WEBVTT 00:00:00.000 --> 00:00:03.480 open up your desktop computer right now or peek 00:00:03.480 --> 00:00:05.919 inside the back of an old television and you'll 00:00:05.919 --> 00:00:09.539 see dozens of these tiny metallic cylinders standing 00:00:09.539 --> 00:00:11.599 on the green circuit board. Yeah they look like 00:00:11.599 --> 00:00:14.220 little miniature green silos basically. Exactly 00:00:14.220 --> 00:00:16.940 and we usually just assume they're you know solid 00:00:16.940 --> 00:00:19.699 boring chunks of metal perfect little mathematical 00:00:19.699 --> 00:00:22.839 monuments of silicon and copper. Right but they 00:00:22.839 --> 00:00:25.469 really aren't. Not at all. No they aren't because 00:00:25.469 --> 00:00:27.489 today on this deep dive we're going to reveal 00:00:27.489 --> 00:00:30.190 how those little towers are actually pressurized 00:00:30.190 --> 00:00:32.609 chemical tanks. We want to show you how they're 00:00:32.609 --> 00:00:35.909 responsible for like historical corporate espionage, 00:00:36.289 --> 00:00:38.810 explosive motherboard failures, and holding these 00:00:38.810 --> 00:00:41.210 weird ghost charges that could actually outlive 00:00:41.210 --> 00:00:44.170 the device itself. It's wild stuff. We're so 00:00:44.170 --> 00:00:45.710 glad you could join us today. Our mission is 00:00:45.710 --> 00:00:49.009 pulling from a remarkably dense, comprehensive, 00:00:50.009 --> 00:00:53.090 and surprisingly dramatic Wikipedia article all 00:00:53.090 --> 00:00:55.570 about the aluminum electrolytic capacitor. We're 00:00:55.570 --> 00:00:57.609 taking you right inside this ubiquitous little 00:00:57.609 --> 00:00:59.909 component to see how it works, how it fails, 00:01:00.109 --> 00:01:02.609 and why the tech industry literally cannot survive 00:01:02.609 --> 00:01:05.189 without it. Yeah. Okay, let's unpack this. Before 00:01:05.189 --> 00:01:07.049 we get into the literal explosions in the skull 00:01:07.049 --> 00:01:08.909 and recipes, we have to understand the magic 00:01:08.909 --> 00:01:11.769 trick that allows these tiny things to pack such 00:01:11.769 --> 00:01:14.489 a massive electrical punch. What's fascinating 00:01:14.489 --> 00:01:17.109 here is that those little towers, the capacitors, 00:01:17.250 --> 00:01:19.409 aren't passive pieces of hardware at all. They 00:01:19.409 --> 00:01:23.250 are incredibly active, delicate chemical ecosystems. 00:01:23.390 --> 00:01:25.890 I mean, they are the wet, messy analog reality 00:01:25.890 --> 00:01:28.790 that keeps our seemingly pristine digital world 00:01:28.790 --> 00:01:30.650 running. Which sounds like an oxymoron, right? 00:01:30.689 --> 00:01:33.049 Wet electronics. Yeah, exactly. But it really 00:01:33.049 --> 00:01:36.189 is a feat of modern material science. To grasp 00:01:36.189 --> 00:01:38.150 what's happening inside that little cylinder, 00:01:38.469 --> 00:01:40.409 you have to look at what a capacitor fundamentally 00:01:40.409 --> 00:01:43.540 does. In its absolute simplest form, capacitor 00:01:43.540 --> 00:01:46.640 stores electrical energy between two conductive 00:01:46.640 --> 00:01:48.560 plates, which are separated by an insulator. 00:01:48.840 --> 00:01:51.500 And we call that insulator a dielectric. All 00:01:51.500 --> 00:01:53.420 right. And the general rule of thumb for storing 00:01:53.420 --> 00:01:55.200 that energy is pretty straightforward. If you 00:01:55.200 --> 00:01:57.120 want to store a lot of energy, meaning you want 00:01:57.120 --> 00:01:59.760 higher capacitance, you need a massive surface 00:01:59.760 --> 00:02:02.099 area for those two conductive plates. And you 00:02:02.099 --> 00:02:05.359 need the insulating gap between them to be microscopically 00:02:05.359 --> 00:02:07.859 thin. So it's basically a real estate problem. 00:02:08.159 --> 00:02:10.520 It is entirely a real estate problem. The capacitance 00:02:10.520 --> 00:02:12.680 formula basically dictates that it's all about 00:02:12.680 --> 00:02:15.780 maximizing the area and minimizing the gap. But 00:02:15.780 --> 00:02:19.159 how do you fit the surface area of a large sheet 00:02:19.159 --> 00:02:22.120 of metal into a component the size of a pencil 00:02:22.120 --> 00:02:25.199 eraser? Which seems physically impossible. Right, 00:02:25.340 --> 00:02:27.860 but that brings us to the first genius move of 00:02:27.860 --> 00:02:31.080 the aluminum electrolytic capacitor. The positive 00:02:31.080 --> 00:02:34.099 plate, the anode, is made of pure aluminum foil. 00:02:34.819 --> 00:02:37.240 But instead of leaving it smooth, manufacturers 00:02:37.240 --> 00:02:39.599 put it through this intensive electrochemical 00:02:39.599 --> 00:02:42.080 etching process. And when you say etching, I 00:02:42.080 --> 00:02:45.120 usually think of like... lightly scratching a 00:02:45.120 --> 00:02:48.360 surface. But according to our source, this is 00:02:48.360 --> 00:02:51.099 vastly more extreme. Oh yeah, very extreme. They 00:02:51.099 --> 00:02:53.819 submerged the foil in an acid bath and run a 00:02:53.819 --> 00:02:55.539 highly controlled electrical current through 00:02:55.539 --> 00:02:58.439 it. The acid literally eats away at the aluminum, 00:02:58.800 --> 00:03:01.520 carving out these microscopic deep tunnels and 00:03:01.520 --> 00:03:04.620 jagged valleys. It turns a flat sheet of metal 00:03:04.620 --> 00:03:07.560 into a microscopic sponge, multiplying the effective 00:03:07.560 --> 00:03:09.939 surface area by a staggering factor of up to 00:03:09.939 --> 00:03:12.360 200. Yeah, think of it like a microscopic dam. 00:03:12.460 --> 00:03:14.699 The rougher and wider the surface of the dam, 00:03:14.800 --> 00:03:17.000 the more water, or in this case electrons, you 00:03:17.000 --> 00:03:19.180 can hold back. By turning that flat foil into 00:03:19.180 --> 00:03:22.020 a sponge -like maze, you provide 200 times more 00:03:22.020 --> 00:03:24.719 physical space for electrons together. Wow, okay. 00:03:24.919 --> 00:03:27.099 So we have our massive surface area, but then 00:03:27.099 --> 00:03:30.020 we need that incredibly thin gap, the dielectric 00:03:30.020 --> 00:03:33.370 insulator. I'm assuming if you want the gap to 00:03:33.370 --> 00:03:36.169 be perfectly uniform over all those microscopic 00:03:36.169 --> 00:03:38.849 jagged peaks and valleys, you can't just slide 00:03:38.849 --> 00:03:40.729 a piece of paper in there. You have to actually 00:03:40.729 --> 00:03:42.909 grow the insulator directly out of the metal 00:03:42.909 --> 00:03:45.810 itself. That is the exact mechanism, yeah. They 00:03:45.810 --> 00:03:48.759 use a process called anodization. By running 00:03:48.759 --> 00:03:51.419 a specific voltage through that spongy aluminum 00:03:51.419 --> 00:03:54.280 in a chemical bath, they force a layer of aluminum 00:03:54.280 --> 00:03:57.280 oxide to grow over every single microscopic peak 00:03:57.280 --> 00:03:59.680 and down into every single tunnel. And this layer 00:03:59.680 --> 00:04:02.400 is thin, right? Insanely thin. For the amorphous 00:04:02.400 --> 00:04:04.879 oxide type, it grows at a rate of about 1 .4 00:04:04.879 --> 00:04:08.009 nanometers per volt applied. To put 1 .4 nanometers 00:04:08.009 --> 00:04:10.490 in perspective for you, a single strand of human 00:04:10.490 --> 00:04:13.610 DNA is about 2 .5 nanometers wide. So this insulating 00:04:13.610 --> 00:04:15.889 barrier, which is actively holding back the electrical 00:04:15.889 --> 00:04:18.310 charge, is literally thinner than a strand of 00:04:18.310 --> 00:04:21.410 DNA. Which is just mind blowing. It is. It's 00:04:21.410 --> 00:04:23.509 like taking a piece of paper the size of a football 00:04:23.509 --> 00:04:26.250 field, crumpling it up at a microscopic level 00:04:26.250 --> 00:04:29.209 and stuffing it into a matchbox, all while keeping 00:04:29.209 --> 00:04:32.149 the layers separated by a gap thinner than a 00:04:32.149 --> 00:04:34.740 DNA strand. It's a brilliant piece of engineering, 00:04:34.839 --> 00:04:37.420 but it immediately creates a massive physical 00:04:37.420 --> 00:04:41.100 problem. You now have this incredibly complex, 00:04:41.199 --> 00:04:44.779 microscopic, sponge -like surface covered in 00:04:44.779 --> 00:04:47.620 a DNA thin layer of glass -like oxide. Right. 00:04:48.000 --> 00:04:51.540 So how on earth do you get a second solid metal 00:04:51.540 --> 00:04:53.860 plate to perfectly press up against all those 00:04:53.860 --> 00:04:55.639 microscopic peaks and valleys to complete the 00:04:55.639 --> 00:04:58.139 circuit? If you just slap another piece of flat 00:04:58.139 --> 00:05:00.699 foil against it, it will only touch the very 00:05:00.699 --> 00:05:03.550 highest microscopic peaks. all that surface area 00:05:03.550 --> 00:05:05.529 just works so hard to carve out. And here's the 00:05:05.529 --> 00:05:07.750 punchline. You don't use a solid plate at all. 00:05:07.970 --> 00:05:10.389 You use a liquid. The actual second plate, the 00:05:10.389 --> 00:05:12.230 cathode, is a liquid electrolyte. Because it's 00:05:12.230 --> 00:05:14.990 a fluid, it effortlessly flows into every single 00:05:14.990 --> 00:05:17.870 microscopic crevice, molding itself exactly to 00:05:17.870 --> 00:05:20.769 that huge rough surface. Exactly. That liquid 00:05:20.769 --> 00:05:23.990 is the true working cathode. The second aluminum 00:05:23.990 --> 00:05:27.009 foil you find inside the capacitor, which most 00:05:27.009 --> 00:05:29.860 people mistakenly assume is the second... is 00:05:29.860 --> 00:05:31.579 really just a terminal. Oh, so it's just there 00:05:31.579 --> 00:05:33.540 to make contact. Right. It's just a piece of 00:05:33.540 --> 00:05:35.939 metal dipping into the liquid to carry the electrical 00:05:35.939 --> 00:05:38.180 current to the outside world. It makes perfect 00:05:38.180 --> 00:05:40.500 sense when you lay out the physics of it. A liquid 00:05:40.500 --> 00:05:42.980 is the only way to make contact with microscopic 00:05:42.980 --> 00:05:46.699 sponge. But it also begs a historical question. 00:05:47.560 --> 00:05:51.500 Who actually looked at a dry, solid electronic 00:05:51.500 --> 00:05:55.300 circuit and thought, you know what this needs? 00:05:55.779 --> 00:05:58.259 A puddle. A puddle. Yeah, well, to find that 00:05:58.259 --> 00:06:00.459 pioneer, you have to go all the way back to 1875. 00:06:01.180 --> 00:06:03.879 A French researcher named Eugene Ducrotet was 00:06:03.879 --> 00:06:06.199 experimenting with different metals and linguids, 00:06:06.360 --> 00:06:08.139 and he discovered what he called valve metals. 00:06:08.339 --> 00:06:10.660 Valve metals. Yeah. He noticed that when aluminum 00:06:10.660 --> 00:06:12.959 is placed in certain liquid baths, it acts like 00:06:12.959 --> 00:06:15.459 a one -way valve for electricity. It blocks current 00:06:15.459 --> 00:06:17.939 in one direction by rapidly growing that oxide 00:06:17.939 --> 00:06:20.300 layer we mentioned, but lets it flow freely in 00:06:20.300 --> 00:06:24.689 the other. Fascinating. And then, in 1896, Carol 00:06:24.689 --> 00:06:27.449 Pollack took that valve idea and patented the 00:06:27.449 --> 00:06:30.490 first actual liquid capacitor. But these early 00:06:30.490 --> 00:06:32.889 versions weren't tiny cylinders, were they? No, 00:06:32.910 --> 00:06:35.490 not at all. They were literally metallic tubs 00:06:35.490 --> 00:06:38.769 or heavy glass jars filled with a sloshing borax 00:06:38.769 --> 00:06:41.750 and water solution with a folded piece of aluminum 00:06:41.750 --> 00:06:43.970 sitting in them. That sounds incredibly dangerous. 00:06:44.129 --> 00:06:47.339 It was definitely clunky. But those giant wet 00:06:47.339 --> 00:06:49.980 tubs were absolutely critical for early telephone 00:06:49.980 --> 00:06:53.379 exchanges in the 1920s. See, early electrical 00:06:53.379 --> 00:06:55.839 grids and telephone lines suffered from terrible 00:06:55.839 --> 00:06:58.779 buzzing noises because alternating current AC 00:06:58.779 --> 00:07:01.620 power has a natural ripple to it. It's not a 00:07:01.620 --> 00:07:03.560 perfectly smooth flow of energy. Right, it goes 00:07:03.560 --> 00:07:06.180 back and forth. Exactly. By hooking up these 00:07:06.180 --> 00:07:09.319 large tubs of liquid borax, the capacitors acted 00:07:09.319 --> 00:07:11.720 like electrical shock absorbers. They soaked 00:07:11.720 --> 00:07:13.600 up the ripples and released the energy smoothly, 00:07:13.939 --> 00:07:16.120 eliminating the buzz. But obviously you can't 00:07:16.120 --> 00:07:18.420 put an open tub of sloshing borax water inside 00:07:18.420 --> 00:07:20.660 a consumer device that might get bumped or tipped 00:07:20.660 --> 00:07:22.980 over in someone's living room. Which brings us 00:07:22.980 --> 00:07:26.779 to 1925 and an inventor named Samuel Rubin. Who 00:07:26.779 --> 00:07:28.839 partnered with Philip Mallory whose company, 00:07:29.040 --> 00:07:31.759 fun fact, eventually became Duracell. Oh wow, 00:07:31.879 --> 00:07:35.300 I didn't know that. But yeah, Rubin invents the... 00:07:35.120 --> 00:07:38.160 quote unquote, dry electrolytic capacitor. He 00:07:38.160 --> 00:07:40.660 took the two aluminum foils, put a spacer of 00:07:40.660 --> 00:07:43.319 highly absorbent paper between them, soaked that 00:07:43.319 --> 00:07:45.980 paper in the liquid electrolyte, and rolled the 00:07:45.980 --> 00:07:47.699 whole thing up into a tight little cylinder. 00:07:47.899 --> 00:07:51.180 And that rolled up compact design is the direct 00:07:51.180 --> 00:07:54.160 ancestor of what sits on your computer's motherboard 00:07:54.160 --> 00:07:56.120 today. I have to pause here, though, because 00:07:56.120 --> 00:07:58.779 looking at the source material, the dry capacitor 00:07:58.779 --> 00:08:00.959 that revolutionized electronics isn't actually 00:08:00.959 --> 00:08:03.759 dry at all. It's literally just a wet paper towel 00:08:03.759 --> 00:08:07.079 rolled up tightly inside an aluminum can. Dry 00:08:07.079 --> 00:08:10.500 is a total lie. It is a complete misnomer. Dry 00:08:10.500 --> 00:08:13.019 in the 1920s context just meant it didn't have 00:08:13.019 --> 00:08:15.680 a pool of free sloshing liquid that could spill 00:08:15.680 --> 00:08:18.459 on your rug. It was still very much wet, relying 00:08:18.459 --> 00:08:21.139 entirely on that liquid -soaked paper spacer. 00:08:21.680 --> 00:08:23.480 But if we connect this to the bigger picture, 00:08:23.860 --> 00:08:26.339 that single, somewhat misleading mechanical shift 00:08:26.339 --> 00:08:29.220 changed the world. By rolling the liquid up in 00:08:29.220 --> 00:08:31.759 paper, Rubin drastically reduced the physical 00:08:31.759 --> 00:08:34.620 size and the manufacturing cost. Suddenly, the 00:08:34.620 --> 00:08:36.840 bulky, expensive electrical buffers needed to 00:08:36.840 --> 00:08:39.600 build an AC -powered home radio could fit in 00:08:39.600 --> 00:08:42.840 a small box and be mass -produced. It's the breakthrough 00:08:42.840 --> 00:08:45.399 that transformed radio broadcasting from a niche 00:08:45.399 --> 00:08:48.629 hobby into a mass market phenomenon. So the Roaring 00:08:48.629 --> 00:08:51.129 Twenties were essentially powered by rolled up 00:08:51.129 --> 00:08:53.929 wet paper towels. But when we take a step back 00:08:53.929 --> 00:08:56.590 and think about the reality of sealing wet paper 00:08:56.590 --> 00:08:59.070 towels soaked in conductive chemicals inside 00:08:59.070 --> 00:09:02.169 airtight metal cans, that sounds less like an 00:09:02.169 --> 00:09:04.110 electronic component, more like a recipe for 00:09:04.110 --> 00:09:06.379 a bomb. It is an inherent danger of the design, 00:09:06.679 --> 00:09:08.879 absolutely. Aluminum electrolytic capacitors 00:09:08.879 --> 00:09:11.539 are polarized. Because that insulating oxide 00:09:11.539 --> 00:09:14.100 layer is grown by applying voltage in a specific 00:09:14.100 --> 00:09:17.059 direction, it only works in that direction. The 00:09:17.059 --> 00:09:19.480 anode must always be positive, and the liquid 00:09:19.480 --> 00:09:22.220 cathode must always be negative. So if a manufacturer 00:09:22.220 --> 00:09:24.980 wires it backward, or someone mistakenly applies 00:09:24.980 --> 00:09:26.899 alternating current where the positive and negative 00:09:26.899 --> 00:09:29.899 keep flipping back and forth, the chemistry revolves. 00:09:30.379 --> 00:09:33.200 Violently. If you apply reverse polarity, the 00:09:33.200 --> 00:09:35.980 chemistry starts working in reverse. That incredibly 00:09:35.980 --> 00:09:38.519 thin natural oxide layer on the cathode foil 00:09:38.519 --> 00:09:41.840 suddenly starts to grow rapidly. And the byproduct 00:09:41.840 --> 00:09:44.360 of that rapid chemical growth is hydrogen gas. 00:09:44.970 --> 00:09:48.029 Remember, this whole rolled -up system is sealed 00:09:48.029 --> 00:09:50.929 tight inside an aluminum can. The pressure has 00:09:50.929 --> 00:09:53.429 nowhere to go. The pressure builds up exponentially. 00:09:53.730 --> 00:09:56.409 If it goes unchecked, the aluminum casing can't 00:09:56.409 --> 00:09:59.350 hold it. The capacitor will literally detonate, 00:09:59.509 --> 00:10:01.409 blowing the metal casing apart like shrapnel. 00:10:01.529 --> 00:10:03.210 Which is why, if you look at the top of almost 00:10:03.210 --> 00:10:05.470 any modern capacitor, you'll see these little 00:10:05.470 --> 00:10:08.990 XK or T -shaped grooves snap directly into the 00:10:08.990 --> 00:10:11.159 metal roof. That's a built -in pressure relief 00:10:11.159 --> 00:10:13.419 vent, right? Exactly. When the hydrogen gas pressure 00:10:13.419 --> 00:10:16.059 hits a critical threshold, the can safely splits 00:10:16.059 --> 00:10:18.279 open along those weak points and hisses out the 00:10:18.279 --> 00:10:20.899 gas, rather than blowing a crater in your motherboard. 00:10:21.259 --> 00:10:23.980 It's a vital safety feature. But sometimes, even 00:10:23.980 --> 00:10:26.200 the vents aren't enough to prevent a total catastrophe. 00:10:26.740 --> 00:10:28.659 Here's where it gets really interesting. Oh, 00:10:28.779 --> 00:10:32.419 the capacitor plague. Yes. If you owned a desktop 00:10:32.419 --> 00:10:35.240 computer, a television, or even a video game 00:10:35.240 --> 00:10:38.220 console between the years 2000 and 2005, and 00:10:38.220 --> 00:10:40.820 it suddenly died with a loud pop and smelled 00:10:40.820 --> 00:10:44.460 like fish, you were likely a victim of this exact 00:10:44.460 --> 00:10:48.190 phenomenon. The tech industry experienced a massive 00:10:48.190 --> 00:10:51.970 unprecedented wave of failures. Millions of power 00:10:51.970 --> 00:10:53.909 supplies and motherboards were just dropping 00:10:53.909 --> 00:10:56.190 dead. Technicians were opening up these machines 00:10:56.190 --> 00:10:58.490 and finding the capacitors bulging, popped open, 00:10:58.610 --> 00:11:00.950 and leaking a crusty brown sludge. And according 00:11:00.950 --> 00:11:03.230 to our source, it wasn't just a random manufacturing 00:11:03.230 --> 00:11:06.029 glitch. It was corporate espionage gone terribly 00:11:06.029 --> 00:11:09.070 wrong. A scientist allegedly stole the recipe 00:11:09.070 --> 00:11:11.889 for a newly developed, highly conductive, water 00:11:11.889 --> 00:11:14.009 -based electrolyte from a Japanese manufacturer. 00:11:14.429 --> 00:11:16.539 They took this stolen recipe and sold it to mass 00:11:16.539 --> 00:11:19.200 market component makers in Taiwan, who then churned 00:11:19.200 --> 00:11:21.299 out millions of these cheap capacitors for the 00:11:21.299 --> 00:11:23.460 global market. But the thief missed a crucial 00:11:23.460 --> 00:11:25.879 detail. The recipe they stole was incomplete. 00:11:25.940 --> 00:11:29.059 It was missing specific, highly proprietary stabilizing 00:11:29.059 --> 00:11:31.460 additives. Why are stabilizers so important if 00:11:31.460 --> 00:11:33.379 the base liquid is just water and some salts? 00:11:33.659 --> 00:11:36.600 Because water and raw aluminum do not get along. 00:11:37.220 --> 00:11:40.139 When water and aluminum mix without those delicate 00:11:40.139 --> 00:11:42.559 stabilizing chemicals balancing the reaction, 00:11:43.059 --> 00:11:45.820 the water attacks the metal. It's not just a 00:11:45.820 --> 00:11:49.639 slow evaporation. It is an aggressive, destructive 00:11:49.639 --> 00:11:53.039 chemical reaction. Yikes. Yeah, the water actively 00:11:53.039 --> 00:11:55.440 corrodes the aluminum, generating massive amounts 00:11:55.440 --> 00:11:57.879 of internal heat and hydrogen gas from the inside 00:11:57.879 --> 00:12:00.419 out. It's a stark reminder that this component 00:12:00.419 --> 00:12:03.000 is fundamentally a contained chemical reaction. 00:12:03.340 --> 00:12:06.440 It's a tightrope walk between achieving high 00:12:06.440 --> 00:12:09.080 electrical conductivity and triggering destructive 00:12:09.080 --> 00:12:12.120 corrosion. And because of one stolen, incomplete 00:12:12.120 --> 00:12:15.440 recipe, it cost the tech industry millions, if 00:12:15.440 --> 00:12:17.620 not billions of dollars to recall and replace 00:12:17.620 --> 00:12:20.059 all those ticking time bombs. It perfectly illustrates 00:12:20.059 --> 00:12:22.759 the fragility of the ecosystem inside that tiny 00:12:22.759 --> 00:12:25.419 can. So a bad recipe turns it into a bomb. Yeah. 00:12:25.500 --> 00:12:27.159 But let's say you have a perfectly manufactured 00:12:27.159 --> 00:12:29.580 capacitor from the best factory on earth and 00:12:29.580 --> 00:12:32.159 it's wired correctly. It still has a fatal flaw. 00:12:32.700 --> 00:12:35.139 Every single wet electrolytic capacitor has a 00:12:35.139 --> 00:12:37.519 ticking clock inside it. Because it relies on 00:12:37.519 --> 00:12:40.240 that liquid -soaked paper and because no rubber 00:12:40.240 --> 00:12:42.500 seal at the bottom of the can is absolutely perfect, 00:12:42.879 --> 00:12:46.759 the liquid electrolyte slowly, inevitably evaporates 00:12:46.759 --> 00:12:49.360 over time. It literally dries out. And as it 00:12:49.360 --> 00:12:52.080 dries out, the physical pathways for the electricity 00:12:52.080 --> 00:12:55.299 to flow start to disappear. The source mentions 00:12:55.299 --> 00:12:58.480 a term here, equivalent series resistance, or 00:12:58.480 --> 00:13:02.049 ESR. That's the critical metric. Imagine a multi 00:13:02.049 --> 00:13:04.809 -lane highway. The liquid electrolyte provides 00:13:04.809 --> 00:13:07.450 millions of microscopic lanes for electrons to 00:13:07.450 --> 00:13:09.889 travel into those etched valleys. As the liquid 00:13:09.889 --> 00:13:12.830 evaporates, lanes start closing. The electrons 00:13:12.830 --> 00:13:15.470 get bottlenecked. That bottleneck is electrical 00:13:15.470 --> 00:13:18.029 resistance, the ESR going up. And whenever you 00:13:18.029 --> 00:13:19.590 have electrical resistance, you get friction. 00:13:19.769 --> 00:13:22.610 You get heat. Exactly. When you have alternating 00:13:22.610 --> 00:13:24.629 current rippling through the capacitor, what 00:13:24.629 --> 00:13:26.789 engineers call ripple current, pushing through 00:13:26.789 --> 00:13:29.740 that bottleneck generates internal heat. That 00:13:29.740 --> 00:13:32.080 heat then speeds up the evaporation of the remaining 00:13:32.080 --> 00:13:34.620 liquid, which raises the resistance even further, 00:13:34.820 --> 00:13:37.779 which creates even more heat. It's a slow, unavoidable 00:13:37.779 --> 00:13:40.440 death spiral. So heat is the ultimate enemy. 00:13:41.059 --> 00:13:43.720 And I'm assuming there's a direct mathematical 00:13:43.720 --> 00:13:46.519 correlation here, like if my computer runs a 00:13:46.519 --> 00:13:49.120 little hot because the fan is dusty, I lose a 00:13:49.120 --> 00:13:52.159 few weeks of lifespan on the hardware. It's actually 00:13:52.159 --> 00:13:54.639 much more dramatic than that. It's an exponential 00:13:54.639 --> 00:13:57.379 curve known in the industry as the 10 degree 00:13:57.379 --> 00:13:59.860 rule, which is derived from the Arrhenius equation. 00:14:00.039 --> 00:14:02.279 Okay, lay that out for me. Well, heat is essentially 00:14:02.279 --> 00:14:05.679 molecular vibration. Adding just 10 degrees Celsius 00:14:05.679 --> 00:14:07.759 doesn't just make the liquid slightly warmer. 00:14:08.139 --> 00:14:11.159 It pushes a massive threshold of the liquid molecules 00:14:11.159 --> 00:14:14.139 past the energy barrier required to break free 00:14:14.139 --> 00:14:16.820 and turn into gas. So what does that mean for 00:14:16.820 --> 00:14:19.779 the lifespan of the component? The rule is remarkably 00:14:19.779 --> 00:14:22.500 brutal. For every 10 degrees Celsius, you drop 00:14:22.500 --> 00:14:24.639 the operating temperature of the rate of evaporation 00:14:24.639 --> 00:14:27.679 halves, meaning you effectively double the capacitor's 00:14:27.679 --> 00:14:30.179 lifespan. Wait, really? Double? Double. So if 00:14:30.179 --> 00:14:32.539 a capacitor is rated by the factory to survive 00:14:32.539 --> 00:14:35.480 2 ,000 hours at a boiling 105 degrees Celsius, 00:14:36.000 --> 00:14:37.419 if you can design your circuit board to keep 00:14:37.419 --> 00:14:40.940 it at a cool 45 degrees, that exact same capacitor 00:14:40.940 --> 00:14:44.299 will last for about 15 years. Wow. So it's like 00:14:44.299 --> 00:14:47.279 keeping a block of ice in a cooler. If you design 00:14:47.279 --> 00:14:49.200 your circuit board with good airflow to keep 00:14:49.200 --> 00:14:52.299 it cool, the capacitor lives for decades. If 00:14:52.299 --> 00:14:54.220 you cram it next to a hot microchip, it dries 00:14:54.220 --> 00:14:57.019 up and dies. The placement of a capacitor on 00:14:57.019 --> 00:14:59.080 a circuit board isn't just about where it fits. 00:14:59.500 --> 00:15:02.620 It's a life -or -death thermal calculation. Yes. 00:15:03.059 --> 00:15:05.679 Engineers literally use this 10 -degree formula 00:15:05.679 --> 00:15:08.240 to calculate the survival probability of your 00:15:08.240 --> 00:15:12.259 electronics, a metric known as FET, or failures 00:15:12.259 --> 00:15:14.799 in time. But the liquid isn't just a flaw that 00:15:14.799 --> 00:15:17.059 causes it to die, right? The source points out 00:15:17.059 --> 00:15:19.919 an incredible paradox. The liquid is also a healer. 00:15:20.039 --> 00:15:22.419 That is the beautiful irony of the design. The 00:15:22.419 --> 00:15:24.419 liquid that eventually evaporates and kills the 00:15:24.419 --> 00:15:26.659 capacitor is the exact same thing keeping it 00:15:26.659 --> 00:15:28.980 alive day to day. How does that work? Well, if 00:15:28.980 --> 00:15:30.840 there's a sudden voltage spike in your power 00:15:30.840 --> 00:15:33.620 supply and a microcrack forms in that ultra -thin 00:15:33.620 --> 00:15:37.039 DNA -sized aluminum oxide dielectric, the capacitor 00:15:37.039 --> 00:15:38.919 would normally short out immediately. The device 00:15:38.919 --> 00:15:41.379 would be dead. But the liquid electrolyte contains 00:15:41.379 --> 00:15:46.120 oxygen. Ah. So it reacts. Exactly. The very millisecond 00:15:46.120 --> 00:15:49.519 a microcrack exposes the raw aluminum underneath, 00:15:49.799 --> 00:15:52.720 the liquid reacts with it, and instantly anodizes 00:15:52.720 --> 00:15:55.600 a brand new layer of oxide over the wound. Cell 00:15:55.600 --> 00:15:58.500 heals in real time. It's like a tiny metallic 00:15:58.500 --> 00:16:01.440 immune system. The fluid rushes to the site of 00:16:01.440 --> 00:16:03.740 the injury and scabs it over with glass. It is 00:16:03.740 --> 00:16:06.240 a perfect self -repair mechanism, and purely 00:16:06.240 --> 00:16:08.620 solid -state components simply cannot do that 00:16:08.620 --> 00:16:10.580 the way these wet ones can. So they heat up, 00:16:10.700 --> 00:16:13.440 they slowly evaporate, they heal their own molecular 00:16:13.440 --> 00:16:16.940 wounds, and sometimes they explode. But what 00:16:16.940 --> 00:16:19.419 happens when they are completely dead? When you 00:16:19.419 --> 00:16:21.100 unplug the device and throw it in the closet, 00:16:21.460 --> 00:16:23.779 the source mentions a phenomenon called dielectric 00:16:23.779 --> 00:16:26.500 absorption, or soakage. It's one of the most 00:16:26.500 --> 00:16:28.899 counterintuitive behaviors in electronics. You 00:16:28.899 --> 00:16:31.120 take a capacitor that has been powered on and 00:16:31.120 --> 00:16:33.519 charged up for a long time, you unplug the device, 00:16:33.600 --> 00:16:36.179 and a technician deliberately shorts the capacitor 00:16:36.179 --> 00:16:38.840 to completely discharge it. The voltage reads 00:16:38.840 --> 00:16:41.679 absolute zero. By all laws of basic physics, 00:16:41.700 --> 00:16:44.620 it should be dead. But it's not. Nope. If you 00:16:44.620 --> 00:16:46.539 remove the short and just let it sit there on 00:16:46.539 --> 00:16:49.179 the workbench, the capacitor will mysteriously 00:16:49.179 --> 00:16:52.500 develop a small electrical voltage all on its 00:16:52.500 --> 00:16:55.409 own. So what does this all mean? Does this mean 00:16:55.409 --> 00:16:57.789 my unplugged PV is still technically holding 00:16:57.789 --> 00:17:00.690 a grudge? In a way, yes. It's also known as battery 00:17:00.690 --> 00:17:03.570 action. To understand what's happening, you have 00:17:03.570 --> 00:17:05.829 to look at the molecules inside the dielectric 00:17:05.829 --> 00:17:08.630 insulator. It's a process called time -delayed 00:17:08.630 --> 00:17:10.710 dipole discharging. Let me see if I can picture 00:17:10.710 --> 00:17:13.349 this. If the molecules in the dielectric are 00:17:13.349 --> 00:17:16.230 dipoles, they have a positive end and a negative 00:17:16.230 --> 00:17:19.009 end, like tiny compass needles. That's a great 00:17:19.009 --> 00:17:22.210 way to visualize it. Imagine thousands of tiny 00:17:22.210 --> 00:17:24.990 compass needles embedded in the material. When 00:17:24.990 --> 00:17:28.009 the power is on for hours or days, the sheer 00:17:28.009 --> 00:17:30.509 force of the voltage forces all those microscopic 00:17:30.509 --> 00:17:33.430 needles to twist and point firmly in one direction. 00:17:33.910 --> 00:17:36.269 They physically stress into alignment. And when 00:17:36.269 --> 00:17:38.470 you unplug it and discharge the energy quickly, 00:17:38.769 --> 00:17:40.890 the energy in the plates is gone, but the physical 00:17:40.890 --> 00:17:43.930 needles themselves take time to relax. Exactly. 00:17:44.410 --> 00:17:46.390 The main electrical charge leaves instantly, 00:17:46.589 --> 00:17:48.849 but those physical molecules are deeply embedded 00:17:48.849 --> 00:17:51.309 in the material. They don't just instantly snap 00:17:51.309 --> 00:17:54.799 back. Over hours or even days, they slowly drift 00:17:54.799 --> 00:17:57.619 and twist back into their natural resting state. 00:17:57.819 --> 00:18:00.859 And that movement generates power. Yes. As they 00:18:00.859 --> 00:18:03.980 physically move, that motion pushes a tiny bit 00:18:03.980 --> 00:18:07.299 of residual ghost electricity back into the conductive 00:18:07.299 --> 00:18:09.819 plates. This raises an important question. Is 00:18:09.819 --> 00:18:13.660 that dangerous? If I touch a circuit board from 00:18:13.660 --> 00:18:15.720 a TV that's been in the attic for a week, am 00:18:15.720 --> 00:18:18.000 I going to get zapped by ghost electricity? In 00:18:18.000 --> 00:18:20.880 a small consumer radio or a TV board, the soakage 00:18:20.880 --> 00:18:23.480 voltage is very low. It might just cause a faint 00:18:23.480 --> 00:18:26.140 surprising pop in the speaker. But in massive 00:18:26.140 --> 00:18:28.420 high voltage industrial capacitors, the kind 00:18:28.420 --> 00:18:30.980 used in power grids or heavy machinery, this 00:18:30.980 --> 00:18:33.019 ghost voltage can actually creep up high enough 00:18:33.019 --> 00:18:35.980 to deliver a lethal shock to a technician days 00:18:35.980 --> 00:18:37.779 after the machine was completely disconnected 00:18:37.779 --> 00:18:40.880 from power. That's terrifying. It is a serious 00:18:40.880 --> 00:18:44.039 hazard. That is why large high -voltage capacitors 00:18:44.039 --> 00:18:46.400 are actually shipped from the factory with a 00:18:46.400 --> 00:18:48.440 thick shorting wire physically wrapped around 00:18:48.440 --> 00:18:51.099 the terminals. You have to manually remove that 00:18:51.099 --> 00:18:53.259 wire right before you install it just to ensure 00:18:53.259 --> 00:18:55.839 that it hasn't soaked up a dangerous ghost charge 00:18:55.839 --> 00:18:58.420 while sitting in the cardboard box. It's a perfect 00:18:58.420 --> 00:19:01.660 example of how the ideal clean physics we learn 00:19:01.660 --> 00:19:04.460 in the textbook where a discharged capacitor 00:19:04.460 --> 00:19:07.210 is a mathematically empty bucket. clashes with 00:19:07.210 --> 00:19:10.490 the messy reality of physical materials. The 00:19:10.490 --> 00:19:12.930 material itself literally remembers the charge. 00:19:13.109 --> 00:19:15.509 The physical world always leaves a trace. What 00:19:15.509 --> 00:19:17.349 we have covered an immense amount of ground today, 00:19:17.750 --> 00:19:20.170 from the genius of taking a flat piece of aluminum 00:19:20.170 --> 00:19:23.230 foil, chemically eating it away to be 200 times 00:19:23.230 --> 00:19:26.150 rougher and coating it in a glass -like barrier 00:19:26.150 --> 00:19:28.990 thinner than human DNA. We've seen how rolling 00:19:28.990 --> 00:19:32.529 up a wet paper towel in 1925 accidentally paved 00:19:32.529 --> 00:19:35.359 the way for mass market consumer radios. And 00:19:35.359 --> 00:19:37.319 we've explored the darker side of the chemistry, 00:19:37.799 --> 00:19:40.019 the capacitor plague, the violent potential of 00:19:40.019 --> 00:19:42.779 missing stabilizers, and the absolute necessity 00:19:42.779 --> 00:19:45.420 of stamping pressure vents into the metal casing. 00:19:45.920 --> 00:19:47.859 We looked at the ticking clock of the Arrhenius 00:19:47.859 --> 00:19:50.779 10 -degree rule, seeing how molecular vibration 00:19:50.779 --> 00:19:53.759 slowly boils away the component's lifeblood, 00:19:53.920 --> 00:19:57.119 even as that exact same liquid desperately rushes 00:19:57.119 --> 00:19:59.200 to self -heal the microscopic cracks inside. 00:19:59.640 --> 00:20:02.480 And we ended with the spooky reality of time 00:20:02.480 --> 00:20:05.519 -delayed dipoles and lethal ghost voltages. It 00:20:05.519 --> 00:20:08.359 is a component defined entirely by its contradictions. 00:20:08.859 --> 00:20:10.920 It is the heart of our electronic world, yet 00:20:10.920 --> 00:20:14.089 it relies on archaic liquid chemistry. It is 00:20:14.089 --> 00:20:16.490 incredibly fragile, yet it is trusted to power 00:20:16.490 --> 00:20:18.769 almost every device on earth. So next time you 00:20:18.769 --> 00:20:20.670 look at a green circuit board and see those little 00:20:20.670 --> 00:20:22.789 aluminum towers standing there, you aren't just 00:20:22.789 --> 00:20:25.089 looking at solid pieces of metal anymore. You 00:20:25.089 --> 00:20:27.170 know you're looking at tiny pressurized chemical 00:20:27.170 --> 00:20:29.430 tanks doing the heavy, dangerous lifting for 00:20:29.430 --> 00:20:32.369 your digital life. They truly are the unsung, 00:20:32.430 --> 00:20:34.970 hardworking heroes of the hardware world. But 00:20:34.970 --> 00:20:37.109 I want to leave you with one final provocative 00:20:37.109 --> 00:20:40.390 thought. Based on where the tech industry is 00:20:40.390 --> 00:20:43.109 desperately trying to head. Because hardware 00:20:43.109 --> 00:20:45.869 manufacturers absolutely hate the evaporation 00:20:45.869 --> 00:20:48.549 problem. They hate that ticking clock. They are 00:20:48.549 --> 00:20:51.829 actively trying to replace these wet liquid electrolytes 00:20:51.829 --> 00:20:55.569 with completely solid conductive polymers using 00:20:55.569 --> 00:20:58.809 advanced materials like PD. And because these 00:20:58.809 --> 00:21:01.109 polymers are solid, they don't evaporate. Right. 00:21:01.150 --> 00:21:02.950 They completely solve the drying out problem 00:21:02.950 --> 00:21:05.750 and eliminate the ESR death spiral. But as we 00:21:05.750 --> 00:21:08.250 learned today, the solid types completely lack 00:21:08.250 --> 00:21:11.630 the incredible instantaneous oxygen -based self 00:21:11.630 --> 00:21:15.450 -healing properties of the wet liquid. So here 00:21:15.450 --> 00:21:18.130 is something to ponder. If material science can 00:21:18.130 --> 00:21:20.430 ever invent a completely solid polymer that can 00:21:20.430 --> 00:21:22.490 perfectly self -heal its own molecular ruins 00:21:22.490 --> 00:21:25.410 the way a sloshing liquid does, we might finally 00:21:25.410 --> 00:21:27.829 cross the threshold into creating consumer electronics 00:21:27.829 --> 00:21:30.509 with literally immortal power supplies. Now that 00:21:30.509 --> 00:21:33.029 is a fascinating prospect for the future of engineering. 00:21:33.289 --> 00:21:35.369 And until someone invents that magical mortal 00:21:35.369 --> 00:21:37.470 polymer, we will just keep relying on our tiny 00:21:37.470 --> 00:21:39.609 pressurized chemical tanks. Thank you for joining 00:21:39.609 --> 00:21:40.349 us on this deep dive.