Scientists at the University of California, Santa Barbara, have devised an advanced field-effect transistor that could enable rapid detection of low-abundance biomolecules including proteins.
The device, called a tunnel field-effect transistor, or T-FET, offers significant improvements in speed and sensitivity over conventional FETs and could potentially detect proteins at attamolar levels, Kaustav Banerjee, director of UCSB's Nanoelectronics Research Lab and leader of the project, told ProteoMonitor.
Field-effect transistors consist of a channel through which charge flows, with the channel's conductivity regulated by a gate. In biosensing applications, the target molecules serve the function of the gate, which allows the presence of the molecules to be detected by changes in the device's electrical properties.
Because of their small size, high sensitivity, and relatively low cost of production, FETs have drawn interest for applications including genomics and label-free protein detection. A number of companies, including Life Technologies and Intel, are investigating their use for DNA sequencing (In Sequence 2/28/2012); and labs including the University of Copenhagen's Nano-Science Center are exploring their potential as tools for proteomics research.
Conventional FETs, however, are inherently limited in their speed and sensitivity due to their reliance on an injection mechanism called thermionic emission, Banerjee said. T-FETs, on the other hand, aren't reliant on this mechanism, which, he said, would in theory allow them to reach higher speeds and lower levels of detection.
The key, Banerjee noted, is how quickly a transistor can switch on and off, which is determined by a parameter called the subthreshold swing. The lower the subthreshold swing, the faster the switching. With conventional FETs, Banerjee said, "the best [subthreshold swing] that you can theoretically achieve is 60 millivolts per decade, a level, he said, that can be bested with T-FET technology.
"The sensitivity of a biosensor is exponentially dependent on the value of the subthreshold swing," he said. "So the lower you go, the higher the sensitivity."
"Imagine that I have a switch in which the current in the switch changes dramatically with a small change in the gate voltage," Banerjee said. And because, in biosensing applications, the gate function is performed by the target molecules, that means "that the small charge associated with [a small number] of biomolecules can now do the same thing for you."
"So any switch that you can switch on and off in a very abrupt fashion will be desirable for biosensing applications," he said. "That's where the advantage of the T-FET comes in. If you are studying proteins … you can do relatively error-free detection. You can also detect at extremely low concentrations, and also the time taken for detection is going to be much lower."
A T-FET device could be up to 10,000 times more sensitive than a conventional FET sensor, Banerjee said.
Banerjee and his team began researching T-FETs five years ago, interested initially in their application to low-power electronics. In 2009, he became curious about the technology's potential as a biosensor, he said, realizing that the same qualities that made it useful in electronics might prove advantageous for biomolecule detection.
"So we started looking into the biosensor side of it, and since then we have done lots of simulations," he said.
This month, the researchers published a paper in Applied Physics Letters in which they outlined the theoretical advantages of T-FET-based biosensors and detailed simulations demonstrating the technique's performance.
They have also, with collaborators in Singapore, managed to construct some T-FET devices, Banerjee added.
"It's not like this is some kind of utopian device that nobody has built," he said. "We have already produced a version of this device."
That version, however, "was designed for electronics applications," Banerjee said, and so the researchers are now working on a prototype for biosensing work.
"If you want to do biosensing you want to build in, for instance, microfluidic channels so that you can put in whatever biological fluids you want to test," he said. The device also needs to be functionalized with affinity reagents "that can attract whatever proteins you want to study and bind them in a stable fashion."
Banerjee said his team aims to have a prototype T-FET biosensor ready for use in the next three to six months, noting that "we are working very actively on that [effort.]" He said they had not yet decided whether they would start with DNA or protein detection.
"We could do either," he said. "The idea is that anything that has charge can be detected, and since all proteins have charge we could easily do proteins."
The researchers have filed an initial patent disclosure for the device and are currently putting together several proposals to obtain additional funding for its development. They are not currently in talks with any commercial outfits about the technology, but, he said, "we are very much open to collaboration."
"There's a lot of work on FET-based biosensors going on around the world, and we believe that this [T-FET technology] could lead to significant advances in all these fields – proteomics, genomics, point-of-care diagnostics," he said.