On-surface synthesis: a bottom-up strategy to low-dimensional carbon structures
This is a modal window.
The media could not be loaded, either because the server or network failed or because the format is not supported.
Formal Metadata
| Title |
| |
| Title of Series | ||
| Number of Parts | 34 | |
| Author | 0000-0001-9589-6855 (ORCID) | |
| License | CC Attribution 4.0 International: You are free to use, adapt and copy, distribute and transmit the work or content in adapted or unchanged form for any legal purpose as long as the work is attributed to the author in the manner specified by the author or licensor. | |
| Identifiers | 10.5446/59091 (DOI) | |
| Publisher | 05jdrrw50 (ROR) | |
| Release Date | ||
| Language | ||
| Producer | 05jdrrw50 (ROR) | |
| Production Year | 2022 | |
| Production Place | Frankfurt am Main |
Content Metadata
| Subject Area | ||
| Genre | ||
| Abstract |
| |
| Keywords |
Beilstein Talks11 / 34
6
14
17
23
00:00
Formation flyingCrystallographic defectMultiple birthNucleosynthesisPlain bearingSubstrate (printing)Cylinder blockSpringer (tank)SensorAdsorptionAtomic force microscopyRing (jewellery)Wedge (mechanical device)Planer (metalworking)PhysisorptionMembrane potentialRulerRadioactive decayElectronBand gapCartridge (firearms)Energy levelSteckverbinderRefractive indexHot workingGroup delay and phase delayWire bondingAtomismSpeckle imagingBending (metalworking)SynthesizerAtomic force microscopyPattern (sewing)SizingRing (jewellery)PolradMaterialDrehmasseCurrent densitySilveringAngeregter ZustandCrystallographic defectElectrical conductorCrystallizationThrust reversalTemperatureMapTape recorderCogenerationOrbitCommunications satelliteAbsorption (electromagnetic radiation)Phase (matter)Linear motorSubstrate (printing)Formation flyingDomäne <Kristallographie>BookbindingToolGaussian beamStripteaseCar tuningGasElektronenkonfigurationCrystal structureVacuum pumpShot peeningTypesettingQuality (business)LimiterModel buildingRedshiftRoll formingOrder and disorder (physics)Controller (control theory)Scale (map)StonewareCocktail party effectForgingVakuumphysikNyquist stability criterionKlemmverbindungLightNeutronenaktivierungBahnelementProzessleittechnikDensityHydrogen atomMeasurementRing strainMetalAutomobileBandstrukturPhase (matter)Field strengthYearAssembly lineConduction bandCylinder blockCombined cycleSheet metalKoerzitivfeldstärkeFahrgeschwindigkeitSingle (music)BiasingSemiconductorSeparation processInterference (wave propagation)Chiralität <Elementarteilchenphysik>MetamaterialSwitchAdsorptionFlexibility (anatomy)Remotely operated underwater vehicleContactorKette <Zugmittel>Effects unitAngle of attackFord ExplorerComputer animation
Transcript: English(auto-generated)
00:05
I would like to thank Ballstein, first of all, that we were able to do this special issue about molecules on surfaces and also for the opportunity to talk today about the results from the last years we obtained in our lab.
00:22
And first of all, I would like to start with an introduction about on-surface synthesis. So on-surface synthesis is a bottom-up strategy to create low-dimensional carbon nanostructures and basically what we do is we are evaporating molecules as building blocks so you can think
00:42
about you're taking here the molecules as Lego brick stones and you bring them on the surface and then these molecules usually form nice self-assemblies. So at this step, the molecules are usually interacting via non-covalent interactions,
01:01
for example, hydrogen bonding and for example, van der Waals interactions and basically to connect the molecules stronger with each other and also to obtain a better electronic communication between them. So we try then to activate the molecules when they are self-ascended on the surface that we can, for example, do by annealing the surface or we can shine light to activate
01:27
the molecules by light or we can also use the electrons, for example, from an STM tip. So if they are activated, we hope that the end groups are splitting off of the molecule, you have radicals and if the molecules are diffusing around on the surface, they eventually
01:43
find other molecules and then they hopefully form nicely or the covalently linked networks or for example, rickets. So that is the process what we do and in this whole thing, the surface basically acts as a catalyst to facilitate this activation of the molecules.
02:04
So this approach has a huge great variety because you can take all kinds of different molecules of different shape, different functionality and then basically create a whole bunch of different materials and the good advantage about it, these molecular layers, they should
02:22
have an improved mechanical and thermal stability compared to self-assemblies because you have then like covalent bonds along the building blocks. Exactly and the other nice thing about it is, so we are working with these on surface synthesis, usually in an ultra-high vacuum environment, reason for that is that we have
02:42
well-defined surfaces and we have also directly approached to measure the structure and the electronic properties of these created networks using STM, for example, or non-contact AFM. Now let me give you some parameters which we can tune to basically create different structures.
03:04
So basically, like the design principle which we can use in on surface synthesis, one of it is definitely the surface. So it depends what kind of termination you are taking, for example, a hexagonal phase or a squared shaped phase. So that will
03:20
help you to create different networks or it also facilitates orientation of structures. Then another point is depending on what kind of substrate you're using, the reactions might proceed in a different way. So if you're looking in the Almond type reaction, this is the most commonly used reaction in on surface synthesis. So where you're using
03:42
basically halogen and you're splitting off the halogen by activating it, for example, on a metal surface by annealing and on gold, usually you proceed directly to the carbon-carbon bond. On silver, normally you take an intermediate step where the radical after the cleavage of the
04:01
bromine will look for an add atom and form metal organic states in between before it then goes to covalent states. So that can be, for example, used this metal organic state to create before some order because the metal organic state is not yet so strongly bound together. And usually
04:22
this is like a reversible bond formation. The other thing with the substrate, you can also tune the reactivity. So if you're taking, for example, a very highly reactive substrate such as platinum, then molecules can also go like in three-dimensional structures and mainly stay as
04:40
individual objects. So the three-dimensional structures are facilitated because you can then create also bonds towards the substrate, which then creates a curvature and eventually you can, for example, also form then like buckyballs and things like this. While if you're taking weakly reacting surfaces, for example, gold,
05:02
then the molecules try to find each other and then you get more polymeric structures, for example, 2D networks or ribbons. So this shows that the surface plays an essential role and we can really use it to tune the reactions and also to create different structures.
05:21
So one aspect this is important to consider, if you're doing on surface synthesis to create these carbon structures, this is basically relying on a kinetic control. So we are taking the molecular precursors which have already a certain design and try then to connect that
05:41
on the surface to form, for example, networks or ribbons. So the structure of the resulting structure is basically given by the precursor. Because if you would try to anneal higher and basically try, for example, to anneal out defects as you would use to do that in
06:00
self-assemblies, basically we'll get to the thermodynamically stable state and then probably you would end up just with a two-dimensional carbon sheet, which is graphene. So that makes the whole thing a little bit more difficult because if you try to create the reaction among the molecules and bind them together, that's usually like a irreversible
06:21
process. So meaning of which if you're trying, for example, to create porous structure, you need to make sure that you have a rigid enough design of the molecule because sometimes you have flexibility in the bond and then you can easily create, instead of a hexagonal network, you can incorporate, for example, pentagons and heptagons
06:41
as defect, as you can, for example, see here in this hexagonal network, you have here one pentagon. And once these defects are formed, there is no way back, so to say, then basically you can not anneal that out and therefore you need to really think about the design of the molecule before to avoid such defects. That brings me to the design of the molecule. I mean,
07:07
you have there different things you can do. First of all, by choosing a certain end group, you can tune different coupling schemes. The most common one, as I already introduced, is the Eimann type coupling where you use bromines or other hydrogen atoms, but they are
07:25
also different reaction schemes. And so this is a quite well studied topic in the sense of to explore also new reaction schemes. Then another thing is how much end groups you attach
07:40
to your molecule. So this is then the symmetry of the molecule that basically then determines what kind of structure you create. So if you have a two-fold symmetry, you will likely create a ribbon. If you have a four-fold symmetry, then you will likely create this clear lattice and with a three-fold symmetry, you probably will create like a hexagonal lattice.
08:03
Then you can incorporate design principle, how to create networks. So you can use this intermediate state with metal atoms to pre-order molecules or you can incorporate multiple reactions. And you can also like add some more functionalities into structures by, for example,
08:21
using different components, use chirality or add hetero atoms to then change, for example, the electronic properties or also change electronic properties by incorporating defects. So defects in this case should be rather be positive. That would mean that you create
08:40
in an hexagonal network, for example, also form membered rings or egg membered rings. So these are just an overview of what you can do. So you see you have great design and you can really create different materials. And just to wrap that a little bit up this topic of on-surface synthesis. So basically it differs to conventional chemistry in several ways.
09:03
So conventional chemistry is done in solution and gas phase. The on-surface chemistry experiments are either done in ultra-high vacuum, you can also do them in air, but ultra-high vacuum is preferable in this sense, if you really want to also do the high resolution characterization of
09:20
these structures, as I will show you later in the talk. So the conventional chemistry, because it's done in solution and gas phase, usually has more simple equipment. You don't need the vacuum pumps and vacuum chamber for this. But it relies on dynamic behavior. So basically in solution the molecules diffuse together and when they come together they do
09:43
the reaction. And so basically the parameters you have there is you can change temperature, concentration, solvents and different catalysts to do the reactions. And on the surface, basically the substrate is the catalyst and is therefore playing an important role.
10:02
And that creates also a template because the molecules are confined then only in two dimensions they can complete the two different reactions. And so that's a bit the the thing which we have to consider is, so reactions as they proceed in conventional chemistry in solution, they don't necessarily need to have the same reaction pathway on the surface,
10:23
which allows us to basically to create science and explore new chemical reactions on the surface. Now what can you do with this on-surface synthesis? What kind of physics can you look at? So if you have the molecules, basically what you're doing is carbon-based structures and
10:44
they're supported on the surface and are usually low dimensional. So either two-dimensional or one-dimensional or single molecules. If you're looking at a sheet of just carbon in two dimensions, then this is graphene. And graphene is a very fascinating material
11:05
because of its very fascinating electronic properties. So if you're looking at the band structure, then we observe at the K point that you have a linear dispersion. So that ensures that you have high charge carrier velocities. But the unfortunate thing is,
11:25
if you would like to use graphene in electronic devices, it does not feature a band gap. So it's different, for example, to semiconductors where you have a band gap here, this does not exist in graphene. But you have different ways how you actually can open up a band gap in carbon
11:45
structures. And one way, for example, is simply to use the dimension. So if you're going from a graphene sheet and you actually try to cut out ribbons, then you can do that in different ways. For example, here in an armchair with an armchair edge, or you can do it with a
12:05
zigzag edge, and then you can create also, depending on the width, different electronic structure of the ribbons. But you can, in this way, open up a band gap. So that's actually a nice way how to tune the electronic properties, and that is also shown that this nicely works in
12:24
the pioneering work of the group of Roman Fassl and Klaus Mühlen, where they used in 2010 on-surface synthesis to create well-ordered graphene anhydrides. And that was done by on-surface synthesis on a gold substrate. They used these precursors with the bromines,
12:45
they cleaved the bromines off the halogens left here on the surface, and then basically the molecules connected together. So these are here. Then linear polymers formed in the first step, and in the second step the material was heated further, then also the hydrogen
13:04
split it off, and the graphene anhydrides were created. So that's a nice example how to really create atomically defined structures, and the structure itself is templated by the structure of the molecule. You can also go to these structures and use the molecules to create
13:27
fascinating lattices, for example, well-controlled 2D polymers, and there basically determines the lattice structures in the end, the electronic properties, because you can, with tight binding,
13:42
for example, calculate out of the geometric structure the band structure of your lattice, and then depending on what kind of molecule you are taking with different symmetries, you can, for example, create a honeycomb network, then you can also create molecular lattices which should feature, actually, Dirac cone features as was expected in graphene. Or you
14:06
can take like Kagome lattices, they have additional flat bands which are interesting, or Lieb lattices which have a square symmetry and would also feature Dirac cones. So to do that, the only thing you need to make sure is that you have a good connection
14:22
among the molecules, so therefore we need on surface synthesis, you need to have a good communication through a CC coupling, you need to have a good design that you get a good crystallinity, and one of the important things is you need to also make sure that basically the electronic structure of the molecular networks is maintained, and for this one
14:42
needs to electronically decouple the molecules from the substrate. And that was actually the special issue we tried to put together, together with Michael Stöhr, and that was about molecular salt assemblies and how to physically and electronically decouple them from substrates.
15:02
So if you're interested in to see different kind of molecular layers, how they can be decoupled, then you can have a look in this special issue. So this promises us if you're creating nice 2D lattices, we can study new carbon-based
15:21
materials with fascinating electronic properties. So what I will show you today is just a couple of examples which we did recently in our lab. So first I wanted to start with a zero-dimensional object. This is a ring which has a very well-defined edge.
15:41
And then I want to talk a bit about these 2D polymers, how to form them, and talk a bit about the electronic structure to just show you some examples how you can basically engineer with these on-surface synthesis molecular structures which have well-defined electronic properties. So all the measurements I want to show you today, they are performed in ultra-high
16:06
vacuum and also at low temperatures. So we are using in our vacuum chamber scanning tunneling microscopy in combination with non-contact atomic force microscopy to get really high
16:20
resolution images of the molecular structures and also at the same time to be able to measure the electronic fingerprint of them. So first to the rings. If you want to create ring-like features from carbon rings, basically you have to connect different granular rings together. And
16:43
this you can do in different ways. So you can do that, for example, in para. Then usually you get linear objects. Meta would be something where you would start to get rings and basically auto would be too tight together. So here it's hard to form rings. So for rings, meta would
17:01
be ideal for a shape purpose. But if you are looking into the electronic properties, actually the para is way better because the para connection is more symmetric. So basically, if you are looking at the electronic pathway here on the left side and on the right side of the molecule, then basically they are the same. And if you're looking on the meta,
17:24
then you will go here on the left side a different way and then on the right side. That leads in para to constructive interference effects and in meta materials basically to destructive interference. So that means the para has the better electronic transport along
17:42
the carbon rings and would be preferred. So we have to find a way somehow how to make out of this para connection still like ring-like features because so far what was used to make rings is always the meta connections. So there are two examples from literature where other groups
18:04
managed to create nice rings, but they have all these meta connections which is not so well in the electronic communication. If you want to have the para connection, there would be the super para phenylene. But this is a molecule that has so much strain
18:21
that you have difficulties to planarize it. So that's actually super challenging. And so we were thinking, is it actually possible to create a ring which has a super para phenylene unit at the rim and inside we need some additional rings to basically support the structure. So that was our dream and we wanted to see if we can synthesize that because
18:44
then we would have a very good electronic communication along the rim and it would be able to planarize this super para phenylene. So the first thing we need to make sure is we need to preserve the electronic properties of the super para phenylene and if we add here
19:03
some more rings, then this does not necessarily mean that basically the orbital picture still stays like that. And so we thought, okay, this could be a precursor unit. If you take four of them, then you could create a ring. And indeed, if you're looking at the orbital picture,
19:23
that still resembles the properties of para phenylenes. So you have for the HOMO, you have this stripe-like pattern and for the LUMO, you have the wave functions going along the molecule. And the trick which we are using is these five-membered rings. So they basically allow us
19:43
to curve this armchair edge because usually this armchair edge you have here in the periphery is only available for linear structures. And the five-membered ring is doing here the job to basically create a certain angle which then gives eventually this ring.
20:04
So we tried that on the surface. We attached some bromines to be able to connect covalently the molecules together. And if you do the on-surface synthesis on a gold substrate, you create different structures. So you create chains, you create trimers, tetramers,
20:22
and sometimes also pentamers. And during which we are interested, this is this tetrameric structure you can see here, you find that on the substrate, the yield is kind of okay-ish, but dominant structure is basically the chains. So here basically the molecules are just
20:44
coupling the other way around compared to here. So we get these rings on the surface. And first of all, we wanted to make sure that they have actually the properties structurally than we think. So we have the five-membered rings and
21:01
the six-membered rings. And for this, we performed non-contact AFM with CO functionalized tips. So this has the advantage that the CO at the tip end is basically flexible. And if it comes into the repulsive contact with the molecule on the substrate, it starts to deflect and that's
21:23
basically what you can record in your AFM images. So this is a technique which was introduced by IBM. And we looked also at the structures by DFD calculations. So indeed, a ring-like structure
21:44
like this is stable. In vacuum, also in gas phase, basically such a structure would be non-planar. That would be the favorite arrangement of these four molecules coupled together. But the planarization energy is very small. So it's less than an electron volt and
22:02
the absorption energy of such a ring on a substrate like the gold surface is nearly 9.5 electron volt. So you can easily planarize the molecules on the surface and therefore, we actually are thinking that these molecules here are nearly planar.
22:21
Then we looked into the electronic structure. So what we expect is that the band gap of these rings should be smaller than in conventional to chloropyrifenylene. The reason for that is because first of all, we planarize it. So then you have a better overlap of the pi-orbitals. And second of all, we have also the pi extension, which should significantly reduce
22:42
the band gap. So in calculation, you would expect a reduction from about 3.5 electron volt to about 2.2. And experimentally in STM, we determined a band gap of about 1.9 fine electrons. So the interesting point now is how is really this communication about these carbon rings.
23:05
And here I come back to this orbital picture I introduced in the beginning. So we hoped to see that for the HOMO, we have like this stripe-like pattern and that the HOMO is going more or less along the ring. And to see if that really is also preserved like
23:23
this in the molecule on the substrate, we performed scanning tunneling spectroscopy and recorded maps at different bias voltages. And so you can see here on this side, the filled orbitals and on this side, the unfilled orbitals. On the top is the experiment
23:41
and on the bottom are calculated images and they are performed with CO tips that you really see also fine structures of the molecule. And indeed, if one is looking here at the filled levels, one sees this stripe pattern of the HOMO and here in the unfilled orbitals,
24:00
it seems to be more states going really along the full ring. So that is actually quite nice. So it seems that we have really like delocalized states among these rings and makes that these rings actually like kind of good electronic structures. So the physical aspect about it would be, I mean, if you have a good conductor,
24:24
which is in a ring-shaped, the question would be what would happen if you will put that in a magnetic field? Because if you have a good conductor in a circular shape, basically, you could induce ring currents in the presence of a magnetic field.
24:41
And so this you can visualize in calculations. To measure at the moment is a bit difficult, but we tried at least to see if our structures by calculation will be able to do so. And for this one can first of all, look at the aromaticity. And second of all, we can look into plots of these induced currents, how they are distributed in the molecule. So basically, we can first
25:08
rely on the Hukliu's rule to predict if this will be the case or not, because you need to have the right amount of electrons to have aromatic structures. So this means we need to have 4 times n plus 2 pi electrons to get aromatic structures or 4 times n pi electrons
25:26
to have anti-aromatic structures. And if we start to calculate the atoms here or respectively the electrons along the rim, so there where we have like this good connection between the molecules, we unfortunately have not the right amount of pi electrons. So we
25:47
have 48. So this means we would have no aromatic character. However, if you would add 2 or take 2 away, then basically, according to the Hukliu's rule, we would be in the case of having
26:02
aromaticity. And so we try to see that if that works. So here in the calculation, you have on the left side the next plot, which gives you an indication about aromatic and anti-aromatic behavior. So this seems to be a mixture of both and it's clearly not fully aromatic.
26:21
And if you're looking in these plots where you see the current density, which is induced in a magnetic field, then you see that there is no particular ring current. It's fairly homogeneously distributed. So this is the uncharged case. And if we add to the molecule two electrons or take them away and look, for example, at the diaions or the dikations,
26:45
then the electronic properties dramatically change. So in the next plot, suddenly everything becomes red. So meaning the molecule gets a really aromatic character. And if one looks into the anisotropy of the induced currents, then one can see now that at the border,
27:02
there is this bigger density and that corresponds to induced ring currents. So that means actually this already fascinating structure, electronically speaking, and that already comes only from the ring-like feature because we did also the calculation for the trimer. And there you can add the right amount of electrons or not. And you see
27:27
basically that in this case, you do not have the ring currents. So that really says that you need to have these closed structures. So that would be a next step to actually measure such properties, to actually measure such induced ring currents. One would need to be able to
27:44
electronically decouple the molecule from the gold substrate. And in the second step, one would need to measure the molecules in presence of a magnetic field. But to summarize this, so this is like a nice example where we have really atomic scale control
28:00
to have a well-defined edge structure, which gives us then exciting electronic properties of the carbon-based object. So going now from the 0D to the 2D networks, there I would like to introduce you quickly two types of networks. So the first thing was this terpenalamine based
28:23
covalent networks. So these are these triangular shaped molecules. And to create the networks, we used a hierarchical design. So we used two different halogens to basically get more order in the structure. So if you do the reaction on the gold substrate, first the iodine will cleave
28:44
off, then you can create macro cycles chains. And if you connect them in the second step, then you have already like form stable elements and we get a decent quality of 2D networks.
29:01
So this terpenalamine covalently linked network you see here, we can obtain them in quite good quality, but at the moment, the size is still a little bit limited. So let's say the single domains are still relatively small. I mean, we can get the higher coverage easily,
29:22
but to improve the quality in one domain, that is something which is not so easy to achieve. So if one is looking from calculations on these networks, then basically these are really 2D layers. They are visisorped on the surface and usually have a planar character.
29:42
And if one is looking at the electronic properties, one can do now exactly this thing that one is looking first at the different lattices which are present. So for example, if you're looking at the nitrogens, then basically we get a honeycomb network. And if you're looking at the newly created carbon bonds, that would be basically a Kagome lattice. Both of these lattices
30:05
have the feature that they should have in the band structure. They are called crossing and linear dispersion around the K point. And that you can also see in the band structure that there are such crossings. The only unfortunate thing is if you create these molecular networks
30:27
on a gold substrate, you basically have also doping or charge transfer from the substrate and that shifts the band structure. So if here basically this band directly going close to zero
30:43
or around zero, this is very much downshifted here to about minus 1.6 electron volt. And basically these bands are also coming down. That's then basically the onsets of the conduction band. So this is the difference. The band gap is roughly similar in calculation
31:01
and in the experiment because we used here an HSE DFD model which can predict the band gap quite well. But because this is in gas phase, the charge transfer from the substrate is not considered. And that makes it then also difficult to measure, for example, the properties
31:22
which are here exciting close to the formula. So another thing which we looked at, how can we, for example, tune the electronic properties of these covalent networks? So how can we, for example, change now the band gap? Because this is still relatively sizable with 2.5 electron volt. And one additional thing apart from the symmetry of networks, which we thought
31:45
this kind of a nice tool is, we have these very nice pores with functional groups there. And we can add molecules into these pores and basically in this way try to further change the electronic properties. And that works actually quite nicely, this concept for this
32:04
strip and the beam networks. So you have here the band structure for the network without molecule inside. So this is this time calculations at the PPE level of DFT. So this means the band gaps are slightly underestimated. And if you add the molecule, the bands or the
32:25
states of the molecule are such that they basically overlap with states from the triphenyl beam network. And therefore, you basically can expect that the band gap of the carbon network is reduced. So that is or will be actually kind of a nice tool to further change
32:47
the electronic properties. And that will be reversible because if you would heat up the substrate, you would expect that the molecules in the pore can leave because they are connected here just by hydrogen bonds. We tried to do that experimentally. So we did it for both
33:02
for the network and for the pores. The pores are just the rings. So there is the connection here not yet done. But it's the behavior for all of them that you can add this dramatic acid into the pores. And you can see that they are taking two different orientations. And in this way, basically, we also showed by the calculation that there is one adsorption
33:26
geometry where this is possible to do so for hydrogen bonds. And then you have really a good charge transfer basically between the molecule and the covalent framework. So as a next step, we also try to do that experimentally to see how the band gap changes. But for this, basically,
33:45
we need to have larger networks to do so. So that is one step. So these hexagonal networks where you can tune the band structures, but this is like everything like SP2 carbon. So this is more inspired by graphene. But there are other interesting like carbon families or carbon
34:07
materials which are less explored or basically cannot be synthesized. And one of it is if you're switching from the SP2 to the SP-SP coordination, then basically then you can
34:22
create these graphene networks. So they contain these triple bonds and single bonds. And these graphene layers compared to graphene were not able yet to be synthesized because they are highly reactive. So there are different of these SP-SP2 graphins. So you can create this
34:46
single triple bonds in different structures. And the exciting property would be that you also have like linear dispersion, but some of them they would also have a band gap. The only difficulty behind it is that you have a relatively high formation energy to actually
35:05
create them. So it means the carbon is rather liking to do the thermodynamic stable structure, which is graphene. And you can see here the formation energy compared to the graphene for these different graphene structures. And you see this really is quite substantial with about
35:26
0.8 electron volt per atom. So this means you really need to have a good precursor design already to convince the carbon to stay in such a graphene network and not go into the thermodynamically stable thing into the graphene. So synthetically, as I said,
35:45
this was not yet achieved. Also in solution, this is very difficult because the tries in solution usually you get then multilayers of these type of materials. Because this is difficult to do with the carbon-carbon bond, it was suggested that what is easier to do
36:06
is to do it kind of with a super molecular approach. So using molecules with triple and single bonds, but basically instead of using the carbon-carbon bond, you could use like soft interactions like hydrogen bonding and halogen bonding. So that was proposed and that was also
36:23
shown that you can do that also in a crystal. And we also showed this is relatively nice to on the surface. So you can either use hydrogen bonds or you can use halogen bonds and you can indeed create supermolecular graphene with very nice order. One is stable up to room
36:42
temperature, the other one you have to create at lower temperatures because otherwise the pro-mines are already cleaving off. Surprising was here actually the halogen bond is even more stable in energy than the hydrogen bond, but both of them show nice structures. So then was the idea if you're staying here with this halogen bonded one, what happens if you're
37:06
annealing these layers to higher temperature and what happens then if the pro-mine is cleaving off, can we then go to kind of graphene like networks and we try that on two different surfaces and different precursors. So here is for example like a precursor which contains
37:25
nitrogen and this is a pure carbon ring, both now demonstrated here on the silver substrate. So if you anneal them at room temperature you already see like nice hexagonal networks, that's already good news, but if one is carefully looking then one always sees that
37:41
they're like round protrusions in between. And also from the dimensions one can extract that actually the molecules do not directly make a carbon-carbon bond, but what they like to do is the radicals try to catch atoms from the substrate and basically create organometallic
38:00
structures. So these are not yet the covalent carbon networks, these are organometallic networks. So this is in the end like the softer interaction depending strength is not so strong but still strong enough that we have kind of a good communication between the molecule building blocks. So we looked at the electronic structures, we found that some molecular states
38:27
and states on the silver atoms are actually shared. So for example if you're looking here at the conduction band plus one then you have it on the molecule here in green and
38:41
on the silver atom in red and on the next molecule also here shown in blue. And you can also see that if you're recording the IDV maps that you basically have like extended states. So that is nice, but also here if you're calculating the band structure in
39:04
gas phase and on the surface basically we have a charge transfer from the substrate and therefore the bands which are in the vacuum band structure around zero and would be actually very interesting are again shifted down and basically therefore we get the semiconducting
39:22
behavior of these networks on the gold substrate or on the silver substrate. So then we tried to look what is the difference on the two surfaces in the reaction to hopefully eventually get like a carbon-carbon network. We did that then on the gold compared to the silver and so what we found is that the reaction is slightly different on the gold. Usually we
39:45
have a preference for linear bonds so there are more atoms involved and on silver basically these radicals rather try to bind to the substrate. So this can be nicely seen if one is looking at the rim of the networks that looks different on gold and on silver.
40:01
And this has then implications for the reaction. So if you're looking basically on the gold substrate so the linear bond is that which would preferably do like the carbon-carbon bond and indeed you can find if you anneal the molecules to about 590k small patches of like covalently bond networks but it's really highly reactive so this triple bond rather
40:25
likes to decompose so therefore it's tough to get large domains of really carbon-carbon bonded networks. And on silver the bonding geometry is different it goes to the substrate so that makes it more difficult to make a bond between the molecules. There we basically stay at the
40:43
same temperature in the organometallic phase but this gives us the advantage that we can heat further without doing the carbon-carbon bond and therefore we can really anneal out the defects in these networks and we get really perfect order or like a
41:01
crystallization of this organometallic network. And we can also nicely show that there's really like a nice crystallization going on by observing the classical defects one would expect in a hexagonal system. So for example like in graphene you observe like the 5-7 defects along the main boundaries or 5-5-8 rings which you also observe along the main boundaries.
41:27
So this is a nice example that is not covalently linked so by a carbon-carbon bond. So this is just organometallic bond but electronically we also see that we have a relatively good communication in the spectroscopy so we have shared state on the molecule and on the network
41:45
but because the interaction energy is less than a carbon-carbon bond we have actually the properties to really get well-oiled structures over larger sizes. So basically what was hindering here the size of the networks is then the steps on the substrate. So you can get
42:04
here easily like 100 or 200 nanometer large networks without any defect. So this has advantage and this advantage is not a carbon-carbon bond but nevertheless it has good electronic properties and you have instead the reversible structure formation which gives you good
42:23
crystallization. So that's another way to go to look in these organometallic networks to create designer made two-dimensional materials. And just to give an outline by a material like this would be interesting. Again basically depending on what we are looking at
42:43
we have hexagonal network or a Kagomi lattice when we are looking for example at the gold atoms or the silver atoms depending on which surface we are. So that gives us again band structures where we have linear crossings around the decay point and if you are looking even more careful
43:03
so these are gas phase calculations we get around the Fermi level also other band crossings. So for example here and here and so that could be for example also suggesting that such networks could potentially be while semi-metals. So that is interesting and
43:23
I think there is a lot more to do with like create using the molecular design to create lattices which then feature very interesting electronic features. So this brings me to the end of the presentation. I would like to thank my team and the collaborators
43:44
for all this great work. So for the chemistry we obtain molecules from Milan-Kiwala, Konstantin Amshaw and Albert Jucks and the theory is in support with the group of Andreas Goering and Bernd Meyer. And I thank you for your attention.