reduction. That means you use the protein native reductase. And one of the more popular ways to study nitrogenase intermediates has been to do cryo reduction. And this means you basically just bombard it with radiolytically generated electrons. So for those of you who've been to synchrotron beams and seen beam reduction, this is kind of a similar idea. And in fact, some of the early E1 states actually were generated at synchrotrons. The reason people are interested in this though is that at least in nitrogenase, it's been shown that this is a means to decouple the electron and proton transfer steps, which in biology you can't do. In cryo reduction, they're related by annealing. So you cryo reduce at 77 Kelvin, anneal up to 200 Kelvin, and it's found that at 200 Kelvin, that's when you see proton motion, at least this is what is believed. But let me talk to you before we get to the XAS about what's known about the these cryo reduce states that are that are EPR silent. So one of the best insights that has come to these clusters is through Mossbauer and what's beautiful about Mossbauer is it's sensitive only to 57 iron. So what you can actually do is, you can extract this Fimo co cluster from nitrogenase 57 iron label, and then put it back in the 57 iron labeling actually happens on on the level of the actual cell cultures. But in any case, it gives you a way to have a unique label on it and see only Fimo co and not the other iron. Now what's interesting though if you compare native reduction and look at how the, the isomer shifts changes so this is what's under native reduction and end to Mr. You see the change in isomer shifts are relatively small on the order of 0.02 to 0.03 millimeters per second. So that's when you look at cryo reduction. So this is MN to the alternative MI or cryo days state. You see that the change in isomer shifts actually double. And this actually led the authors in this 2000 Jack's paper to conclude that perhaps native reduction occurs at the molybdenum and cryo reduction occurs at the iron. And this would be really important because it would indicate that you can't use cryo reduction to understand the mechanism and this is something we also wanted to investigate. So the first people to study E1 by excess. In fact, 25 years ago, the group of Steve Kramer reported excess on the E1 state. What was somewhat surprising about their excess, however, is that they actually showed the militant oxygen distance is actually contracted. So this wouldn't imply reduction at molybdenum, but a contraction 0.06 to 0.07 angstroms likely looks more like oxidation. The same, however, were reported in that paper to assess whether or not it looked oxidized. And we nevertheless thought this warranted three investigating the excess as well. So what is the question we're asking, we really want to know where does the first electron go when we're thinking about the mechanism, does it go to the molybdenum, or does it go to the iron, and where does the proton go, does it perhaps actually form a hydride which is thought to be mechanistically or does it maybe protonate a bridge and can we tell. So, we'd also like to understand whether or not cryo reduction is really related to native reduction simply through annealing. This was something that the initial Mossbauer paper didn't look at what happened to the Mossbauer when you annealed. So this was something we also wanted to go back to and really figure this out. Now, one of the places we thought we had the absolute best handle was to use molybdenum high energy resolution fluorescence detected X-ray absorption. So again, like the selenium perf we're now going to use molybdenum perf. So we only have one unique molybdenum. And so if any redox is happening at the molybdenum, these edges should be sensitive. And so I just want to point out here, I'm showing Mofi protein in black where the molybdenum is molybdenum three. I'm showing some molybdenum three model complexes in red and green. And then I'm showing a molybdenum four in blue. And of course, as we oxidize them, you would expect the edge to go up. And I just want to show that even with these relatively large core whole lifetime broppings at molybdenum, you still expect to see significant changes. So again, only one molybdenum. So if something's happening there, we should see it. So what did we see? Here's the molybdenum perf of E0 in purple, E1 in green, and the various cryo reduced states. I'm just showing different doses. And then also what happens after annealing at 200 kelvin. And you can see within error, there's nothing happening in molybdenum. So certainly nothing to say a full redox event occurs in molybdenum. Now what happens when we instead look at the iron. Here again, everything looks very similar. But remember, unlike the Mossbauer, where I could iron 57 label and just look at the homoco. In this case, I have 15 irons that are all contributing in this mofy spectrum. And so I have to actually look at different spectra subtract E0 from E1. And I'm showing error bars here just to show that we got good enough signal to noise in both E0 and E1 spectra to believe these differences. And what you see is that E0 and the cryo reduced really are different. These are the green for the natively reduced and the black for the cryo reduced. But when I anneal it, you get the red spectrum, which shows that it's moving back towards something that looks a bit more like the native reduced. And so this at least provides some confirmation that it appears reasonable that the native and cryo reduced are actually related by annealing, and that in fact, no redox chemistry is happening at molybdenum. But we wanted to go back and revisit the Mossbauer as well. And so we actually contacted the authors from the 2000 paper. And believe it or not, they actually had their original Mossbauer samples still stored in liquid nitrogen. We measured them and reproduced values very close to their original paper. And then we did the one key experiment that wasn't done we annealed to 200 Kelvin and what you can see if you look at the table, the bottom line. Of course E0 subtracted from E0 gives a delta in the isomer shift of zero. But for the cryo reduced we saw a larger change in 0.07. But when we annealed it, we saw a change of only 0.02. So it was identical then in terms of the change in the isomer shift to the natively reduced. So we've confirmed the electron doesn't go to the molybdenum. It's clearly going to the iron but affecting only a modest change in the isomer shifts. So the next question is, can we tell where the proton goes? And here we ask an exhaust and QMM help us find this out. And so here I'll just go through this somewhat briefly. These are the molybdenum k-edge exhausts of E0 and E1. And recall I told you already that we get at most 50% E1. So the bottom spectral that's labeled just E1, this comes from doing a subtraction. And we do this just because we're looking to see trends. That trend should increase even more if we look at the subtracted data. And what I'm showing in the right is the difference between the mixture in E0 versus our pure E1 and E0. And we don't see any particularly pronounced trends on going from black to red except perhaps the slight change in the molybdenum iron component. But there was no evidence for significant structural changes of molybdenum. In contrast to iron, you can already see it if you look at the Fourier transform on the bottom left. There are two clear features in E0. They start to coalesce in the mixture. And by the time we get to the pure E1, they're even closer together. And if you look at the distance analysis, this is coming from the iron sulfur distances elongating slightly, while the short iron-iron distances within the cubane are contracting slightly. So these two result in the Fourier transform coalescing ever so slightly. So even though it's only a one electron effect, and there are many irons, the excess still evidence this change. Now, the next question is, can we take our excess structure correlated to QMM studies and get an idea of where the proton is going on that basis? And so this is the 1000 atom active region that we used with 133 atom quantum mechanics region. And I just want to point out that we looked at many different program symmetry solutions and also we modeled protonation in different ways. So what Albert did was to look at protonations that could happen at the 3A, 5A or 2B position. He also looked at remote protonations from the central sulfide bridges. He looked at the formation of possible hydrides or even possibly protonating the central carbide, which has been proposed. And in terms of the surrounding amino acids, he also looked at different protonations, for instance, in this histidine either at the delta or epsilon position and how that could affect the structures. So what we're showing here is our evaluation of the QMM models relative to the average distances we got from excess. And so this is just an RMSD. And you can see for most of the RMSD is there between sort of 0.015 and 0.025. The only real outlier we found structurally was when you protonate that central carbide that structure seemed way off. But in addition to looking at the structures, we also evaluated the total energies. And we found with the histidine, if we protonated the delta position, we certainly favored than protonating the 2B. If we protonated the epsilon position, we saw that we couldn't really distinguish either putting a proton on the 2B or the 5A was equally favorable. So based on this, let me just summarize. So on the right panel is the summary for this part. What we saw from looking at the E1 studies is that reduction of E1 occurs at iron. It actually does occur on the side of the cubane closer to molybdenum. So this may explain really subtle perturbations of molybdenum. And what we see experiment and theory taken together favor protonation of a belt sulfide, either the 2B or the 5A position. I've also in the first part showed you how selenium substitution gives us a means to get more selective information. It surprisingly shows there's a large degree of electronic asymmetry in FimoCo and contrast to what we know about iron sulfur model complexes. And the selenium hearth also shows that CO bound nitrogenases provide evidence for redox reorganization upon reduction. And so what we're doing next is to look at other E1 states, possibly combine my selenium substitution, although those are really tough experiments. So I have to again applaud all my group members for their tremendous efforts they've put into this. But I hope all in all whether you're interested in biological catalysis or other kinds of catalysis, I hope I've shown you how much advanced spectroscopic approaches are needed to understand complexity really in every area of catalysis. And today I chose to focus on biological catalysis, but I want to just give out a shout out to one of my recent PhD students, Abbas Mahesh D. Eskari, who just last week got the cover of ACS catalysis for some of his work on heterogeneous catalysis using in situ and stick some nano reactors. So if you're interested in more heterogeneous, I wanted to advertise that and finally I just wanted to thank all of my group members, I've tried to acknowledge them throughout the course of the talk. We have wonderful collaborators that I also tried to mention, and all the synchrotron facilities that gave us beam time, and you for your attention. And I'm happy to answer questions. Thank you very much, Serena, it was really great talk. So I think, and you're also our first speaker really stay on time. So thank you very much. Excellent. So it's time now to go through the questions. We received some few questions. So maybe I'll know what to start. Yeah, the first question is for hybrid devices comprised of nitrogenesis and photo electrodes. Is it possible to differentiate or understand charge transfer at various interfaces? In a hybrid device that uses the nitrogenase? Yes. Okay, you know, this is an excellent question and it's been a controversial question. So for a long time, I would say people would tell you that you couldn't build a hybrid device with nitrogenases because it always needed its native reductase. But very recently, there have been hybrid devices or hybrid like materials using cadmium sulfide to trigger electron transfer. But it's very hard to evaluate those systems in a controlled way, because although you can photo activate it, it gives an uncontrolled cascade. It's very difficult for us to look at it stepwise. But I'm sure there are other groups also working on this. I know Shelly Manteer's group has made a lot of progress. I'm also getting this into polymer like films. So I think it's on the horizon, but we're not there yet. Thank you. We have another one. Is it possible to determine what type of bond breaks or form during transition from one state to another? I mean, ultimately, this is kind of our dream experiment, right, is to actually watch bond making and breaking. And so we hope that someday we'll be able to watch the NN bond become activated. And we've shown this in model complexes that you can do this with valence to core X-ray emission, but it hasn't been done yet for nitrogenase. So these are hard experiments, but I'm sure they'll be exciting and keep us busy for years to come. Yeah, it would be very interesting to see how this bond breaks and forms. Yeah, I think for many people this would be a major goal. Okay, so we have other questions. So this is more like, yeah, so what's the time scale of the observed changes in between different energy states? Yeah, so it's a little bit hard to give an absolute number because it will depend on how you you tune the electron efflux, right? So that's, you're kind of changing it by tuning parameters, but these are proteins that aren't that fast. So there's many intermediates that are of interest that are generated on the millisecond to even second time scales. The bigger problem we have is reaching a configuration where we accumulate enough of a given intermediate to be able to probe it, right? And that's, that's the challenge. Thank you for the answers. So we continue with the next question. So what are the minimum information we need to from our typical labs before going, before seeing these pieces with x-rays? Yeah, I mean, in the audience, there might be people that are not familiar with this type of technique. So I suppose they are interested in knowing what they need to do before. Yeah, it's a great question to ask before you think about going to a synchrotron. I mean, I think that it will depend a bit on your system, of course, but I think that whenever you have, you know, open shell paramagnetic metals, it's nice to look at EPR when possible for almost anything that's an inorganic system and colored. It's good to verify how things change by UVViz. I think that people understand that different laboratories have access to different levels of infrastructure and information, right? So I think it's good to reach out and collaborate, not just with a synchrotron scientist, but with other spectroscopists that can help you, right? I think that's, it's, these are complex questions that are best tackled by a global community. And I don't think there's a box you can tick off and say, I minimally require this because it's different for every system, right? Yeah. Okay, so we have the last question, which is also the longest. So I would break it in a short question. So, first of all, he says, thanks for a very interesting talk. It's always amazing to see what can be achieved from the synergy of different expertise. So I'm wondering how one of your synchrotron experiments would look like. So what's the state of the aggregation of your sample and its medium? First question. Okay. Maybe you can answer to this question. Yeah, so that's an excellent, an important question for us. For most of our biological samples. They're prepared as solutions. So they might be freeze quenched if it needs to be trapped on a short time scale or if it's just a native resting, we would just freeze it in liquid nitrogen and then it's transferred from liquid nitrogen to a helium cryostat and measured at low temperatures, typically either rastering or moving many spots because of a possible damage. We are looking towards doing flow studies in solutions with these proteins that brings obviously other challenges. Yeah, in fact, the following questions are, do you actively change any sample conditions besides temperature? Also, is radiation damage a problem? So radiation damage is, for anyone who works in the biology realm, it's a nightmare for us. But interestingly, it varies a lot by protein as well. So if you look at something like the photo system, that's probably the nightmare of photo reduction problems. And interestingly nitrogenase we find is less susceptible to reduction. It will eventually change in the beam so we don't dwell. The dwell time will depend on the dose per spot. But we have, I would say, dwell times that are order of magnitude longer than what's possible for photo system.