Research into nanomechanics and nanoelectronics is very fashionable, as evinced by the growing number of research groups working in these areas. The aim is the development of ultra-rapid, highly sensitive construction elements with very low energy output. The transition from the micro- to the nanoscale enables systems not only to become smaller, but also promises their wider application, e.g. in the life sciences. But however compelling the scenario, the transition is anything but trivial. The manufacture of sensor systems on a nanoscale, in particular, is still in its infancy.

ETH graduate students Christoph Stampfer, Thomas Helbling and Alain Jungen of the Chair of Micro and Nanosystems under Christofer Hierold are therefore doing pioneering work (1). They are investigating the conditions under which carbon nanotubes, with their special properties, can be produced and integrated into construction elements or systems. Together with research colleagues in Switzerland, Germany and the USA they have now developed a prototype for a pressure sensor. Details of its manufacture and tests carried out were recently reported in Nano Letters(2).

Curling direction determines properties

The sensor developed at ETH Zurich consists of an ultra-thin aluminium oxide membrane, to which the carbon nanotubes are attached. These nanotubes - macromolecules composed of carbon atoms - are the actual sensor elements. A nanotube may be imagined as a seamlessly-welded graphite layer rolled up into a cylinder. The cylinder walls consist of the hexagon typical of graphite. Depending on the spatial curling direction of the graphite layer, nanotubes with varying structures appear. The electrical conductivity and in particular the electromechanical properties are strongly dependent on this curling direction and on the diameter of the nanotubes. These can show both metallic and semiconducting behaviour.

The tubes typically have a diameter of a few nanometres and a length of some few to 100 micrometres. The nanotubes are manufactured either at very high temperatures (around 3000°C) in a carbon plasma produced, for example, by light passing between two graphite electrodes, or are precipitated at less high temperatures (around 900°C) in the presence of suitable catalysts.

But how can these painstakingly-produced nanotubes be used in a sensor? “When a difference in pressure moves the sensor membrane, it is also stretched. We assume that this stretching of the membrane is directly transmitted to the carbon nanotubes. Because of the major alteration in the nanotubes’ electrical resistor when the membrane is stretched, they can be evaluated and determined via electrical measurement,” explain Christoph Stampfer and Thomas Helbling.

Layering and etching

For the manufacture of the prototype the researchers combined conventional materials and microsystem processes with new methods developed to integrate nanotechnologies. As substrate material a 300-micrometer-thick silicon wafer was used. Aluminium oxide, the material used to cover the membrane of the pressure sensor, was then applied to the silicon base using a method known as “atomic layer deposition”. The great advantage of this procedure is its very good control over the membrane layer thickness, which in this case is 100 nanometres. This step in the process was carried out by partners at the University of Colorado at Boulder, USA.

Pressure sensor in mini-format: view of a sensor membrane (dark grey), in whose centre carbon nanotubes, themselves too small to be detected in the image, are electrically connected. Cf. footnote 2 for more information (Copyright American Chemical Society, 2006). large

To obtain the pressure sensor membrane, the scientists had to etch away part of the silicon wafer with photolithography, using methods of micromechanics. The membrane then attained a diameter of 100 micrometres. Further steps in the procedure led to the placement of nanotubes in the centre of the membrane and their connection to electrodes composed of a titanium-gold alloy.

The most difficult part in the entire process was the step-by-step integration of the nanotubes and the membrane, recall Stampfer and Helbling. The placing of the individual tubes, the production of electrical and mechanical contacts and the handling of the very thin and fragile structures had proved very exacting. “For several steps we therefore used the FIRST Lab, an ETH Zurich cleanroom laboratory. It offers excellent possibilities for combining nano- and microtechnologies to produce our prototypes (3).”

After manufacture the sensor naturally had to be tested. First the mechanical properties of the membrane and the movements brought about by pressure differences were measured. From these data researchers could calculate the location-dependent stretching of the membrane. This information was in turn used to evaluate the electromechanical properties of the nanotubes via simultaneous measurement of tube resistance. The scientists observed that the electrical resistance of the nanotubes was a function of the membrane stretching due to the pressure difference - meaning that the sensor worked as they had hoped.

Significant development potential for even smaller sensors

The pressure sensor uses, as electromechanical signal transformers, metallic carbon nanotubes, which with their sensitivity already slightly outperform other so-called ‘piezoresistors’ made of conventional materials. A commentator in the specialist journal Nature Materials writes of the ETH development: “This prototype sensor already performs to current state-of-the-art technical standards. The larger gauge factors ascertained for stretched nanotubes in other experiments seem to imply that their sensitivity can even be significantly improved.”

Future pressure sensors may not only be more sensitive,, but also much smaller, however. For according to Christoph Stampfer and Thomas Helbling they could significantly reduce the size of their sensor without any great conceptual alterations. A smaller version is only one of the goals that the ETH scientists have set themselves, however. In further experiments they wish to discover whether the carbon nanotubes can be integrated precisely and directly into the sensor membrane via a growth process. If they succeed in generating targeted growth this would be the basis for mass production of the tubes. Such an achievement would, with certainty, attract companies involved in sensor development and production.