Compact Carbon Nanotube Field-Effect Transistor Model (CCAM) is a semi-physical carbon nanotube compact model that accurately describes the shape of DC and small-signal characteristics of fabricated carbon nano-tube FETs (CNTFETs). The model allows, for a given gate length, geometry scaling from single-finger single-tube to multi-finger multi-tube transistors. The features include parasitics, ambipolar transport, dynamic behavior and trap model. The model shows excellent agreement with the data from both the Boltzmann transport equation and measurements of Schottky-barrier CNTFETs and has been implemented in Verilog-A, making it widely available across circuit simulators.

• CCAM Compact Carbon Nanotube Field-Effect Transistor Model 2.2.0 Verilog-A(VA | 28 KB)
• CCAM Compact Carbon Nanotube Field-Effect Transistor Model 2.2.0 Benchmarks(ZIP | 29 KB)
• CCAM Compact Carbon Nanotube Field-Effect Transistor Model 2.2.0 Parameters(ZIP | 19 KB)
• CCAM Compact Carbon Nanotube Field-Effect Transistor Model 2.2.0 Experimental Data(ZIP | 183 KB)
• CCAM Compact Carbon Nanotube Field-Effect Transistor Model 2.2.0 Manual(PDF | 377 KB)
• License terms

Researchers should cite this work as follows:

• Michael Schroter; Annamalai, M.; Max Haferlach; Martin Claus (2022). CCAM Compact Carbon Nanotube Field-Effect Transistor Model. (Version 2.2.0). nanoHUB. doi:10.21981/5E9F-2S90

  BibTex | EndNote

1. carbon nanotubes
2. CNTFETs
3. Compact Model

Major updates:

• Bugfix for discontinuity in linear empirical current model at VGiSi = 0 V and non-convergence in harmonic balance simulation at large input power.
• Re-implementation of thermal network and self-heating mechanism and a flag is added in order to switch on and off the self-heating effect  (similar to the implmentation in HICUM/L2 ver. 2.4.0).
• Implmentation of width and tube density scalable model.
• A type specifier parameter is added to switch the model between p-type and n-type.

Minor updates:

• $simparam function is used only for supported simulator via a conditional check for predefined macro - ’__VAMS_COMPACT_MODELING__’.
• $vt function is used with device temperature (Tdev_s and Tdev_m) as an input arguement instead of using ambient temperature of the circuit.
• The device temperature is can now be changed by the user via parameter "dt" and/or by the self-heating effect based on the value of the self-heating flag.
• if the resistance between the node is zero, then the nodal voltage is updated as zero. This leads to node collapsing and this feature is supported in ADS and spectre but not in QUCS. In order to simulate CCAM in QUCS, statements like "V(node) <+ 0" is replaced with "I(node) <+ V(node)/‘Rmin" where Rmin = 1e-6 Ohm (Fix proposed by Y. Zimmermann).
• A parameter to mention the temperature change with respect to nominal temperature at which parameters are extracted.