Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
IEC 62258-6: Thermal Simulation Requirements for Semiconductor Die Products
A technical guide to the information requirements for thermal modeling of bare and minimally packaged semiconductor die, covering data exchange, simulation models, and engineering best practices
Introduction to IEC 62258-6: Thermal Simulation Data for Semiconductor Die Products
IEC 62258-6:2006 is part of the IEC 62258 series that facilitates the production, supply, and use of semiconductor die products — bare wafers, singulated die, die with connection structures, and minimally packaged die. This specific part defines the information requirements for thermal simulation of bare and minimally packaged semiconductor die, enabling accurate modeling of thermal behavior in electronic systems.
Thermal simulation of bare die is fundamentally different from packaged device simulation. Without the thermal ballast of a traditional package, junction temperatures in bare die assemblies can rise 30–50% faster under transient loads. IEC 62258-6 provides the data framework needed to capture these effects accurately.
The standard applies to all stages of the supply chain — die manufacturers, distributors, system integrators, and end users — and is intended to be read in conjunction with IEC 62258-1 (procurement and use requirements) and IEC 62258-2 (exchange data formats). The thermal data and models specified in this standard are essential for verifying the correct functionality and reliability of electronic systems that integrate bare semiconductor die.
Key Information Requirements for Thermal Simulation
| Category | Required Data | Application |
|---|---|---|
| Operating conditions | Temperature range, max junction temperature, extended junction temperature range with reduced lifetime | Design margin assessment |
| Power dissipation | Maximum, typical, and minimum values under specified conditions | Thermal budget calculation |
| Heat source distribution | Die surface plot showing position and area of heat sources, type and power per source, surface vs. bulk heat generation | FEM model boundary conditions |
| Thermal material properties | Thermal conductivity (all materials), specific thermal capacity (for transient simulation) | Steady-state and transient analysis |
| Package thermal resistance (minimally packaged) | RθJA and/or RθJB values, test method description, ambient conditions | System-level thermal verification |
| Encapsulation properties | Thermal properties of encapsulant, adhesive, substrate dielectric | Multi-material thermal path modeling |
| Simulation model metadata | File name, creation date, model description, source, compatible simulators, program version, compliance level, scope/limitations | Model traceability and reuse |
When requesting thermal data from a die supplier, always specify whether you need steady-state or transient simulation data. Transient simulations require specific thermal capacity values for all materials — a parameter that is often omitted from standard data sheets. A transient-aware thermal model can be the difference between a design that survives power cycling and one that fails after 1,000 thermal cycles.
Engineering Best Practices for Die-Level Thermal Simulation
Building Accurate Thermal Models for Bare Die
The standard identifies three categories of die products, each with distinct thermal modeling requirements. For bare die without connection structures, the key input parameters are the die geometry (thickness and area), the thermal conductivity of the semiconductor material (typically 130–150 W/m·K for silicon at 300 K, but degrading to ~80 W/m·K at 400 K), and the heat source distribution on the die surface. The strong temperature dependence of silicon thermal conductivity is a critical factor — using room-temperature conductivity values for a die operating at 125 °C can underpredict the junction temperature by 10–15 °C.
For minimally packaged die (e.g., chip-scale packages or overmolded die), the standard requires junction-to-ambient and junction-to-board thermal resistance values measured per JEDEC JESD51 standards. However, these package-level thermal metrics must be supplemented with die-level data (junction temperature, power distribution) to create accurate system-level thermal models. The interaction between the die, the adhesive layer, the substrate, and the system board must be modeled as a coupled thermal network, not as independent resistances.
Use compact thermal models (CTMs) — such as the DELPHI or Cauer network models — for system-level simulations rather than detailed finite-element models. CTMs reduce simulation time by 95% while maintaining accuracy within 5% of full FEM results, making them practical for iterative design optimization.
Simulation Model Documentation and Exchange
IEC 62258-6 places strong emphasis on model documentation and metadata. Every thermal simulation model must include: the file name and creation date, a detailed scope description including limitations, the source and originator of the model, the name and version of compatible simulation programs (e.g., ANSYS, FLOTHERM, ICEPAK), and the compliance level. This metadata framework ensures that thermal models can be reliably exchanged between organizations and reused across different projects without ambiguity. The standard recommends using the XML schema defined in IEC 62258-7 for electronic data exchange, which provides a standardized format for conveying thermal, electrical, and mechanical die data.
Q: Can IEC 62258-6 be used for GaN or SiC die thermal simulation?
A: Yes, the methodology is material-agnostic. However, the thermal conductivity of GaN (~130–210 W/m·K depending on substrate) and SiC (~350–490 W/m·K) differs significantly from silicon, and their temperature coefficients must be applied correctly in the model.
A: Yes, the methodology is material-agnostic. However, the thermal conductivity of GaN (~130–210 W/m·K depending on substrate) and SiC (~350–490 W/m·K) differs significantly from silicon, and their temperature coefficients must be applied correctly in the model.
Q: What is the minimum data set required for a basic thermal simulation?
A: At minimum, you need: die dimensions, maximum junction temperature, maximum power dissipation, and thermal conductivity of the die material. For any practical simulation, the operating temperature range and heat source distribution are also essential.
A: At minimum, you need: die dimensions, maximum junction temperature, maximum power dissipation, and thermal conductivity of the die material. For any practical simulation, the operating temperature range and heat source distribution are also essential.
Q: How does die thickness affect thermal performance?
A: Thinner die (50–100 μm) have lower vertical thermal resistance but higher current density per unit cross-section. A 50 μm thick silicon die has roughly half the thermal resistance of a 200 μm die, which can significantly reduce the junction-to-case temperature differential.
A: Thinner die (50–100 μm) have lower vertical thermal resistance but higher current density per unit cross-section. A 50 μm thick silicon die has roughly half the thermal resistance of a 200 μm die, which can significantly reduce the junction-to-case temperature differential.
Q: Does the standard address die attachment (bondline) thermal resistance?
A: Yes, indirectly. The thermal properties of all materials in the thermal path must be stated, which includes die attach materials (solder, sintered silver, or adhesive). The thermal conductivity and thickness of the bondline layer should be included in the material properties data set for accurate simulation.
A: Yes, indirectly. The thermal properties of all materials in the thermal path must be stated, which includes die attach materials (solder, sintered silver, or adhesive). The thermal conductivity and thickness of the bondline layer should be included in the material properties data set for accurate simulation.
Practical Workflow for Die-Level Thermal Simulation
Implementing a thermal simulation workflow compliant with IEC 62258-6 involves several practical steps. First, the die supplier must provide a thermal data sheet that includes all parameters listed in the standard: die geometry, material thermal properties, power dissipation ratings, operating temperature range, and maximum junction temperature. This data sheet should be in a machine-readable format (ideally XML per IEC 62258-7 or a standardized spreadsheet template) to facilitate direct import into simulation tools.
The second step is model creation. For initial design studies, a simplified single-resistor compact thermal model (one-dimensional heat flow assumption) can provide rapid temperature estimates with approximately 15–20% accuracy. For detailed design verification, a multi-resistor Cauer or DELPHI network model should be constructed using the detailed material property data and heat source distribution information. The most accurate approach — full three-dimensional finite-element modeling — is reserved for final design sign-off and for analyzing complex multi-die assemblies where thermal coupling between adjacent die significantly affects junction temperatures.
The third step is model validation. IEC 62258-6 recommends that simulation results be correlated with physical measurements using thermal test vehicles or specially instrumented prototypes. Key validation metrics include steady-state junction temperature (measured via the temperature-sensitive parameter method per JEDEC JESD51-1) and transient thermal impedance curves. A validated model with correlation within 5 °C or 10% (whichever is tighter) is considered acceptable for design purposes. When such correlation cannot be achieved, the standard recommends a sensitivity analysis to identify which input parameters have the greatest influence on the results, guiding further characterization efforts.
