Graphene was first isolated by Andre Geim's group at the University of Manchester in 2004 using an exfoliation technique that involves peeling off single layers of the material from graphite crystals using sticky tape. Although the researchers produced pristine graphene, the approach cannot be used to produce industrial scales of the material because it is incredibly time consuming and labour intensive and results in yields of just milligrams.

Although other, chemical, methods to produce graphene exist – for example, fragmenting oxidized graphite onto sheets of graphene oxide, which is then reduced to graphene with hydrazine – these always produce defective graphene. This is because the chemical processing disrupts the regular hexagonal carbon lattice in the material.

Harsh process
An alternative method is to split graphite into thin graphene pallets by applying ultrasound to the graphite crystals in solvent. Such sonication-assisted dispersion methods are inexpensive and simple and so could be used to produce large amounts of graphene according to some researchers. However, sonication is a relatively harsh process, says team member Elena Polyakova of Columbia, that might produce high local temperatures and pressures. Indeed, previous work has already found that graphene's conductivity decreases after sonication, most probably because of defects appearing in the sample.

To understand this process further, the team, led by George Flynn of Columbia, decided to study sonication-produced graphene in more detail using Raman, X-ray photoelectron spectroscopy (XPS), scanning tunnelling microscope (STM) and atomic force microscope (AFM) techniques. The researchers were able to visualize defects by STM for the first time and XPS helped to highlight the chemical nature of the defects.

The samples studied came from Andrei Geim's group and were made by placing crystals of graphite in a bath of dimethylformamide (DMF) and then sonicating these with ultrasound for several hours. Graphite is hydrophobic, which means it tends to clump together in water, but, in principle, the DMF allows it to "dissolve" into flakes. Next, the mixture is centrifuged for 10 minutes to remove thick flakes from the monolayer flakes of graphene that are then sprayed onto a glass slide. Finally, the slides are annealed for two hours at 250 °C in an atmosphere of hydrogen and argon gas.

Lost local bonding
AFM images of the samples showed that the majority of the graphene flakes still seem to be clustered together – a phenomenon that has already been observed for carbon nanotubes that tend to form bundles thanks to strong van der Waals attraction between the walls of neighbouring tubes. Raman microscopy, which is a fast, non-destructive way to measure graphene's thickness – as well as structural damage in various carbon materials – revealed a spectrum typical of disordered carbon in which the sp2 character of local carbon bonding was partially lost.

"All of these observations suggest that local disorder has been introduced into the graphene planes as a result of sonication," said Polyakova. "The graphene films also seem to be heavily contaminated with impurities that are found between the graphene layers." The defects can either be chemically bound to graphene or just intercalated when the film forms.

Too much oxygen
STM, which is good at revealing how crystalline graphene films are, also showed strong local buckling in the graphene sheets. Finally, XPS studies revealed that the graphene films contain a lot of oxygen. This finding was surprising because no deliberate oxidation was actually involved during the sonication process and, normally, such a rich oxygen signal is only found in graphene oxide that has been subjected to harsh acid treatment to separate the graphene planes. Polyakova believes that it is this oxygen that is responsible for reducing the conductivity and increasing the resistivity of sonication-produced graphene. The oxygen most probably comes from carboxyl, hydroxyl or epoxy groups (that contain oxygen) attached to the sp2 planes.

"Graphene films produced by sonication have more defects than expected," she told "We conclude that direct methods to produce graphene such as chemical vapour deposition would be much better suited for applications that require low resistance and high conductivity."

Polyakova also tells us that this was her last project at Columbia. She has since founded Graphene Laboratories Inc, a company that sells various graphene materials via

The present work was reported in ACS Nano.