The direct answer: Thermal resistivity is an intrinsic material property measuring its opposition to heat flow, expressed in °C·m/W. However, on the electronics workbench, we almost exclusively use thermal resistance ($R_{\theta}$, measured in °C/W), which accounts for the specific geometry of a component. For a power semiconductor, thermal resistance dictates exactly how many degrees the silicon junction temperature rises per watt of dissipated power. If you do not calculate this path, your silicon will cook itself into an early grave.
The Thermal Resistivity Definition vs. Thermal Resistance
Amateurs often confuse thermal resistivity with thermal resistance. The distinction is identical to electrical resistivity vs. electrical resistance. Thermal resistivity ($\rho$) is a bulk material property—copper has a low thermal resistivity, while air has a high one. Thermal resistance ($R_{\theta}$) is what you actually measure across a specific physical object, calculated as $R_{\theta} = \rho \cdot (L / A)$, where $L$ is thickness and $A$ is cross-sectional area.
When reading a datasheet for a TO-220 voltage regulator or a D2PAK MOSFET, you will not see thermal resistivity. You will see thermal resistance. This is the practical metric for thermal management. Think of it as Ohm’s Law for heat:
- Voltage ($V$) becomes Temperature Difference ($\Delta T$)
- Current ($I$) becomes Power Dissipation ($P_D$)
- Resistance ($R$) becomes Thermal Resistance ($R_{\theta}$)
The governing equation is $\Delta T = P_D \times R_{\theta}$. If your component dissipates 10W and the thermal resistance from the silicon junction to the room air is 5°C/W, the junction will be 50°C hotter than the room.
Thermal Path Math: Junction to Ambient ($R_{\theta JA}$)
Heat travels from the silicon die (Junction) through the package (Case), across a thermal interface material (Sink), and into the surrounding air (Ambient). The total thermal resistance is the sum of these interfaces:
$R_{\theta JA} = R_{\theta JC} + R_{\theta CS} + R_{\theta SA}$
- $R_{\theta JC}$ (Junction-to-Case): Fixed by the manufacturer. Found in the datasheet.
- $R_{\theta CS}$ (Case-to-Sink): Determined by your thermal interface material (TIM). Bare metal-to-metal is ~1.5°C/W; quality thermal paste drops this to ~0.2°C/W.
- $R_{\theta SA}$ (Sink-to-Ambient): Determined by your heatsink and airflow. This is the only variable you control.
Datasheets list a Maximum Junction Temperature ($T_{J(max)}$), often 150°C or 175°C. Never design to this number. At 175°C, silicon leakage currents skyrocket and packaging epoxies degrade. For reliable, long-life designs, cap your target $T_J$ at 105°C to 125°C.
Worked Example: Sinking an IRF540N MOSFET
Let’s say you are switching a 15A load with an Infineon IRF540N MOSFET in a TO-220 package. The $R_{DS(on)}$ at 10Vgs is 0.044Ω. Power dissipation ($P_D = I^2 \times R$) is $15^2 \times 0.044 = 9.9W$. Let's round to 10W.
- Target $T_J$: 115°C
- Ambient $T_A$ (inside a poorly ventilated enclosure): 55°C
- $R_{\theta JC}$ (from datasheet): 1.5°C/W
- $R_{\theta CS}$ (using Arctic Silver 5 paste): 0.3°C/W
First, find the maximum allowable total thermal resistance:
$R_{\theta JA(max)} = (T_J - T_A) / P_D = (115 - 55) / 10 = 6.0°C/W}$.
Now, solve for the required heatsink ($R_{\theta SA}$):
$R_{\theta SA} = R_{\theta JA} - R_{\theta JC} - R_{\theta CS}$
$R_{\theta SA} = 6.0 - 1.5 - 0.3 = 4.2°C/W}$.
You need a heatsink that performs at 4.2°C/W or better in natural convection.
Derating Curves and Thermal Failure Signatures
How hot is too hot? The datasheet's Power Derating Curve answers this. Above a case temperature of 25°C, the maximum allowable power dissipation drops linearly. If a MOSFET is rated for 100W at 25°C, but you are running the case at 100°C, the derating curve might show you can only safely push 40W through it. Ignoring this curve is the number one cause of bench-top magic smoke.
When you push silicon past its thermal limits, it doesn't just instantly melt. It exhibits specific failure signatures:
- Electromigration: At sustained temperatures above 125°C, high current density causes metal atoms in the silicon interconnects to physically migrate. This increases the MOSFET's $R_{DS(on)}$ over time, creating a positive feedback loop of more heat and higher resistance until the part fails open.
- Thermal Runaway (BJTs): Bipolar Junction Transistors have a negative temperature coefficient for $V_{BE}$. As they get hot, they require less base voltage to conduct, drawing more collector current, which generates more heat. Without emitter degeneration resistors, BJTs will thermally runaway and short out.
- Solder Joint Fatigue: Repeated thermal cycling (heating up under load, cooling down when off) causes the die-attach solder to crack due to the coefficient of thermal expansion (CTE) mismatch between silicon and copper. This manifests as intermittent faults that only appear when the board is hot.
Heatsink Selection and Airflow Headroom
Returning to our IRF540N example, we need an $R_{\theta SA}$ of 4.2°C/W. If we look at the Wakefield-Vette thermal catalog, a standard stamped aluminum sink like the Wakefield-Vette 577304B00000G offers roughly 5.5°C/W in natural convection—not good enough. We must step up to an extruded aluminum heatsink.
The Wakefield-Vette 641K (extruded aluminum, black anodized, approx 1.5" x 1" x 0.5") provides an $R_{\theta SA}$ of roughly 3.8°C/W in natural convection. This safely clears our 4.2°C/W requirement.
What Airflow and Enclosure Changes Buy You
If your enclosure design changes and ambient temperature spikes to 70°C, your math breaks. You have two choices: buy a massive, expensive heatsink, or introduce forced air. Adding a small 40mm fan, like a Noctua NF-A4x20 5V PWM, pushing 200 Linear Feet per Minute (LFM) of air across the Wakefield-Vette 641K will drop its effective $R_{\theta SA}$ from 3.8°C/W down to approximately 1.8°C/W. This buys you massive thermal headroom, allowing you to either push more current or shrink the physical size of the heatsink.
Always choose black anodized aluminum heatsinks over raw silver aluminum for passive (fanless) cooling. Black anodization significantly increases the surface emissivity, improving radiative heat transfer by up to 20% in still-air environments.
The Thermal Decision Tree: What to Buy
Stop guessing and use this decision matrix to select your thermal management hardware based on your calculated power dissipation ($P_D$). This path terminates in concrete hardware picks for a standard TO-220 or TO-247 package.
| Calculated $P_D$ | Thermal Strategy | Concrete Hardware Pick | Interface Material |
|---|---|---|---|
| < 1.0W | Bare package or PCB copper pour | None (Leave TO-220 tab bare, or use SMD D2PAK with thermal vias) | N/A |
| 1.0W - 5.0W | Stamped aluminum finned sink | Aavid Thermalloy 577304B00000G (Clip-on) | Silicone thermal pad (e.g., Bergquist Sil-Pad) |
| 5.0W - 20W | Extruded aluminum + natural convection | Wakefield-Vette 641K (Bolt-on) | Thermal paste (Arctic Silver 5 or Noctua NT-H2) |
| 20W - 50W | Extruded aluminum + forced air | Wakefield-Vette 641K + Noctua NF-A4x20 5V Fan | Thermal paste + mechanical spring clip |
| > 50W | Active liquid cooling or massive fin arrays | Koolance VLX-3 or custom cold plate | Liquid metal (Thermal Grizzly Conductonaut) if bare copper |
If your math from the Junction-to-Ambient equation lands you in the 5W to 20W bracket, your default pick is the Wakefield-Vette 641K paired with Noctua NT-H2 thermal paste. Apply the paste using the 5-dot method, torque the mounting screw to the manufacturer's spec (typically 0.5 N·m for TO-220 to avoid cracking the package), and verify the case temperature with a thermocouple after 30 minutes of continuous load. Thermal management is not a guessing game; it is a calculated physics equation. Do the math, buy the right extrusion, and your silicon will outlive the rest of your circuit.
