Boron nitride is a promising path to practical graphene devices

May 30, 2011 by Paul Preuss
Graphene (top layer) is a hexagonal arrangement of carbon atoms. Hexagonal boron nitride is a similar arrangement of boron and nitrogen atoms whose lattice constant is just 1.7 percent larger. Boron nitride’s attributes make it an excellent substrate for preserving graphene’s intrinsic properties.

( -- Graphene is a two-dimensional honeycomb of carbon, just one atom thick, whose intriguing electronic properties include very high electron mobility and very low resistivity. Graphene is so sensitive to its environment, however, that these remarkable attributes can be wrecked by interference from nearby materials. Finding the best substrate on which to mount graphene is critical if graphene devices are ever to become practical.

Groups led by Michael Crommie and Alex Zettl, scientists in the Division at the U.S. Department of Energy’s Lawrence Berkeley National Laboratory (Berkeley Lab) and professors of physics at the University of California at Berkeley, have joined forces to examine the best candidates for preserving ’s intrinsic properties. Results of their  research on graphene’s interaction with a boron nitride substrate recently appeared in Nano Letters.

“Any substrate influences the properties of graphene, so if you want to study its intrinsic properties the best way is to work with suspended graphene,” says Régis Decker, a former postdoctoral fellow in the Crommie group, now at the University of Hamburg, Germany, and lead author of the Nano Letters report. “However, suspended graphene is quite unstable when investigated with scanning probe techniques like scanning tunneling microscopy” – STM – “because the graphene membrane can vibrate under the tip. So the idea is to find a substrate that mimics the case of suspended graphene.”

A group based at Columbia University reported, in October 2010, that graphene supported on a boron nitride (BN) substrate had dramatically better than graphene mounted on the most common semiconductor substrate, (SiO2).

The researchers deposited boron nitride flakes on a layer of silicon dioxide, grown on a layer of doped silicon. The doped silicon was used as a gate electrode for doping the graphene during scanning tunneling microscopy. Graphene was applied to both the boron nitride flakes (under the STM tip) and the bare silicon dioxide; the graphene (dark and light purple) was grounded by an electrode of gold/titanium (gold). The STM could scan across both substrate systems.

“The Columbia group showed that the mobility of electrons in graphene on boron nitride is much better than graphene on silicon dioxide, but there were many questions that their macroscopic measurements didn’t answer,” says the Crommie group’s Yang Wang, co-lead author of the Nano Letters report. The Crommie and Zettl groups compared the two systems side by side to find out just why boron nitride works so well. “To investigate BN on the atomic scale we used STM to build up a picture of the topography of the system and measure its local electronic states.”

Searching for what makes boron nitride special

Says Decker, “For a graphene-substrate system to be able to mimic suspended graphene, the substrate needs a large electronic band gap and no dangling bonds, so as to avoid any change in the electronic structure of graphene. The substrate would also need to be very flat, as suspended graphene would be. Boron nitride is a good candidate because it fills these requirements.”

What first attracted investigators to boron nitride’s potential as a graphene substrate were its unusual structural properties. In its hexagonal structure (h-BN), alternating nitrogen and boron closely mimic the way carbon atoms are arranged in graphene. Boron and nitrogen atoms in BN compounds are paired equally, and together their valence electrons (three and five, respectively) equal those of a pair of carbon atoms (four each). Although the h-BN lattice is some 1.7 percent larger than graphene’s and not commensurate with it, the two honeycombs laid one on the other can be aligned much more closely than graphene on silicon dioxide. Unlike graphene, which normally has no band gap, h-BN has a wide band gap, due to the alternating boron and nitrogen atoms in its lattice.

To create graphene/BN devices, the Zettl group first reduced boron nitride crystals to tiny flakes by the tried-and-true method of “exfoliating” them between strips of Scotch Tape. The BN flakes were deposited on a layer of SiO2, which was grown on a layer of doped silicon which, in turn, was used as a gate electrode to tune the charge concentration – a way of “doping” the graphene layer above – during scanning tunneling microscopy.

The Crommie group’s Qiong Wu created graphene by means of chemical vapor deposition on copper; on copper, carbon atoms self-assemble into a lattice a single atom thick. The graphene sheets were transferred from the copper to soft plastic and then placed on top of the boron nitride flakes by pressing the plastic onto the BN. The whole assembly was annealed at high heat.

The graphene layer was grounded by depositing a titanium gold electrode on it. Three graphene/BN systems were made this way, ready for direct STM comparisons with graphene on silicon dioxide. The STM tip could scan across the graphene layer, measuring topography and local charge concentrations at various doping levels determined by the silicon-layer gate electrode.

Boron nitride versus silicon dioxide

“A couple of things were thought to interfere with electron mobility in graphene on silicon dioxide,” says Victor Brar of the Crommie group. “One is impurities that dope the graphene and locally alter the concentration of charges.”

A sure way to shorten the mean free path of electrons (or their positively charged counterparts, electron absences called holes) is to strew the path with obstacles known as charge puddles, which are fluctuations in local concentrations of charge. In graphene on SiO2, charge puddles are common.

“We had previously studied the properties of graphene/silicon dioxide systems in detail,” says Michael Crommie, “and showed that charge puddles aren’t caused by ripples or corrugations in the graphene sheet, as had been suggested, but rather by impurities below the graphene layer.”

Results of measuring graphene on a boron nitride substrate are at left, graphene on silicon dioxide at right. The STM mapped both the topography of the systems (back) and the local charge densities (front). Graphene on boron nitride is extraordinarily flat, and inhomogeniety of local charge statesis is signicantly reduced compared to silicon dioxide.

One source of those impurities could be foreign matter trapped between the graphene and the substrate when the graphene layer is applied. Tiny air bubbles or water molecules or other foreign matter could act as dopants.

“When we made the graphene on boron-nitride devices we looked for atmospheric impurities, but we didn’t see any evidence of their effects,” says Brar. “For making practical graphene devices, that’s good news, because it means they don’t have to be assembled in a vacuum.”

Another source of graphene doping and subsequent charge concentrations is dangling bonds in the substrate. A valence electron available for bonding with another atom is a recipe for chemical reactivity, and silicon dioxide has a high concentration of dangling bonds. Boron nitride, however, has no leftover electrons to form dangling bonds.

STM comparisons of the two systems vividly displayed the differences between them. Topographically, graphene on boron nitride is much less rough than graphene on silicon, with height differences on the scanned surfaces reaching only around 40 picometers (trillionths of a meter). Height differences with the silicon dioxide substrate were up to 30 times greater.

Electronically, variations in charge density were reduced dramatically in the BN substrate. Compared to the nearly unvarying values of the boron nitride system, graphs of the silicon dioxide systems resemble modern-art color-field paintings.

Finally, Decker says, “because its lattice constant is very close to that of graphene, theorists predicted that this induces a band gap in graphene, which would be interesting for applications” – if not for maintaining graphene’s intrinsic properties. The Crommie group investigated how might vary according to the orientation of the graphene sheet on the boron nitride substrate. The two, not-quite-commensurate lattices betrayed their alignment by exhibiting changing moiré patterns with different orientations.

Says Wang, “We saw many different alignments, including alignments that were almost perfect. But the graphene still showed no band gap.” In sum, how graphene is oriented on a boron nitride substrate makes no detectable difference in its excellent electronic properties.

Michael Crommie says, “The graphene/BN system is really much nicer than any other substrate for a range of applications. There are many fewer impurities, much less charge inhomogeneity, much less bumpiness, and much more stability – in all, a much cleaner environment for studying graphene’s intrinsic properties. is a really fabulous system for practical graphene devices.”

Explore further: Toward a better understanding of bilayer graphene

More information:

“Local electronic properties of graphene on a BN substrate via scanning tunneling microscopy,” by Régis Decker, Yang Wang, Victor W. Brar, William Regan, Hsin-Zon Tsai, Qiong Wu, William Gannett, Alex Zettl, and Michael F. Crommie, appears in Nano Letters and is available online at

After the Crommie and Zettl groups submitted their paper, related work by Xue et al appeared in Nature Materials, April 2011, available online at

Early results measuring the properties of graphene on a boron nitride substrate were published by Dean et al in Nature Nanotechnology, October 2010, available online at

Related Stories

Toward a better understanding of bilayer graphene

October 26, 2010

( -- "Graphene is a very exciting material with a number of interesting possibilities, including for use in electronic devices," Pablo Jarillo-Herrero tells "However, all graphene systems are electronically ...

Damaging graphene to create a band gap

November 22, 2010

( -- "Graphene offers a lot of interesting potential applications for nanoelectronics," Florian Banhart tells, "but there is no band gap. This is a well-known problem. Without the band gap, switching ...

Boron Nitride Nanotubes More Amenable Than Carbon

May 17, 2004

Carbon nanotubes get a lot of press attention, but boron nitride (BN) nanotubes might have superior properties. K.H. Khoo and his colleagues form University of California performed first-principles calculations on BN nanotubes ...

Two graphene layers may be better than one

April 27, 2011

( -- Researchers at the National Institute of Standards and Technology have shown that the electronic properties of two layers of graphene vary on the nanometer scale. The surprising new results reveal that not ...

Researchers make graphene hybrid

March 1, 2010

Rice University researchers have found a way to stitch graphene and hexagonal boron nitride (h-BN) into a two-dimensional quilt that offers new paths of exploration for materials scientists.

Recommended for you

Nano-decoy lures human influenza A virus to its doom

October 25, 2016

To infect its victims, influenza A heads for the lungs, where it latches onto sialic acid on the surface of cells. So researchers created the perfect decoy: A carefully constructed spherical nanoparticle coated in sialic ...

New method increases energy density in lithium batteries

October 24, 2016

Yuan Yang, assistant professor of materials science and engineering at Columbia Engineering, has developed a new method to increase the energy density of lithium (Li-ion) batteries. He has built a trilayer structure that ...

Nanofiber coating prevents infections of prosthetic joints

October 24, 2016

In a proof-of-concept study with mice, scientists at The Johns Hopkins University show that a novel coating they made with antibiotic-releasing nanofibers has the potential to better prevent at least some serious bacterial ...


Adjust slider to filter visible comments by rank

Display comments: newest first

1.5 / 5 (2) May 30, 2011
Boron nitride is a promising path to practical graphene devices.. which, I guess, is yet another way of saying that we'll probably never see any practical commercial product for any of these hundreds of technologies we see bandied about here all the time for years on end.
5 / 5 (2) May 31, 2011
You won't see it until it's understood and cheap to produce. Expect graphene applications to be mainstream after 2020. Silicon can carry us along Moore's Law until then.
5 / 5 (2) May 31, 2011
which, I guess, is yet another way of saying that we'll probably never see any practical commercial product for any of these hundreds of technologies we see bandied about here all the time for years on end.

Research always comes years (mostly even decades) before commercialization. This is a site which reports on research results.

Research is not like in the movies. It's hard, patient work that may or may not lead to results. and even then there has to be a favorable economic environment (and sometimes the political will) to bring these things to market. Stuff isn't just built because it's cool. Someone needs to realize that he can make a buck off it before we'll see any of this in houshold items.

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more

Click here to reset your password.
Sign in to get notified via email when new comments are made.