# How Chips Use On-Chip Microbatteries for Autonomous Sensor Power

An episode of [The Hardware Podcast with Fexingo: Chips, Devices, and Electronics Engineering Conversations](https://invisiblenetwork.ai/podcast/1896799982), 20 Jul 2026, 13 min.

Audio: https://audio.fexingo.com/business/the-hardware-podcast/episode-0123.mp3

In this episode of The Hardware Podcast, Lucas and Luna explore the emerging technology of on-chip microbatteries—tiny energy storage units integrated directly onto silicon chips. They discuss how researchers at MIT and imec are developing solid-state microbatteries with energy densities up to 100 microwatt-hours per square millimeter, enabling self-powered IoT sensors that could run for years without battery changes. The hosts examine the key challenges: thermal integration during chip fabrication, electrolyte materials that withstand semiconductor processing, and the trade-off between capacity and footprint. They also highlight a real-world application: a smart building sensor prototype that uses an on-chip microbattery to transmit temperature data every 15 minutes for over a decade. The episode concludes with a look at how this could impact medical implants, environmental monitoring, and ultra-low-power edge computing. #OnChipMicrobatteries #EnergyHarvesting #IoT #Semiconductors #SolidStateBatteries #MIT #Imec #AutonomousSensors #ChipIntegration #LowPower #MedicalImplants #EdgeComputing #BatteryTechnology #SmartBuildings #Nanotechnology #Technology #FexingoBusiness #BusinessPodcast Keep every episode free: buymeacoffee.com/fexingo

## Transcript

Lucas: Alright Luna, we've talked a lot about on-chip sensors that measure everything from sweat to terahertz radiation. But there's a problem that's been nagging at me: how do you power these things when they're embedded in a wall or inside someone's body?
Luna: Yeah, you can't exactly run a wire to every sensor in a smart building, and changing coin cells in a thousand sensors is a nightmare.
Lucas: Exactly. And that's where on-chip microbatteries come in. Instead of attaching a separate battery to the chip, you integrate the energy storage directly onto the silicon die. We're talking solid-state batteries that are just a few micrometers thick, deposited using semiconductor fabrication techniques.
Luna: I've seen some papers from imec on that. They're using thin-film lithium-based electrolytes that can be sputtered onto the chip without damaging the underlying circuitry.
Lucas: Right. And the key metric here is energy density per square millimeter. Researchers at MIT recently demonstrated a microbattery that delivers about 100 microwatt-hours per square millimeter. That doesn't sound like much, but for a tiny sensor that only needs to wake up once an hour and transmit a few bytes, it's a game-changer.
Luna: So how do they actually make these? I imagine the thermal budget is a huge constraint. You can't bake the chip at 800 degrees Celsius after you've already built the transistors.
Lucas: That's the clever part. Most on-chip microbatteries are fabricated using low-temperature processes—typically below 200 degrees Celsius. They use materials like lithium phosphorus oxynitride, or LiPON, as the electrolyte, and thin films of lithium cobalt oxide or lithium manganese oxide for the cathode. The anode is often lithium metal or silicon nanowires.
Luna: And the whole stack is only a few microns thick, so it doesn't add much to the chip's footprint.
Lucas: Right. The total thickness is typically between 10 and 30 micrometers. So you could stack it on top of a sensor die or even integrate it into the interposer layer of a multi-chip module.
Luna: But what about cycle life? If the sensor is supposed to run for years, the battery needs to handle thousands of shallow charge-discharge cycles.
Lucas: Solid-state microbatteries actually excel there. Because the electrolyte is solid, there's no dendrite formation like in liquid lithium-ion cells. Imec has demonstrated over 10,000 cycles with minimal capacity fade. At one cycle per hour, that's over a year of continuous operation, but in practice, the sensor might only cycle a few times a day.
Luna: And you can pair them with energy harvesting—like a tiny solar cell or a piezoelectric vibration harvester—so the battery just acts as a buffer.
Lucas: That's the ideal setup. Take the smart building example: a sensor that monitors temperature and humidity, with a small photovoltaic cell on top. During the day, it harvests enough light to charge the microbattery, and at night, the battery powers the sensor's wake-up and wireless transmission. That system could run practically forever without maintenance.
Luna: And the wireless transmission itself is getting incredibly low-power. There are Bluetooth Low Energy chips that can transmit a packet with just a few nanojoules.
Lucas: Right, so the bottleneck is really the standby power of the sensor itself. That's where chip designers are focusing: ultra-low-leakage transistors and deep sleep modes that draw picoamps. Combine that with a microbattery that can trickle charge for years, and you've got a truly autonomous sensor.
Luna: There's a prototype from MIT that I saw at ISSCC last year. They built a temperature sensor on a 180-nanometer CMOS process with an integrated solid-state battery. The whole thing is about one square millimeter, and they demonstrated continuous operation for over a decade under indoor lighting conditions.
Lucas: That's exactly the kind of result that gets system designers excited. And it's not just sensors. Medical implants are another killer application. Imagine a pacemaker that never needs a battery replacement surgery because it's powered by a microbattery recharged by the patient's own body heat.
Luna: But the body heat harvesting is still pretty low power—on the order of tens of microwatts. So the microbattery needs to be efficient both in storing that energy and delivering bursts for things like pacing pulses.
Lucas: And that's where the internal resistance matters. Solid-state thin-film batteries have very low internal resistance compared to conventional coin cells, so they can deliver high current pulses without a big voltage drop. That's critical for defibrillators or neurostimulators.
Luna: Let's talk about the elephant in the room: manufacturing integration. How do you add a battery to a chip without breaking the fab?
Lucas: This is the big challenge. Most fabs are not set up to handle lithium or other reactive metals used in batteries. So the approach is to do the battery fabrication as a back end of line process, after the CMOS transistors are made, but before the chip is packaged. Some groups are even doing it after dicing, using a post-processing step at the wafer level.
Luna: And the packaging itself has to be hermetic. If any moisture gets into the solid-state electrolyte, it can degrade the lithium and cause short circuits.
Lucas: Exactly. That's why you see a lot of research into atomic layer deposition of alumina or silicon nitride as encapsulants. Just a few nanometers of ALD coating can provide a sufficient moisture barrier to last decades.
Luna: There's also a cost question. Adding a few extra mask steps and a dedicated thin-film deposition tool isn't free. For high-volume applications like smart building sensors, the cost per chip has to be under a dollar for the whole system.
Lucas: But the volume argument works in its favor. If you're making tens of millions of these sensors, the incremental cost of the microbattery might be just a few cents per chip. And you save the cost of a separate battery holder, wiring, and assembly. So the total system cost can actually be lower.
Luna: Right, plus you eliminate the maintenance cost. For a building manager, not having to replace batteries in 10,000 sensors every two years is a huge operational saving.
Lucas: And that's the kind of value proposition that gets the business side interested. We're already seeing startups like Enfucell and Jenax working on printed flexible microbatteries, but the on-chip integration is still mostly in R&D labs. I think we'll see the first commercial products within three to five years.
Luna: Before we go deeper into the applications, I want to mention something that makes this possible for us to keep having these conversations. If today's discussion about self-powered sensors gave you something useful, it's worth knowing that listener support is what keeps this show ad-free and focused on the engineering details.
Lucas: That's a good point. We don't run commercials, so we rely on people who find value in the deep dives. If you'd like to help, you can support the show at buy me a coffee dot com slash fexingo. Every bit goes toward research and production.
Luna: And it genuinely lets us spend more time on topics like on-chip microbatteries. So thanks to anyone who does.
Lucas: Alright, back to the tech. One area where on-chip microbatteries could really shine is in environmental monitoring. Think about sensors deployed in remote forests or oceans to track climate data. You can't go change batteries there.
Luna: And the energy harvester might not be reliable—solar panels get covered in dust, piezoelectric harvesters need vibration. So the battery has to store enough energy for long periods of darkness or stillness.
Lucas: That pushes the energy density requirement. For a sensor that only wakes up once a day to report data, you might need a few hundred microwatt-hours of storage. That's achievable with a microbattery of about 10 square millimeters.
Luna: And the self-discharge rate becomes critical. If the battery loses 10 percent of its charge per month, that's a problem over a five-year deployment.
Lucas: Solid-state batteries have very low self-discharge—typically less than one percent per year because the electrolyte is stable and there are no parasitic side reactions. That's a huge advantage over conventional lithium-ion coin cells.
Luna: Another angle is the form factor. Because the battery is on-chip, you can design the sensor package to be extremely thin—like a credit card or even a sticker. That opens up applications in logistics, like tracking individual packages through the supply chain.
Lucas: Or in healthcare, a wearable patch that monitors glucose or lactate and transmits data to a smartphone. The whole patch could be disposable and cost pennies if the chip and battery are integrated.
Luna: But there's a safety consideration: lithium is flammable. If the patch gets damaged, could the battery catch fire?
Lucas: Solid-state batteries are inherently safer than liquid electrolyte cells because there's no flammable solvent. The lithium metal content is also extremely small—micrograms, not milligrams. Even if the battery shorts, the energy released is tiny. It's not going to cause a fire.
Luna: That's reassuring for medical and consumer applications. So what's the current state of the art in terms of commercial availability?
Lucas: Right now, companies like Cymbet, STMicroelectronics, and Murata offer off-the-shelf solid-state microbatteries that are about the size of a grain of rice. But those are separate components, not integrated onto the same chip as the sensor. The true on-chip integration is still in the prototype stage.
Luna: There's a project at imec where they've co-integrated a temperature sensor, a microcontroller, a wireless transmitter, and a solid-state microbattery all on one chip. The total power consumption is under 10 microwatts average.
Lucas: And they've shown that the chip can run indefinitely under typical indoor lighting with a small solar cell. That's a landmark demonstration. I think within five years, we'll see this kind of chip in commercial building automation systems.
Luna: What about the performance limitations? The energy density is still orders of magnitude lower than a conventional AA battery.
Lucas: That's true. An AA battery has about 2,000 milliwatt-hours. To match that with on-chip microbatteries, you'd need an area of 20,000 square millimeters—that's the size of a credit card. So for high-power applications, you still need external batteries. But for the ultra-low-power sensor niche, the on-chip approach is ideal.
Luna: And that niche is huge. Billions of IoT devices are expected in the next decade. Even if each one only needs a tiny battery, the total market is enormous.
Lucas: Exactly. And the environmental impact is also positive: fewer disposable batteries means less toxic waste. So there's a sustainability angle too.
Luna: It seems like the stars are aligning: low-power chips, efficient energy harvesting, and now the ability to store that energy on the same piece of silicon.
Lucas: The key enabler has been advances in thin-film deposition and solid-state electrolytes. Without those, we'd still be stuck with external coin cells. So it's a great example of materials science enabling a new system architecture.
Luna: And I'm curious to see how the industry standardizes the integration. Will we see an open interface for on-chip power management?
Lucas: Probably not an open standard, but each foundry will develop its own process to add a microbattery layer as an optional module. TSMC already offers a 'microbattery on chip' option in some of their advanced nodes, though I think it's still in the early stages.
Luna: We'll have to keep an eye on that. For now, it's an exciting area that bridges materials science, chip design, and system engineering.
Lucas: Definitely. And as always, we'll track the progress and bring you updates as these chips move from the lab into products.

Shows are listed from Apple Podcasts' charts and directory, and Podcast Index where it is used. Every show and episode belongs to its publisher and plays from the publisher's own feed; this site hosts none of them.