Championed by the University of Surrey, SGTs can be made on the same thin-film processes as FETs, but are far less sensitive to the geometric errors inherent in printing and other non-lithographic fabrication techniques – because SGTs are hardly sensitive to drain-source distance, while FETs are extremely sensitive.
For this reason, advocates are proposing them for printed electronics and displays, particularly on flexible substrates.
Like for like, they also exhibit lower saturation voltage and higher gain.
The downside is that they are slower and can handle less current than similar FETs.
In a paper published in Nature’s Scientific Reports, Surrey scientists led by Sporea modelled two-transistor inverters made 4µm source-drain gap poly-silicon FETs or SGTs – using models calibrated from real devices.
Operating with a 5V supply, SGT inverters showed improved gain and noise-margin.
“Coupled with the almost-standard fabrication process, this circuit robustness recommends SGT as reliable cost-effective implementations for applications such as remote sensing and data processing,” said the University.
According to Surrey, even though switching is slower than with FETs, the high gain of SGTs means the inverter spends less time in the linear region, where both top and bottom transistors are partially-on. “Switching power is lower, as is the power-delay product”.
Previously, along with Philips, the Surrey team has put a lot of work into SGT analogue circuits.
Gain, linearity and dynamic range are high.
The second characteristic makes them particularly suited to OLED pixel drivers where, said Sporea, organic FETs might need a 10V overhead to work well with a 2V OLED, while SGTs need only 1.5V. “The 10V can be reduced using sub-threshold FETs, but by the time you have added sub-threshold compensation you need 10x transistors in the pixel.”
OLEDs need current drive, which favours FETs.