Photokatalytische CO2‑Reduktion mit Diamantkatalys

Okay.

OK, then let's stop. OK, here was supposed to be a part of my name, but something happened. The presentation, how to fix that. And then I will just talk about my project, and then the results that I had in the last months. So my project is to make the photocatalytic reduction of CO2. Some of the changes that we have today is that this is currently kind of low-efficient. We usually use toxic and critical materials that are hard to find, or toxic, or both of them. And many approaches requires large amount of water, and it's not that good to escape the--

So my project is a direct reduction of CO2 using solar light and to achieve this project is a ferment method or CO and use sustainable and no critical catalyst that you can have in larger scales and not have so much problem with that. And enable the application of the regions. So places that have a lot of solar light are using deserts. are not able to have a lot of water.

So if you can move those to the desert and use less water, you can increase the efficiency of the CO2 reduction. And for that, the project is used a non-seroidal diamond as a catalyst in the ionic liquids, because they can capture CO2, can have a higher CO2 solubility. As well, can capture the water from the humidity from the air. So we could capture the water from the desert have enough water in the air to

use as the DLT lipids. So the diamonds, they have, I think by now, have a negative electric field. That means that when it's in water it can emit solid electrons, that these electrons are responsible to the CO2 reduction. And as I said before, one goal is to use the water from air to make those electrons and then

reduce the CO2 to other problems. So one part that I'm going to talk a lot about today is the CBD optimization that I'm doing. So the diamond that we use are the CBD growth diamond, and also the CBD growth diamonds, that are the BDDs. With the CBD, a lot of things matter. So we always see a substrate, and we grow from these diamond seeds.

Currently I'm using the same as the Amelie Amelie, that's the batch 0062. They are mostly big diamonds before milling, I have not seen the mill diamonds yet. The size of the sitting diamonds matter because when we have a lot of big diamonds, when we grow with the CBD, we first have to fill those gaps before going up and have thicker electrodes. And when we go to smaller lines, we have

to fill a little bit less area because also the surface area here is much bigger so we can fill those gaps quicker than here and this is my little thinner films as we're going to show after that i also working with the sp2 cargo because i draw a stick uh one thing that we can do actually is to increase not increase but introduce new states that you can use to produce the solid letters one of these ways out can be using sp2 carbon that's

we always have in the CBD but you can increase this by increasing the amount of carbon in the reactor I'll also be talking a little bit about nanosurf diamonds that I did not finish but I started learning how to make the nanosurf with the hedgehogs with your hands we did not finish yet but we hope to finish soon and after some time I talked a little bit about sanitation that's one ideas for the project is to add also nanoparticles and utilize the user in the

dyes that can also introduce new states, the band gap, so we can make more sort of electrons. One of the ideas with nanoparticles is that we can have the ability fumes, the structure of the diamond, and we can try to add the particles to different places, and also maybe grow another layer of diamond over, so we would have like this structure of nanoparticles inside, or maybe over the point and also over it.

And so that's some ideas of the project. One thing that I started doing, I think in the middle of May, or around May, that we talked in the subgroup meeting of Cateris, is that is a Python script for the evolution of the random spectrums, because much I did with QX QTR plot, and it was pretty hard worked with QTR plot. Then I made a Python script that works basically this way, You have the configuring settings.

and select our data. And these are the only parts that we have to do something that the script do, itself. So it takes roughly about five minutes to set up and run script. If you have set up as I have for my samples, it takes less than one minute to have the . All libraries are open source. If you want to learn some how I can see, by opening it. Then,

I wanted to use these two equations, the Galson and the Fevon. I can add more if necessary, but I think this is a lot of stuff. And I also do a manual how to use it. I have some stuff in the presentation, but I forgot to do the proper manual, a software on how to use it. But it's on the server, so we can take a look. The configure settings are pretty simple. We have a simple name, the labels, because it It's not the best image, but you can

So here's some setup for the image, some setups for the parameters of the population. All the export files are exported with the sample name plus the information. So it would be like BR06B as parameters, then a CSV or an image and - And then, oh.

or curves, that curves have all the curves that you plot in a problem or stuff you can also do that. Then there's a little bit of setup that's the number of peaks that's going to be convoluted, the range that we're working. I would say that the hardest part that's not that hard is that you have to set up the initial values for the evolution, and the script works in a way that took me some time to realize which was the best. So we have these two equations as we don't

any place that we choose which equation you use we choose by the initial values so whenever that we put four initial values it choose the fenn equation because we have like four variables and when you input three values it goes to the galaxy equation that has three variables so i tried to take the make the setup the smallest possible and this was one way to do that and we also have to label our peaks that we just label in the order they appear here

And right now it's set up to seven, making seven peaks, but you can also just enter more because there's no limit. I don't know how computation takes more than seven. I just tried with five, but with five, it takes less than one minute, and it's fine. Then we have to input limits, so they have gotten the script don't go crazy for the collusion that sometimes happens. But you can see the leaves are pretty much very broad.

to zero from 1,000 and they actually use just 0.8. Just so some values that are not supposed to be negative goes to negative values and sometimes if I'm not linked to 1,000 to go to 1 billion something that doesn't make sense. So it's just safety measures to go get like great values. Can I ask something? So here was it possible to treat B1 and B2 as the funnel? Yeah.

Yeah. This is when you use the four variable equations. Oh, sorry. Let me stop. And then read the script. When you read it, it will modify the explorer, and then you can select all your files. This regard is very important. I recommend you have a couple of the files, because when I was in the script, there was a book that overwrote all of the CSV files with some image, so it was not like in the readable feed, but just like a lot of random characters. So this is like, it's corrected. I don't know if it would have another book. So I will. Better work on a copy. Yes. Good idea.

After that we will open another file explorer and select the folder that the data will be exported. We have these three files, that's one figure, two sets in files, one of the parameters, one is to plug the curves in another software. So, what's basically this, I will also show how it looks like in my run spectra. Also, as I said, the

size of the sitting part, sitting suspension, spartments and spending matters. The one that I use for the show today, I did this one here that's about 110 nanometers with some of them being 35. So I did some centrifugation of the original batch because I don't forget to bring here but my other presentation might remember that we had a lot of stuff around Montalvo so it was not that good to use those ones. I basically did some slow centrifugation for 10 minutes, took the super latent,

use, then I concerted with the hot preparation, and I determined the concentration with hitting one milliliter of the sample because the Frisdar was not working when I did this, but I did it two times, the same procedure, and got basically the same results, so I'm using currently this one because I have a little bit more. So for the CVD I had this initial plan, that's 12 samples. I did 10 of them because as written here, around 21 degree Celsius I felt that

The cooling trap that we have in the CVD was almost going to the pump, the carbon pump. And we have just one cooling trap on two there. So I thought for the one with higher temperature, the 23 and 27, I have to use two cooling traps to make sure that we do not damage the pump. But my initial plan is basically changing just the temperature in the bubbler. I think I have So this is the setup of our CBD.

our bottom source is the 10mgT in a bubbler that's controlled by a thermostat. And by changing the temperature in the bubbler we increase or decrease the bottom to carbon ratio in the gas phase. So my idea was to change about 20,000 ppm in each of them and also change the flow of methane. So here we have 1% of methane in the gas phase, Here you have 2% of methane in gas phase. And this is the amount of carbon

I doubled to try to increase the SP2 concentration in the BDDs. So it's a pretty high range of temperature. Going to almost a little bit less than 10 times the change of the bottom carbon ratio. And I noticed that we can see how much the TNBT, because they also have carbon in the stream,

carbon in each molecule. Also as a secondary carbon source for our system. So if you look in the high temperatures like 20 degrees, we have 37% of carbon coming from CNBT in the gas phase and then go to the lower temperatures. So we also have very small contribution of carbon from the 10 BT. And this might matter because with lower temperatures, we might also expect maybe

thinner fumes because it would also decrease the amount of carbon about 130% less. Maybe even change the sp2 to sp3 ratio, right? So that you have a difference in how much non-diamond carbon is formed. That could also be influenced by that. Because the carbon from the TNBT, the decomposition in the plasma is different than from methane. So it might actually exist in larger clusters and so this might actually affect the sp2 ratio. So we will see from the Raman tab. Yeah. Yeah.

What I realized here is that actually when you increase the carbon concentration, not for your old experiments, the bubbler temperature was kept constant, so you have two reviables in some of the experiments. So I think it would be better to keep the bubbler temperature as a constant when you increase the carbon concentration so that you have a clearer comparison. I also like to compare the amount of water in the death space, like the

for 36,000, about 36,000. It's not exactly the same, but-- But closing. --the important part, it's close. But there might be a difference, because the . Just to talk a little bit about the environment, I have this-- I'll talk a little bit more about this, but in the environment of the BADs, we have four. And we'll see that sometimes it has five bands. That's the B1. That's the rate to the bottom. The B2, that's the rate to the diamond, but also to the bottom, because this is a diamond for being raised.

the diamond peak and the G-band. And as I said, some samples might be very thinner to the side of the particles. Sometimes I found the silicon, because I grow my sample on silicon. And maybe half of the samples, we that is signal, signal.

But that means that they have somewhere a hole in the film, right? Or is the Boran film so thin that it shines through? It shines through. Okay. So it's very thin then. Yeah. Yeah. Growth time was four hours? Four hours, yeah. So here is one of the places I compared the taper to fixing the bubbler. So fixing 7% but with 1% of the thing and 2% of the thing.

you can see that as the edges so in the edges we usually have less borders open that's pretty common even in the literature and they're quite thinner and you can see that this happens with the sample so here it's pretty much as B1 B2 and the diamond so here's B1 silicon B2 diamond you see the

where it shoots the diamond peak, it's about the same intensity. And here, the diamond peak becomes way less intense than those ones. So we have different, maybe in the doping on the amount or the size of the film. And when it fixes the kind of fix the bar to carbon saturation, we also have something like that. And I forgot to do this in the last one, but you can see the difference in the G-band as well.

because this one with 1% of the jbane is higher in the low, a little bit above the size of the baseline, and with the 2% it's more permanent, meaning that we have more sp2 carbon. And comparing like this one with 1% of methane, I also have the silicon, when we go to the 2% of methane, we don't have silicon. And this is the thing that happens with all temples, all of those that have silicon with 1% don't have 2% because

change the thickness these are way less bright as well so they reflect very less light the percent can i ask something about this um so maybe for the the convolution here you might want to increase the number of scans for your experiment so that you decrease the level of noise because for the two percent i'm not sure if you tried to fit the silicon peak already or was it not just included but maybe there is something hidden there so

So I think if you have less noisy effects, it's going to be a clearer deconvolution. Just to make sure. I don't know for this one, but I know that whenever I do this revolution, I do the R-square is around 99%. The ones that we have to add to the system always go a little bit under, so that age? It's normal. Yeah, because maybe it's harder to fit. Because the contribution of that peak is so small, and then the fitting with those very unbalanced values that you have to do generates usually R-values a bit less.

spectacular than without such parasitic peaks. I mean, it's a parasitic peak that sits on top of a larger peak and that makes it a bit difficult to fit it. But through the noise in the spectra, especially at the lower values, could probably be a bit improved. On the other hand, I think this fitting works really, really nicely, especially it saves us from using stupid QTI plot. And it does it better and faster. And then you can put the data in whatever

presentation software you want and then you can make really beautiful looking spectra also for publication fittings that we want to do because it seems quite versatile how you set it up So this is I think a major step forward for many of us needing to fit multi-peak thingies so Now I actually have one idea so in our revision of the carbon paper they wanted us to also determine the boron concentration from the raman

And Maciej found some equations and he applied and then we obtained the same concentration, basically the same order of magnitude as we obtained from the SIMS measurements and the and so on. So it worked pretty well. So I think what we can do is also include that in our script so that we have the boron concentration. Right away. And now we have some benchmarks. Maybe we do a few more samples with SIMS also and then we have like this and this signal corresponds to this so that we have like this benchmarking and then we can maybe because the Raman is much quicker than anything else we could at least have a

regular look at our diagnostics of the boron content and only if something looks weird we then go for the more demanding mochotki I mean you need to do a lot of impedance measurements for that and so once you know how to do it it's okay but still it's time-consuming the raman is straightforward and then you put it in your Python script and then it's good and sims is only when we ask somebody who has a SIMS. So this is the most difficult to get.

So that would be a good idea to include this so that we have a permanent diagnostics of the boron content if something goes wrong. Then we can maybe even use this to monitor the well-being of our CVD machine because the incorporation of the boron will differ when there are, for instance, some deposits or we have fluctuations in the vacuum or whatever. So it could actually help us to keep track on

And we monitor this that with the same condition, the boron content in the sample itself goes somewhere, be it higher or lower, then you could actually see, oh, I think it's time for cleaning. Well, we can see it. We can see that, but maybe we can see it a bit earlier than when we have diagnostics. Yeah, very nice. So, and we see more SB2 as expected when we increase the methane concentration, so this is...

I think not a surprise. And somebody decided we want sunlight now. I don't know why. Maybe we can say no to this. But maybe the other one is the best. Yeah. So I'm Lankan. Yeah, very nice. Thank you, Bernardo, for doing this. It's going to save us a lot of time. Yeah, it's very helpful for the entire group. Yeah. Then I start to work with the diamond. It takes a little color.

the center and the asymmetric value. So the asymmetric value actually means that when the absolute value of the film is closer to zero, that's in case it has more dopamine. And when you see that the edges always have the center, it always has more absolute values than the edges. So in the center, it always has more dopamine than the edges of the film. film also we can see that we have some shifts in the

the center of the peak, the biggest one's been about 10 And I tried to make a calibration curve that's not working, that it's not even fit yet. So this is the center. And you can see that maybe here we have some tendance to go up, but it's not fixed. Not very pronounced. There's no real trend. No real trend. So that maybe because the suspension it's very small and have a lot of difference inside of each of these samples. Or maybe--

is simple linear fit is not the best way, but you just use the equation that you set, it would be probably better to do that. - Maybe you can try this on the same set of sample and see what it gives, and then maybe compare this. And this could be quite interesting and also give us more knowledge about our process. So maybe use the equation that Matthe applied for the BDD paper, and then see if this gives you a better

or if the changes are so subtle that you don't see them, that can also be, right? So that maybe we need to ramp up the differences a bit to actually see a trend. That we are here in the region where the variations are within the natural variations. - Electrochemistry also would be a nice way to see. - Have a look, yeah. - So the rest of the presentation is something that I present to Professor and I don't know all the way from you as well, and also to you did a lot.

So I'll talk a little bit about the CA, that's principal component analysis. That today it's very on type because of AI. But it was actually a quite old tool from the last center that helped us simplify complex data sets to single variables trying to use the data. The way that we do that is the PCA identified some directions in our data that have more than various. So here you have various in this way. And this is not the best one, but also have this one. And this is blue.

that's not looked like an arrow. What does it rotate the coordinates to like the direction of matching variance as it is like this. So the matching variance here also align and copulates the angle of this difference. And with that we can reduce the dimensionality of our data set. So whenever we have a lot of data, so around have like let's say 1,000 data, I'm going to shift the values and we have done samples.

have about 10,000 points and this try to reduce those 10,000 points to some more human visible and human that you can like really see and try to work with that but there's a lot of advantages but have some disadvantages that the PCs so the values that don't have a real world meaning it all assumes linear relationship between the features and it's very sensitive to outliers and scaling

So, due to the scaling, we have to do some preprocess in the REMA data. So I do the normalization of the data that I do already for the evolution. I do this term that we subtract the mean and then divide by the center. The subtract the mean is something that we have to do for PCA and divide by the center deviation. It's optional, but for spectroscopy, which are recommended, there is this paper on atrial protocols that's really good about it. That's an amazing agreement.

It goes into the machine learning, but I stopped it a little bit earlier. And also doing smoking because in some places, as we said, you'll see the salt, the have a lot of noise. And the PCA will just like put noise as data, so that also a little bit of smoking. And so this is not my data, but just so you know how to look at it. We don't look really about the values, but we look how the data are spread. So divide the data in groups. So here, an example we have.

the group 2 that this I want that above zero and both directions or axis so we have this one that are above zero with the x axis but below the y axis and this group so the PCA tries to the separate things in groups and you can use these groups as some data some five control groups or to try to predict other properties let's say all of these groups are semiconductors and all in terms of these groups are metallic conductors you can

try to predict values, so that's why it's quite high right now because of AI. This is something that machine learning techniques does. So my goal was to try to do something like that because if you have some good separation of groups in Raymond and also all groups have different characteristics, you can use them in Raymond to predict other characteristics. It's pretty nice. Then I did

So here is the pre-process factor, there's some smoothing and some other mathematical operations to have the variance in size. Here you can kind of see two groups, but here we also have this big silicon in some samples. I did not show you the worst ones, but I have some extra slides. And what PCA is doing right now is just separating things that have silicon to the things that don't have silicon.

Here you can see which, so we have PC1 and PC2, and which big it refers to. Here is not, here are the places that they're different from each other, so you have one big in PC2, and one that's the silicon, and then the diamond, the B1 and the J-Ben, around here. And this data is just separate the ones that have silicon from the, they don't have. And what's actually kind of good, you can use that to see our sample is thinner or not, but not the bigger.

best way so professor said well just try to fit from 1000 to 1600 and I did it and have like a better results so here we can see that PC1 is basically the diamond and the negative values like here is different between the boron the beach and the diamond peak but PC2 is the one that I

have the GBM and we can see that these two can separate very good are samples of 2% of methane from the ones that 1% of methane in groups that have less SPG carbon, from groups that have more SPG carbon. The edges here looks very spread. This is normal I would say. It's the edges. Yeah. Yep. But then I did just the centers.

And we can see a better separation here. Even you can see one outlier. This outlier, I know that is the one that-- maybe not the only one, but it was the worst one, the 1% that had more signal in the silicon. So maybe that's why it's a very good outlier here. But when I also cut from 1,000 to 1,600, I got better separation from the edges, so it may

Here we have some very good group that's 2% in the edges, but with the 1% they also have a little bit spread and some stuff that's not that good, but the edges are not the best, so I was kind of surprised that we could have this nice group here and this kind of nice group here, so I was kind of actually surprised about that. So as I said, we will see if maybe all of these have some characteristics that just don't have, so we can

predict some maybe chemistry or maybe XRD by the way and try to correlate those steps. Also, when I do XRD, I try to do the same thing, the same thing with the XRD to see, like combine two tools with that, like some, predict another one, so. - I mean, it will be a, like, a machine learning guided design of experiment, what you can do in the end. This is, I mean, this is the plan, to have then a very clear outlook when we want to tune the properties of a sample that we don't need to, like,

then probably have something that is straight to the so with very few experiments then we can grow directly the the desired properties so yeah and so some summer the series of the non-mill then the Ibiski finish I have to grow two more samples but I have also the last blower so they have

that the two pin traps are the same side of our line. The CBD growth is actually going very well. The PCA is going to be a good tool for this data set. I just copied and forgot to change it. So Outlook, our previous series of the new diamonds, I know that the BLS is still kind of messy, but let's say it's about seven nanometers. And even if it's like 15 millimeters, it's still almost 10 times.

smaller than the ones I'm using right now. Doing some electrochemicals characterization, SEM and XRD of some of the samples. You have to choose some of our samples because I have a silicon, and it's a very thin film sample. The silicon might be a problem for XRD, for example. - Well, we can separate it quite a bit, even if you see it. You can actually make a baseline with the silicon alone, and then normalize to the main silicon peak, and then actually even do a baseline subtraction to really get out the silicon peak. So this should be no problem. - Because even if it's a very thick film, you will still see the silicon substrate. - But if it's really overshining the diamond, even then you can do a neat silicon and then...

do baseline subtraction with normalization. This works. Maybe one suggestion for the investigation of the influence of the size of the seeding particles. Maybe it would be good to have a look in AFM on the seeded but not yet CVD-grown substrates so that we have an idea homogeneously they are distributed. This is maybe also one of the things that influences the homogeneity until the edges. If the seeding is not homogeneous or even the seeding is less dense at the edges, what can happen? I mean, I guess you do spin coating.

Or drop casting? The coating is not operational yet. Why? Because it is operational. Is it? Yeah, for a long time already. No, I don't think it was installed. Maciej told me he installed it. No, I think he did not have the time. But he told me that it was operational. I don't remember. The air connection? Which is set this week for the furnace. So it was not...

But without the air it's working. - What I heard from Matry is that this is needed for this. - He told me it's operational. Anyway, so you are doing what as the deposition right now? - We doing ultrasound, so we place some of the same suspension in a bag. - And then you just put it in the, so you sonicate in the dispersion with the. So then maybe it's good to see for different coating techniques also and see how we manage to get a very homogeneous distribution of the seeds until the edges of the substrate. But for that we need AFM. You could probably talk to,

We have an AFM in the 8th floor. We used it before. I think Yasmina used it before. So she can give you the contact in the 8th floor. You did it with Sally or? AFM? Yeah? Yes, with Sally and with Bettina from the Physical Institute. Yeah, in the Zeta-Quant then. Or in the physics department. Ah, in the Dressel Group. Yeah. In the Dressel Group. Yeah, we used the AFM there as well, right? Yeah. But,

I'm not sure if the AFM here is working again. - We know it should be working. I saw somebody measuring. - Right, I heard also that it's working again. So then maybe talk to, with Yasmina, to some people from the Franz Lageren group, and then we can use the AFM. - Can I take a long and also measure some stuff? - Well, we first try with one and then go step by step. And then, yeah. Okay.

There are two costs to determine the emissions. The most way is the transient spectroscopy. That will help. But you can sort of do a remand filtration to determine the electrons. But we need to probably then start making contact with a group that can help us to do the transient spectroscopy. I think now is the time that we have samples that would be interesting. And so then probably we should start looking for someone who is willing to do that. Yeah, in my old institute.

Yeah, but they are probably more in molecular transient spectroscopy. We need somebody who has experience with these solid samples. It's not the same setup, actually. It's very tricky. So it's probably even synchrotron transient spectroscopy that we need. So I will talk to these guys at BESSI where we did it the last time for the advanced functional materials paper. If they are still interested in this, they changed a bit topics. I'm not sure because

one of the guys left but this is something we need to maybe launch so that it's ready the collaboration when we need it yeah and as a month but I'll select some when I have both series that some of the best ones to the headshots that are here and found also find suitable in the particles and see you to old, has been used there.

like this paper from John Ford and you can see that he did the same idea. Those are the parts that he used was like kind of very big so they just stayed above of the the structure of the diamonds and not really in between or I think in between it would be kind of hard but it also can be possible So I have some inspiration to play, please, too.

to some nanoparticles that you can use. And thank you. The material that they used for this cuprous oxide composite was so-called black diamond from the Fraunhofer Institute in Freiburg. I mean, this was running in the EU project that I coordinated. So this is Peter Knittel, the first guy. He's the one who made the diamond. And there are these guys from the Synchrotron. And then John Ford did the...

the deposition of the cuprous oxide. The black diamond is a bit of different material than what we do. It's not needles, it's actually... - You say it's a coral-like structure? - Yeah, yeah, exactly. So it's a different type of etching process that they use. So it's very porous and not like straight needles like ours, because we do really like the mask and then we do needle formation. So I have some hope that with the right solvent, we could actually get something in between the needles, because they don't have actually this possibility.

corrugated it's not a like hedgehog structure that they had and maybe we can think of I mean the wetting will be a problem because the distance between the needles is small so the surface tension of depending the solvent could actually make it impossible to wet the space between so we need to think about how to reduce the surface tension but then we might actually be able to somehow dip it into such a solution and do some precipitation so that we have a copper solution and that we do maybe electro deposition or something like that that we deposit when it's wetting the gaps between the needle so that we get it also in between because that would be really nice

And then your idea of overgrowing a protective layer, that could also be interesting. So let's see. So there's a lot of interesting things ahead. What could also be is that because here you actually increase the carbon concentration quite a bit and increasing the sp2 carbon quite a bit. In what? In the films, in the BDD films. So the carbon concentration and the sp2 carbon he...

increasing right? Maybe you have two percent of methane. Yes but maybe he was saying that he would go higher even more. So if that is the case you can actually also do some oxidation so you can oxidize your sp2 carbon and also obtain this nanostructured film. It would also increase the capacitance and then would give you a higher surface area that could also be a nanostructuring option. Yeah?

The question then is if this is closed pores or open pores because they need to be accessible for the solvent.

This is something because the grain bar is embedded somewhere in the film. So when you do this oxidation, for instance, electrochemical, you can also oxidize stuff that is then forming closed pores or disrupt the structure. This is something, I mean, Stefan Roosevelt has done something like that. I remember they did an oxidation of such SP2, SP3 hybrid films and they got like a kind of foam. It's like a foam that they had. and it's conductive

And this is maybe something we have to have a look how controlled we could do something like that. Could be, for example, you grow your film and then at the last hour you increase the sp2 carbon and then... Make like a top layer. Yes. On top of a solid and conductive area that is not prone to oxidation. So that you make directly a hierarchical structure. Something to think about.

Lots of options we have and we have to think, well, which ones will give us good electrochemical properties and especially electron emission properties. So what my dream would be that with the PCA, we could one time at some point then predict if our electron emission properties, how they depend on the CVD parameters. I mean, that would be very, very nice to then go for a directed optimization of that property that is the most important for, I mean, your catalysis and also for key.

Yolen for now so for everybody working with those electrode materials. A little question. In the beginning you said you want to use a catalyst, it's a CO2 reducing catalyst. But isn't actually the Faraday efficiencies of hydrogen terminated BDDs, with SP2, SP3 combined structures have already quite high yields as far as I know. There are over 90% to release such a catalyst.

CO2 reduction on the BDDs because the efficiencies are quite high. It's photochemistry, no. You mean for the photo? Yeah, yeah, yeah. Electro-synthesis, this is something you can do easily on a BDD. No, no, it's with sunlight. The problem is that there's so far only electro-catalysis. This is A-NAGA. Yeah, yeah, yeah. The SP2, SP3 mix. Yes, exactly. And they are quite good, but they are not very long-lived, to be honest. So they decompose those electrodes because they have quite high voltages and currents.

And then this is probably not really scalable. So, and they need this pulsed operation to get those high efficiencies. So this is also something we need to keep in mind that maybe continuous operation is not the way to get best results because the interrupted electrosynthesis turned out to give much higher efficiencies apparently because the products then disappear and then you get new available surface and such things. Yeah, yeah. And...

So this A. Naga work is really worth reading. I think NAS has a big collection of those papers. And there you're right. They have really high faraday efficiency, but for the photocatalysis, this is not yet the case. And I mean, the ideal would be to use as little as possible electricity. We probably will need a little bias to overcome the surface barrier because over time the hydrogen termination is not 100% stable. So we probably will need some bias, but hopefully much below the water splitting level.

stuff potential and then so the ideal would be to have as much energy from the sunlight yes so and for that the nanostructuring is good because it broadens the absorption spectrum of the material so that we get more excitation also from the visible light part and then of course co-catalysts could probably also play a role to to increase efficiency yeah and electro catalytic use of BDDs they also you lose their surface termination quite fast yes yeah yeah yes yes they are not always not always talking about it but it's obvious from the data it's not fully being gone no you can directly see how it after a few cycles of a

applying higher potentials, it's then changed. Exactly. Yeah. And it also changes efficiency quite a bit, also in the case of electrocatalysis. So probably we need to go with something that is not sensitive to that. So maybe... A safer bias between minus O5 to plus O5, for example. Yeah, something maybe 1 volt is still okay. Above that, it's going... Above, it's going down. Problematic, yes. So, um...

So the goal is to replace the electricity by sunlight and still get those efficiencies. The deep UV excitation has something, it's difficult to calculate it really as a Faraday efficiency, but John Ford tried when we had the EU project and the best we ever got with the ruthenium functionalized stuff was something around 50%, but that also is not very long lived because then at some point the ruthenium complex goes away after a while.

which is not very sustainable and definitely not scalable, and with all that ruthenium anyway, not scalable. So we need the iron alternative or the manganese alternative because otherwise such a surface sensitization approach will not work with ruthenium, nobody will buy this. As a proof of concept, it's fine, but after that, then you need an alternative that is also economically and environmentally viable.

And this is not, both is not the case for ruthenium. We still will try with your stuff at some point with ruthenium to get proof of concept data because we know it's working. Yeah, but the ultimate goal is not ruthenium. Yeah. Good. Yeah, thank you very much. Do we have literature today? I didn't see any papers. So it was the second NAS in the list?

And you skipped both? No, so you told me to empty my whole week for the paper revision and I talked to Johannes and he told me that he sent an email to you about this but he did not get a reply so I didn't do anything. Okay, then we don't have literature. Then please prepare this for next week. Then we have two literatures next week. Next week is Daniel I think. For progress. Yes, exactly. And then literature is who? Me. Then we have two literature. Okay.

Then we have lab cleaning tomorrow. I have the list here and it seems that some people have not yet entered what they will do so please finish this. And there I see a slight disbalance between some people having a lot of tasks and others not so much. And I mean I will not tell you what to do but this is maybe something you should consider to share in a fair manner with everybody. And then who did not yet tell me what you want to drink and eat on Saturday please enter this also in that list.

We start around three downstairs. One thing that we still need to do is to assemble our new barbecue. We have the new barbecue. I bought one, big one with a hood and everything so that we can properly barbecue our stuff. I already got some Franconian wine that Rocio is interested in. Ilka will also come. So yeah, yeah. And quite a few old guys. The legendary Thomas will come who invented all the R-relation chemistry and Yvonne who did the TBTQ stuff.

will come and Daniel who did all the diazonium and Prato reaction stuff, they will also come. So some of the old heroes will come who started all this. Okay, good. Then I put this back in the coffee room, so please if you plan to attend, I need to know what you want to drink and eat that everything is catered for. What else? seem to work

oven thing with Ennio Knagfus is... It's all working. I met him when he came to fix the last bit. Briefly talked to him. So that's all good. Maybe a quick update on the glove box. The manometer arrived, it was installed and already distilled off some solvents inside the glove box. So all good. I've also updated this in the standard operating procedure. Nice. How to use that. Dorox Fume Mode has electricity again. They apparently just changed the defect lamp.

and the funeral as itself is still dead. I know. Some people saw me when I came with the people from the university construction office, which is the Universitätsbauamt yesterday. There is some hope on our construction work in the night room. It will start somewhere around October. So I have a flow plan. And when which stuff will... It will mean that...

will have this room not available until March. So we need to get an arrangement with another group that we have a that we can put something somewhere else overnight and also we will then heavily need to use the rooms upstairs and I will try to find some solution here on the fifth floor as well because there are some labs that are currently unused and I will ask my colleagues if I could borrow them until March so that we don't need to move all this stuff upstairs. Yeah, but there's hope that we finally get it and I could convince them that they will replace those two fume hoods in the main lab.

while doing that. So when everything goes well, we not only will have the new night room and the analytics room finished and done, but also in that lab get two new fume holes. So let's see. I tried my best and the guy from the construction office was okay with it. So they will tell me next week how the timeline will change by that, but that means they're doing detailed work and it's not that they question the whole thing as such.

So, yeah, the new person from the UBA seems to be quite motivated and communicative. So we had a very good discussion yesterday also about technical details and how things, he has apparently a bit more pragmatic outlook. We also talked about maybe because they will not replace all the humans in the entire building. I mean, there is a lot of things that would need to be, that would be a multi-ten million euro investment and in a building that will be destroyed in 10, 15 years, they will not do that.

So we need something else that makes our fume modes more operational again. And one thing that is always not working is this control box, the little one, that tells us if the fume mode is correctly working or not. And so now they are thinking of actually equipping all the fume modes with a separate sensor so that we are at least safe knowing if the fume mode is currently operational or not. The whole system at the moment runs on manual operations, also is not controlled by the controlling units because the main controller was broken last Monday. You remember there was this message then on Tuesday that no fume mode is available in part of the building.

And that was due to the failure of the central control unit. And this one apparently has not yet been repaired, and it's even not clear if it can be repaired, but they were able to switch the whole system to continuous manual operation. So it's now running on full force the entire time, which I think is even better because it's not doing this and this all the time. So we have the flow meter now, right?

Is that functioning? - Yeah, it's working. - So I would say that what we do maybe a weekly check of the fume hoods so that we have a better monitoring of how our fume hoods look. And then what we had in Wurzburg in the old building because there the fume hoods were even 20 years older, we had little, in German we called it Flutterfähnchen. So it's actually a little piece of Teflon ribbon that you stick with duct tape or something else on the bottom of your fume hood.

shutter and then it tells you because when there's no airflow it needs to be something very light Kim vibe is also good but that's not so stable teflon ribbon is better when it's vertical there's no air when it's sticking out of the fume hood air is coming out from the fume hood so it should always be a little bit bent inside then you are sure that the fume hood is putting the stuff inside so this is a very low level but very clear and safe way of monitoring your fume hood

ventilation so that if there's any problem you immediately see it. So you will look for a very lightweight Teflon ribbon so that we can install this. It's the solvent proof version that a drop of acetone will not or water will not directly just destroy your setup. And then you put it, I mean the fuel mode is closed like this, then you put it on one of the corners or on both and the top and then let it hang until like one centimeter above the

bench so that you have maximum exposure when the fume hood is closed like with this gap between the bench and the shutter and then you can measure your ventilation constantly and then if you have doubts Mattis has the flow meter then you can properly measure how much it went down or up or whatever and I think this is a way how we can protect ourselves here properly in the group so that we don't have issues because I see a problem also with information here that they that there is some problem and we get the information way too late

And so we need to do something for us to deal with that properly. And I think this is something that will not disturb your work, that keeps you safe, and that is easy to install. And I think we have to work with such solutions here, be pragmatic, but not take risks that are unnecessary. So this is, I think, the way we will go. So maybe you could have a look at the Teflon ribbons that are available in the Kimmel bar, if one of those is really light enough to do that. We had very lightweight...

"like the ones that plumbers use for installations" dieses Dichtband was man beim Klempner hat. Genau das könnte Herr Knackwus vielleicht sogar haben oder Herr Ulrich. Wir haben das um die Dichtung umgedreht, ich weiß das ist das. Oh ja, dann können wir das mal ausprobieren, weil dann kann man ein Stück Ducktape nehmen, dann Teflonband rein machen, also so runter hängen lassen und dann installieren wir das mal testweise in einem Abzug, ob das funktioniert und ansonsten müssen wir noch, Es gibt noch so ein anderes, ein bisschen festeres, ganz leichtes Teflonband, aber vielleicht geht auch das,

Das wäre natürlich die simpelste Lösung. Die ist dann Wasser und Aceton proof. Das Chemwipe ist jetzt aber, dass jeder Wasser Wasser das machen würde. Es wird definitiv sein, dass die Füße nicht arbeiten. Was machen wir dann? Ja, aber in Druk's case, das machen wir nichts.

Das Problem ist, dass sie keine Spare-Dienste haben. Das ist das Hauptproblem. Das ist das Problem, was ich gerade gelernt habe, ist, dass wir es manuell öffnen können. Da ist eine Klappe drin oben, die kann man wohl mit der Hand öffnen. Im Timot? Ja, oben in dem... Da sollten wir vielleicht nochmal mit der PC reden, die haben das nämlich überall schon gemacht. Und dann kann man den Abzug sozusagen selbst wieder in Betrieb nehmen.

Das würde ich sagen, ist der Workaround. Ja, auch dieses Dezernat Technik da unten, das kann man glaube ich in der Pfeife raufen. Aber die ganz besonders, wir hatten ja gestern diese Sitzung mit denen über diese Abzugsschose von letzte Woche. Es war sehr interessant, wie viele Ausreden man haben kann, damit man nichts unternehmen muss.

So kann man das, glaube ich, zusammenfassen. Das geht jetzt zur Kanzlerin. Also das spielen wir jetzt von oben. Da waren sich auch tatsächlich mal alle Dozenten in der OC über was ganz einig. Und das will was heißen. Also das geht jetzt mit geballter Force. Also die Dekanin ist auch an Bord. Also wir machen da jetzt ein bisschen größere Welle von oben. Weil das geht so nicht weiter. Das ist inzwischen, also ich würde das Arbeitsverweigerung nennen.

kann man dazu eigentlich nicht mehr sagen. Die hatten Personaleckpässe, die haben inzwischen aber neue Mitarbeiter. Ja, die müssen eingelernt werden, aber nichtsdestotrotz kann man die ja auch auf der Baustelle einlernen. Das heißt, man könnte mit denen mal vorbeikommen und ihnen zeigen, was hier so die typischen Probleme sind. Das große Problem ist, wenn wir eine zu große Welle machen, kann es auch sein, dass uns die Berufsgenossenschaft das ganze Gebäude zumacht, von einem Tag auf den anderen. In Berlin ist das passiert und das wollen wir natürlich auch gerne vermeiden. Klar. Also das ist so ein bisschen

Wir müssen eine mittelgroße Welle machen, ja, das ist ein bisschen diplomatisch, ein bisschen tricky, aber wir arbeiten daran. So, if you have problems with your fume hood and you are working with something that is cancerogenic, then go to another fume hood. The ones in the night room are properly working, so those are always an option. And I will try to secure some other places where hopefully the fume hoods are working. Normally,

They work, but there are these occasions when they don't. And then, depending on what you are trying to do, I mean, if you do an ether extraction, it's okay. The ventilation, even with the chimney effect and the fume hood, will be enough. But if you use methyl iodide or whatever, something that is volatile and dangerous, then, of course, you have to be absolutely sure that the thing is working. So if you have something that is not volatile, the fume hood is not relevant. then gloves are the relevant things.

is also something you need to then properly consider. As soon as you have things that are volatile and dangerous for your health, only working fume should be used. This is clear. And for non-volatile things, the protective measures are others. Then, of course, exposure to skin and eyes and so on needs to be avoided. Yeah. So we will work on it. And I think the...

proper monitoring so that we are knowing if it's working or not is the first step because we cannot rely on information by the department six because they come very late and so better have a continuous monitoring then I will let you know as soon as I talked to the other guys if we can have alternatives here on the on this floor if not we need to then really put stuff upstairs and so

start working there with higher intensity. Part of the fume hoods there are also problematic, right? Yeah, the left one on the center floor is working quite good, and the other one-- It's so-so. So-so, yeah. But maybe-- It highly depends on how close you do the fume hood, right? Yes, but this is due to the performance of the entire ventilation engine. The closer you have the fume hood, the better the ventilation is.

Humults that have a bit further away from their central engines have lower flow. That is... So what could also help in those cases is not opening a shutter entirely, but use those sliding glass doors and work from the side. That actually improves the ventilation substantially. Because the...I think in one of the safety instructions we had an image on the slide where they showed how the ventilation is going. It's actually a...

like a tornado that is vertically like this through the fume hood. So when you come from the sides, it disturbs this tornado much less than when you open the shutter and then it has turbulences at its lower bottom, that like ventilation tornado. So the opening from the side for the ventilation is much better, maybe less convenient for handling stuff, but it's much better for the

ventilation in the fume hood. You can try this when once you have these little ribbons hanging, opening the window will basically not affect the ventilation whereas moving it up you will see that it's less. So this is also one safety measure especially in fume hoods that are a bit weaker than to work with the windows open from the side and not opening the shutter up to the maximum.

Well, is the fume hood in the small room upstairs working or not? What is the small room? There are dozens, 200 kaboves, links and rechts. Well, I didn't test them to be honest, but the left one is so full of stuff you can't use the room actually. Well, this is mostly our stuff, so we just need to tidy up. That's a thing that needs to be done at some point. Anyway, because this could be an alternative night room. I can test it. How long will be the construction? Until March.

Six months. Yes. And then also half of the big lab is... Yes. Yeah. Yeah, so the lab where Doro, Gaspina are currently, this will be then closed anyway because there they will have the... How do you call this? It's like the entrance for the construction site because they need to change into some protective gear and so on. So that's the changing area. It's a Schleuse, yeah, in German. Entry chamber. Entry chamber, yeah. Entry chamber. Ah, yeah. Okay.

Like for the clean room, but the other way around. So the entrance to the dirty room. So that they don't bring the stuff to our lab. So, yeah. But at least progress. Finally. I mean, I only believe it when they do it. Because last year we also had a plan already. with these details.

And even some of the stuff had been ordered already and is somewhere now. So and then the main reason was apparently that the ministry made a household stop. And then projects that had not yet really started were just dead ended. That was the reason why it didn't eventually start last year, although all the planning and all the preparations had been done already. But there's hope that we have to be.

will have this at some point early next year but seeing is believing before let's see but this is apparently a problem everywhere even Max Planck is complaining about construction stuff they also have problems to find companies we are coming now to the point where companies are problem that we don't find enough companies anymore for the construction world.

Nobody wants to do this job. Yeah. Okay, so this is update on that part. Then. I would have a small question regarding the archive fridge. Yes. I was defrosting this one other freezer we had because it was full of... The archive freezer will not be defrosted. No, no, no. I'm just asking because there were two trays full of archived samples in there. Yes. Where are they now? In the archive freezer. Where was their space?

because the reason why they were in the other one was that there is no space in the archive freezer. No, no, the old archive, the actual archive freezer, which is not locked, I am talking about this one, and I put it in there while defrosting, I will put it back today. Yeah, right, the actual archive freezer is working again, right? And so we can now, now we have two archive freezers. Yes. Yeah, the one that is in the glove box room, where everything was put when it broke down. Yes, and there's no space, that's right. That's correct, yeah. Yeah, yeah. And now you open the archive freezer, the other archive freezer again, and that is now where you put those stuff. Actually, leave them there and keep it on. Okay, but then I have to change everything because the boxes are from the other freezer, and there are no boxes, but I can do that. We can have a look together. Yes, we can do that. I think that's the best.

I'm asking because then we have space in the freezer again and use it for chemicals. Yeah, yeah, exactly. And also the archive stuff. That was an emergency usage, but it should not be mixed up with current work because the risk of killing archive material then is too big. So this should always be in separate freezers. So then we opened archive freezer number two now, and then you have more space again, and we have a look together. Yeah, we can do this. That is, I think, the best way to do it, to keep track. Yeah, yeah.

I think it's from Julia and Basti. And so this is the stuff that I actually drove from Wurzburg here. Yeah. - I think now we also have Ellie's stuff. - Ellie's stuff is also there, yes. - Ellie's stuff, yes, yeah, yeah, yeah. - Exactly, and so. - But that's on one of the fridge, which is on our side, not in the. - These are the dispersions, which are not freezes. - Yeah, but here is also freezer stuff. And that is probably in the same freezer as the ones from Julia and Basti. And yeah. - Like normal chemicals, molecular stuff. - And then Ellie has a lot of things that shouldn't be frozen, and those are in one of the fridges. - Okay, yeah. - And this is probably the stuff that you are talking about.

So we probably also need some kind of archive fridge at some point for those things that cannot be frozen. There is this extra fridge in the analytics room that is even unplugged. Maybe we can plug this in and use it as an archive fridge because this is also a place where we don't store chemicals anyway. That's maybe a good idea. My fridge is also quite empty. I use some space but there is also... Yeah, but this is a fridge that is used and maybe even will be more used when you have stuff. So I prefer to have archive fridges separated from current usage fridges. So I think the analytical room is better and you keep your fridge for more stuff that you make. And maybe Bernardo also needs at some point.

I mean right now we don't need necessarily space but I do right it's better to call them apart. Yes. There's no confusion and accidental. Oh I don't know what it is. Lab cleaning. I throw it away. There are four fridges. Two of them are full with chemicals. Two of them are with archived samples and the two chemical fridges are about to get full. Yes so it's probably good to move some of the stuff there to empty one of the fridges in the lab where people are really working with chemicals because then you have...

flexibility there for chemicals. Yes. So it's okay if we plug the fridge that is in the analytics room, we just keep it there? And if it's working, then we move it? Yes, that's a good idea. Exactly. I'd see if it works. So it's running, but it looks rather new. I think it looks rather new. So I have some hopes for that one. You never know. Yeah. What else?

Gibt es Neuigkeiten vom Pumpenstand? Also Johannes hat den Wagen geholt. Und ich würde jetzt den Herr Ulrich, also den Wagen haben wir gestern ausbekommen, dann würde ich den Herr Ulrich jetzt fragen, welche Pumpe am besten noch funktionieren würde. Ja. Von denen, die wir haben, also wo er die meist davon hat. Ja, dann soll er die mal angucken. Genau. Machen Sie das am besten noch diese Woche, der geht irgendwann in Urlaub. Ach so, okay. Ja, ich habe ihm vorhin schon eine Mail geschrieben. Okay, gut. Dann. Ja. Gut. Dann.

cleaning tomorrow

I think we can continue with the small subgroup meeting. We have a meeting today with the guys from the technical chemistry. So maybe it's at 2:00, right? So maybe you can come five minutes before 2:00 to my office and we go there together because you don't know where. So then we go together. Okay, good. Do you want to start tomorrow? Tomorrow I will finish this.

- Thank you.