Accurate, Efficient and Reliable Small-Signal Modeling Approaches for GaN HEMTs

At a Glance

Metadata	Details
Publication Date	2023-01-01
Journal	IEEE Access
Authors	Saddam Husain, Anwar Jarndal, Mohammad Hashmi, Fadhel M. Ghannouchi
Institutions	University of Sharjah, Nazarbayev University
Citations	12
Analysis	Full AI Review Included

Technical Documentation & Analysis: GaN HEMT Modeling on Diamond Substrates

Executive Summary

This document analyzes the research on small-signal modeling of Gallium Nitride (GaN) High Electron Mobility Transistors (HEMTs) grown on diamond substrates, highlighting the critical role of high-quality diamond material and connecting the findings directly to 6CCVD’s advanced MPCVD diamond capabilities.

Core Achievement: Successful development and validation of accurate Small-Signal Equivalent Circuit Models (SSECMs) for GaN HEMTs operating across a wide frequency range (0.1 GHz to 40 GHz).
Material Foundation: The Device Under Test (DUT) utilized a 500 µm thick diamond substrate, confirming diamond’s essential role in thermal management for high-power RF applications (5G, 6G, mmWave).
Thermal Superiority: Diamond substrates provide 4 to 5 times higher thermal conductivity compared to conventional SiC or Si, enabling lower self-heating and better performance in high-power RF Power Amplifiers (RFPAs).
Modeling Validation: Both scanning-based systematic and Optimization Algorithm (OA)-based hybrid extraction methodologies (MPA, POA, TSA) achieved excellent agreement between measured and simulated S-parameters.
Physical Relevance: OA-based hybrid approaches yielded more physically relevant and reliable SSECM parameters, although the Tunicate Swarm Algorithm (TSA) demonstrated the fastest execution time among the optimization methods tested.
6CCVD Value Proposition: 6CCVD provides the necessary high-quality, custom-dimensioned MPCVD diamond substrates (SCD/PCD) with ultra-smooth polishing (Ra < 1nm) required for high-reliability GaN epitaxy and subsequent device fabrication.

Technical Specifications

The following table summarizes the key physical and performance parameters extracted from the research paper, focusing on the GaN-on-Diamond HEMT structure and operational limits.

Parameter	Value	Unit	Context
Substrate Material	Diamond	500 µm	Base material for GaN HEMT epitaxy
Device Geometry	4 x 125	µm	Gate width (W_g) x Gate number (N_g)
Gate Length (L)	0.25	µm	Critical dimension for 40 GHz operation
Barrier Layer (AlGaN)	20	nm	Epitaxial layer thickness
Buffer Layer (GaN)	2	µm	Epitaxial layer thickness
RF Characterization Range	0.1 to 40	GHz	Wideband S-parameter measurement range
Maximum V_DS Bias	30	V	Operating range for Drain-Source Voltage
Minimum Modeling Error (TSA)	0.060343	-	Error at Cold-FET Pinch-off (V_GS = -3 V, V_DS = 0 V)
Parasitic Inductance (L_g)	64.3591	pH	Extracted value for 4x125 µm HEMT
Transconductance (G_m)	251.8868	mS	Extracted value at V_GS = -1 V, V_DS = 10 V

Key Methodologies

The research employed a rigorous two-pronged approach to extract the Small-Signal Equivalent Circuit Model (SSECM) parameters for the GaN-on-Diamond HEMT.

Device Preparation and Characterization:
- Epitaxial Growth: GaN HEMT structure grown on a 500 µm diamond substrate, including AlN nucleation (1 nm), GaN buffer (2 µm), GaN channel, and AlGaN barrier (20 nm).
- RF Measurement: S-parameters recorded using a N5245 Vector Network Analyzer (VNA) and RF wafer probe station.
- Calibration: VNA calibrated using the Line-Reflect Match technique (Cascade Microtech standard 104-783).
- Operating Conditions: Measurements taken across 0.1 GHz to 40 GHz, with V_GS ranging from -3 V to 0 V and V_DS ranging from 0 V to 30 V.
Scanning-Based Systematic Extraction:
- Cold-FET Pinch-off: Initial total capacitances (C_gst, C_gdt, C_dst) calculated at V_GS = -3 V, V_DS = 0 V (no channel current).
- Systematic Search: A systematic search was performed by assigning incremental values to parasitic pad capacitances (C_pga and C_pgi) within a defined range (0.01xC_gst to C_gst).
- Extrinsic Parameter De-embedding: Parasitic inductances (L_g, L_s, L_d) and resistances (R_g, R_s, R_d) were extracted using unbiased measurements (V_GS = 0 V, V_DS = 0 V).
Optimization Algorithm (OA)-Based Hybrid Extraction:
- Algorithms Used: Marine Predators Algorithm (MPA), Pelican Optimization Algorithm (POA), and Tunicate Swarm Algorithm (TSA).
- Hybrid Approach: OAs were combined with direct extraction to optimize the SSECM parameters, minimizing the error between measured and simulated S-parameters (E_ab).
- Performance Metrics: Algorithms were evaluated based on accuracy, convergence behavior, complexity, and execution time (TSA proved fastest).

6CCVD Solutions & Capabilities

The successful realization of high-performance GaN HEMTs on diamond, as demonstrated in this research, relies fundamentally on the quality and customization of the diamond substrate. 6CCVD is uniquely positioned to supply the materials and engineering services required to replicate and advance this technology.

Applicable Materials

To achieve the high thermal performance required for GaN HEMT applications (5G/6G, Radar), researchers need diamond substrates optimized for epitaxy.

Optical Grade Polycrystalline Diamond (PCD):
- Recommendation: Ideal for large-area GaN growth due to high thermal conductivity and availability in large dimensions.
- 6CCVD Capability: Plates/wafers available up to 125mm diameter, supporting industrial-scale RF device fabrication.
High-Purity Single Crystal Diamond (SCD):
- Recommendation: Suitable for smaller, ultra-high-performance devices where maximum purity and thermal properties are paramount.
- 6CCVD Capability: SCD wafers available in thicknesses from 0.1 µm up to 500 µm (matching the thickness used in the paper).
Boron-Doped Diamond (BDD):
- Recommendation: While not used as the substrate here, BDD is available for researchers exploring integrated diamond sensor or electrochemical components within the HEMT structure.

Customization Potential

The paper highlights the need for specific device geometries (4x125 µm HEMT) and precise material interfaces. 6CCVD offers comprehensive customization services to meet these exacting requirements.

Customization Service	Relevance to GaN HEMT Research	6CCVD Capability
Substrate Thickness	Matching the 500 µm thickness used in the study.	SCD/PCD available from 0.1 µm to 500 µm (wafers) and up to 10 mm (substrates).
Surface Finish	Essential for high-quality GaN epitaxial growth (AlN nucleation layer).	Ultra-Precision Polishing: Achieves surface roughness Ra < 1nm (SCD) and Ra < 5nm (Inch-size PCD).
Custom Dimensions	Providing substrates in specific sizes for wafer-level processing.	Custom plates and wafers available up to 125mm diameter.
Metalization	Required for Source, Drain, and Gate contact pads (e.g., Ti/Au).	Internal Metalization Capability: Deposition of Au, Pt, Pd, Ti, W, and Cu for integrated device contacts.

Engineering Support

The successful modeling of GaN-on-Diamond HEMTs requires deep expertise in material properties and RF performance. 6CCVD’s in-house PhD team specializes in the material science of MPCVD diamond and its integration into high-frequency electronics.

Material Selection: Our experts can assist researchers in selecting the optimal diamond grade (SCD vs. PCD) and surface preparation technique for specific GaN epitaxial recipes.
Thermal Management Consultation: We provide guidance on how diamond thickness and quality impact the thermal resistance and overall reliability of high-power RFPA projects, directly addressing the core advantage exploited in this research.
Custom Specification Design: Assistance with defining precise dimensional tolerances and metal stack requirements for similar mmWave Transistor Modeling and RF Power Amplifier projects.

For custom specifications or material consultation, visit 6ccvd.com or contact our engineering team directly.

View Original Abstract

This article presents accurate, efficient and reliable small-signal model parameter extraction approaches applied to Gallium Nitride (GaN) High Electron Mobility Transistor (HEMT). Firstly, a scanning-based systematic model parameter extraction methodology is developed. Then, newly reported Optimization Algorithms (OAs) namely Marine Predators Algorithm (MPA), Pelican Optimization Algorithm (POA) and Tunicate Swarm Algorithm (TSA) in combination with direct extraction method are utilized to develop hybrid model parameter extraction methodologies. Lastly, both the scanning-based systematic and OA-based hybrid modelling procedures are thoroughly validated and demonstrated on a GaN HEMT grown on diamond substrate to identify their pros and cons in distinct application settings. Moreover, reliability, accuracy, convergence behavior, complexity and execution time of MPA-, POA- and TSA-based hybrid extraction procedures are also discussed. We found that both classes of the approaches are able to produce an excellent agreement between the measured and modelled S-parameters for a wide frequency range up to 40 GHz. However, OA-based hybrid modelling procedures are more physically relevant.

Tech Support

Original Source

DOI: https://doi.org/10.1109/access.2023.3317530