Vishay ThermaWick: A Chip Part That Conducts Heat but Not Electricity
イチケン / ICHIKEN
In this video, Ichiken tries out Vishay's ThermaWick. It looks like an ordinary ceramic chip capacitor, but its job is to carry heat, not signals. It is electrically insulating and passes only heat. The presenter asks how the part is meant to be used and how much it actually helps. They answer by reading the datasheet, taking a part apart, and running two thermal-imaging experiments.
The problem it addresses: copper pours that are also electrical connections
ThermaWick mounts on a circuit board like any chip component. The presenter starts with how heat normally leaves a board. Heat from resistors and ICs flows into the board or a heat sink, and designers often use the board's copper foil as a heat sink. Heat moves from the hot part into the copper and the substrate, then leaves through convection and radiation.
The obvious fix for a hot part is therefore more copper. In the presenter's illustration, enlarging the copper area around a hot component brought its temperature from 80°C down to 50°C, which "kind of feels like 'We did it!'"
The catch is that extending copper to reach a larger area creates an electrical connection as well as a thermal one. The ground plane is usually the largest copper area on a board. Tying a hot component's copper to it would force that node to ground potential. The presenter notes that ground potential is sometimes fine and sometimes isn't. When it isn't, ThermaWick can bridge two copper areas thermally while keeping them electrically insulated, even if they sit at different potentials.
Last resort or deliberate design element?
The presenter's first reaction was that this is "definitely an interesting product." Their suggested use, though, is modest. Rather than build a layout around ThermaWick from the start, they would reach for it when a design is truly struggling with heat, as a kind of last resort. They add that designs which deliberately incorporate the part would also be interesting in their own way. In their view, the decision will come down to weighing the added part cost against the cost of reworking the design.
What the datasheet says: aluminum nitride
The datasheet describes the part as an electrically insulated thermal conductor made from aluminum nitride, a material that insulates electrically and conducts heat well. The presenter notes that aluminum nitride is also used as an insulating substrate for mounting chips in power semiconductor modules.
As a comparison, they recall a Toyota Prius inverter they took apart earlier. They say they are not sure the material there is aluminum nitride. In that inverter, a thin ceramic plate sits between the power card and the heat exchanger, providing insulation and heat transfer at the same time. They mention that such ceramic feels noticeably cool to the touch, which they attribute to its good thermal conductivity.
The presenter reports the insulation resistance as 999 MΩ or higher. When they broke a part open, the interior appeared to be aluminum nitride and nothing else. A multimeter reading across the part showed "Overflow," meaning the resistance was beyond the meter's range.
Capacitance: DC-insulated, but not invisible to AC
Because the part consists of electrodes on a ceramic plate, it is electrically a small capacitor. The datasheet capacitance depends on chip size and falls roughly between a few hundredths of a picofarad and a few tenths of a picofarad. The exact figures are unclear in the transcript. The presenter tried to measure it but found the value too small for their equipment.
In their view, this is negligible for something like linking a ground plane to a power line. It is worth keeping in mind where potentials swing sharply or when connecting separate ground planes. The part blocks DC, but a small AC current can flow through its capacitance. The presenter sees this as another reason the part has interesting potential applications depending on how it is used.
Package sizes and thermal resistance
The presenter bought parts from DigiKey. Sizes run from 0603, the smallest, to 2512, the largest, and there is also a version with electrodes on the long sides. They quote these thermal resistances:
- 0603: 14°C/W. The presenter describes it as suited to targeting a heat problem with pinpoint precision.
- 2512: 15°C/W. The long terminal spacing gives more insulation distance, useful across large potential differences, but raises the thermal resistance.
- 1225 (long-side electrodes): 4°C/W. The wide, short heat path conducts heat well, but the insulation distance is shorter.
As with chip resistors, parts can be placed in parallel. Two in parallel halve the thermal resistance, three cut it to a third, and so on.
Experiment 1: bridging two copper areas
The first test used an evaluation board with two copper areas, A and B. A resistor sits on area A. When voltage is applied, the resistor heats up and heat spreads into A. A ThermaWick can be fitted between A and B, and the board has one version with the part and one without. With the part installed, heat should move from A into B, lowering the resistor temperature while warming B.
The presenter drew this as a thermal circuit. P is the resistor's heat output, T_R is the resistor temperature, R_A is copper A's thermal resistance, R_TW is the ThermaWick's, and R_B is copper B's. Without the ThermaWick, the branch through R_TW and R_B is open. With it, that branch is connected. They acknowledged that some heat also reaches B through the board substrate, but they ignored that path in the model.
On the thermal camera, heat visibly crept into copper B on the version with the ThermaWick. Over time the whole board warmed, including the version without the part, because heat also travels through the substrate. The presenter said the resistor temperature did not drop dramatically with the ThermaWick present. Their tentative explanation was that copper A's thermal resistance R_A is already fairly small and the ThermaWick's is higher than copper's, so the extra path made little visible difference. They added that they could still see heat moving into the neighboring copper, and wondered whether heat was actually transferring as specified.
Experiment 2: a more extreme setup
To make the effect clearer, the presenter soldered a resistor directly to a ThermaWick in two configurations. One had no copper foil attached to the ThermaWick. The other had copper foil soldered to its far side.
On the thermal camera, the version without copper heated up faster from the start. With copper attached, heat escaped through the ThermaWick into the foil and the resistor warmed more gradually. At near-steady state, the resistor reached about 100°C without the copper and about 70°C with it.
The presenter explained the result with the same thermal circuit. In this setup, R_A is very high in both cases, so little heat escapes that way. Without copper, the R_B branch is effectively open and the path through R_TW leads nowhere useful. With copper, heat flows through R_TW and R_B to surroundings at about 25°C. Connecting the copper gave that heat a real exit, and the resistor temperature dropped substantially.
Conclusion
The presenter summarizes ThermaWick as a chip component that is electrically insulating but lets heat through, and recommends it to designers who want more options for heat dissipation. The two experiments show a mixed picture. When the part bridged two copper areas that already dissipated heat reasonably well, the improvement was hard to see. When it gave an otherwise isolated hot resistor a path to extra copper, the resistor ran about 30°C cooler.
This video is brought to you by DigiKey.
Hi, I got my hands on an interesting heat-dissipation component today, so let me introduce it to you. Here it is. This is a product called ThermaWick, sold by Vishay. At first glance, this component looks like a ceramic capacitor, but its function is completely different; it serves as a component for transferring heat. These are components designed for heat dissipation. Furthermore, it is electrically insulated and allows only heat to pass through.
How do I use this part? And just how effective is it? We'll verify this as we conduct the experiment. But first, let me introduce our sponsor.
I'm sure you've probably thought about making small electronic devices equipped with Wi-Fi or Bluetooth before, haven't you? But designing wireless antennas for those kinds of applications is quite difficult. If you try to make it using a PCB layout, without knowledge of electromagnetic field simulation, prototype testing, and RF design, it's hard to get it to work properly.
Here's the solution to that problem. This is a chip antenna from TE Connectivity. Actually, I didn't even know antennas came as chips. Believe it or not, the smallest model is a tiny 2 × 1.2 mm. Solder it to the edge of the circuit board, and if you design it according to the recommended circuit and layout in the datasheet, you can use it as an antenna right away.
The lineup ranges from the most basic 2.4 GHz Wi-Fi model to the latest Wi-Fi 6E-compatible triple-band models. In addition, we even carry models that combine GPS and Wi-Fi in a single component. "I don't have any experience with RF design, but I need to incorporate wireless functionality." "I want to make the circuit board smaller." This product is perfect for people like that. For more details, please click the link in the description.
As for how to use it, basically, it is mounted on a circuit board and used in the same way as chip electronic components. In standard electronic circuits, the heat generated by resistors and IC chips dissipates to the circuit board or heat sink. In some cases, by actively dissipating heat into the copper foil on the circuit board, the copper foil section is used as a heat sink.
Heat escapes from the component generating heat into the copper foil, and of course, heat is also transferred to the printed circuit board. Heat is dissipated through convection in the surrounding airflow or radiation. To dissipate heat into the copper foil on the circuit board, it would be better to make the copper foil larger as well. As expected, the temperature of the components that generate heat also goes down.
In that case, the idea is to make the copper foil that dissipates the heat as wide as possible. I think that's probably the first thing that comes to mind. I'll remove the ThermaWick for now. If you add copper foil at this stage of the design, this will increase the surface area of the copper foil, allowing heat to dissipate more effectively. The temperature of the part, which had been 80°C, dropped to 50°C, a drop of 30°C. It kind of feels like, "We did it!"
But if I do that, it will cause thermal coupling, and it will also be electrically connected. Generally speaking, I think the ground plane is usually quite large, but if you connect it there, the potential in this area will also become ground. I suppose there may be cases where it's fine even at ground potential, but I suppose there are cases where that isn't the case.
In such cases, if you use ThermaWick, this allows heat to be dissipated to the other copper foil while maintaining electrical insulation. Of course, it's fine even if the potential of the copper foil is different. When I saw this product, I thought, "This is definitely an interesting product."
When using it as a heat-dissipating component, instead of designing circuit layouts and artwork based solely on this component, I thought it might be a good idea to use it as a sort of last resort when you're really struggling with heat dissipation. However, I think designs that actively incorporate this component are interesting in their own way. It's like weighing the cost of replacement parts against the cost of design rework. With that in mind, I guess this is where they'll decide whether to use this part, right?
Let's also take a look at the product datasheet. The product's structure is described here, and it says at the bottom, "It gets this cold, you know." One of its features is that it is an electrically insulated thermal conductor. I hear they use aluminum nitride as the material. Aluminum nitride is an insulator and also conducts heat well. This material is suitable for this application. Aluminum nitride is sometimes used as an insulating substrate for mounting chips in power semiconductor modules.
I'm not sure if it's aluminum nitride, but as a similar example, something similar is used in the power card section of the inverter from the Prius I disassembled earlier. This is also made using ceramic-based materials. It provides both electrical insulation and thermal conductivity. A thin ceramic plate is sandwiched between the power card and the heat exchanger. Actually, when you touch this ceramic material, it feels quite cool to the touch. That's because it conducts heat well, isn't it?
It says "electrically insulated." I've heard the resistance value is 999 MΩ or higher. When I actually broke it open to see what was inside, as you can see, it consists solely of aluminum nitride. Since it is electrically insulated, copper foils with different potentials can be thermally bonded together. When I actually measured the resistance with a multimeter, as shown here, it displays "Overflow." This is outside the measurable range.
Also, the capacitance is low. If you take a closer look, you'll see that electrodes are attached to the ceramic plate, so electrically, it looks like a capacitor. Here is the actual capacitance. Although the capacitance varies depending on the chip size, it ranges from 0.0* pF to about 0.* pF. For example, when connecting the ground plane to the power line, it's hardly something you'd even notice. If you use it to connect areas where the potential is fluctuating wildly to ground planes, you might find this interesting.
I actually tried measuring the capacitance as well, but it's so small that I can't measure it with the equipment I have. The capacitance is small. So, while it is DC-insulated, in terms of AC, it seems a small amount of current is flowing through the capacitance component. Depending on how you use it, this heat-dissipation component seems to have some pretty interesting potential applications.
Here's what it looks like in person. I bought it from DigiKey. Sizes range from 0603, the smallest, to 2512, the largest. There are even ones with long-side electrodes. Regarding this 0603 size, does it really feel like it targets heat issues with pinpoint accuracy? The thermal resistance is 14°C/W.
Since the 2512 size has a long distance between terminals, it seems like it would work well in areas with a large potential difference. That way, you can maintain a greater insulation distance. However, please note that because the distance between the terminals is long, the thermal resistance will be correspondingly higher. The thermal resistance is 15°C/W.
There is also a 1225 model with a long-side electrode. Because it conducts heat well, its thermal resistance is low. For the 1225 size, the thermal resistance is 4°C/W. However, the insulation distance will be shorter. After that, just like when using a regular chip resistor, you can also use them in parallel to reduce thermal resistance. If you connect two in parallel, the thermal resistance is halved. If you connect three in parallel, the thermal resistance will decrease to one-third, and so on.
Now we're moving on to the experiment. We will use the experimental board provided here. There are two sections of copper foil on the circuit board, labeled A and B, respectively. The resistor is mounted on the copper foil labeled A, and when a voltage is applied to it, the temperature of the resistor rises. Heat flows into the copper foil at point A, causing its temperature to rise.
A ThermaWick can be installed between copper foils A and B, and it's actually already mounted on the circuit board at the bottom. If you have a ThermaWick, heat is dissipated through the ThermaWick from A to the copper foil at B, so the temperature of the resistor is expected to decrease. However, the temperature of the copper foil in section B will rise.
In terms of thermal design, the circuit is configured as follows. In fact, some heat is also transferred to copper foil B through the substrate of the printed circuit board, but I'm ignoring that this time. Here, P represents the heat generated by the resistor, and TR represents the temperature of the resistor. RA is the thermal resistance of the copper foil in A, and RTW is the thermal resistance of the ThermaWick. If you haven't installed ThermaWick yet, this circuit is open. If you've already installed it, this is the circuit. There is the thermal resistance of the ThermaWick, and then there is the thermal resistance of copper foil B.
You can already tell the difference, can't you? With a ThermaWick, heat is being transferred to the copper foil in B, making it gradually warm. As time passes, since heat is transferred through the substrate of the printed circuit board, the temperature of the circuit board without ThermaWick installed is also rising overall.
Just because there's a ThermaWick, it doesn't feel like the temperature has dropped that drastically. This is just a thought, but the thermal resistance RA—that's the portion for copper foil A, right? Since it's fairly small, and ThermaWick has a higher thermal resistance than copper foil, it's possible there were no visible results. Just as I was starting to get into it, since I could see the heat gradually transferring to the copper foil next to it, I do wonder if the heat is actually being transferred as specified.
To make it even easier to understand, let's try a more extreme experiment. The resistor and ThermaWick are soldered directly together. In addition, we've applied a copper foil to the ThermaWick. We have two versions: one with no copper foil at all, and here is the ThermaWick with copper foil soldered to it.
Here is the thermal imaging footage from the experiment. As you can see from these experimental results, immediately after the experiment began, the side without copper foil heated up more quickly. With copper foil, since the heat is escaping through the ThermaWick to the copper foil, the temperature rise across the resistor is gradual. Once a near-steady state has been reached, for the version without copper foil, the resistor temperature is approximately 100°C. For the version with copper foil, the resistor temperature is approximately 70°C. You can see that the heat is dissipating well.
As for my thoughts on this experiment, in both cases, the thermal resistance RA is very high. Heat doesn't escape from here. If you don't have copper foil, RB is open, and the temperature at the point just to the right of RTW will be 25°C. If you have copper foil, heat escapes through RTW and RB, through the copper foil to a location at 25°C. As a result, heat dissipates better when the copper foil is connected. The results showed that the resistor temperature also dropped significantly.
So, I tried using some interesting chip components that are electrically insulated but allow only heat to pass through. For designers looking to expand their heat dissipation design options, be sure to check it out.
If you found this video helpful, please give it a thumbs up. If you have any questions or concerns, please feel free to leave a comment. Thank you for watching until the end.
We are currently selling electronics kits, such as breadboard circuits made by Ichiken and pre-wired LED products. For more details, see the description section.
Article published