InGaAs HEMTs - the successor to CMOS?
A research team at the Massachusetts Institute of Technology (MIT) says that its improved InGaAs HEMT process could make the compound material a serious contender for advanced logic applications - if and when the silicon CMOS roadmap runs out of steam.
Continued shrinking of silicon MOSFET transistor size by advanced lithography has so far ensured that logic devices have kept pace with Moore s Law. However, a fundamental barrier to the continuation of this approach is expected to hit in 10-15 years and, as yet, a solution has not been found.
"Unless we do something radical pretty soon, the microelectronics revolution might come to a screeching halt," said Jesús del Alamo from MIT.
Del Alamo and colleague Dae-Hyun Kim say that although InGaAs has always been regarded as a very attractive material for use in logic transistor channels, HEMTs based on this compound have previously reached their scaling limit at a 100 nm gate length.
But in new research presented at this week s IEDM conference in San Francisco, Kim and del Alamo have discovered that the scaling problem can be overcome by using a much thinner InAlAs barrier layer. The device epilayer material was grown on InP substrates at the Singapore-based wafer foundry MBE Technology.
With a barrier thickness of 3 nm instead of 11 nm, their In0.7Ga0.3As HEMTs showed excellent drive current and gate delay characteristics, and a transconductance that scaled nicely with gate length.
Leakage current trade-off
On the down side, some of the devices featuring an even thinner layer of InAlAs (0.5 nm) suffered from excessive gate leakage. Despite this trade-off, del Alamo reckons that the gate leakage problem can be solved, and that with optimized contacts, InGaAs HEMTs could be "the technology of choice when the CMOS roadmap comes to an end".
Compound semiconductors aren t the only option, of course. Development of so-called "high-k" gate dielectric materials are thought by many to offer the best solution, but there is no guarantee that this approach will work.
As a result, many leading research teams are now trying to develop compound materials, which are intrinsically "faster" than silicon due to higher carrier mobilities, for logic applications.
Indeed, it is the silicon logic industry s biggest and best-known exponent - Intel - that has funded the del Alamo team at MIT.
Intel s director of transistor research and nanotechnology Robert Chau said that the MIT team s work showed some exciting results at the low supply voltage of 0.5 V. "[This] is a very important research milestone," Chau commented.
InGaAs HEMTs are just one of the III-V options for post-CMOS logic, however. Freescale Semiconductor in particular has been working on GaAs-based MOSFETs, where the main difficulty has been to develop a suitable oxide material for the gate of the transistor.
The Freescale team uses gadolinium in a GdGaO3 gate oxide compound for its MOSFET structures.
