 Good evening everybody, my name is Janik Verbele from Brussels University. I'm a researcher finishing a PhD in energy harvesting and renewable energy for electronic systems later this year. You may wonder why I'm giving this talk on a conference like this one. Well, obviously there is a lot of potential in renewable energy. You all know renewable energy from large-scale applications, but there is also a huge potential for small applications like electronics. And I'm going to try to motivate you to look into this technology in the next hour. So if you just look around you, you all have this kind of wireless devices. Some of you are wearing one. The badge of the camp is one of them and they're currently powered by batteries, obviously. So batteries are historically a very interesting solution because a modern battery mimics the properties of an ideal power supply quite well. So that means that the voltage remains stable until the battery is nearly empty and then it falls off really quickly. So that means that you have a nearly perfect voltage source, which is very convenient for electronic systems. But batteries have a lot of drawbacks which we'll get into later. So the question is now if we all have these devices, are there alternatives? And obviously the alternative is renewable energy. So if you think about the Netherlands, what do you think about first? Windmills, of course. Yes, windmills. So windmills are the primary energy harvesting methods already for long, long time. And they are pretty renewable. Wind is being produced by the Sun indirectly. So if you harvest wind, you're indirectly harvesting sunlight. So it's very convenient since there are a lot of areas around the world where there is a lot of wind. There are also, unfortunately, more areas where there is no wind. So it's not a solution for everywhere, but you could use this already since ancient times. If you think about the Netherlands, what else do you think about? Water and bicycles. Bicycles and windmills. And bicycles are also a very convenient type of renewable energy because they have these generators on the bicycle. If you drive the bicycle, you attach the generator, then you basically convert some of your physical power into electrical power. And then there's also still the food in the Netherlands, which I, as a Belgian, don't consider suitable for human consumption. So that's biomass, which is also a form of renewable energy if you think about it, right? So why not batteries? You could ask. Well, the badge does have a battery. So what's so bad with batteries? Well, first of all, batteries have a limited lifetime. They run empty after a while. If they're primary cells, which means non-rechargeable, then they have a limited lifetime. You can calculate this out, but they can start leaking, for example. If you have a large, long lifetime device, then you open it up after five more years, then you may have noticed that the battery starts leaking and it leaks its electrolytes into your device, which corrodes the PCB and everything is ruined. Rechargeable batteries don't have this problem that they run out eventually. By the way, you could also recharge a non-rechargeable battery exactly once. That's a heck. But even rechargeable batteries have only a limited number of charged discharge cycles. For a cell phone or a laptop battery, these types of cells are commonly used, lithium-ion, lithium-polymer batteries, 18650 form factor. You can recharge them around between 200 and 600 times, and then their capacity drops below 80 percent of their original capacity. And that's the point where people start wondering, why is my cell phone battery not lasting the entire day anymore? So we have this amazing technology where you're having a wireless device and we have all the infrastructure in place, we have powerful processors, we have very neat screens with high resolution, we have high-speed internet connections, you can stream porn on your way to work in the morning, but probably you won't be able to stream porn again in the evening because your battery will have run empty during the day. Which is really a shame. Yes, I totally agree about that. So batteries have a limited capacity, but they also create a lot of environmental problems. First of all, the natural resources that are required to make these batteries are often dug up from the earth in countries where we don't really know that the environmental problems exist, usually because it's either large corporations which control the entire mining operation, or the local governments have such stakes in this operation that everything is being controlled. So you see here a mining operation in Brazil. These are the rare earth metals typically being used in batteries, think about cobalt for example. This is the scale of the village there in the lower right corner that gives you an idea of the scale of the pollution that is going on. You may have known from ancient times already that to extract gold from ore, toxic metals like mercury are being used to create amalgams. It can be used to extract gold from the ore surrounding it, but the concentrations are ever decreasing. So the rich ore deposits are now close to being depleted, which means that more ore needs to be extracted from the earth to harvest the same amount of metals. So you need ever increasing huge amounts of natural resources being dug up to extract ever decreasing amounts of natural resources. And that's a problem because you have waste being produced during the manufacturing of batteries, but you also have waste being produced when they are being recycled. And that's also a huge problem. For example, Belgium has one of the world leader rankings in recycling batteries and world leader ranking means that around 65 to 70 percent of batteries is actually being collected and recycled. That seems quite a lot, but also means that 30 percent of batteries still end up in landfills. And in many parts of the world the ratings are much much worse than that. So that means that the efforts required to dig up these minerals from the earth and convert them into batteries are actually going to waste because these same minerals are not being recycled properly. And at this moment already it is more profitable to mine from a waste dump than to actually mine from a quarry because the concentration of raw metals is higher in a garbage dump than it is in most quarries around the world, which is quite bad if you think about it. So what is the solution then? Well renewable energy. We have many different types of them. We all know light of course solar energy. It's all around us, solar panels I'll get into this later. Very convenient light source because the sun is nearly indepletable. But there's also thermal energy, not always available on the large scale, but definitely sub-bootable for energy harvesting purposes. Kinetic energy, we all know that from the wind turbines, but there are many other sources like your bicycle for example is a nice example of a kinetic energy source. Radiation is upcoming. I'll dig into that as well. And finally chemical energy. There are many organisms living in deep under ocean trenches that have never seen sunlight just like most engineers and nerds around us and that only live on chemical energy, just like us. So energy harvesting, to put it in the scale, there are many different size categories. So you know energy harvesting from the megawatt installations. So these are commercial installations producing megawatt ranges. The ones on your roof go to a few kilowatts at most. But what we're really interested for electronics is the sub-watt range. So if you can harvest a few watts or microwatts or even nanowatts, they can already do a whole lot with modern technology. And that is where the interesting parts start because then the question comes how do you scale these existing technologies? So when Nikola Tesla invented AC generators, that's already a long long time ago, then the technology didn't substantially change since that time. But in recent years we want to shrink this technology down. We want to make them portable. We want to integrate them in wearable devices in a lot of portable autonomous smart systems. So making something on a megawatt scale is actually quite easy. The larger you make something, the more efficient it becomes, the more easy it comes to construct. But if you scale it down to a watt range, for example, that's already a six-order of magnitude difference, then it becomes slightly more complex to design something. You see this kind of bicycles in stations in Belgium, railway stations, where people sit on and then they start to pedal and then they can plug in their cell phone or their laptop and then they get the impression that they're actually generating renewable energy. The question is, well, how do you scale the same technology? Another six orders of magnitude down to the micro-watt range. Is it possible to make a kinetic energy harvester that is small enough to fit into the pockets of your pants, for example? That is the question we're trying to answer in energy harvesting. So if you're thinking about energy harvesting, it's more like looking at the environment, what is around you and which types of energy can we actually harvest electrical power from. That is the baseline. And there are a lot of different approaches. For example, assume that you're going hiking in the mountains. In summer, of course, you want to be able to have Twitter updates about your progress in the hiking. So you need a connection, of course. You need a phone to be able to make selfies from the mountains because nobody has seen a mountain before. So obviously you'd need a solar cell on a backpack to recharge your phone while you're hiking. This is actually, from an energetic point of view, a very interesting approach. Why? Because the sunlight that is otherwise falling on this backpack doesn't have any useful purpose. It just heats up the black backpack and this heat then just flows off with the wind. You don't have energy from it. So by putting a solar cell on a backpack, you can actually have a net gain in energy. This is some energy that would otherwise not be used in a useful way. On the other hand, if you're using that same generator on a bicycle, then that's a whole different story. There's something like the law of conservation of energy, which means that energy cannot be destroyed nor created. So if you want to extract mechanical energy from the system to convert it to electrical energy, it means that you will have to pedal harder. So it is still a form of energy harvesting, but you have to put more energy into it to convert a small fraction into electrical power. And always keep in mind that these energy harvests no matter if it's a solar cell or an electrical harvester for mechanical energy or a thermal harvester or an RF harvester, none of them have an efficiency even close to 100 percent. So it means that you have to put a lot more mechanical energy into the system to get a tiny fraction of electrical power out. Right? The completely ridiculous type of energy harvesting also exists, that if you have active beacons, less use, for example, active RFID, you all know these scanners at shops. There are many easy ways to bypass these, by the way, aluminum foil. I won't dig into this. This is out of scope of the talk. But these basically transmit an active field and the same rule applies there. The efficiency of such an RF harvester is between the two and the five percent. So it means that for each watt of energy you want to get out in your wireless system, you have to put in 20 watts of electrical power. Quite a low efficiency. So the right side is to be avoided, the left side is to be preferred. Now there are a lot of applications already existing. What I'm now telling is not anything new. If you look around you, there are the many commercial available implementations. For example, environmental monitoring on the top left side. There's a lot of interest in environmental monitoring. Indian air pollution, for example, is quite a hot topic, but also pollutants for bees, for insects. These are a growing concern. If you're living in a city, then you sacrifice two years of your life just because of air pollution. So if you're getting 80 years old, then two years is quite a substantial amount of your life that you're missing out on. So there is a growing concern among the population that for example air pollution is something that needs to be monitored and autonomous solar power stations are ideal for that. So it's one of the applications that are at the moment being developed. Wearables are another interesting approach. So if you have small wearable applications externally or internally, that is to be determined, then you can harvest electrical power from your body. Either through heat or through motion, a lot of different possibilities. Vibration energy harvesting is also being used. For example, to prevent trains from colliding, to monitor all kinds of industrial installations like compressors, generators, pumps, systems that vibrate. There are very specifically designed harvester for this, that vibrates together with the system. If a compressor is close to requiring maintenance, then the frequency of its vibration starts deviating because the bearings are wearing out, for example. An energy harvester cannot only harvest energy from this vibration source, but can also detect the vibration frequency. They can do some smart processing on this and the energy harvest can send a text message to some maintenance monkey and saying, hey, my bearings need to be replaced because otherwise next week I'll be failing. Which is preventive maintenance, which can save a company a lot of money. So especially in industry people are interested in this kind of technology. Public domain is interested in for the health monitoring. So there are a lot of different stakeholders that could benefit from energy harvesting. Now there are five factors that play an important role and now I come to the essence of this talk. Why are we having it here? Well, quite frankly, because there are five factors that continuously improve technologically and make new possibilities as they go. For example, the harvester themselves are getting better each day and if harvester are getting better, that means that you get a better efficiency or it could harvest from new power sources. It means that you could get much higher input power to your system. Sensors are also improving. There are a lot of new sensors available. Think about wearable sensors, for example. Diabetes is a growing problem in our society. So glucose sensors are becoming a more and more hot topic. But these things are at the moment requiring a lot of power. So if you can make them low power through applying MEMS technology, for example, then there are many new applications possible. If we're going to processing, then microcontrollers are getting increasingly efficient. For example, your batch, your your cam batch now has a dual core microcontroller running at 180 megahertz with lots of memory, RAM, flash, memory on board. If I would have told you ten years ago that wearable device would have a dual core microcontroller running at 180 megahertz, would have all have laughed me. Probably. And, yeah, rightfully. This is just one way to illustrate how fast this technology is evolving. So each day there are new applications being made possible because each of these five pillars are improving. Storage and power conversion, I'll take them together in the stock, are also important factors. Forget about batteries for the moment. There are new technologies, super capacitors, solid state batteries that make new domains. In energy storage and which may enable long time autonomy. And finally actuators. I'm putting here an antenna. Also goes a lot broader than that. Some of you may have attended the Laura talk earlier today illustrating the possibilities of a long range radio. You can now with 15 milliwatts of power reach a distance of 15 kilometers in range, which is huge compared to a few years ago. Bluetooth low energy is covering the lower side of the spectrum with a lot larger data rate, but also much higher power consumption. So whether you're going for a long range and high throughput or a long range and a small throughput, there are different wireless solutions possible, widely available in the market, commercially available. There are chips that it could readily available in your application. So just to give you an idea of how fast this technology is progressing, this is an enrol graph of solar energy harvests, basically solar panels. And you see of course that they are steadily going up. So there are many different types of solar cells that are under active development. You know the silicon type solar cells because you see them on roofs. You see them in all kinds of wearable products, but there are a lot of exotic types as well. So whenever you have a great idea, you implement it on a prototype, you go to enrol, they put it on a one kilowatts light bulb at a distance of one meter and illuminating one square meter and they're testing the efficiency. And then you see that there are a lot of interesting things coming out. You all know polycrystalline silicon cells, for example. You can easily recognize them by holding them at an angle and you see the different crystals. These are at the moment reaching an efficiency around 20 to 21 percent. So they're not too bad, but they're really easy to manufacture and easy is relative for a solar cell. If you're going to the better solar cells that we're talking about, monocrystalline silicon cells, these are mapped like these, so you don't see any crystals in them anymore. And these have a much higher efficiency. So these are the kind of cells that are being sent to space in satellites and spaceships. You also occasionally see them in consumer applications, but they're quite expensive. So difficult to manufacture because they need to be a single crystal, as the name suggests. Amorphous silicon cells, you know them from calculators, for example, are the red cells or the darker blue cells. They have a lower efficiency, but they are also much cheaper to produce. And as a nice add-on, it can make them flexible. So if you're thinking about wearables, for example, then these kind of solar cells are really nice because they are watertight. So they are rain resistant and they just have two electrodes on the back, which you can directly solder on. So if you're thinking about hiking in the mountains, well, this is a nice cell with a border around it. So you can stitch it to your backpack and hook it up to an energy harvestor, for example. These cells are not commercial, these are not prototypes. These are actually commercially available cells. You can buy them off eBay, off Aliexpress, and they only cost a few dollars per piece. Something worth mentioning is a historical mistake that are cadmium tellurium solar panels. Anybody have any idea why cadmium tellurium solar panels are not a good idea? Because they have cadmium, obviously. So tellurium is a bottleneck. There are no tellurium mines in the world. You can look that up. So where is tellurium coming from? Well, it's actually collected as a side product in zinc mining. And for around 40 years, nobody knew what to do with tellurium. So for example, the Russians collected huge stockpiles of tellurium because they were interested in the zinc. They didn't know anything to do with the tellurium. So they put them in warehouses and then suddenly there was a guy who said, hmm, hey, I have a genius idea. What about making a really toxic type of solar cell that we can put this tellurium in? That sounds like a wonderful idea. So they combined two really toxic elements together, combining them into cadmium tellurium solar cells. And the result is that this stockpile of a tellurium ran out pretty quickly, obviously, because you can't mine it. It's tied to the zinc mining supply. So there are now a lot of cadmium tellurium solar panels on roofs. And they are not being maintained anymore because there is no tellurium enough to make commercially available cadmium tellurium solar panels, just to give you an idea of what not to do on a long term scale in energy harvesting. There are a lot of emerging technologies also available, multi-junction cells, for example. That's basically just stacking different types of solar cells on top of each other that harvest in different frequency bands. CIGS cells, dye-sensitized solar cells that are dye-sensitized solar cells. Basically, what is a dye-sensitized solar cell? Well, if you have a piece of glass, you coat it with titanium dioxide and you pour beetroot juice over it, then you have a dye-sensitized solar cell. So it doesn't have to be extremely complex. The only thing that makes these slightly more difficult is because these are chemically stable over a long time. And that is a recurring problem for most of energy harvesters in the experimental range, that is how to make them last for a long time. Most of them are really poor at resisting UV light, so they decay after a long time. And obviously, the sun has a lot of UV light in its spectrum. So many of these cells will decay after a few years. That's why these cells are commonly used indoors, where there is no UV light from artificial light sources, but outdoor, these are still a problem. The good thing is that these photosensitive dyes are everywhere. There are other flowers that have them. There are a lot of vegetables and red fruits that have them. Perovskite is a mineral that can be found in quarries just across the border with Germany. In quarries, you can just dig it up and process it in a clean way into solar panels. So in contrast with the heavily polluting silicon cell solar industry, these technologies have a much better yield and are much more environmental friendly. So I think these are the emerging technologies to watch out for in the next years. We can also harvest vibrations, vibrations and in the sense of mechanical energy. As soon as something is vibrating, then you can convert these vibrations into electrical power. This is also not something new. Those of you who are old enough to have known the era of watches when we were watches around our wastes. These watches were initially driven by a quartz crystal. Why is a quartz crystal interesting? Well, if you apply a voltage, then it vibrates at a very constant frequency. Everything works vice versa as well. So if you make it vibrate, it will produce a voltage. And quartz is not the only material that does this. Sugar also exhibit this property and bones as well. So if you break your leg, then you will, for a fraction of a second, create some renewable energy. Unfortunately, breaking bones repeatedly is not a durable, sustainable way of energy harvesting, but it is possible. So I want. Yes, well, there are some ethical restrictions to that. Of course, you can try, but. Yes, we should not go too deeply into that. Fortunately, you don't have to break all the bones of your body just for the sake of renewable energy. We have a lot of synthetic materials as well. PZT is one of the oldest ones, but PVDF is upcoming. It's polyvinyl defluoride. It's a compound that can be easily spread out over any surface. And as soon as you bend it, there is a voltage appearing over its terminals. So the advantage of this is that you can bend it and rebund it and reshape it a hundred of thousands of times. So this creates harvesters with a long lifetime. So how do this practically, just to give you an idea, you start coating cantilever-shaped base material or substrate with these PVDF films, you turn them into a disc which can be pressed on and then you make a nice cover for it and you have a button. And that's really convenient because if you press the button, you actually mechanically deform this film and you create a pulse of energy. So when people press this button, they don't realize they are actually providing the power to power the system. They are controlling with it, so you can directly harvest human power with a button. Unfortunately, these kind of devices are very frequency constrained. So they work really well if you're aiming them in their resonance frequency, in their mechanical resonance frequency. Unfortunately, if you deviate from this resonance frequency as you show in the graph, then the amount of power you can harvest drops very quickly. So the difficulty and the holy grail in energy harvesting vibrations is designing a broadband vibration energy harvester. If I have to walk around at a very constant speed to make this harvester work, people are going to think it's pretty retarded. However, if I can make this harvester work at any speed I'm walking at, then this is a very convenient way of powering wireless devices. Of course, the trivial way of doing it is just using electromagnetics. If you move a magnet in and out of a copper coil, then it induces a current. Just already old technology, but it's very, very difficult to make this technology small enough. And that is what a lot of companies are working on right now. For example, these wireless switches that control your ceiling lights or your shutters or the ceiling fan in the summer. How do we take this existing technology that's already well known that is ancient, but how do we make it small enough to fit into portable devices to make it convenient enough to use in everyday applications? Finally, covering heat harvesting. You all learned about the Peltier cells probably in school. They look like this. They're basically semiconductor material pressed between two ceramic plates. If you apply a voltage to one side, then one side becomes cold and the other becomes hot. It also works in the other direction. If you apply thermal gradient over this device, then it will create an electrical voltage. This voltage is very small. It's typically in the order of microvolts per cell per degree centigrade. But if you put a lot of them in series, then you can get a voltage out that is quite usable. So we now have thermal harvesters. They come in all kinds of shapes. These are also commercially available devices, for example. These are, for example, being used in these USB fridges. We're all engineers, so we need a constant supply of cool drinks to keep us going. So you need such a small fridge next to your laptop where exactly one can of coke fits in. And these are used to cool this can. So they can easily run from a USB port. You can't mean lots of different shapes. So you want a big can of coke, then you can use the big ones. But it can be scaled down arbitrarily small. So this is also a heat harvester that easily fits on a fingernail. So they don't have to be very big. They don't have to be very small either. So you can make them any arbitrary size you want for your application. The nice thing about it is that they work both ways. So they can harvest from a heat source, but can also harvest from a cold source, as long as you can maintain this temperature difference. They have a very high power density. But of course, they heat up your cool object or they cool down your hot object. Fortunately, there are interesting applications. For example, in industrial applications where you have steam pipes that are constantly at a very high elevated temperature, but also mechanical systems like compressors that heat up tremendously much. Domestic environments like your heater at home, you need a thermostatic valve, can be controlled by a thermal energy harvester. And of course wearables, a human body produces a constant output of around 200 watts of thermal energy. That's quite a bit. Of course, you can't cover your entire body in these energy harvests. Again, it's possible it looks pretty silly if you're walking over the street like that. But for example, you could use your mother-in-law in matrix style to harvest thermal energy from. So that's perfectly possible. And then your mother-in-law also has a future use. Again, I'm not obliging you to do these are just suggestions, just for clarity. Sensors, very important. That's one of the other pillars. Sensors are very interesting devices because they can sense the environment around us. But to have a performance, you need to power them with a constant current or a constant voltage. And keeping them powered is often a very complex matter. For many sensing applications, the sensor needs to be turned on for a specific period of time. For example, if you want to measure air pollution, then you need to run the sensor for around five to 10 seconds before you can get a reading out. So it means you can't run them at a very small duty cycle. And it means that you have a tremendous amount of power going into these things. MEMS technology, micro-electromechanical systems, offer a nice opportunity here because they shrink down the sensor, make them mechanically smaller, but also reduce the power consumption. You can use them for both sensing applications and for harvesters. So there are now vibration harvesters, for example. They're also made with MEMS technology. And these will also immerse in the market in the next years. So let's look at a real-world environment and see how we can apply energy harvesting in this application. So if you look around you, it's pretty cold here. It's pretty dark. So the types of energy you could harvest are pretty limited. So there is some artificial lighting. There is a lot of structures that are now cooling down from being heated up during the day. So you have temperature gradients, you have light. There's also a lot of radiation available because we all need Wi-Fi to stream that porn, as I explained earlier. So there are many different energy sources, but you have to quantize them. You have to know how much is available, when is it available, and how can I harvest from them. What is the energy density? So to do this, you need to do some benchmarking. And that is the part that is often overlooked. If many people are designing such an application, for example, in an Israel environment, people are being commissioned in electronic engineering to retrofit an existing application with energy harvesters and then the boss says, here, we have already this product. There are thousands of them on the market, but batteries and are leaking, etc., etc. Please turn this into an energy harvesting application. Yeah, okay. It's something more complex than just throwing a battery out and putting a solar cell on. So it is a quite complex matter to model an environment and then estimate how much power is available and then dimension the size of your harvester accordingly. So what you see in commercial applications is that either harvests are used in uncritical applications, like all these solar-powered garden lights, for example, or the energy harvesters are over-dimensioned in just to be sure they're big enough. Unfortunately, as I explained earlier, if you over-dimension the harvester, it means you need to harvest and mine a lot more raw resources to construct these harvesters. They're also getting a lot more expensive and they're physically getting bigger. So it's very beneficial to know exactly how much power and energy is available to dimension a harvester optimally. So to do that, we designed these boards. They are equipped with a lot of sensors to do some benchmarking. A few interesting ones. I won't go over all of them, but they have a broadband light sensor, for example, that allows to measure all kinds of lights, both in the UV spectrum, in the visible light spectrum, in the infrared spectrum. There are temperature sensors to measure temperature differences. There are even pressure sensors to measure the ambient surroundings. We have particle sensors to give an idea of the air quality, to get an airflow ID. And a lot of other hardware, which I won't discuss in detail, which you can come talk to me about after the discussion. Of course, this is just a first version. So we developed it later on in a project called the ambient energy monitor. The ambient energy monitor has a goal to provide electronic engineers an easy interface on quantizing the amount of ambient power available in a certain environment. So there are, at the moment, three of these modules finished, one for vibrations, one for thermal energy, and one for lights. And a fourth one for RF energy is still upcoming. So how do they look? Well, like this. That's one of the current prototypes. So they have microcontroller on them equipped with a lot of sensors. This is the version for vibrations. It gives you an idea of any vibration that happens in the immediate environment of the device. It has accelerometers. It has gyroscopes on board. It has shock sensors. It has tilt sensors. It even has a PVDF transducer on board to capture any kinds of airborne sound waves, which you could also harvest energy from. So what this thing does is eventually spitting out a file with samples for each type of environmental data. And these can then be processed in a way that allows to extract profiles from this. So it tells you how much of a certain type of environmental energy is available at any given time. What can you do with that? Well, you can use it to correlate with existing energy harvesters. For example, each of these solar cells have a different spectral characteristic. Some work better outside, some work better inside. If you look at your calculator, then you will notice that most of these use these red cells. These are amorphous silicon cells. Why? Because amorphous silicon has a better characteristic, for example, called fluorescent lights. While polycrystalline and monocrystalline silicon cells perform better outside. So assume you want to construct something in the hallway of your building. And the building has a really small external window, for example. But led lighting, then the question is, well, which type of solar cell is the best suitable in this corridor? This is the way to do this. So you have the profiles of each type of energy harvester. You measure the spectrum that is available in this environment, and you just correlate them together to select the most optimal type of solar cell. So that not only gives you an idea of which type of solar cell is the most optimal, but it also gives you an idea of price already, of availability. To make your system working is basically just matching the energy balance. As I said before, you can't generate energy out of thin air. You can't destroy it either. So there is a balance that needs to be kept. You need to produce as much energy as you consume. If you have a large microcontroller running, a touchscreen interface, then you will have to harvest a considerable amount of power as well. So the ambient energy benchmark gives you an idea of how much energy is available. You put in a flow chart, and you dimension the harvester appropriately. On the other side, of course, you have your application. You have to measure it. Just check how much power it consumes over time. You can average it out if you have local energy storage. And this balance needs to be matched. There are a few of these interesting applications available. You all know these, for example, in Newson's next to the road. These are radar systems often powered by solar panels on top of them. And these are actually very nice examples of proper pattern matching. When are most cars driving? Well, during the day, because most people are not engineers and they live during the day, apparently. So that means that your system will also need to be operated most of the time during the day. So one is your solar cell producing the most power, obviously, when the sun is shining. So when you match these two together, then you see that you have a peak in solar output at noon, when the sun is in the Senate and producing most of power. For traffic, you have two peaks. You have the morning rush hour and evening rush hour. During noon, everybody is either at work or at school. And then you have a dip again. But overall, you see that these are pretty well matched. And this is basically your ideal situation. If you can do this, then you can minimize the local energy storage that is required. And it means that you can scale down on a battery, for example. There are also new possibilities becoming possible. For example, super capacitors, electric double-air capacitors. They are now growing in capacity. When most of you are at school, the teacher will have said, well, a capacitor goes up to a few dozen microfarads or a few thousand microfarads. And you already have a big one. Well, at this moment, we have super capacitors, electric double-air capacitors with a capacitance of up to 1,600 farads. And they are as big as this wireless presenter, a stall and maybe a bit bigger in diameter, but not much bigger. So that means they can put a lot of energy into these super capacitors. And super capacitors don't have this problem that they start leaking. They don't have this problem that they wear out after a certain number of charges start cycles. So matching the patterns is actually quite beneficial for your system design. It is already being used in Mars servers, for example, as well. These also need to be able to communicate 24-7. So during the day, you can harvest from the sun. At night, there is a very large temperature gradient on Mars because it doesn't have a proper atmosphere. So it means that your rover heats up during the day, cools down at night. You have a temperature gradient which you can harvest energy from. So this is called complementary balanced energy harvesting. You have a requirement. You have a budget. And you try to match the power requirements of your electronics with the supply of your energy harvesters. So it can be a single energy harvester or it can be multiple energy harvesters working together to generate the required power. That is the complementary balanced aspect of the thing. Another example are, for example, roller coasters. We all love them when we get out. There are trains running over the tracks. These are very interesting properties because the track vibrates quite heavily. So if you mount such a vibration energy harvester on the track, then it vibrates when a train is passing by. And that's what you're interested in because you want to know where the trains are. So the moment that the train is passing by, your energy harvester is producing a power output. And that is the moment when you want to make a wireless transaction. Then you want to send a status update to the control center saying, OK, I'm here on this track. I'm sensor 432. And I am detecting a train passing by. And if you dimension it correctly, then your wireless transaction will be completed by the time the train is gone. Of course, you also need a standby control so that your system can actually give a live pulse because otherwise you don't know if the train is actually there or just not being detected because the sensor failed. So you need a standby power. And that could be a solar cell, for example. So by combining a small solar cell with a vibration energy harvester, you could get a much more performance system because you could eliminate the necessity for local storage altogether. And if you don't have this local storage, you could design a system that runs entirely on energy harvesting, on renewable energy, and doesn't have a battery. If it doesn't have a battery, it basically doesn't need maintenance because all these energy harvests are solid state devices. So if you design them properly, they last a very long lifetime. And you can basically install these sensors on the track and you never have to look at them anymore. That's, of course, a hypothetical situation. If you're clever, you will still check them regularly. But there are a lot of situations, for example, in industrial environments, in a nuclear reactor where it's physically impossible to do maintenance or very expensive to do maintenance. And then this kind of application becomes very profitable. The power path is something to keep in mind. You can't directly connect an energy harvester to an electronic system. Hooking it up is not a great idea. Why? Because, for example, the output voltage of many harvesters will greatly vary. And this may damage your electronics if you're overvolting them. So basically, you will always put a voltage regulator between them. Linear voltage regulators have very low efficiency. So if you're clever, you're using a DC-DC converter. There are not many DC-DC converters specifically designed for energy harvesting applications with very high efficiency and very low quiescent current. After which, you have the choice where to put that energy. So you can either put it in a capacitor, a supercapacitor. They look like this. So small cylinders, just like other ordinary capacitors. You could use it directly if your system is complementary balanced matched. Or you can store them in solid state batteries. Solid state batteries also exist already for a while. This is a development module from Simbets. And the batteries actually are just chips. They are batteries on a dial, literally. And they're also manufactured with a conventional semiconductor process. Unfortunately, the capacitance is quite small. We're talking in milliamp hours here. A few milliamp hours usually less than that. So the capacity is very small. But these cells have actually a duty cycle of over 100,000 charge. These charge cycles. So if you compare that to the classic lithium ion, lithium polymer cells, which rarely go over 1,000 cycles, then you may see the benefits of this kind of technology. Unfortunately, it's very difficult to make them stable. Tin energy was a startup a few years ago. They attempted it. They made cells like these, very thin, there for the name. We're talking about less than half a millimeter in thickness here. A capacity is up to a few milliamp hours. But they vanished from the market, just because it's not profitable enough. And that's immediately also the bottleneck of energy harvesting. And I'll have to crush your expectations on that. Energy harvesting is, for most applications, not more economical than classic batteries. And the inconvenient truth is that batteries are ridiculously cheap. If you buy batteries in bulk in China, for example, where it being manufactured, then you're talking about cents per battery. And an energy harvester or a solid state cell could never get this low within a few years. So it is sadly still a truth that a classic battery is such a cheap and easy solution that many manufacturers don't even consider energy harvesting. Luckily, there are a few guys who did see the light. An ocean is one of them. They already have commercial energy harvesting applications available on the market. Wireless switches, for example, they're already for sale. They're interfacing with their own wireless network. So you can just buy a kit from them online. You mount their wireless sensors on your ceiling, wherever you want. You mount a wireless switch against your wall. You press the switch. It generates the power from the pressing event. And you never have to worry about changing batteries. Unfortunately, it's still quite expensive. And that's why they still have to take off. But they are now working together with a lot of manufacturers. Next stream is a manufacturer commercially manufacturing these kind of potea devices for industrial applications, for example. My D is a manufacturing vibration energy harvester. I think everybody knows what first solar will manufacture for a type of solar cells. But the efficiency becomes a huge issue. Consider that a lot of these devices have a very low efficiency. And we shouldn't joke around about that. Solar cells reach up to 45% to 48% for multi-junction cells. But in real world applications, it's much lower than that. If I look at these polycrystalline cells, I'm talking about 21% efficiency. So that means that only a fifth of the solar power that falls on them is actually being converted into electrical energy. But you can't do anything with that directly. You need to first have a DC-DC converter stage to stay in the maximum power point of these solar cells. And such a DC-DC converter also has an efficiency well below 100%. If you have to store this energy into a battery as a temporary storage, charging a battery has its losses. You all notice if you plug in your smartphone to charge it, it will heat up. What is heating up? That is the battery that is having a very low efficiency. So whether you're storing it in a battery or in a super capacitor, you're basically wasting energy. If you're retrieving this energy, then you're losing again because you always have losses over the internal resistance of your battery or your super capacitor. And these batteries and super capacitors never match the voltage of your systems. So you need a second DC-DC converter stage, which again loses a lot of power. So if you look at the end-to-end efficiency from your 100% solar power or heat or vibrations, you only have a very tiny amount left. To give you an idea of common efficiencies for solar cells, we already talked about it, between 20% and 25% for crystalline silicon cells. Thermal energy harvesters, their efficiency is even lower. We're talking about 7% to 8% electrical efficiency. Vibration harvesters are even lower. They were talking 4% to 6%. And RF energy harvesters I'm not even talking about because it's embarrassingly low. They were talking 1% to 2% electrical efficiency. So if you're not thinking about finding the holy grail and setting up a radiation energy harvester next to your telcos, a wireless power tower, well, then they're going to get at your door pretty quickly. Because for each watt you get out, they have to put in 100 watts of wireless transmission power. So that is quite inefficient. The harvester coverage is also quite limited. This is a common RFID tag. People ask, well, are there any limitations on the amount of energy I can harvest? Well, the answer is no. If you could make your harvester infinitely large. But in practical applications, the size of your harvester is limited, obviously. Well, this is the size of this harvester. Well, this is the harvester coverage. It only captures energy in the flux that it's covering. And it is pretty small. So do you want more power? Well, then either you have to increase the power source density or you have to make your harvester physically bigger. That are the two options. There is no such thing as making a tiny harvester that outputs 100 watts of power that doesn't exist and won't be existing any time soon. So why do you want durability? Why do we want energy harvesting? Then you may wonder, well, because you do not want to be the guy who has to climb up to that chimney to replace the batteries every two years. Believe me, you don't want to be that guy. So if you install something on that chimney, then it better works reliably. And then it's your job to design something that needs as little maintenance as possible. And it's usually least material, where you're being a contract maintainer. You have to climb up the tower. It costs you a lot of money just on manpower, but also on insurance, for example, because these are dangerous working conditions. So if you can make this reliable just by design, by only using off the shelf, but commercially validated components, by using solid state components, then the autonomy will increase and you won't have any problems with reliability. Storage, a small thing about it. I already discussed the secondary chemical cells. These are the ones used on the batch. You already know them. They have very interesting properties. They are lithium polymer batteries. So you stick a knife in them and they burst into flames. Don't do that while you're wearing the batch, by the way. Electrolytic double-air capacitors don't have this problem, but they have a very high short-circuit current. So if you want to look at the internals, it's basically just a capacitor with modified electrolyte. Very high short-circuit current, but doesn't have this environmental hazard because it's just basically the same technology. The solid state cells, as manufactured by Simbat, they use a classic semiconductor technology. So they're more polluting and lower capacity, so they're not really being used in any commercial product that I am aware about. If I'm mistaken, you know if any please feel free to join in the discussion afterwards. DC-DC converters, I think many of you will be familiar with the concept, so I won't elaborate too far on this. There are two types. You have the buck converters, which decrease the voltage. You have boost converters, which increase the voltage. The interesting thing is that you can use them as charge pumps so it could control the rate at which power is being extracted from a harvester to design a maximum power point tracker. If you have a solar cell, then it outputs its maximum power in one single point, and that is the largest area under the curve. You want to operate the solar harvester in that point, and that's why you have these converter devices that are being sold for hundreds of euros usually. Well, what they do is exactly operating the solar panels on your roof in that maximum power point. You also have these for energy harvester. They are a lot smaller. Linear technology is one of the companies that manufactures them, but Texas Instruments, analog devices, microchip are also manufacturing them. The interesting part about it is that they are now highly integrated. So many of you know these cheap Chinese bug boost converters. They come with a chip and an inductor and a few capacitors. Well, many manufacturers now have integrated all these components. It's possible to create inductors with a standard silicon manufacturing process. So we're talking about a planar inductor on the chip itself. You don't need an external inductor anymore. They're operating at frequencies between two and 10 megahertz switching speeds. So the ripple is low and the footprint gets really, really small. So that means that you get a really high power density and quite a good efficiency for the price being. I have to warn for myths, who of you have heard about solar frigging roadways? Yeah, most of you. So this was actually a Kickstarter a few years ago. Just for those wondering, the Moose is actually Photoshopped in. If you were wondering about that. So this was two Americans who thought they had a genius idea. What if we could code all the roads in Alaska with solar panels? Then we could harvest energy from them. And then in winter, we could melt the snow and the ice using the power that is being produced by the solar panels. Now, nobody apparently considered the problem that there is no power being produced if there is snow on the solar panels. Nonetheless, people didn't seem to realize it. And the campaign gained over $2.5 million in crowdfunding alone while it has been proven by Dave Jones and others in Meanwhile. They're just not feasible. So if you just made a calculation, it doesn't add up. It's practically very difficult to make something that is resistant against all kinds of external damage. So a solar panel belongs on the roof or somewhere aimed towards the sun. It doesn't belong on the road. But still, there are a lot of people who see a miracle solution and the kind of things. And then throw their money at it. Well, if any of these people are in the room, please throw your money at me. I highly appreciate it. But many of these things just will never work. And that's also why people are losing faith in all these crowdfunding campaigns because many of these things are already flawed from the beginning. There are a lot of opportunities, though. Retrofitting is one of them. If you're having existing applications that are running on batteries and you're tired of your smoke detector starting to beep in the middle of the fucking night because its batteries are empty, well, then energy harvester are obviously the most interesting way to go. If you want to install systems in a building for CO2 monitoring, for smoke detection, for humidity control, and just want to install them and you only want to look at them when the building needs maintenance in another 25 years, then basically energy harvester are the only solution. There are now government regulations and demands for smart metering applications, smart water meters, smart electricity meters, smart gas meters, that these meters need to have an autonomy of at least 16 years. So that means that the company needs to be able to install these meters into your basement. You need to be able to pile boxes on top of it and forget about it for the next 16 years. That's really, really hard to do with conventional batteries. So also in these applications, energy harvesters are gaining and increasing importance. Now, development cycle is always the same. It's trial and error as most of things in engineering. As you start with a prototype, you throw some math at it. There are the models that I discussed before. You test if it works and if it doesn't, you adapt the models on the real harvester and power budgets you have. So there are a lot of applications available, weighting scales with solar harvesters, body heat sensors, wearables. They are still expensive, but they are there. Then finally, state awareness. That's one of the remaining holy grails in energy harvesting. So your system needs to know what state it is in. For example, what is the time? If you want to correlate data, then that's very important. But you also need to know the state of charge. So if I have a supercapacitor on board or a rechargeable battery or a state battery, then how much energy is still available there? The batch does this by simply measuring the battery voltage with one of the analog pins of the ESP32 and converting this with a lot of math and a lot of models into a charge between 0 and 100%. You can do this for all kinds of energy storage devices, but for some, it's more complex than for others. And finally, the state of health. Why is this important? Well, because if you're talking about long autonomy systems, 10, 15, 25 years, then of course, the system starts deviating. Batteries start to behave differently. They lose capacity. But also sensors start to decay. They need recalibration. So you need to start tracking the health of your own system to periodic recalibrations, keep track of how much energy can be stored, because otherwise you're running into trouble there. So the basic principle of these long autonomy, smart systems, is more complex than what you know from the classic topologies, because there are a lot more problems in, for example, these long-time health monitoring. Two concrete, very simple applications to round up the talk and to illustrate this for you. Here are these mountains again, which some of you may have seen already. If you're going to a mountain, you're going skiing, then you want to know how much snow is there, of course, because it's pretty pointless to go to a mountain. You're skiing if there is no snow. So what you do is build a snow logger. And you stick it on a pole. You know these HC04S ultrasonic distance sensors. You put them on a pole. You put a solar cell on top of it. The sensor aims down. It measures the distance. And then you know what the snow height is. You can transmit them to a base station using Laura, for example. Laura has a range up to 15 kilometers in line of sight. That's a really neat application. This is a very interesting application from an energy harvesting perspective. Why? Because it's really surprisingly easy to do. You have a low sample frequency, because except when there is an avalanche, the snow height usually doesn't change that fast. So you could get away with sampling each quarter of an hour, or half an hour, or even each hour. You could have local storage, not a problem at all. You could log this data and then transmit it into a block, for example. And there are low reliability demands, except one tourist or one skier that wouldn't come if there is no data available. Nobody is going to die if there is one of these sensors failing. OK, then it's maybe one slope that doesn't have the snow height on it, but nobody is really going to give a shit about it. On the other hand, I'm from Belgium, and our nuclear reactors are as stable as the mental state of our politicians, which is not a lot, believe me. So we have a constant threat of nuclear catastrophe. And this is another scale of energy harvesting constraints, of course. Because if you want to put a harvester power detector around these dangerous power stations to detect a leak, then obviously you want to know about that leak as fast as possible. So you need to do some continuous sampling and sample as fast as possible, because you want to be the first one to pick up. You want to have a continuous wireless communication with the base station, not only to transmit this data, but also to send a skip alive pulse. And you need high reliability, of course, because you don't want any of these sensors failing. If such an event occurs, and you don't notice it, then you will have a zombie apocalypse on your hand before you even realize it. So it's really important to take measures as fast as possible if such a leak is being detected. So it gives a really nice contrast. One is quite simple to implement. The other is quite high-demanding to implement, although they're both outdoor applications. They can both use solar cells. But depending on the system, the requirements are completely different. Now, for those of you who are wondering, neither of these are actually existing at this moment. So if you have the time, you could build them and you could get rich with them, because these are nice opportunities. So to round up, before we go to Q&A, what is energy harvesting? Well, basically, I think it is a really nice opportunity to equip existing electronic applications with renewable energy. It's already available all around us. You look around, there is light, there is heat, there is electromagnetic radiation. It's just there waiting for you to use, to power your own system. It's scalable, it's reliable, if you design it properly, and it could potentially, and I say potentially, lower the environmental cost of your system as well. If you do proper recycling from solar panels, for example, then you get the potential to really decrease the environmental cost. I really would like to thank you for staying up this late with me. If you have any questions, I would be happy to answer them. If you have anything to add to the discussion, please feel free to do so. And if you're too tired, then I wish you a very good night, and I hope you found it interesting. Thank you. Any questions? Please, sir. Yes, it is. So this is what we're currently working on. Okay, so the question is, how could we basically adapt the behavior of wireless systems to incorporate the available environmental energy into routing wireless signals in a more efficient way? I think there was a question. Ah, oh, you're thinking of, you're thinking about hacking an RF-powered wireless system. Yeah, because well, Wi-Fi access point is always powered from the mains, and well, if you're a sole energy harvesting device, you want having as much power beamed along your part of your antenna. So is there a way to actually, well, just fake that you are a device doing lots of traffic and well, you only want to receive lots of traffic because well, you want to transmit as, well, you might need to transmit a little bit of power to the next point just to make it send bogus packets to you. You could, but it will be very inefficient. So the efforts of doing that will be offset by the extremely low efficiency of such a setup. So there are actually modules available. I have them with me actually. Company called the Power Cost is building a base stations that transmit at 433 megahertz in the ISM band with quite a bit output power, 100 milliwatts, which is the maximum allowed by EU regulations. And then they sell these kind of modules with an integrated antenna on them that immediately capture this RF energy. What they don't tell you is that the efficiency is below 1%. Well, it would help if you having an antenna size to the frequency that you want to receive. Yeah, yeah, yes, absolutely. So the larger you make the antenna, the more power you can capture, of course, but the energy density really depends on the environment you're in. So ideally you could harvest in the 2.4 gigahertz band, for example, where Wi-Fi and ZigBee and Bluetooth Low Energy and ZigFox and all these other protocols are operating in. But yeah, it really depends on the environment. Other questions? Yes? Is that a question? Hold on. Microphone coming up. Well, thank you for the speech. It was really nice. I just thought of curiosity. Can you give us maybe some numbers? How much energy you could harvest, let's say from a piezoelectric device. Let's say from walking, you know, you sometimes you hear about trying to harvest the energy of walking, yeah. It all depends on how much strain you want to put on the person producing this energy. So it's really a trade-off. If you're looking at the wrong numbers, then the human body can produce around 200 to 250 watts of mechanical energy. Of course, if you would hypothetically harvest all that energy, that means you're not moving anymore because there is no mechanical energy left to move. So essentially, what you want to is achieving some kind of compromise where you design a harvester that harvests a fraction of that mechanical energy without feeling like a burden to the user. What is being researched at the moment by Tom Kruppenkin, for example, and other researchers are energy harvester embedded in the shoes because every time you're walking, you create a mechanical deformation of the shoe and you could harvest energy from that. That is energy that is otherwise just being wasted. There are other applications in military body armor, for example, where they use the movement of the knee to apply energy harvester for cantilever-based systems on. But you really feel that there is something blocking your natural movement so that it's not so convenient. And also you need to put all kind of straps around your legs to physically mount this on your body and if you have too many pizzas and it doesn't fit anymore. So there are a lot of constraints that don't make it very pleasant to apply this. So what they're looking at, PaveGen, for example, is another company that is doing this. They are retrofitting baseball fields in the slums around Brazil where there are lots of young kids playing soccer and baseball in the middle of the night. So they now have equipped these fields with subterrain energy harvesters. If you run over it, then you create a physical deformation that actually powers the lighting around the field. So this already exists. This is actually subsidized by Elon Musk as one of the leading technologies. So this is something that is not really visible as energy harvesting. It doesn't interfere with what people are doing. And I personally think that this is also the way to go. If you need to intrude in people's daily habits too strongly, then they will not adopt the energy harvesting topologies. Just let's say a normal average person put in piezoelectric generator and down under the shoes. Yes, electric energy harvests are quite inefficient. So then we're talking about the order tons of microwatts to a few milliwatts. So just some LEDs, maybe we can just... Yes, yes. Basically, yes. Cool shoes. A few microwatts is already enough for a low power microcontroller doing a virus transmission, for example. Okay. So it really depends what you want to do in your application. Cool. Thanks. Can the microphone get in the back, please? You said that solar panels also generate heats and residual energy. Could that's the residual heat to be harvest in order to increase the efficiency of the solar panel? That's actually a very good question. So what is being done now in research is actually combining solar panels with a black surface with thermal energy harvesters in the back. So if you basically glue them together, then the front harvest light, the panel heats up in the process. But by heating up, the efficiency of the solar panel actually decreases. So you ideally want to cool your solar panel. So instead of putting a heat sink on, well, you can just as well mount a thermal energy harvester on with the heat sink attached in the back and combine both. So that's one of these examples of complementary balance energy harvesting where you're harvesting from two sources simultaneously. Now why it's not being used commercially yet is because the efficiency of thermal harvesters are a lot lower than from solar cells. So it's not really economically interesting to do this at this point in large-scale applications. But for energy harvesting in low-power wireless devices, this could be an interesting opportunity, yes. All right, thank you. Anyone else? Yeah, please. Hi, in your monitoring device, you have a MAM spectrometer or something like that. Yeah. And you have also something to measure the thermal radiation spectra. We have infrared spectrometer, yes. Infrared, the medium infrared or? The spectrometer ranges from 200 nanometers to 1,100 nanometers. Yeah, okay. So that's covering from infrared to UV. Yeah, okay. So no thermal radiation? No, no, no, no, directly, no. Okay, no. Just thanks. Are those monitor devices available? At the moment, they're actually, these are actually the only ones in existence. Yeah. So I'll have to disappoint you. No, they're not commercially available yet. These are research prototypes that are still in development. So I'm still doing an effort to get the price down. They're not completely finished yet either. So if any of you is interested in collaborating, I am absolutely open to any kind of ideas. The hardware is pretty finished by now, but there's still a lot of embedded software development to be done, writing driver software for the sensor, for example, doing the data processing. So if anybody wants to help out in finishing them up, then I would be delighted to talk to you after the presentation. To give you an idea of the price, it's actually still quite expensive technology. So as a prototype, we're talking about 80 to 100 euros per module. So that's not something you could just put on any device because it's quite expensive. But I hope to bring them to market once they are mature enough. These are all open source hardware and software, by the way. So both the circuits, the calculations, the software it's all available on CircuitMaker, on GitHub. So if anybody would like to participate, you can. You can fork them already. Take a look at the circuitry. It's up to you. Anyone else? Any more thoughts or ideas? One more? Yeah. Yeah, about the frequencies, what's the most efficient frequency to harvest? Because all the HF bands are less and less AM transmitters. AM transmitters output an awful lot of power. Yeah, yeah, that's true. So at the moment, the most interesting bands are the GPRS bands because they're still being used quite a lot for cellular communication and 2.4 gigahertz ISM bands, which are being used for Wi-Fi, BLE, SIGFOX, SIGB and many others. SIGFOX is 868, I think. It's double frequency, I believe. Okay, go ahead, quick. But yeah, to answer your question, it really depends on the environment for example, in an outdoor environment, you will have GPRS coverage for sure, but yeah, somewhere in the middle of nowhere, you will not have any Wi-Fi networks, obviously. So it really depends on where you are, which kind of environment you're in, indoor, outdoor, behind a concrete wall or just in a wooden shed. It's quite a complex question, please. I have a question. You're talking about 2.4 gigahertz and also the GSM bands, GPRS. The power of FM radio and TV transmitters is orders of magnitude higher. That is true, but the problem is that coverage is quite unevenly distributed. Yeah, okay. But there is an subtle other issue. In the early 1980s when radio piracy was quite popular, people put out sometimes a truck load with a TL tubes like we see here today in the tent. And then they spontaneously started to illuminate because they were sapping the power from the ether. That was a trick for one of the pirate radios to get silencing their competitors. There is some regulation and prohibition about bringing devices, which set off electromagnetic spectrum power too close to transmitters. Do you expect something to happen when you start to use it for mobile? Because there is a license holder who put up the power in the air. Yeah, so that's what I meant earlier. If you're harvesting, if you're living next to a wireless base station and you're going to surround it with antennas to harvest power from, somebody is not going to be too happy with you, of course. On the other hand, if you're harvesting that power on a very small scale, then that power is otherwise not being used anyways. For example, the 77 kilohertz band that transmits the DCF 77 time signal across Europe is a very low frequency band, but contains a comparably large amount of energy. So it could actively harvest from that and nobody is going to really care about it. The problem is that everybody starts doing that on a large scale. Then of course, the amount of power you need to put in these transmitters to reach everybody is going to increase as well. And that is what I meant with one of my early slides. There is a difference between energy harvesting in general and energy scavenging. So if you're harvesting from an otherwise unused source of energy, like solar power falling on your backpack, nobody is going to care about it. If you're going to tap into an otherwise usable sort of energy, then that's going to be resulting in conflicts. Yes. Final question, no? Can I have a final applause? For Janneke. Thank you very much. Okay, that concludes the program for tonight, I believe. I'm not a herald, I'm just sort of, I'm an angel. I used to be, I'm normally a herald, but not for this talk. So I just jumped in. So applause for the volunteer who just jumped in. Thank you.