Articolo in rivista, 2021, ENG, 10.1109/TED.2021.3084547
Valletta, Antonio; Mussi, Valentina; Rapisarda, Matteo; Lucibello, Andrea; Natali, Marco; Peroni, Marco; Lanzieri, Claudio; Fortunato, Guglielmo; Mariucci, Luigi
IMM CNR; Leonardo Co
The electrothermal behavior of gallium nitride (GaN) HEMTs has been simulated by using a hybrid approach in which the problem is solved by coupling together an effective model (for the electrical part) and a 3-D finite element model (for the thermal part). The effective model relies on the estimation of the channel current at different temperatures in the absence of thermal gradients. This regime occurs in real devices only during the very initial stage of bias pulses, when self-heating effects are not yet developed, for time intervals shorter than 1 ns. Virtual output electrical characteristic, in which self-heating effects are negligible, have been derived from pulsed measurements of the electrical output characteristics and electrothermal transient simulations. The maximum temperature because of self-heating evaluated by using the virtual output characteristic are substantially higher than those obtained using the short time-pulsed measurements directly. The results have been validated by a comparison with temperature measurements obtained using Raman thermography. This approach has proven to be numerically very efficient and fast, allowing the analysis of realistic complex structures and circuits.
I.E.E.E. transactions on electron devices 68 (8), pp. 3740–3747
Temperature measurement, Computational modeling, Logic gates, HEMTs, Solid modeling, Pulse measurements, Gallium nitride, Electrothermal simulations, gallium nitride (GaN) HEMT, hybrid simulations, self-heating, virtual characteristics
Fortunato Guglielmo, Mariucci Luigi, Valletta Antonio, Mussi Valentina, Rapisarda Matteo
ID: 481689
Year: 2021
Type: Articolo in rivista
Creation: 2023-05-19 17:28:33.000
Last update: 2023-07-20 12:26:07.000
External IDs
CNR OAI-PMH: oai:it.cnr:prodotti:481689
DOI: 10.1109/TED.2021.3084547
ISI Web of Science (WOS): 000678349800007