The influence of edge structure on the electronic properties of graphene quantum dots and nanoribbons

Kyle A. Ritter, Joseph W. Lyding (2009). The influence of edge structure on the electronic properties of graphene quantum dots and nanoribbons Nature Materials, 8 (3), 235-242 DOI: 10.1038/nmat2378

Characterization/electronic properties testing paper. These authors did mechanical exfoliation of graphite, but used a silicon wafer (Si(100)-2X1:H) instead of SiO2 and made sure to pulverize their sample as much as possible. They were able to find both graphene quantum dots (GQD, basically small square or round pieces of graphene) and graphene nanoribbons (GNR's). They then used Scanning Tunneling Microscopy (STM) and it's cousin Scanning Tunneling Spectroscopy (STS) to determine different pieces morphology, edge state (zigzag vs. armchair), and band gap (HOMO-LUMO gap, for us chemists; a band gap of 0 means the material is metallic). Their results suggest that zigzag GQD's are metallic, while armchair GQD's are semiconducting. The nanoribbons, it turned out, were semiconducting no matter what the edge state, but zigzag ribbons had a lower band gap (0.14eV, meaning they were more conductive) than armchair ribbons (0.38eV), as expected.

The large disparity in conductivity between edge states in carbon nanotubes is a big reason why they didn't pan out to be the wonder materials everyone thought they would, since most synthetic procedures gave a mixture of zigzag and armchair (metallic and semiconducting) nanotubes. This paper shows us that GQD's would have exactly the same problems. Although the difference is not as big in graphene nanoribbons, it would still be very difficult to build a device where some of your connecting elements would be twice as conductive as others.

Moral of the story: Zigzag-edged graphene structures are more metallic than armchair structures, with zigzag GQD's being metallic, zigzag GNR's semiconducting, and armchair GQD's and GNR's semiconducting.

This article also got some press, which (in my humble opinion) gets the quantum dot and nanoribbon results a little bit mixed up:
EE Times
Science Daily

EN#7

Chemical Modification of Epitaxial Graphene: Spontaneous Grafting of Aryl Groups

Elena Bekyarova, Mikhail E. Itkis, Palanisamy Ramesh, Claire Berger, Michael Sprinkle, Walt A. de Heer, Robert C. Haddon (2009). Chemical Modification of Epitaxial Graphene: Spontaneous Grafting of Aryl Groups Journal of the American Chemical Society, 131 (4), 1336-1337 DOI: 10.1021/ja8057327

This paper detailed attempts to modify the surface of epitaxial graphene with a diazonium salt, shown below in the paper's graphical abstract:


This reaction has been used before to modify carbon nanotubes, and is a very logical next step to tuning the electronic properties of graphene (since the reaction essentially creates defects which lower the conductivity). The authors proved they actually functionalized their graphene by IR (which clearly showed the N02 group) and XPS (giving a different C1s region and a brand new peak in the N1s region). The resulting material was about half as conductive as the starting material and the conductivity showed a much greater dependence on temperature.

Moral of the story: if you want semiconducting graphene (which would be very helpful in making actual devices), one way to get it is to modify the surface of the graphene with plain old organic chemistry.

Flexible Graphene Films via the Filtration of Water-Soluble Noncovalent Functionalized Graphene Sheets

Yuxi Xu, Hua Bai, Gewu Lu, Chun Li, Gaoquan Shi (2008). Flexible Graphene Films via the Filtration of Water-Soluble Noncovalent Functionalized Graphene Sheets Journal of the American Chemical Society, 130 (18), 5856-5857 DOI: 10.1021/ja800745y

This quick communication features a standard Hummers-style graphene oxide (GO) reduction, but uses a water-soluble pyrene derivative (1-pyrenebutyrate, PB-) as a stabilizer to keep the GO soluble in water after hydrazine reduction. They find that depositing films from this solution gives a PB-/reduced graphene complex that's one layer (1.7nm) thick in some regions. Filtering the solutions instead of depositing them gives a black, flexible film; one 30 micrometer thick film, for example, had a tensile strength of 8.4 MPa with a modulus of 4.2 GPa. The conductivity of these films was around 2X10^2 S/m, similar to our previous GO paper and about 7 orders of magnitude higher than the pre-reduced GO. The authors also use this material in some TiO2 solar cells and claim significant improvement.

Moral of the story: this paper is another incremental step in finding a nice way to make reduced GO films by using different stabilizing/solubilizing agents

EN#32

Synthesis of graphene-based nanosheets via chemical reduction of exfoliated graphite oxide

Sasha Stankovich, Dmitriy A. Dikin, Richard D. Piner, Kevin A. Kohlhaas, Alfred Kleinhammes, Yuanyuan Jia, Yue Wu, SonBinh T. Nguyen, and Rodney S. Ruoff (2007). Synthesis of graphene-based nanosheets via chemical reduction of exfoliated graphite oxide Carbon, 45 (7), 1558-1565 DOI: 10.1016/j.carbon.2007.02.034

This is a mostly characterization-based paper looking in more depth at traditional hydrazine-reduced graphene oxide (GO). After oxidizing using Hummer's method (sulfuric acid, potassium permanganate) and reducing with hydrazine, the researchers characterized the before-and-after transformation with:

SEM: showed crumpled up sheets of the reduced GO
Surface area: 466 m2/g for reduced GO, rather high but still way below "real" graphene (2620 m2/g)
Elemental analysis: C/O ratio of 2.7 before reduction and 10.3 after, which still seems like a lot of oxygen after reduction. The C/N ratio after reduction was 16.1, meaning a good bit of the nitrogen from the hydrazine ended up in the final product.
Water content: 25 wt% before reduction; 2.8% after
TGA: for GO before reduction, mass loss started below 100 C, but main mass loss came around 200. Reduced GO was thermally stable up to 800 C.
C13 NMR (MAS): GO showed peaks for epoxides, hydroxyls and carbonyls, as expected; these peaks were absent in the reduced sample.
XPS: GO had 4 peaks for "oxygenated" components. These 4 peaks were smaller in the reduced sample, but a 5th peak for a C-N bond appeared.
Raman: Best shown with the picture below, borrowed under the auspices of fair use:

The first spectrum is pristine graphene showing a sharp "G" peak at 1581 cm-1. Second spectrum is GO, showing a much broader "G" peak in addition to a new "D" peak at 1363 cm-1, showing disorder. Third spectrum is reduced GO, which interestingly shows an even larger "D" peak, implying that the reduced film might have a higher disorder than the GO.
Conductivity: GO was least conductive (on the order of 10-3 S/m), with the reduced GO being much more conductive (around 10^2 S/m), a conductivity close to that of graphite (around 10^3 S/m).
The authors also briefly discuss the mechanism of hydrazine reduction, but essentially say that any explanation they can come up with doesn't fully explain their observations.

Moral of the story: As we knew, "graphene" made from reduced graphite oxide can be useful, but has so many defects that it's properties are very different from pristine graphene.

EN#41

Electronic structure and band-gap modulation of graphene via substrate surface chemistry

Philip Shemella, Saroj K. Nayak (2009). Electronic structure and band-gap modulation of graphene via substrate surface chemistry Applied Physics Letters, 94 (3) DOI: 10.1063/1.3070238

This is a theory paper using DFT to determine the effect of the substrate on the properties of graphene. According to the paper's calculations, O-terminated SiO2 strongly interacts with a graphene layer, making divots in the graphene layer to give a higher surface roughness and resistivity (with a band gap around 0.35eV, similar to armchair graphene nanoribbons). Putting a second layer of graphene on top of this lowers the band gap some, but doesn't eliminate it altogether (0.10 eV, similar to zigzag graphene nanoribbons). If you passivate your SiO2 with some hydrogen, giving hydroxyl groups on the surface instead of dangling oxygen, you get very little interaction with graphene. This graphene would still be metallic (band gap = 0) . The moral of the story is to be mindful of what you put your graphene on, since even different forms of SiO2 can have a big impact on properties.

EN#39

Highly conducting graphene sheets and Langmuir–Blodgett films

Xiaolin Li, Guangyu Zhang, Xuedong Bai, Xiaoming Sun, Xinran Wang, Enge Wang, Hongjie Dai (2008). Highly conducting graphene sheets and Langmuir–Blodgett films Nature Nanotechnology, 3 (9), 538-542 DOI: 10.1038/nnano.2008.210

Sequel to Hongjie Dai's nanoribbons article. The researchers make "expandable graphite" (basically graphite oxide, GO) by pumping in sulfuric and nitric acid. They expand/partially reduce it by heating to 1000 C, then throw in more sulfuric acid (oleum) and tetrabutyl ammonium hydroxide (TBA) in DMF to further intercalate stuff into the layers. Apparently, the graphite oxide approach works because the oxygen functionalities allow all kinds of stuff to sneak in between the layers.

Dai and friends then make films of their GO-esque material before and after annealing/reducing at 800 C in H2 gas. The films are compared to GO made with a more traditional Hummer's method by AFM, TEM, electron diffraction (ED), IR, and X-ray photoelectron spectroscopy (XPS). They find that their pre-annealed films had less oxygen functionalities, higher conductivity, and were more hydrophobic than the traditional GO. Post-annealing films of graphene (which they could deposit with organic solvents onto surfaces) were shown to have a lower incidence of defects (basically holes) than traditional reduced graphite oxide. Essentially, these authors found a more gentle way to make graphene from the graphene oxide approach.

EN#40

Graphene Literature Reviews Express

I just wanted to announce that Graphene Literature Reviews is being turned back on, but switching to a more Spartan format. I've still got to read and summarize after all, so I'll be posting much shorter summaries in the future, with minimal links, probably poor grammar, less explanation, and few pretty pictures. I hope you are still able to learn and enjoy.

Gone Fishin (or Synthesizing)

Hope everyone's been doing well out there. As you might have noticed, there haven't been any posts here for a few months; I'm pretty much done with blogging graphene literature for a while. I learned a lot about graphenes and about reading/understanding literature, and I appreciate everyone coming to check it out. If anyone would like to use my platform to write similar research reviews, please feel free to email me at RobWtzl@gmail.com.

Thanks again to everyone who came and read some stuff, and thanks especially to the kind folks at Carbon-Based Curiosities, Research Blogging, The Chem Blog, and Blogger for providing publicity, inspiration, and technical support.

Highly oriented monolayer graphite formation on Pt(111) by a supersonic methane beam

Highly oriented monolayer graphite formation on Pt(111) by a supersonic methane beam.

Hirokazu Ueta, Morihiko Saida, Chikara Nakai, Yoichi Yamada, Masahiro Sasaki, and Shigehiko Yamamoto

Surface Science, 2004, 560(1-3), 183-190.

Ok, so really I just wanted to post this because the authors use a supersonic methane beam. I mean, are you kidding me? Supersonic? Methane? Beam? I want one, even if all it does is leave a thin layer of graphite upon mine enemies.

So let's cut to the chase- the background portion of this paper mentions that "single layer graphite" is made by thermal decomposition of hydrocarbons on a metal surface (usually using CVD)- I posted on this before, and there are supposedly papers going back to 1975 that have done similar things. This is how one makes the fancy graphite used for "Scotch tape" graphene exfoliation, and closely related techniques are used to make carbon nanotubes. The authors of this paper mention that the graphene layers made by this method are often quite heterogeneous and that other carbon structures are present, making characterization difficult. Their solution is to use kinetic instead of thermal energy to convert hydrocarbons (methane) into graphene.

So, a quick word on their setup: they shoot methane onto a substrate that's connected to a mass spec, then take the substrate out and take some STM images. Here's their schematic of the device:


So now you can go home and build one yourself. After blasting their platinum substrate with methane, they look at it with STM (using the local tunneling barrier height technique, LBH, which I'll skip for now) and, as expected, find a Moire pattern, suggesting few- or single-layer graphene. They make two major finds:

First, their domain size is much larger than in contemporary (2004) thermal graphitization techniques, meaning they get a more homogeneous film. For this to happen, the graphene layers must be reasonably mobile on the surface. Graphene made with higher kinetic energy methane was the most homogeneous.

Second, they imply that their graphene layers actually grow over the step (like a blanket on a pillow; a similar effect was found in the previously mentioned post) instead of stopping and starting on steps.

What does these two things have in common? The authors suggest that the mobility of the graphene layers and their independence from surface features mean that the film doesn't interact very much with the metal surface, both during graphitization and afterwards. This is a good thing, since surface defects won't be translated into the final product, and the smaller interaction may also give different electronic (or catalytic) properties. The authors finish with a discussion of how the kinetic energy they use effects the film, and also make a comment about how the graphene must be distorted to give a lower-energy conformation (this seems a bit dated), but we'll wrap it up knowing that our graphene just has a lower interaction with the metal surface.

All in all, the graphene layers made here are not very uniform (compared to current 2008 techniques) and probably won't be used for industrial purposes. It's still a nice paper that uses a technique equal in coolness to sharks with laser beams on their heads.



UETA, H. (2004). Highly oriented monolayer graphite formation on Pt(111) by a supersonic methane beam. Surface Science, 560(1-3), 183-190. DOI: 10.1016/j.susc.2004.04.039

I Need a Physicist

I have a confession to make- I'd like to know some graphene physics.

I can't tell a Dirac fermion from a Klein paradox.

I don't know about the ambipolar field effect or the K' point of the graphene band structure.

I'm not sure why one would need a low interface trap density, and if I was trapped in the Brilllouin zone, I'd never get out.

Can anyone recommend any good resources for this information, written at the level of a knuckle-dragging organic chemist? I've been unable to find any good primers on the subject, and every introduction I read names some effect I have no idea about. If anyone with some knowledge in the field would like to guest post (or open their own blog), that would also be great. You can leave feedback in the comments, or email me RobWtzl@gmail.com if you have any information as to the whereabouts of my physics blindspots.

A study of graphenes prepared by different methods: characterization, properties and solubilization.

A study of graphenes prepared by different methods: characterization, properties and solubilization.

K. S. Subrahmanyam, S. R. C. Vivekchand, A. Govindaraj and C. N. R. Rao

J. Mater. Chem., 2008, 18, 1517 - 1523

This paper is from a special theme issue in Journal of Materials Chemistry on carbon nanostructures; if you're in to that sort of thing, check it out here. There's also a feature article by organic "graphenes" guru Klaus Mullen in there that I'll be reviewing soon, but you should go ahead and read it yourself since it's pretty nifty. I love the concept of the paper we're reviewing today; it's a head-to-head comparison of graphene made by different methods. Unfortunately, two of the three methods they use are pretty obscure, and the authors do not test "Scotch-tape" style graphene. Fortunately (and accidentally), I've already reviewed the relavent methods.

The paper evaluates graphene made by the dreaded camphor method (they call this CD), the well-respected method of oxidizing graphite to graphene oxide (EG, since they make a big deal of exfoliating it once it's oxidized; to me that's too easy to confuse with our Scotch tape method), and the circa-2003 method of thermally converting nanodiamond (they call this DG). I'd like to point out that that last sentence is more self-referential than most Wikipedia articles, and I don't care who knows it. The authors also mention graphene made from arc evaporation of SiC, but only take TEM images of it.

Anyway, this is an amazing amount of data, so I'll just break this down by characterization method and skip the barely-mentioned TEM.

X-Ray Diffraction (XRD): Gives number of layers (two of the samples gave two different sets of layered structures) and size of the crystallites:
CG: 51 layers, 6.1 nm crystallites
EG: 3 and 16 layers, 4.7 nm crystallites
DG: 6 and 87 layers, 5.0 nm crystallites

AFM: Also number of layers:
CG: 20
EG: 3-6
DG: 3-6

Raman: Gives multiple sizes of crystallites, in nm:
CG: 7, 10, 12
EG: 4, 6, 7
DG: 3, 4, 5

Magnetic susceptibility, in emu g^-1:
CG: No (publishable) data
EG: -3.5*10^-6
DG:-4.4*10^-6 (both EG and DG show Curie behavior, which I can't find a good summary for)

TGA: supposedly gives the oxidation temperature, but I think in the graphene oxide (EG) case it might better reflect the decarboxylation temperature (in degrees C):
CG: 730
EG: 520
DG: 700

Surface Area (in m^2 g^-1)
CG: 46
EG: 925
DG: 520
Lit value for single-layer graphene: 2600

Hydrogen Uptake (at low pressure and temp, then high pressure and temp, in wt%):
CG: No data
EG: 1.38, 3.1
DG: 0.68, 2.5 (these are comparable with carbon nanotubes)
Department of Energy Target: 6.0

Electrochem:
CG: acts like basal plane of graphite
EG, DG: better kinetics than CG

Supercapacitor measurements, in F g^-1:
CG: No data
EG: 117 (supposedly good)
DG: 35


Then they chemically modified these things with the nitric/sulfuric treatment or with an amidation treatment to make them more soluble.

Well, there you have it folks. I'm frankly blown away that they put all of this data into one paper, and although I wish they would have included the epitaxial "Scotch tape" graphene, it was interesting to see that the nanodiamond (DG) sample had properties similar to the graphene oxide. If you're making graphene and want to see how your samples compare- here's a whole battery of tests to match up to. Expect this paper to get a lot of citations in the future.

Also, the camphor graphene pretty much sucked. Told you.



Subrahmanyam, K.S., Vivekchand, S.R., Govindaraj, A., Rao, C.N. (2008). A study of graphenes prepared by different methods: characterization, properties and solubilization. Journal of Materials Chemistry, 18(13), 1517. DOI: 10.1039/b716536f

Defect formation in graphene nanosheets by acid treatment: an x-ray absorption spectroscopy and density functional theory study

Defect formation in graphene nanosheets by acid treatment: an x-ray absorption spectroscopy and density functional theory study.

Coleman, V.A., Knut, R., Karis, O., Grennberg, H., Jansson, U., Quinlan, R., Holloway, B.C., Sanyal, B., Eriksson, O.

Journal of Physics D: Applied Physics, 2008, 41(6), 062001

And now, for something completely different. You might remember when I griped about a theoretical paper called "graphene nanoribbons with chemically modified edges"; apparently, those authors weren't alone. Many physicists (or P-Chemists) were excited enough about broken graphene to do a lot of calculations on the subject, since this may give metallic as opposed to semi-metallic properties (or something). Our current paper details an attempt to actually put holes in graphene sheets and check out a variety of the properties.

Let me begin by saying that the authors use a form of graphene I haven't encountered before, which they call carbon nanosheets (CNS). They use "radio frequency plasma enhanced chemical vapor deposition" to get a structure best described by the picture from the text:


The authors claim (or imply) that this structure will be an adequate analog for actual graphene sheets, and proceed to treat it with HCl for a few hours. They then did the same treatment on another sample using distilled water instead of HCl and took some SEM images to make sure they still had carbon nanosheets. The authors moved on to analyze the sheets with X-ray absorption spectroscopy (XAS, pdf summary from the University of Calgary here) which, in analogy to XPS, gives information on what kind of bonds are present in the sample (pi* C-C, sigma* C-O, pi* C=O, etc.). The authors find that, compared to a control, their acid-treated samples have a significantly higher peak that they ascribe to C-O bonds and a lower peak in the area attributed to C-C pi* and sigma* bonds. They theorize that the acid broke C-C bonds, which were replaced by hydroxyl groups, and back this up by mentioning the higher wettability (read: lower hydrophobicity) of the film and by XPS measurements showing a much higher proportion of C-O bonds (we've seen this before).

After establishing that acid does indeed add oxygen (and probably hydroxyls) to their films, the authors start doing some calculations of broken graphene layers (essentially calculating the situation where some C-C bonds are absent, which they say gives the same results as having pendant hydrogens where the bonds broke). I'll spare myself the details, but the energy states they find are consistent with their XAS measurements, and the calculations predict a metallic state in the graphene. They conclude that although they did not attempt to measure any metallic properties in the graphene, their calculations are probably valid since they correctly predicted their XAS results.

Maybe I'm missing something, but treating graphene/graphite with acid and finding that it oxidized isn't exactly new. The XPS and XAS data presented look very close to the relevant data for graphene oxide, and although it's surprising that graphene can be oxidized without a strong oxidant like HNO3, there's no reason to think oxidized defects introduced in this manner will be different than oxidized defects induced by more conventional acids. And although I'm out of my league here, I don't understand how they would get away with modeling the defects as simply breaking the C-C bonds (or using hydrogens), instead of having electron-donating hydroxyl groups present.

All in all, Coleman et al. did a good job communicating what they did, and I'm always glad when people try to test theory with wet lab research. These defects will have a huge role in modulating the electronic properties of graphene, and more work needs to be done on the subject. It just seems like the end result (and a lot of hard work, I'm sure) went into a study that told us what we already knew.



Coleman, V.A., Knut, R., Karis, O., Grennberg, H., Jansson, U., Quinlan, R., Holloway, B.C., Sanyal, B., Eriksson, O. (2008). Defect formation in graphene nanosheets by acid treatment: an x-ray absorption spectroscopy and density functional theory study. Journal of Physics D: Applied Physics, 41(6), 062001. DOI: 10.1088/0022-3727/41/6/062001