Abstract
As semiconductor scaling approaches the sub-10nm regime, maintaining high drive current ($I_{on}$) while suppressing short-channel effects (SCEs) becomes increasingly difficult. This study proposes a novel Double-Channeled Nanotube (DC NT) Gate-All-Around (GAA) FET. By integrating an additional channel within the vacant inner core of a standard nanotube, the DC NT structure achieves a 64.8% $I_{on}$ boost over conventional NT GAAFETs and a 1.7x increase over Nanowire (NW) GAAFETs within the same physical footprint. Despite a slight increase in leakage current, the $I_{on}/I_{off}$ ratio is maintained or slightly improved, demonstrating superior electrostatic control.
Introduction
The transition from planar MOSFETs to FinFETs and eventually to Gate-All-Around (GAA) architectures has been driven by the need for better electrostatic control to mitigate short-channel effects (SCEs). While Nanowire (NW) and Nanotube (NT) GAAFETs offer excellent gate controllability, the drive current density remains a bottleneck for high-speed applications. Increasing $I_{on}$ typically requires widening the device, which conflicts with the relentless push for higher integration density. This paper explores a "Double-Channeled" approach to maximize current density without expanding the device footprint.
Device Architecture and Simulation Methodology
The researchers compared three primary structures: the standard Nanowire (NW), the single-walled Nanotube (NT), and the proposed Double-Channeled Nanotube (DC NT). The DC NT design leverages the inner space of the nanotube—which usually contains only gate material or isolation—to house a second concentric silicon channel.
Simulations were performed using Technology Computer-Aided Design (TCAD) with a gate length of 50nm and a nanotube thickness of 8nm. The models incorporated high-k dielectrics ($HfO_2$) and accounted for quantum confinement effects using a quantization gradient density model. This ensures that the simulation accurately reflects the physics of carriers at the nanometer scale, where wave-like behavior becomes significant.
Performance Evaluation
The primary advantage of the DC NT structure is its significant boost in drive current. By effectively doubling the available channel width within the same footprint, the DC NT GAAFET provides a substantial performance leap.
As shown in the transfer characteristics, the DC NT structure (red curves) consistently outperforms both the standard NT (blue) and NW (black) designs. Specifically:
- $I_{on}$ Gain: The DC NT achieves a 64.8% improvement over the NT and a 1.7x improvement over the NW.
- Electrostatic Integrity: The Subthreshold Swing (SS) remains nearly constant at approximately 61 mV/dec for both NT and DC NT structures, indicating that the addition of the second channel does not compromise the gate's ability to switch the device off.
- $I_{on}/I_{off}$ Ratio: While the leakage current ($I_{off}$) increases by 61.8% due to the increased channel volume, the massive $I_{on}$ gain results in a net 2.4% improvement in the $I_{on}/I_{off}$ ratio compared to the single NT.
Physical Insights
The performance boost is not merely a result of increased channel area. The nanotube geometry offers higher electron mobility due to reduced surface scattering compared to the solid nanowire. Furthermore, the "gate-all-around" nature of both the inner and outer channels in the DC NT structure ensures that the entire silicon volume is efficiently modulated by the gate electric field. This leads to a more uniform current distribution and suppressed short-channel effects compared to footprint-matched nanowires.
Conclusion
The Double-Channeled Nanotube GAAFET represents a promising path for scaling beyond the 10nm node. By utilizing the internal volume of the nanotube, it breaks the traditional trade-off between device footprint and drive current. While fabrication remains a challenge, the theoretical benefits—particularly the 65% boost in $I_{on}$ with maintained electrostatic control—make it a compelling candidate for future high-performance logic and Complementary FET (CFET) architectures.
Summary updated