Modeling and analysis of circuit performance of ballistic CNFET

Bipul C. Paul, Shinobu Fujita, M. Okajima, Thomas C. M. Lee2006 · DOI

TL;DR

This paper introduces a new, circuit-compatible analytical model for ballistic Carbon Nanotube Field-Effect Transistors (CNFETs) that can be readily integrated into standard SPICE simulators. Crucially, it highlights that the superior intrinsic performance of CNFETs is severely hampered by parasitic fringe capacitances, which are strongly dependent on device width, suggesting that significant fabrication advancements are needed to realize their full potential.

What Problem Does This Paper Address?

As silicon technology approaches its fundamental limits, researchers are exploring alternative materials and devices for future integrated circuits. Carbon nanotubes (CNTs) have emerged as a promising candidate due to their unique electronic properties, including near-ballistic transport, which theoretically promises extremely high speeds. However, evaluating the real-world circuit performance of CNFETs is challenging because existing device models are often numerical and not directly compatible with widely used circuit simulators like SPICE. Furthermore, the theoretical speed advantage of intrinsic CNFETs might be overshadowed by parasitic effects in practical circuit implementations. This paper aims to bridge this gap by developing a practical analytical model for CNFETs and analyzing the impact of parasitic capacitances on their circuit performance.

Methods

The authors propose a "quasi-analytical" device model for intrinsic ballistic CNFETs. This model is derived by making approximations to the fundamental device equations, allowing for closed-form expressions that can be implemented in SPICE. The approach involves:

  • Approximating Carrier Density: The carrier density in each sub-band of the nanotube is approximated using a density-of-states function and Fermi-Dirac statistics.
  • Analytical Solution: The resulting integral equations for charge are solved analytically by dividing the operating conditions into "below threshold" and "above threshold" regimes and employing series expansions.
  • Polynomial Approximation: The charge (QCNT) is expressed as a polynomial of the channel potential, enabling the derivation of an analytical relationship between gate voltage (Vgs) and channel potential.
  • Parasitic Capacitance Analysis: The paper analytically calculates the fringe capacitance (Cfr.) of state-of-the-art CNFET geometries using established formulas for fringe fields. This analysis considers various geometric parameters like gate metal thickness, oxide thickness, and source/drain length.

Datasets

Not discussed in the paper. The authors compare their model's results against a "physical model" (numerical solution of device equations) and experimental data where available.

Key Insights and Ideas

  • Bridging the Simulation Gap: The development of a SPICE-compatible analytical model is a significant step towards enabling widespread circuit-level design and analysis of CNFET-based circuits.
  • Parasitics are the Bottleneck: The paper strongly emphasizes that the superior intrinsic performance of CNFETs (theoretically in the Terahertz range) is severely limited by parasitic fringe capacitances. These parasitic capacitances can be orders of magnitude larger than the intrinsic gate capacitance.
  • Width is King (for Parasitics): The analysis reveals that parasitic fringe capacitance is highly sensitive to the device width (W). To achieve the theoretical speed advantage, the width of CNFETs needs to be drastically reduced, which presents a significant fabrication challenge.
  • Geometric Optimization Limitations: Optimizing parameters like gate metal thickness, gate oxide thickness, or source/drain length has a limited impact on reducing fringe capacitance.
  • High-K Dielectrics Offer Hope: Using high-K gate dielectrics could potentially reduce effective gate capacitance by allowing for larger oxide thicknesses while maintaining good gate control.

Key Results

  • The proposed analytical model shows good agreement with the physical model for both below and above threshold conditions, as demonstrated in Figure 2.
  • The I-V characteristics predicted by the proposed model closely match experimental data and physical models, outperforming a previously proposed model (Figure 3).
  • A SPICE simulation of a 2-input NAND gate using the developed CNFET model shows a significantly faster intrinsic delay (2.52 ps) compared to conventional 45nm CMOS technology (165 ps).
  • However, when parasitic capacitances are included in the simulation, the delay of the CNFET NAND gate increases dramatically to 344 ps, highlighting the detrimental impact of parasitics.
  • The analysis of fringe capacitance (Cfr.) shows it can be as high as ~700 aF/µm width, significantly exceeding the intrinsic capacitance of ~1.5 aF for a 50nm gate.

Limitations and Future Work

  • Fabrication Challenges: The paper repeatedly points out that achieving the necessary reduction in device width (on the order of 10nm) for effective performance is extremely difficult with current lithography technology.
  • Interconnects Not Considered: The analysis does not account for interconnect delays, which could further limit circuit performance. The authors suggest that a quantitative analysis of interconnect impact is premature due to a lack of information on CNFET circuit layout configurations.
  • Future Work: The authors suggest that future work should focus on improving nanotube fabrication techniques to enable smaller device widths and exploring advancements in lithography processes. Further investigation into the impact of interconnects will also be crucial.

Summary updated

Abstract

With the advent of carbon nanotube technology, evaluating circuit and system performance using these devices is becoming extremely important. In this paper, we propose a quasi-analytical device model for intrinsic ballistic CNFET, which can be used in any conventional circuit simulator like SPICE. This simple quasi-analytical model is seen to be effective in a wide variety of CNFET structures as well as for a wide range of operating conditions in the digital circuit application domain. We also provide an insight how the parasitic fringe capacitance in state-of-the-art CNFET geometries impacts the overall performance of CNFET circuits. We show that unless the device width can be significantly reduced, the effective gate capacitance of CNFET will be strongly dominated by the parasitic fringe capacitances and the superior performance of intrinsic CNFET over silicon MOSFET cannot be achieved in circuit.
BibTeX
@misc{paul2006modeling,
  author = {Bipul C. Paul and Shinobu Fujita and M. Okajima and Thomas C. M. Lee},
  title = {{Modeling and analysis of circuit performance of ballistic CNFET}},
  year = {2006},
  doi = {10.1145/1146909.1147092},
}

Loading related content...

Figure