Issue



Carbon nanotubes promise a simple approach to making gas sensors


09/01/2008







by Roger Allan, contributing editor

Low cost is imperative when manufacturing any kind of sensor???particularly one used in mass-market applications such as automotive, home and consumer electronics???and this means keeping the sensing structure as simple and practical as possible. Researchers at YTC America Inc., Camarillo, CA, a wholly owned subsidiary of Japan’s Yazaki Corp., think they have a good idea based on strands of micrometer-sized multiwalled carbon nanotubes (MWCNTs). Their work was also discussed at the recent Sensors Expo & Conference (June 9-11, 2008, Rosemont, IL), and at this year’s and last year’s Materials Research Society meetings.

Their work, ongoing for about three years, involves assembling and testing a simple two-electrode ionization gas sensor, which has yielded good performance results. They believe this sensor may find main applications in factory and dwelling facilities as well as automobiles, for detecting nitrogen oxide (NOx) gases such as nitrogen dioxide and other binary compounds of oxygen and nitrogen.

The sensor is essentially assembled as a sandwiched structure of an anode and a cathode which break down gases within these structures using a high voltage. One electrode consists of a conducting substrate, which can be either silicon or metal, on which CNTs are grown. The other electrode is a covering conducting plate separated from the substrate by insulators.


Fig. 1: The gas sensor’s configuration consists of a gas manifold (a), the electronics to generate the high voltages (b), and the sensor electrodes and a test chamber (c).
Click here to enlarge image

So far, the sensor has been able to detect gases in one part per million (ppm) or better and has done so in <50ms, a response time fast enough for many applications, including detecting corrosive gases such as NO2.

“The biggest challenge this type of ionization gas sensors has is selectivity,” said Dr. Alex Moser, a senior member of YTC America’s technical staff (other collaborating technical staff members included Robert Gump, Dan Niebauer, and Leonid Grigorian). “They can be made to be very sensitive, but do not have the selectivity levels needed for many applications,” though that is something they are investigating and hope to improve on, he added.

In their work, the researchers discovered that altering the areal density of CNTs could enable observation of breakdown behavior for different gas mixtures, and thus improving the sensor’s selectivity. For example, three different electrode pairs will each respond differently to a different gas mixture, depending on the spacing between the CNT elements on a given substrate.

Because it is an ionization sensor, it requires a high voltage to ionize the gases it is detecting (see Figure 1). Normally voltages of more than 1kV are needed for gas ionization sensors. The researchers also hope to lower the normally high voltages needed for ionization-type sensors down to =100V, potentially allowing for sensor operation from a battery. This can be achieved by optimally narrowing the dielectric space between the two anode and cathode electrodes where the gas is ionized.


Fig. 2: Instead of measuring the current at the breakdown voltage as done in previous work, the shift in the breakdown voltage (∆Vb) is proportional to the analyte concentration [X] in the host gas.
Click here to enlarge image

A lower voltage means that the sensor may be able to operate from a battery. During the Sensors Expo presentation, they showed a graph that provided gas ionization breakdowns around 500V-600V, at a vacuum level of 200Torr (see Figure 2). The lower the vacuum levels, the lower the need for a high breakdown voltage.

“The approach we’re using has merits for use in detecting automotive NOx gases in many applications. It is a simple and robust structure that lends itself to good performance levels and low-cost manufacturing,” said Moser.

One of the applications cited was a car’s interior passenger compartment, as well a car’s exhaust system. The sensor could potentially be manufacturable at a cost on the order of $25 including the electronics, but that might take some 3-5 years to accomplish, Moser said. On the other hand, he noted that trucks can avail themselves of this type of sensor for about $50, because the higher cost is outweighed by the advantages gained by trucking applications???for example, the mean time between failure requirements is substantially longer than in the consumer automotive application.

Exploiting the inherent advantages of advanced CNT materials is the company’s main target of their CNT research effort, with the goal of developing commercially viable products coming out of this work.

For further information, contact YCT America Inc., 3401 Calle Tocato, Camarillo, CA 93012, www.ycta.com.


Roger Allan is a 40-year electronics journalism veteran, and served as Electronic Design’s executive editor for 15 of those years. He has covered topics from semiconductors, components, packaging and power devices, to communications, test and measurement, automotive electronics, robotics, medical electronics, military electronics, and industrial electronics. His specialties include MEMS and nanoelectronics technologies. Email: [email protected]