Fabrication of carbon nanoscrolls

Easy fabrication of carbon nanoscrolls could speed their use in nanotechnology applications

Carbon nanotubes have long been recognized as a promising material for the storage of hydrogen (read for background: "New carbon nanotube hydrogen storage results surpass Freedom Car requirements"). Back in 2003, researchers first synthesized carbon nanoscrolls – another carbon nanomaterial similar to multi-walled carbon nanotubes – that was reported to be promising for hydrogen storage

Graphene, multi-walled carbon nanotube, carbon nanoscroll

Carbon nanoscrolls (CNS) can be obtained by rolling up a graphene sheet into a tubular structure. In contrast to multi-walled carbon nanotubes (MWCNTs), with CNS one can vary the distance between layers, a property that might be crucial for gas storage applications. CNS are also expected to be useful in other nanotechnology applications, for instance in nanoelectronics, since they inherit some properties from both graphene and carbon nanotubes, e.g. high mechanical strength and carrier mobility. However, theoretical calculations also predict some unusual electronic and optical properties of CNS due to their unique topology.

"Although there are abundant theoretical predictions, few experimental investigations on CNS have been reported" says Kaili Jiang. "For instance, up to now its electronic structure remains experimentally untouched, either through optical methods or transport measurements. The challenges are how to fabricate high-purity and high-quality CNS and integrate them into devices."

Previously, several methods have been developed to make CNS. However, they were hard to control, difficult to purify, and the fabricated scrolls were found to possess poor morphology. Now, Jiang and his team at the Department of Physics and Tsinghua-Foxconn Nanotechnology Research Center at Tsinghua University in Beijing have developed a simple and effective technique for fabricating high-quality CNS. This work is based on the development of fabricating and characterizing single-layer graphene on substrate. It will be helpful for exploring both the physical properties and applications of CNS.

"During our fabrication of a graphene-based device using electron-beam-lithography, we found that some 2D graphene is likely to shrink to a 1D fiber-like structure when spinning the photoresist," Xu Xie tells Nanowerk. "At first we tried our best to avoid it because we actually had hoped to obtained 2D graphene devices. However, Professor Jiang thought the shrunk graphene might have interesting properties to explore, so we tried to fabricate it intentionally and then characterize it. Since we do not want to work with a photoresist (which is hard to clean and might affect the physical properties of graphene) we experimented with some chemical solutions. Finally we found that isopropyl alcohol (IPA) is a good candidate and it is the interaction between IPA solutions and the graphene surface that causes the graphene to roll up."

Xie, a graduate student in Jiang's group, together with Long Ju, is first author of a recent paper in Nano Letters that reports the team's findings (Controlled Fabrication of High-Quality Carbon Nanoscrolls from Monolayer Graphene). Xie and Ju outline the four steps they followed for fabricating the CNS:

1) First, graphene was extracted by mechanical exfoliation of natural graphite on a silicon wafer covered with 285 nm silicon dioxide and then immersed this chip with graphene into IPA solution. Now, the upper and lower surfaces of the graphene were in contact with IPA and silicon dioxide respectively, resulting in a surface strain in the graphene. This surface strain served as one of the driving forces to bend the graphene sheet.

2) Following the immersion into IPA solution, the edges of the graphene sheet were lifted up from the substrate by the surface strain. Then IPA molecules entered into the space between graphene and the substrate, which also facilitated the detachment of the graphene.

3) In the third step, the detached parts of the graphene curved due to perturbations in the solution environment. Once the free end touched another part of the graphene by chance, the π-π interaction of the overlapped parts decreased the total free energy of graphene, even though the curvature energy increased due to bending.

4) The scrolling process proceeded continuously in the final step, that is, the graphene became detached from the substrate and rolled up bit by bit until a scroll was formed with a decrease in total free energy.

"Based on this 4-step analysis, we optimized the scrolling process in three aspects" say Xu Xie and Long Ju. "Firstly, we found that pure IPA was too harsh for scrolling – if the graphene detaches from the substrate more quickly than it scrolls, direct folding might occur – but that we could affect the scrolling speed with different concentrations. Secondly, the shape of the graphene sheet is important; a long smooth edge was beneficial for scrolling and the sheets tended to scroll along this edge. And lastly, we noticed that the scrolling process is sensitive to contamination and that heavily contaminated graphene is difficult to roll. Therefore, in order to fabricate high-quality CNS, the graphene needs to be carefully protected"

a) Schematic representations of current conduction at low bias. For CNT, only the outermost shell (which is in direct contact with the metal electrodes) sustains current. For CNS, the current extends to the whole scrolled graphene layer. (b) Current-voltage (I-V) curve measured under ambient conditions showing the high-bias electrical characteristics of a CNS about 50 nm in diameter. The breakdown point, which shows the diameter thinning after breakdown, lies midway between the two electrodes.

The researchers also conduct an interesting comparison of the current conduction characteristics of carbon nanoscrolls and multi-walled carbon nanotubes.

"For MWCNTs, only the outermost shell participates in conduction at low bias; the inner shells can only couple to the external electrodes through a barrier composed of the outer graphene layers," explain the Tsingua scientists. "In the CNS, however, current flows through the whole scrolled graphene layer."

An interesting aspect of this work is that, since the CNS are fabricated directly on SiO2/Si substrate, it facilitates further device-making. Jiang, Xie and the team are also intrigued by the variations in properties that could be achieved by controlling diameter and chirality of the CNS. To that end they are planning further experiments to clarify the roles and morphologies of the surfaces and edges in the scrolling process.

Integrating nanotube

Integrating nanotube-based NEMS into large scale MEMS

The fundamental issue of large-scale carbon nanotube (CNT) device fabrication remains the biggest challenge for effective commercialization of CNT-based nanoelectronic devices. For CNT electronics to become a reality requires manufacturing techniques to simultaneously and reproducibly fabricate a very large number of such devices on a single chip, each accessible individually for electronic transport. Conventional nanotube growth and device fabrication techniques using chemical vapor deposition or spin-casting are unable to achieve this, due to a lack of precise control over nanotube positioning and orientation.

Researchers are working on exciting projects – for instance through dielectrophoretic deposition of CNTs (see "Carbon nanotubes to the rescue of Moore's law") – to bring CNT electronics closer to reality. Another recent example is a chemical vapor deposition (CVD) CNT growth method developed in Israel to achieve the integration of carbon nanotubes into micro-fabricated devices.

"Past studies demonstrated integration of suspended single-walled CNT into silicon based devices," Yael Hanein tells Nanowerk. "However, fabrication processes needed in order to place the tubes in the right place and to electrically contact them are not compatible with simple fabrication procedure commonly employed by the industry. We were motivated by the need to develop a simple method to integrate taut carbon nanotubes with standard micro-fabrication processes."

Schematic representation of CNTs bridging neighboring pillars.

Hanein is a senior lecturer at Tel Aviv University and a member of the university's Center for Nanoscience and Nanotechnology. The research in her group is focused on developing novel micro and nanofabrication techniques and systems. Previously, we have reported on her group's technique for making complex networks of carbon nanotubes that can be stamped onto circuit boards. The scientific core of her new work is the complex dynamics of carbon nanotubes during chemical vapor deposition growth.

"By understanding the mechanisms governing carbon nanotube growth and by exploiting their interactions with various surfaces, we were able to grow well aligned electrically-contacted single wall carbon nanotubes," she says. "The ability of carbon nanotubes to grow at specific positions pending proper chemical and mechanical design of the micro-fabricated device offer a simple approach to form devices which otherwise require very cumbersome fabrication approaches."

The team – which included Slava Krylov, a senior lecturer at the Department of Mechanical Engineering at Tel-Aviv University and graduate students Gabriel Karp, Assaf Ya'akobovitz, and Moshe David-Pur – utilizes the CVD growth of CNTs over pillar-patterned silicon substrates to facilitate the formation of devices with taut and aligned CNTs grown exclusively at desired positions with built-in electrical contacts.

"Our approach builds on common silicon-on-insulator microfabrication techniques combined with conventional thermal CVD of CNTs," explains Hanein. "We exploit the selectivity of CNT growth on different substrates – i.e. silicon versus oxidized silicon – to confine the CNT growth to particular locations. The interaction between the CNTs and the silicon surface affects the tube alignment and tautness. We circumvent the poor compatibility of many conducting materials with the CVD growth process by positioning the electrode material next to the growth site rather than directly under it. The contact between the electrode and the tubes is obtained as a result of the tube interaction with the device surfaces."

Hanein points out that they were specifically looking for a method which can be implemented in a manufacturing environment. Therefore, any approach relying on intensive manual work had to be ruled out.

"We decided instead to exploit the mechanisms that affect carbon nanotube growth to do the work for us. Specifically, by using the CNT's sensitivity to different materials and surfaces during the growth, a simple method to integrate suspended, taut and defect-free carbon nanotubes into complex devices became feasible."

One restriction of this technique currently is the inability to control the chirality of the CNTs. This results in great variations of the electrical properties of the tethered nanotubes. The scientists feel confident, though, that the variability between CNT devices can readily be overcome during the testing stage. "In fact" says Hanein, "Raman spectroscopy can readily be implemented to characterize the tubes and to identify their metallic or semi-conducting properties."

A test MEMS device with a single CNT bridging two electrodes. The fabrication process is a conventional silicon-on-insulator MEMS process, and the placement of the tube is achieved seamlessly during the CVD CNT growth

By using this fabrication process, CNTs can be integrated into future micro electro-mechanical systems and may serve as ultrasensitive force and displacement sensors with enhanced performance. The method is suitable for building various carbon nanotube based nano electro mechanical devices (NEMS). These devices could be used in highly sensitive linear and angular accelerometers, as well as flow and heat detectors.

The Israeli team have so far integrated carbon nanotubes into simple devices. The next thing they are planning to do is to tailor them into fully functional MEMS devices. According to hanein, they will first use specially designed test devices but the next challenge is to use real devices designed for operation with carbon nanotubes as displacement sensors such as MEMS accelerometers and gyros.

The team has published their findings in a paper in the July 15, 2009 online edition of Journal of Micromechechanics and Microengineering (Integration of suspended carbon nanotubes into micro-fabricated devices).