WEBVTT 00:00:00.000 --> 00:00:03.459 You plug your laptop into the wall, or you turn 00:00:03.459 --> 00:00:05.980 on your flat screen TV. Right. Or even just tap 00:00:05.980 --> 00:00:07.860 the screen of your smartphone. Yeah, exactly. 00:00:08.199 --> 00:00:11.539 Every single day, you rely on countless electronic 00:00:11.539 --> 00:00:14.480 devices. And you probably trust that they are 00:00:14.480 --> 00:00:17.920 these solid permanent fixtures of modern engineering. 00:00:18.140 --> 00:00:20.179 I mean, most people just imagine the inside of 00:00:20.179 --> 00:00:23.920 their tech as this pristine, dry cityscape of 00:00:23.920 --> 00:00:26.820 silicon, gold, and copper. Right. But hidden 00:00:26.820 --> 00:00:29.320 inside almost all those power supplies and motherboards 00:00:29.320 --> 00:00:32.719 is a tiny liquid -filled canister doing the heavy 00:00:32.719 --> 00:00:34.560 lifting. Which is pretty terrifying when you 00:00:34.560 --> 00:00:37.880 think about it. It really is. And if that little 00:00:37.880 --> 00:00:40.719 canister gets plugged in backward, it won't just 00:00:40.719 --> 00:00:43.439 fail quietly. It will literally explode. Yeah, 00:00:43.439 --> 00:00:45.640 it is the ultimate contradiction of modern electronics. 00:00:46.079 --> 00:00:48.640 We expect our digital world to be dry and static 00:00:48.640 --> 00:00:51.579 and immortal. Oh, for sure. But the reality is 00:00:51.579 --> 00:00:54.039 that our most advanced circuits are utterly dependent 00:00:54.039 --> 00:00:57.100 on what is essentially a microscopic roll of 00:00:57.100 --> 00:00:59.859 wet paper and chemically treated metal. Welcome 00:00:59.859 --> 00:01:02.460 to today's Deep Dive. I'm your host and I'm joined 00:01:02.460 --> 00:01:05.459 by our resident expert. Today, we are looking 00:01:05.459 --> 00:01:09.140 at a comprehensive, highly detailed Wikipedia 00:01:09.140 --> 00:01:13.250 article all about aluminum electrolytic capacitors. 00:01:13.310 --> 00:01:15.450 It's a surprisingly dramatic topic, honestly. 00:01:15.569 --> 00:01:18.909 It is. Our mission today is to demystify this 00:01:18.909 --> 00:01:22.030 ubiquitous but totally hidden component. We are 00:01:22.030 --> 00:01:24.329 going straight into its history, the violent 00:01:24.329 --> 00:01:26.189 chemistry that happens when the manufacturing 00:01:26.189 --> 00:01:29.390 goes wrong, and exactly why our modern tech relies 00:01:29.390 --> 00:01:32.129 so heavily on a tiny can of sloshing liquid. 00:01:32.170 --> 00:01:34.250 It's going to be a wild ride. OK, let's unpack 00:01:34.250 --> 00:01:37.010 this. We all know capacitors dump energy instantly 00:01:37.010 --> 00:01:39.450 by separating two conductive plates with an insulator. 00:01:39.469 --> 00:01:41.670 right? A dielectric. Yeah, that's the basic physics 00:01:41.670 --> 00:01:44.409 of it. But the engineering of this specific dielectric 00:01:44.409 --> 00:01:47.790 is completely wild. When you strip away the outer 00:01:47.790 --> 00:01:51.170 aluminum casing, you don't find high -tech solid 00:01:51.170 --> 00:01:53.730 state physics. Not at all. You find two incredibly 00:01:53.730 --> 00:01:56.609 thin foils of pure aluminum sandwiched around 00:01:56.609 --> 00:01:59.170 a paper spacer that is completely saturated in 00:01:59.170 --> 00:02:01.489 a liquid electrolyte. Right, and that wet sandwich 00:02:01.489 --> 00:02:03.590 is rolled up tight almost like a miniature sleeping 00:02:03.590 --> 00:02:06.290 bag, and then it's sealed inside the metal can. 00:02:06.489 --> 00:02:09.449 A wet sleeping bag? That sounds terrible. Well, 00:02:09.629 --> 00:02:12.289 yes. But for electronics, it's magic. The magic 00:02:12.289 --> 00:02:14.069 really happens on just one of those aluminum 00:02:14.069 --> 00:02:16.689 foils, the anode. The positive side, right? Exactly. 00:02:16.969 --> 00:02:19.669 During manufacturing, that positive foil is put 00:02:19.669 --> 00:02:22.409 through an electrochemical bath to grow a layer 00:02:22.409 --> 00:02:25.449 of aluminum oxide directly on its surface. And 00:02:25.449 --> 00:02:27.990 that oxide layer acts as the insulator. Yes, 00:02:28.110 --> 00:02:30.439 the dielectric. And the source material highlights 00:02:30.439 --> 00:02:33.900 just how insanely thin this layer is. It's measured 00:02:33.900 --> 00:02:37.479 in nanometers per volt. Nanometers? Yes, specifically 00:02:37.479 --> 00:02:41.419 about 1 .2 to 1 .5 nanometers of thickness for 00:02:41.419 --> 00:02:43.819 every volt the capacitor is rated to handle. 00:02:43.919 --> 00:02:45.620 Okay, wait. To put that into perspective for 00:02:45.620 --> 00:02:47.620 you listening, if you have a standard 10 -volt 00:02:47.620 --> 00:02:50.360 repasseter in a household device, that insoluting 00:02:50.360 --> 00:02:53.960 wall is maybe 15 nanometers thick? Which is almost 00:02:53.960 --> 00:02:56.180 impossibly small. Right, because a single strand 00:02:56.180 --> 00:03:00.770 of human hair is roughly 80 ,000 n - Wow. So 00:03:00.770 --> 00:03:02.810 we are talking about an insulating barrier that 00:03:02.810 --> 00:03:05.689 is functionally atomic scale, yet somehow strong 00:03:05.689 --> 00:03:07.990 enough to block the voltage from crossing over. 00:03:08.289 --> 00:03:11.770 And because the formula for capacitance dictates 00:03:11.770 --> 00:03:13.509 that a thinner insulator allows you to store 00:03:13.509 --> 00:03:16.469 more energy, that microscopically thin oxide 00:03:16.469 --> 00:03:18.969 layer is doing an immense amount of work in a 00:03:18.969 --> 00:03:22.069 tiny footprint. But an ultra -thin insulator 00:03:22.069 --> 00:03:24.289 is only half the equation for massive energy 00:03:24.289 --> 00:03:27.090 storage, isn't it? It is. You also need a massive 00:03:27.090 --> 00:03:29.469 surface area for those conductive plates. You 00:03:29.469 --> 00:03:31.689 can't just use a flat sheet of aluminum foil. 00:03:31.909 --> 00:03:33.270 Because it just wouldn't hold enough charge. 00:03:33.389 --> 00:03:36.009 Right. So before they even grow that oxide layer, 00:03:36.469 --> 00:03:38.889 the engineers run the raw aluminum foil through 00:03:38.889 --> 00:03:41.449 a highly corrosive chloride acid bath. While 00:03:41.449 --> 00:03:43.310 pumping and alternating current through it, yeah. 00:03:43.509 --> 00:03:45.969 Exactly. This electrochemical etching literally 00:03:45.969 --> 00:03:48.349 eats away at the metal, pitting it and tunneling 00:03:48.349 --> 00:03:51.770 into it. It sounds so destructive. It does, but 00:03:51.770 --> 00:03:54.590 it transforms a smooth metallic surface into 00:03:54.590 --> 00:03:57.710 a microscopic, sponge -like mountain range of 00:03:57.710 --> 00:04:00.349 ridges, valleys, and deep canyons. That actually 00:04:00.349 --> 00:04:03.150 brings up a great analogy. Imagine you have a 00:04:03.150 --> 00:04:05.750 perfectly flat piece of standard notebook paper 00:04:05.750 --> 00:04:08.169 taking up space on your desk. Okay, I'm picturing 00:04:08.169 --> 00:04:11.659 it. If you crumple that paper into a tight, jagged 00:04:11.659 --> 00:04:14.719 little ball, it takes up a fraction of the physical 00:04:14.719 --> 00:04:17.199 footprint. Oh, right. But if you were a microbe 00:04:17.199 --> 00:04:19.920 trying to walk across every single fold, ridge, 00:04:20.040 --> 00:04:22.839 and inner valley of that crumpled ball, the actual 00:04:22.839 --> 00:04:25.720 surface area you'd have to cover is huge. Exactly 00:04:25.720 --> 00:04:28.279 how it works. The source notes this etching process 00:04:28.279 --> 00:04:31.019 increases the active surface area of the aluminum 00:04:31.019 --> 00:04:34.980 foil by a staggering 200 times. 200 times. That's 00:04:34.980 --> 00:04:37.459 incredible. Yeah, you are packing 200 times more 00:04:37.459 --> 00:04:40.709 capacity. area into the exact same physical cylinder. 00:04:41.029 --> 00:04:43.149 What's fascinating here is how the design actually 00:04:43.149 --> 00:04:46.430 uses the liquid, and it explains why the liquid 00:04:46.430 --> 00:04:48.870 is even there in the first place. Because you 00:04:48.870 --> 00:04:51.629 might assume the second aluminum foil, the one 00:04:51.629 --> 00:04:53.850 on the other side of the wet paper, is the negative 00:04:53.850 --> 00:04:56.250 plate of the capacitor? Right, but it's actually 00:04:56.250 --> 00:04:59.189 not. That second foil is merely a terminal. Wait, 00:04:59.310 --> 00:05:01.470 really? Just a terminal? Yeah, it's just a piece 00:05:01.470 --> 00:05:03.850 of metal used to carry the current out to the 00:05:03.850 --> 00:05:06.910 pin on the bottom. The liquid electrolyte itself 00:05:06.910 --> 00:05:09.290 acts as the true cathode, the second electrode. 00:05:09.509 --> 00:05:11.449 Oh, I see. So the liquid is actually bridging 00:05:11.449 --> 00:05:14.430 the gap. Because if you tried to press a flat 00:05:14.430 --> 00:05:18.250 solid piece of metal against that etched 200 00:05:18.250 --> 00:05:21.170 times magnified microscopic terrain, a solid 00:05:21.170 --> 00:05:23.389 would only bump against the highest peaks of 00:05:23.389 --> 00:05:25.350 the metal. Exactly. You would instantly lose 00:05:25.350 --> 00:05:27.550 99 % of the surface area you just spent all that 00:05:27.550 --> 00:05:30.149 time etching. Ah, because a solid cannot conform 00:05:30.149 --> 00:05:33.620 to a sponge. But a liquid flows. The solvent, 00:05:33.980 --> 00:05:35.959 which is often something like ethylene glycol 00:05:35.959 --> 00:05:38.779 mixed with conductance salts, seeps into every 00:05:38.779 --> 00:05:42.259 single microscopic 15 nanometer canyon and pour 00:05:42.259 --> 00:05:45.000 that etched aluminum. It perfectly molds to the 00:05:45.000 --> 00:05:47.300 rough terrain. It matches the surface area completely. 00:05:47.360 --> 00:05:49.980 Without the liquid acting as a flexible flowing 00:05:49.980 --> 00:05:52.720 electrode, that massive energy storage in such 00:05:52.720 --> 00:05:55.600 a tiny can would be physically impossible. So 00:05:55.600 --> 00:05:57.500 the liquid is literally the only way to make 00:05:57.500 --> 00:06:00.100 physical contact with the entire crumpled surface. 00:06:00.360 --> 00:06:02.769 Exactly right. But, you know, knowing that our 00:06:02.769 --> 00:06:05.269 most advanced hardware relies entirely on a liquid 00:06:05.269 --> 00:06:07.769 electrolyte perfectly coating a microscopic surface 00:06:07.769 --> 00:06:09.790 makes me wonder about the origin of this. How 00:06:09.790 --> 00:06:12.550 they even came up with it. Yeah. Because water, 00:06:13.050 --> 00:06:15.029 liquid solvents, and high -voltage electricity 00:06:15.029 --> 00:06:17.750 are famously a terrible combination. Well, the 00:06:17.750 --> 00:06:19.910 fundamental chemistry actually predates modern 00:06:19.910 --> 00:06:23.110 electronics by decades. Really? How far back? 00:06:23.310 --> 00:06:27.129 Way back. In 1875, a French researcher named 00:06:27.129 --> 00:06:30.050 Eugène Ducretet discovered that certain metals 00:06:30.430 --> 00:06:33.089 which he called valve metals, including aluminum, 00:06:33.529 --> 00:06:36.410 could form an oxide layer. And that layer blocked 00:06:36.410 --> 00:06:38.389 electric current in one direction, but let it 00:06:38.389 --> 00:06:40.629 flow freely in the reverse direction. Right. 00:06:40.750 --> 00:06:43.350 It acted like a mechanical one -way valve, but 00:06:43.350 --> 00:06:46.769 for electricity. Then in 1896, Carol Pollack 00:06:46.769 --> 00:06:48.930 took that observation and realized something 00:06:48.930 --> 00:06:51.730 crucial. He saw that the oxide layer stayed stable 00:06:51.730 --> 00:06:55.089 over time if the aluminum was continuously submerged 00:06:55.089 --> 00:06:57.949 in a neutral or slightly alkaline liquid. Even 00:06:57.949 --> 00:07:00.360 when the power was turned off. And so he patented 00:07:00.360 --> 00:07:02.420 the first electric liquid capacitor. He did. 00:07:02.959 --> 00:07:04.819 But the early industrial applications of this 00:07:04.819 --> 00:07:07.000 are just staggering. Yeah. The source notes that 00:07:07.000 --> 00:07:09.639 in the 1920s, these early capacitors were heavily 00:07:09.639 --> 00:07:12.060 used in massive telephone exchanges. Right. Because 00:07:12.060 --> 00:07:14.819 early DC power supplies for phone lines had a 00:07:14.819 --> 00:07:17.399 lot of AC ripple from the generators. Which created 00:07:17.399 --> 00:07:20.180 a constant loud humming noise on the line. So 00:07:20.180 --> 00:07:22.500 they needed massive capacitance to filter out 00:07:22.500 --> 00:07:25.089 that background noise. And remember, they didn't 00:07:25.089 --> 00:07:27.850 have the technology to electrochemically etch 00:07:27.850 --> 00:07:30.810 the aluminum yet. Oh, right. They couldn't crumple 00:07:30.810 --> 00:07:33.410 the surface area on a microscopic level. So to 00:07:33.410 --> 00:07:35.949 get enough surface area to filter a city's telephone 00:07:35.949 --> 00:07:38.870 exchange, they had to scale up physically. Wait, 00:07:39.769 --> 00:07:43.410 like physically larger containers? Yes. They 00:07:43.410 --> 00:07:46.230 used literal metallic boxes, or vats, acting 00:07:46.230 --> 00:07:48.910 as the cathode, filled with gallons of a liquid 00:07:48.910 --> 00:07:52.009 borax and water electrolyte. Gallons? Yeah, they 00:07:52.009 --> 00:07:55.290 called them wet capacitors. You had a giant aluminum 00:07:55.290 --> 00:07:57.910 plate folded up and suspended inside an open 00:07:57.910 --> 00:08:00.980 tub of swashing liquid. That implies the entire 00:08:00.980 --> 00:08:03.319 telecommunications industry was basically relying 00:08:03.319 --> 00:08:06.019 on open buckets of water and borax sitting next 00:08:06.019 --> 00:08:07.980 to high voltage equipment. That's exactly what 00:08:07.980 --> 00:08:10.339 it was. That's insane. I mean, the vibration 00:08:10.339 --> 00:08:13.920 of a passing train or just a heavy footstep could 00:08:13.920 --> 00:08:16.379 ripple the liquid and disrupt the capacitance. 00:08:16.560 --> 00:08:18.959 Not to mention the nightmare of an actual spill 00:08:18.959 --> 00:08:21.060 shorting out the entire building. Right. It must 00:08:21.060 --> 00:08:23.300 have made early electronics incredibly bulky, 00:08:23.399 --> 00:08:26.500 heavy and, well, geographically limited. It was 00:08:26.500 --> 00:08:28.839 a massive bottleneck for consumer electronics. 00:08:28.879 --> 00:08:31.959 You couldn't exactly put a sloshing vat of Borax 00:08:31.959 --> 00:08:34.139 into a household appliance. No, I wouldn't want 00:08:34.139 --> 00:08:36.419 that in my living room. So the breakthrough happened 00:08:36.419 --> 00:08:41.059 in 1925. An inventor named Samuel Rubin, who 00:08:41.059 --> 00:08:43.740 actually teamed up with Philip Mallory, the founder 00:08:43.740 --> 00:08:45.720 of the company that eventually became Duracell. 00:08:45.820 --> 00:08:49.000 Oh wow, Duracell. Yeah, he patented the dry stacked 00:08:49.000 --> 00:08:51.740 construction. Yeah. Rubin realized he didn't 00:08:51.740 --> 00:08:54.000 need a literal bucket of liquid. You just needed 00:08:54.000 --> 00:08:56.539 the liquid to stay in continuous contact with 00:08:56.539 --> 00:08:59.200 the metal plate. Exactly. So we introduced the 00:08:59.200 --> 00:09:02.480 paper spacer. By soaking highly absorbent paper 00:09:02.480 --> 00:09:05.039 in the electrolyte and sandwiching it between 00:09:05.039 --> 00:09:08.340 the foils, capillary action traps the liquid 00:09:08.340 --> 00:09:10.820 in the paper. And that completely eliminated 00:09:10.820 --> 00:09:13.399 the need for a structural tub of liquid. That 00:09:13.399 --> 00:09:16.340 one mechanical change must have drastically reduced 00:09:16.340 --> 00:09:18.899 the size and manufacturing price of capacitors. 00:09:18.919 --> 00:09:21.399 It changed everything. It was this exact invention 00:09:21.399 --> 00:09:23.779 that allowed manufacturers to build compact, 00:09:24.039 --> 00:09:26.820 stable power supplies. Which made household AC 00:09:26.820 --> 00:09:29.639 radios affordable for a broader group of consumers 00:09:29.639 --> 00:09:33.460 in the late 1920s. Precisely. But the engineering 00:09:33.460 --> 00:09:36.539 quest for better, cheaper, and more conductive 00:09:36.539 --> 00:09:39.240 liquid electrolytes obviously didn't stop with 00:09:39.240 --> 00:09:41.960 borax and paper in the 1920s. Right. Over the 00:09:41.960 --> 00:09:44.440 decades, the industry moved to organic solvents, 00:09:44.740 --> 00:09:47.179 always chasing lower electrical resistance. And 00:09:47.179 --> 00:09:49.200 that relentless pursuit of the perfect liquid 00:09:49.200 --> 00:09:53.059 formula is exactly what led to a massive explosive 00:09:53.059 --> 00:09:56.039 technological disaster a few decades later. Yeah, 00:09:56.220 --> 00:09:57.980 the drive for cheaper components with better 00:09:57.980 --> 00:10:00.279 specifications triggered one of the most infamous 00:10:00.279 --> 00:10:02.340 periods in modern hardware history. It really 00:10:02.340 --> 00:10:04.919 did. And it all revolves around a metric called 00:10:04.919 --> 00:10:08.000 ESR, or equivalent series resistance. Okay, think 00:10:08.000 --> 00:10:11.559 of ESR like friction in a water pipe. If the 00:10:11.559 --> 00:10:13.899 pipe has high resistance, the water struggles 00:10:13.899 --> 00:10:17.159 to get through, creating friction and heat. By 00:10:17.159 --> 00:10:20.240 the late 1990s, computer CPUs were hitting the 00:10:20.240 --> 00:10:22.700 gigahertz range. They were switching incredibly 00:10:22.700 --> 00:10:25.399 fast and needed massive instant gulps of power 00:10:25.399 --> 00:10:28.240 in nanoseconds. And high ESR in a capacitor means 00:10:28.240 --> 00:10:30.679 it resists that sudden draw of current. Which 00:10:30.679 --> 00:10:33.279 causes the voltage to drop and generates a massive 00:10:33.279 --> 00:10:35.419 amount of internal heat. Right, so motherboard 00:10:35.419 --> 00:10:37.820 manufacturers desperately needed capacitors. 00:10:38.409 --> 00:10:41.450 with the lowest friction, the lowest ESR possible. 00:10:41.590 --> 00:10:45.909 So to achieve that ultra -low ESR cheaply, researchers 00:10:45.909 --> 00:10:49.370 in Japan developed new water -based electrolytes. 00:10:49.669 --> 00:10:52.250 Because water is incredibly cheap, and it's a 00:10:52.250 --> 00:10:54.690 phenomenal solvent that significantly improves 00:10:54.690 --> 00:10:56.669 the conductivity of the electrolyte. But there's 00:10:56.669 --> 00:10:59.149 a catch, right? From a chemical standpoint, water 00:10:59.149 --> 00:11:02.190 is highly aggressive toward aluminum. Very aggressive. 00:11:02.250 --> 00:11:04.750 It hydrates the protective aluminum oxide layer. 00:11:04.860 --> 00:11:07.980 Essentially dissolving the very dielectric barrier 00:11:07.980 --> 00:11:11.019 the capacitor relies on. Exactly. So to make 00:11:11.019 --> 00:11:13.620 these water -based liquids work without eating 00:11:13.620 --> 00:11:16.279 the capacitor from the inside out, the engineers 00:11:16.279 --> 00:11:19.340 had to invent highly specific chemical additives. 00:11:19.679 --> 00:11:21.639 Stabilizers. Right. These are chemical buffers 00:11:21.639 --> 00:11:24.139 that passivate the aluminum, essentially repairing 00:11:24.139 --> 00:11:26.620 the oxide layer continuously before the water 00:11:26.620 --> 00:11:29.019 can degrade it. Here's where it gets really interesting. 00:11:29.440 --> 00:11:31.639 What happens if those stabilizers are missing 00:11:31.639 --> 00:11:34.279 or incorrectly formulated? The chemistry goes 00:11:34.279 --> 00:11:36.779 into a runaway state very quickly. And according 00:11:36.779 --> 00:11:39.899 to the source, between the years 2000 and 2005, 00:11:40.340 --> 00:11:43.460 a recipe for an advanced water -based electrolyte 00:11:43.460 --> 00:11:46.460 was actually stolen through corporate espionage. 00:11:46.799 --> 00:11:50.059 Yes. Someone lifted the formula and sold it to 00:11:50.059 --> 00:11:52.930 cheaper manufacturing plants in Taiwan. But... 00:11:52.750 --> 00:11:56.169 The stolen recipe was incomplete. It was entirely 00:11:56.169 --> 00:11:58.590 missing those crucial stabilizing additives. 00:11:58.690 --> 00:12:02.149 So this flawed, highly reactive liquid made its 00:12:02.149 --> 00:12:05.029 way into mass production. Billions of them. It 00:12:05.029 --> 00:12:07.009 got rolled up into millions of capacitors that 00:12:07.009 --> 00:12:09.230 were then soldered onto the motherboards, graphics 00:12:09.230 --> 00:12:11.389 cards, and power supplies of the world's leading 00:12:11.389 --> 00:12:14.049 PC manufacturers. And the physical consequences 00:12:14.049 --> 00:12:17.129 were catastrophic because of the specific chemical 00:12:17.129 --> 00:12:19.389 reaction that occurs when water attacks aluminum. 00:12:19.529 --> 00:12:21.669 It's a violent exothermic reaction, right? Very 00:12:21.669 --> 00:12:24.309 violent. The water transforms the metallic aluminum 00:12:24.309 --> 00:12:27.250 into aluminum hydroxide, and as a byproduct, 00:12:27.330 --> 00:12:30.269 it releases hydrogen gas. Oh man, and all of 00:12:30.269 --> 00:12:32.850 this intense heat and expanding hydrogen gas 00:12:32.850 --> 00:12:34.889 is happening inside a tightly sealed aluminum 00:12:34.889 --> 00:12:36.950 can plugged into a wall outlet. If we connect 00:12:36.950 --> 00:12:39.049 this to the bigger picture, this is why the mechanical 00:12:39.049 --> 00:12:41.450 encapsulation of these devices is so critical. 00:12:41.710 --> 00:12:45.690 The actual design of the can itself. Right. Internal 00:12:45.690 --> 00:12:48.669 gas pressure is a known risk, even in perfectly 00:12:48.669 --> 00:12:51.840 healthy capacitors under heavy load. Because 00:12:51.840 --> 00:12:54.639 of that, the aluminum cans are engineered with 00:12:54.639 --> 00:12:56.919 specific safety features. Oh, like the little 00:12:56.919 --> 00:12:59.559 markings on top. Exactly. If you look at the 00:12:59.559 --> 00:13:02.100 top of a modern electrolytic capacitor, you'll 00:13:02.100 --> 00:13:04.960 see a cross or a K -shape stamped into the metal. 00:13:05.159 --> 00:13:07.059 I always thought those were just decorative. 00:13:07.299 --> 00:13:09.740 No, they are pressure relief vents. They are 00:13:09.740 --> 00:13:12.519 pre -determined breaking points scored into the 00:13:12.519 --> 00:13:14.559 metal. So if the internal gas pressure builds 00:13:14.559 --> 00:13:17.340 up beyond a safe threshold, the metal splits 00:13:17.340 --> 00:13:19.960 at those weakened scores. Allowing the hot gas 00:13:19.960 --> 00:13:22.940 to hiss and vent out safely rather than the entire 00:13:22.940 --> 00:13:25.500 can rupturing and fragmenting like shrapnel. 00:13:26.159 --> 00:13:29.200 Wow. But the hydrogen gas generation from that 00:13:29.200 --> 00:13:32.399 stolen recipe was so massive and so fast that 00:13:32.399 --> 00:13:34.460 even the vents couldn't always handle the pressure 00:13:34.460 --> 00:13:36.820 spike. And unfortunately, in a lot of the cheaper 00:13:36.820 --> 00:13:39.179 knockoff capacitors, the vents were sometimes 00:13:39.179 --> 00:13:41.639 just cosmetic. Wait, they weren't actually scored 00:13:41.639 --> 00:13:44.000 deeply enough to break? Right. The result was 00:13:44.000 --> 00:13:47.240 the infamous capacitor plague. Millions of computer 00:13:47.240 --> 00:13:49.980 components turned into tiny bursting bombs. It 00:13:49.980 --> 00:13:52.159 was terrible. People would be sitting at their 00:13:52.159 --> 00:13:54.700 desks and their computer would emit a loud pop, 00:13:55.340 --> 00:13:58.000 hiss violently, and completely die. Oozing a 00:13:58.000 --> 00:14:00.399 brown corrosive liquid all over the motherboard. 00:14:00.620 --> 00:14:03.059 It was a massive financial blow to the tech industry, 00:14:03.419 --> 00:14:05.259 forcing widespread recalls. But you know what 00:14:05.259 --> 00:14:07.460 the capacitor plague really highlights isn't 00:14:07.460 --> 00:14:09.740 just the danger of stolen intellectual property, 00:14:10.039 --> 00:14:12.840 it proves the inherent daily vulnerability of 00:14:12.840 --> 00:14:15.379 these components. Yes, even with a perfect chemical 00:14:15.379 --> 00:14:18.559 recipe, wet capacitors are constantly at the 00:14:18.559 --> 00:14:21.940 mercy of thermodynamics. So if a flawed liquid 00:14:21.940 --> 00:14:25.120 boils and explodes under stress, what happens 00:14:25.120 --> 00:14:28.299 to a perfectly healthy liquid under normal daily 00:14:28.299 --> 00:14:31.740 operating heat? Well, unlike solid -state components 00:14:31.740 --> 00:14:34.580 like a silicon microchip, which theoretically 00:14:34.580 --> 00:14:36.820 could sit in the server rack and last indefinitely 00:14:36.820 --> 00:14:39.960 if kept in good electrical condition, these wet 00:14:39.960 --> 00:14:43.580 electrolytic capacitors have a definitive, unavoidable 00:14:43.580 --> 00:14:46.100 wear -out failure. It's an unstoppable chemical 00:14:46.100 --> 00:14:48.500 clock, and it all comes back to that liquid. 00:14:48.879 --> 00:14:52.120 Over time, no matter how well crimped and sealed 00:14:52.120 --> 00:14:55.759 the aluminum can is. the liquid electrolyte slowly 00:14:55.759 --> 00:14:58.139 evaporates. Because at the bottom of the component 00:14:58.139 --> 00:15:00.279 where the metallic pins come out, there is a 00:15:00.279 --> 00:15:03.139 rubber elastomer seal. And the liquid slowly 00:15:03.139 --> 00:15:05.559 diffuses right through that rubber molecule by 00:15:05.559 --> 00:15:08.100 molecule. And as the liquid dries out, there 00:15:08.100 --> 00:15:10.759 is less solvent to make contact with the microscopic 00:15:10.759 --> 00:15:13.360 aluminum canyons. Exactly. The active surface 00:15:13.360 --> 00:15:16.279 area shrinks, the capacitance drops, the internal 00:15:16.279 --> 00:15:19.159 resistance skyrockets, and eventually the power 00:15:19.159 --> 00:15:21.480 delivery becomes so unstable that the device 00:15:21.480 --> 00:15:23.730 it's powering starts to glitch and fail. And 00:15:23.730 --> 00:15:26.049 the speed of that diffusion is entirely dictated 00:15:26.049 --> 00:15:28.909 by heat. Yes. The source details the 10 degree 00:15:28.909 --> 00:15:31.149 rule, which is based on the Arrhenius equation. 00:15:31.389 --> 00:15:33.570 Essentially, thermal energy gives molecules the 00:15:33.570 --> 00:15:36.450 kinetic energy they need to break bonds and escape 00:15:36.450 --> 00:15:39.169 as a gas. The rule of thumb in the industry is 00:15:39.169 --> 00:15:41.730 brutally simple. for every 10 degrees Celsius 00:15:41.730 --> 00:15:44.269 drop in operating temperature. The evaporation 00:15:44.269 --> 00:15:46.509 rate of the liquid halves. Which effectively 00:15:46.509 --> 00:15:48.850 doubles the capacitor's lifespan. The math on 00:15:48.850 --> 00:15:51.210 that 10 degree rule is staggering when applied 00:15:51.210 --> 00:15:54.230 to real -world engineering. It really is. Manufacturers 00:15:54.230 --> 00:15:57.830 rate capacitors for a specific lifespan at their 00:15:57.830 --> 00:16:00.360 absolute maximum temperature rating. Let's say 00:16:00.360 --> 00:16:03.460 a capacitor is rated to survive for 2 ,000 hours 00:16:03.460 --> 00:16:06.539 at 105 degrees Celsius. Right, and 2 ,000 hours 00:16:06.539 --> 00:16:09.240 is barely three months of continuous use. That 00:16:09.240 --> 00:16:11.480 sounds like a terrible lifespan for a router 00:16:11.480 --> 00:16:14.220 or a TV. But nobody runs their living room TV 00:16:14.220 --> 00:16:17.809 at 105 degrees Celsius. Thankfully, if you operate 00:16:17.809 --> 00:16:20.529 that exact same device in a well -ventilated 00:16:20.529 --> 00:16:22.690 entertainment center where the internal capacitor 00:16:22.690 --> 00:16:25.769 only reaches 45 degrees Celsius, you apply that 00:16:25.769 --> 00:16:27.929 Arrhenius 10 degree rule. Let's do the math. 00:16:28.110 --> 00:16:29.990 You are dropping the temperature from 105 down 00:16:29.990 --> 00:16:32.870 to 45. That's six increments of 10 degrees. So 00:16:32.870 --> 00:16:35.889 you double that 2 ,000 hour lifespan six consecutive 00:16:35.889 --> 00:16:39.450 times. Exactly. 2 ,000 becomes 4 ,000, then 8 00:16:39.450 --> 00:16:43.360 ,000. Then 16 ,000, 32 ,000, 64 ,000. All the 00:16:43.360 --> 00:16:46.200 way up to 128 ,000 hours. That is roughly 15 00:16:46.200 --> 00:16:49.100 years of solid life simply by keeping the component 00:16:49.100 --> 00:16:52.080 cool and reducing the kinetic energy of the liquid. 00:16:52.299 --> 00:16:54.820 It proves mathematically why heat is the ultimate 00:16:54.820 --> 00:16:56.899 enemy of your electronics. But heat isn't the 00:16:56.899 --> 00:16:58.799 only way to kill them. We mentioned the explosions 00:16:58.799 --> 00:17:01.500 in the introduction. If you are building a circuit 00:17:01.500 --> 00:17:04.380 and you solder one of these wet capacitors in 00:17:04.380 --> 00:17:07.460 backward, which is called reverse polarity, you 00:17:07.460 --> 00:17:10.720 trigger an immediate disaster. The source explains 00:17:10.720 --> 00:17:13.079 that applying reverse voltage actively attacks 00:17:13.079 --> 00:17:16.119 the chemistry. It forces a current that dissolves 00:17:16.119 --> 00:17:19.000 the ultra thin protective oxide layer on the 00:17:19.000 --> 00:17:21.769 anode. And without that insulator, a massive 00:17:21.769 --> 00:17:24.190 current surges through the raw metal, generating 00:17:24.190 --> 00:17:28.109 a huge spike of hydrogen gas that causes a catastrophic, 00:17:28.230 --> 00:17:31.190 spectacular burst. The chemical balance is completely 00:17:31.190 --> 00:17:33.750 inverted. The protective wall is stripped away, 00:17:33.930 --> 00:17:36.289 the liquid boils instantly from the short circuit, 00:17:36.490 --> 00:17:39.430 and the can ruptures. It's violent and instant. 00:17:39.529 --> 00:17:41.849 Wait, if the 10 degree rule means that ambient 00:17:41.849 --> 00:17:43.990 heat is constantly killing them by drying them 00:17:43.990 --> 00:17:46.130 out, and if simply soldering them in backward 00:17:46.130 --> 00:17:49.450 causes literal explosions, I... I don't get it. 00:17:49.630 --> 00:17:51.609 What do you mean? Well, solid -state polymers 00:17:51.609 --> 00:17:54.410 exist. We use solid capacitors in high -end smartphones 00:17:54.410 --> 00:17:56.589 and ultra -thin laptops. Why are we still risking 00:17:56.589 --> 00:17:59.210 our desktop motherboards, power grids, and daily 00:17:59.210 --> 00:18:01.630 appliances with a technology that relies on sloshing 00:18:01.630 --> 00:18:03.930 liquid that wants to evaporate or explode? This 00:18:03.930 --> 00:18:06.569 raises an important question. It's a fundamental 00:18:06.569 --> 00:18:09.509 engineering trade -off, and it points to a hidden 00:18:09.509 --> 00:18:12.069 superpower of the liquid design. A superpower? 00:18:12.190 --> 00:18:15.430 Yes. Solid polymer electrolytes do exist, and 00:18:15.430 --> 00:18:18.210 they don't evaporate. But the liquid electrolyte 00:18:18.269 --> 00:18:21.289 possesses an ability that solid materials fundamentally 00:18:21.289 --> 00:18:24.789 lack. Which is? The ability to self -heal. Self 00:18:24.789 --> 00:18:27.490 -healing? Like a biological scab? Very much like 00:18:27.490 --> 00:18:31.009 that. In any active electrical circuit, you constantly 00:18:31.009 --> 00:18:34.009 experience microscopic voltage spikes, or transients, 00:18:34.450 --> 00:18:36.769 that exceed the component's rating. OK. These 00:18:36.769 --> 00:18:39.029 spikes can punch microscopic holes right through 00:18:39.029 --> 00:18:41.710 that 15 nanometer thin aluminum oxide layer. 00:18:41.809 --> 00:18:44.849 Oh, I see. If a voltage spike punctures a purely 00:18:44.849 --> 00:18:47.329 solid capacitor, the component is permanently 00:18:47.329 --> 00:18:50.369 damaged. A short circuit is created, and the 00:18:50.369 --> 00:18:52.890 component is dead forever. But the wet capacitor 00:18:52.890 --> 00:18:55.589 reacts differently to the puncture. Right? The 00:18:55.589 --> 00:18:58.170 liquid electrolyte is rich in oxygen. When a 00:18:58.170 --> 00:19:00.309 voltage spike punches a microscopic hole in the 00:19:00.309 --> 00:19:02.589 oxide layer, the electricity arcs through the 00:19:02.589 --> 00:19:05.650 gap. And that tiny spark creates intense localized 00:19:05.650 --> 00:19:09.349 heat. Exactly. That sudden heat. drives a chemical 00:19:09.349 --> 00:19:11.849 reaction between the liquid electrolyte and the 00:19:11.849 --> 00:19:14.890 newly exposed raw aluminum. Oh, wow. It instantly 00:19:14.890 --> 00:19:17.029 cooks the oxygen in the liquid into a fresh plug 00:19:17.029 --> 00:19:20.089 of aluminum oxide, sealing the hole. So the sheet 00:19:20.089 --> 00:19:22.609 of the short circuit spark itself is what triggers 00:19:22.609 --> 00:19:25.970 the localized anodization. Yes, the liquid sacrifices 00:19:25.970 --> 00:19:29.470 a tiny microscopic fraction of its own volume 00:19:29.470 --> 00:19:33.009 to instantly grow a fresh scab of aluminum oxide 00:19:33.009 --> 00:19:35.890 over the puncture. It constantly actively repairs 00:19:35.890 --> 00:19:38.750 the insulating wall. We accept the slow drying 00:19:38.750 --> 00:19:41.509 out of the vital fluids over 15 years in exchange 00:19:41.509 --> 00:19:44.450 for immense cheap power storage that can actively 00:19:44.450 --> 00:19:47.569 heal its own microscopic wounds against the daily 00:19:47.569 --> 00:19:50.200 relentless assaults of electrical spikes. That 00:19:50.200 --> 00:19:52.339 reframes the entire way I look at my electronics. 00:19:52.539 --> 00:19:54.380 I mean, we view our tech through the lens of 00:19:54.380 --> 00:19:56.920 software, right? Infinitely replicable, perfectly 00:19:56.920 --> 00:19:59.940 preserved, immortal data. But the hardware holding 00:19:59.940 --> 00:20:02.819 that data up is fighting a constant losing battle 00:20:02.819 --> 00:20:05.099 against chemistry, diffusion and thermodynamics. 00:20:05.380 --> 00:20:07.279 Every time you look at a heavy power brick under 00:20:07.279 --> 00:20:09.279 your desk, you now know what's really going on 00:20:09.279 --> 00:20:11.640 inside. There are tiny canisters of liquid in 00:20:11.640 --> 00:20:13.900 there. They are self -healing. They are breathing 00:20:13.900 --> 00:20:16.579 through rubber seals. And they're acting as a 00:20:16.579 --> 00:20:19.799 literal silent chemical clock, determining the 00:20:19.799 --> 00:20:21.859 exact moment your device will finally give out. 00:20:22.019 --> 00:20:24.359 So what does this all mean? The digital realm 00:20:24.359 --> 00:20:26.500 feels permanent, but its physical foundation 00:20:26.500 --> 00:20:29.559 is tied to the inevitable evaporation of a microscopic 00:20:29.559 --> 00:20:32.440 drop of solvent trapped in a crumpled aluminum 00:20:32.440 --> 00:20:35.059 can. It's incredibly fragile when you really 00:20:35.059 --> 00:20:38.339 zoom in. It makes you wonder, as we push to build 00:20:38.339 --> 00:20:42.059 massive heat -generating AI data centers to power 00:20:42.059 --> 00:20:44.480 the sprawling future of software, machine learning, 00:20:44.539 --> 00:20:48.039 and cloud computing, are we fundamentally bottlenecked 00:20:48.039 --> 00:20:50.819 by the fragile, century -old chemistry of wet 00:20:50.819 --> 00:20:53.880 paper and aluminum? We are trying to build infinite, 00:20:54.039 --> 00:20:56.960 immortal digital minds, but their lifeblood still 00:20:56.960 --> 00:21:00.259 relies on a physical liquid that is slowly, inevitably 00:21:00.259 --> 00:21:01.019 drying out.