Large-Scale Growth, Clean Transfer, and Nanosandwiching of Low-Dimensional Nanomaterials
Sep 5, 2023
"Large-Scale Growth, Clean Transfer, and Nanosandwiching of Low-Dimensional Nanomaterials"
Josh Wood, Electrical and Computer Engineering
CNST Nanotechnology Workshop 2013
Well, good morning everyone. Thank you for giving me this opportunity to talk about some of our recent work
Involving the large-scale growth, clean transfer, and nano-sandwiching of low-dimensional nanomaterials. I'm Josh Wood. I work with Professor Joe Eisen
In the Department of Electrical and Computer Engineering
So the first talk, Dr. Mozeni gave a nice integration of 3-5 with graphene
And graphene's got a lot of interest lately because of the fact that it's a very good conductor
One of the best conductors that we know of. It has a great mobility of about a thousand
It's metallic
But it doesn't have a band gap and since it doesn't have a band gap, all this talk of
Replacing silicon-based CMOS with graphene is, well frankly, kind of a lot of talk. So you have to think about
What are alternative applications that graphene might be good for?
But a more possible, a more logical choice is using it as a membrane. Here is a touchscreen
This is a Samsung report
And in this membrane idea, gases, liquids, things like that are pretty much trapped in the membrane
So graphene as a membrane seems to be a more fruitful field with a better application suited specifically only for graphene
You can also think about other
Structures such as tall transistors and exotonic constates. I won't talk about these here
But these are other possibilities in which you can use graphene with say molybdenum disulfide, hexagonal boron nitride
and other low dimensional nanomaterials to try to achieve a goal
So what I'll talk about today is three main things. How we've improved the growth of graphene by chemical vapor deposition on copper
How we can transfer graphene and other related nanomaterials off of copper to arbitrary surfaces
And how we can nanoscan, that is layer, these low dimensional nanomaterials for different applications such as
integration of graphene with biomaterials
So on the first point, you need a large area synthesis method to make graphene, boron nitride, etc.
And normally the one of choice is chemical vapor deposition
We found in our recent report that single layer high quality graphene grows on the
1-1-1 surface, the hexagonally ordered surface of copper
by last matching diffusion. The other growth surfaces are more multi-layer graphene and more defective
And engineering the copper substrates is pretty imperative to control all the defective areas of boron nitride on copper
So there's been a lot of reports in the graphene community about how to
They get better quality films by chemical vapor deposition on copper
For example, you can polish the foil and seed growth in order to eliminate defects
You can get different shapes, different flowers, dendrites based upon the hydrogen gas you have in the chamber
In this case different single crystals or for a polycrystalline foil
So if you change the hydrogen flow rate, you can get a control of that
You can lower the amount of carbon feedstock and go to ambient pressure to try to get hexagonally ordered shapes due to the surface diffusion
But there's been very little discussion as to what will happen if you put it on these polycrystalline foils
What are the growth dynamics in the system? Can we control them? So what are we going to do?
Well, what we do in our study is we grow graphene on copper
We use standard crystal graphene to measure electron backscatter diffraction
It gives us the facets of the copper surface, this one, a 1-1-1 dominant surface
And some high-end surfaces like 3-1-0, 4-1-0, etc.
We do scanning electron microscopy
We find that there are islands that form on the high-end surfaces
And the 1-1-1 surface gives us a very high quality film
And the fact is that's due to this lattice match that exists on the 1-1-1 surface
But how do we know it's graphene?
Well, we use a standard large area method known as Raman spectroscopy
Which probes the local vibrational structure of the material
It's sensitive to the overall crystalline order of the material
There are two major bands that we're interested in
In this case, the G-band, which is the actual vibration of the copper
And then this 2-D band, which arises due to the electronic structure
And there's the value of the intensity of the band that the G-band should be about to
And if we look at different facets, we notice that the 1-1-1 surface gives us a value that's an overstatement
And surprisingly, the copper 1-0-0 of the square lattice
Gives us a low value for the 2-D-to-G ratio
Suggesting lower quality materials and also a higher layer number
So it's a special point spectrum
But whole populations give us more information about the entire sample
If we take a look at the amount of 1-1-1 surface
And what I mean by that is, if a step-high index surface has 1-1-1 facets
And 1-0-0 terraces or a high surface set
Then you take that, plot that out, and you find the 1-1-1 surface D
Gives you monolayer
You can also look at the constrain and doping level
Between the 4-1-1 and 5-1-1
Confirming the fact that 1-1-1 in copper is the best for growing graphene
The growth rate is also the highest
So you get high quality growth and also a high plastic growth rate
If you want high quality films, you need to engineer copper
To have a predominance of 1-1-1 texture surfaces
If you do a good job of controlling your growth
You get grain boundaries
And these grain boundaries are very good for scattering
It has a lot of strong carrier scattering
Which we proved at the atomic level
In this case, it's the scanning tunnel spectroscopy
It's the state at the grain boundary
So if you made a device
You need to be able to engineer these
By controlling the surface structure of the copper
Which in turn affects the graphene quality
We can also use other low dimensional nanomaterials
Like boron nitride
This is a growth process
We can also do Raman spectroscopy
And selected area electron diffraction
And boron nitride
We note that if we change the surface growth conditions
If we go to low pressure for hydrogen
That we get planar growth
We get a grain boundary boron nitride
If you go to high pressure for hydrogen
The growth dynamic changes dramatically
And if you go to atmospheric pressure
You get dendritic features
These are indeed insulating films
You want thinner films so you can engineer
Hatter structures
Which show that these films are insulating
They're streaky because this is an insulator
And you get a significant amount of band bending
But the band gap here for our CVD films
Is about 4.1 electron volts
We can also look at other nanomaterials
We add fluorine
They have a stoichiometry of about C4F
They have a regular hexagonal structure
In this case the last constant is about 2% larger than graphene
They have a 3.2 electron volt band gap
By angle resolved photomesh and spectroscopy
And we confirm that with scanning tunneling spectroscopy
We have the ability to combine graphene with other nanomaterials
The question is how do we get it off the copper surface
And that ties me into my next talk
Which is how to transfer these off the copper surfaces
And as I mentioned you have to remove that copper surface
To get to do any sort of reasonable application
And you want to transfer it normally to silicon dioxide
You can go to mica and other surfaces like graphene
The normal transfer method is using a polymer known as polymethyl methacrylate
It's an e-beam resist
If you don't really do a good job removing it, this is what you get
The copper surface is contaminated
And this is going to be bad when you want to make clean interfaces
So we've looked at other polymer supports like polydisphenolate carbonate
Also known as polycarbonate for large scale transfer
And these are clean to the atomic level
And these allow for us to do layering of different nanomaterials
So how do we transfer our films?
Well, we spin on our polymer support on the graphene on copper
We etch back the copper, we float it onto water
We transfer it to our substrate, we dissolve it in solvent
We have some PMMA residue, we do a forming gas anneal
It works pretty well, but there's still a lot of room to improve
One thing that comes out of that is you trap water
In this case we did heavy water with Fourier Transform Infrared Spectroscopy
We have the vibrational mode of heavy water
So this process encapsulates water and traps remnant PMMA on the top surface
Raman also shows the trapped water
And if we go down to the atomic level
In this case we have graphene draping over water layers
Using carbonate to determine how many water layers there are
We see that indeed, there's trapped water here
We published this in a recent report
We also look at different grain boundary effects
And a lot of larger water add layer layering
We notice the water is pretty much trapped there forever
It doesn't go anywhere, you can push it around a little bit
With the STM tip, and we can kind of move it around
The system is very stable
Which is very surprising, if you do a wet transfer process
You have to worry about this water being stuck at your interface
But I mentioned that the PMMA was also a problem
That there was remnant PMMA on the top surface
This is well known in literature
That if you try different annealing conditions and different dissolution conditions
You still don't really eliminate all the residue
Now if you switch the polymer to polycarbonate
This is due to the fact that the polycarbonate is from a different polymer family than PMMA
It's an aromatic condensation polymer
So perhaps the surface chemistries are playing a dramatic role
In how the transfer works
So how do we prove that it's cleaner with polycarbonate versus PMMA?
If we do a good job annealing, you get a really nice surface
The roughness here is basically the silicon dioxide's roughness
In this atomic force microscopy image
If you don't do a good job, you get these contaminants
If you don't do any annealing, this is what you get with PMMA
But if you don't do any annealing with polycarbonate, you basically have nothing
It's the pristine graphene
These are for the same growth, not the same time, everything's the same
So it suggests that the polycarbonate is much more of an ideal transfer support
We also look at a bunch of other different polymers
In this case polylactic acid, polyethylaldehyde
And bilayer PMMA and PC
We notice that these two are very brittle
They're not very good for transferring graphene
PC and PMMA are equally good in terms of how structurally sound they are
But if we look at x-ray photoelectron spectroscopy
We find that the amount of carbonyl and carboxyl residue from the PMMA
Is significantly worse compared against PC
So again, it's clean
If we look at the roughnesses over different samples
We notice that the polycarbonate transfer, again, is as clean as the anneal
If we do a scanning telemicroscopy image, in this case at 54 Celsius
We get the atomic resolution of the graphene lattice
We do electrical measurements
We notice that the on-off ratio is high
So PC versus PMMA is again a clean surface
And if we do a bunch of different statistics on these samples
Our mobility is low because of the low quality growth
But generally speaking, the polycarbonate has higher electron and hole mobilities
So this enables some new stuff
We can then transfer graphene on top of graphene
And we get this rich mosaic of water layering between the layers
We can fluorinate the bottom layer
And we can trap water between the two layers
And you get different sort of decoration of water
Due to the superhydrophobic substrate for fluorinated graphene
You can do the same thing with boron nitride
And here you get similar point-like defects
So now we're just starting to open the field of different layering
And looking at hydration and clean interfaces using this polycarbonate transfer method
So this is my final part of my talk
Where I'm going to mention how we've nano-sandwiched
Different low-dimensional nanomaterials for different applications
In this case, heterogeneous integration of biomolecules and graphene
So if we combine these together
We get these novel layered structures, HOMO and hetero structures
Graphene can act as a metallic conformal cover slip
And we can nano-sandwich proteins, viruses, DNA
And we can look at different hydration and temperature characteristics
And the preliminary scanning tunnel microscopy data
Indicates the possibility of doing atomic-level assessment
Of these nano-sandwiched biomaterials
Which has been an elusive goal for about two decades
So basically to summarize this
We want to combine things like DNA and mica
Low hydration characteristics, a rod-shaped virus known as the tobacco mosaic virus
The methyl-bonded domain protein
Combine them all together with different nanomaterials
And different substrates
And we can imagine any sort of combination which would be of interest
So before we actually make the nano-sandwich
We put down one layer of graphene
We put our, say in this case, tobacco mosaic viruses
And we measure how tall they are
So they're about a little shorter than we expect
But still they're roughly in the ballpark
We put a layer of graphene on top, they drop about four-fold in height
And we measure their static pressure on the overall structure
And that seems to suggest that these conditions are a little extreme
They might be good for modeling systems
Where proteins or viruses go very, very awry
We also again looked at proteins
In this case the methyl-bonded domain covalently linked to DNA
And we find that if we put it underneath
Between two graphene sheets, we have this weird hydration pattern
We have these craters, these areas of influence around the protein
These are reproducible across any number of samples
We see them for all of them
And the heights match up quite well with what we'd expect for the protein
And the methylated DNA attached to it
But we also prove that it is indeed proteins, amino acids
In this case by doing large area XPS, 400 microns
There's enough amino acids on the surface of the graphene nano-sandwich
For us to see the bonds corresponding for the amino acids
One thing that we notice is that when we heat these samples up
In this case this is the room temperature measurement
We go close to the melting point of the protein
In this case we're looking at this protein right here
We notice the shape of this doesn't change at 45C
We go to 50C and it still doesn't change
Which is very odd, we would expect them to start to denature
At these temperatures close to the melting point
And so we did some simulations to try to bear out exactly what was going on
In this case this is an MBD protein nano-sandwich between two graphene sheets
It moves around pretty violently and that's because it's immediately denaturing
As soon as it's encapsulated between the two graphene layers
If we look at it at elevated temperature
In this case, a very extreme case, 200C
The thing immediately denatures, which is not terribly surprising
You're four times higher than the melting point of the protein
So it suggests that the nano-sandwiching process
Puts a huge amount of hydrostatic pressure
On these proteins causing them to unfurl
As well as their hydrophobic core
To show that schematically, here we have a protein, the MBD
Nano-sandwiched between two graphene sheets
Which have also pi clouds
And the pi cloud is going to interact with the hydrophobic inner core of the protein
That electrostatic interaction plus the 1GPa hydrostatic pressure
Causes this thing to go inside out
And when it denatures like that
It unfolds and shows its protein exterior
This causes the hydrophobic shell to propagate
For about 100-200 nm beyond the extent of the protein
This changes the local configuration of the water in the vicinity of the protein
This basically means there's going to be less water there
It's going to be less dense, it's going to change its volume and configuration
So finally, I'm going to show a slide on the provisional STM measurements
That we've done on this case
Mica with Nutravitin, a different protein
With graphene encapsulated on top, with water trapped between
And we notice these tendril-like features
It's a little streaky because of the low degas conditions
But these features are unusual, they're not what we saw previously
For graphene on water on mica
So these tendrils seem to template to the actual NA proteins
As they've denatured under the hydrostatic pressure
We again get atomic resolution, you can see the honeycomb with the graphene here
And strangely you have these weird tendril features
That extend beyond and then become gravely unstable
Which suggests that this is much taller than the vicinity
To our knowledge, this is the first report of biomolecules
Entrapped in this way under graphene
Allowing us to understand some of the unusual hydration characteristics
That exist in the system
So I've talked about three things
How do we make better graphene, better nanomaterials
How do we cleanly transfer them, how do we sandwich them all together
To make unique structures that allow us to probe phenomena
And the three major conclusions is that graphene on copper
Grows best and fastest on the copper 1-to-1 surface
You can get clean transfer using polybisphenol-like carbon
As opposed to PMMA and other alternative polymer scaffolds
And that you can nanosandwich these low dimensional nanomaterials
With different biomaterials to enable unique hydration
At atomic scale studies
And with that, there's many people, this is a short list
That make this happen, our funding agencies
And you for your attention, thank you very much
Can you bring back that slide
With the six gray squares across
Six gray, oh you mean the SEM
Sure
Maybe
This one?
I think it was a slide with trapped water
Trapped water, okay
Now
Okay, the upper row, middle
Where water is, or the lower left
Are you able to estimate how many water molecules
Are there in that?
Within the protrusion, that's a layer three of water
There are four angstroms in thickness
So based upon
I don't know exactly the number of molecules off hand
That would assume that we know
Exactly the state of water underneath the graphene
Which is actually a point of quite a bit of controversy
It seems that since it's stable
It's not liquid, but since it's at room temperature
It should be ice, yet we're able to
Nanomanipulate it, so it's some sort of weird
Viscous water, we would need to know exactly what state
It would be to estimate the number of molecules
That protrusion, but we know it's the third layer of water
Based upon the roughness and the overall step heights
Yes
All these STS measurements
Are done at room temperature, we don't have
Our low temperature STM is under construction
So they're all done at room temperature
Typically with a tungsten tip
Minus three to three volts
Normally with constant
I think those were constant spacing STS
So we keep the height the same in those measurements
But I'd have to go back to the data set and double check
It's different from the DIDV
You would measure in physics
We apply a dither for those measurements
And we lock and detect the dither
Can you extract any information?
The electronic information, the band gap, etc
The band edges are very much legitimate
That tells us the local electronic structure
Is right underneath the tip
Provided you have atomic resolution
If your tip is in a bad state
The spectrum is going to be more suspicious
The higher the band gap
The ability for you to do variable spacing STS
In order to try to get that band edge just exactly right
With the fact that we have a copper support
The charging is less
Larger band gap materials are harder
To measure exactly the band gap
But we can at least get a conservative underestimate
Of what the band gap should be
So for boron nitride it was 4.1, that was low end
You talked about mobility
Those were worst case scenarios
That was pretty bad growth
And I'm doing the worst case
In order to demonstrate the merits of the polycarbonate
Over and against the PMMA
More typical mobilities are on the order of 3000
After we extract the contact resistance very carefully
Thank you
Thank you very much

