WEBVTT 00:00:00.000 --> 00:00:04.419 Welcome to today's deep dive. And I got to say, 00:00:04.480 --> 00:00:06.660 I am so excited to jump into this one with you. 00:00:06.780 --> 00:00:09.119 Yeah, me too. It's a really fun topic today. 00:00:09.259 --> 00:00:12.259 It is. So our mission today is based on this 00:00:12.259 --> 00:00:15.519 incredibly detailed encyclopedic article all 00:00:15.519 --> 00:00:18.719 about a single ubiquitous component. It's called 00:00:18.719 --> 00:00:21.800 the aluminum electrolytic capacitor. Right, which 00:00:21.800 --> 00:00:24.890 sounds incredibly dry at first. Totally. I mean 00:00:24.890 --> 00:00:26.989 before your eyes just glaze over the technical 00:00:26.989 --> 00:00:29.289 name think about this for a second. Yeah, you 00:00:29.289 --> 00:00:32.670 look at your smartphone Your laptop the TV mounted 00:00:32.670 --> 00:00:34.929 on your wall and you probably think of them as 00:00:34.929 --> 00:00:38.170 solid Yeah, like static blocks of glass and plastic 00:00:38.170 --> 00:00:41.119 exactly you know, dependable permanent solid 00:00:41.119 --> 00:00:43.619 state stuff. Which is, you know, a very comforting 00:00:43.619 --> 00:00:45.880 illusion. We really like to believe our technology 00:00:45.880 --> 00:00:48.840 is just pure solid state, just electrons zipping 00:00:48.840 --> 00:00:51.479 through rigid metal pathways. Right. But what 00:00:51.479 --> 00:00:53.600 if I told you that hidden inside almost every 00:00:53.600 --> 00:00:56.759 electronic device you own is this vibrant, slightly 00:00:56.759 --> 00:00:59.219 dangerous and honestly endlessly fascinating 00:00:59.219 --> 00:01:02.359 microscopic world? Because underneath your fingertips, 00:01:02.619 --> 00:01:04.700 right this second, there isn't just solid metal 00:01:04.700 --> 00:01:08.439 and wire. There's actually a highly active liquid 00:01:08.439 --> 00:01:10.840 chemical ecosystem working furiously to keep 00:01:10.840 --> 00:01:13.319 your devices running. It's wild. If you could 00:01:13.319 --> 00:01:15.739 shrink down and look inside one of these capacitors, 00:01:15.819 --> 00:01:17.859 you'd see something that behaves a lot more like 00:01:17.859 --> 00:01:20.760 a living biological system than a piece of static 00:01:20.760 --> 00:01:24.459 hardware. OK, so to understand why there is literally 00:01:24.459 --> 00:01:27.500 liquid sloshing around inside our dry, solid 00:01:27.500 --> 00:01:29.739 electronics, we really need to start with what 00:01:29.739 --> 00:01:31.939 a capacitor actually is. Like, what does it do? 00:01:32.030 --> 00:01:34.730 Sure. So at its most basic level, a capacitor 00:01:34.730 --> 00:01:37.390 stores electrical energy. But not like a battery, 00:01:37.390 --> 00:01:40.629 right? Right. Exactly. A battery relies on a 00:01:40.629 --> 00:01:43.290 slow chemical reaction to give you a steady drip 00:01:43.290 --> 00:01:46.790 of power over hours or days. A capacitor is more 00:01:46.790 --> 00:01:50.650 like a pressurized water tank. OK, a water tank. 00:01:50.650 --> 00:01:52.670 I like that. It stores energy in an electric 00:01:52.670 --> 00:01:55.250 field and can just dump its entire payload almost 00:01:55.250 --> 00:01:57.870 instantly. So if your computer suddenly needs 00:01:57.870 --> 00:02:01.010 a massive spike of power for, say, a fraction 00:02:01.010 --> 00:02:03.840 of a second, to render a graphic or access the 00:02:03.840 --> 00:02:06.079 hard drive? The battery is just way too slow 00:02:06.079 --> 00:02:08.240 to provide that. Right. So the capacitor steps 00:02:08.240 --> 00:02:11.080 in, dumps its stored energy to handle the spike, 00:02:11.340 --> 00:02:13.740 and then immediately refills itself. It smooths 00:02:13.740 --> 00:02:16.800 out the power delivery. And the physical construction 00:02:16.800 --> 00:02:20.319 of an aluminum electrolytic capacitor sounds 00:02:20.319 --> 00:02:24.000 incredibly simple at first glance. Like, according 00:02:24.000 --> 00:02:26.400 to the source, if you were to crack open one 00:02:26.400 --> 00:02:28.639 of these little cylinders and unroll the inside, 00:02:29.080 --> 00:02:32.060 you'd essentially find a very long metallic sandwich. 00:02:32.219 --> 00:02:34.139 A metallic sandwich, yeah. That's a good way 00:02:34.139 --> 00:02:36.960 to put it. You have two highly purified aluminum 00:02:36.960 --> 00:02:39.819 foils. One is the anode, the positive side, and 00:02:39.819 --> 00:02:41.639 one is the cathode, the negative side. Right. 00:02:42.139 --> 00:02:44.520 And separating them is this highly absorbent 00:02:44.520 --> 00:02:46.979 paper spacer. You wind that whole sandwich up 00:02:46.979 --> 00:02:50.139 tightly, soak the paper in a liquid or gel electrolyte, 00:02:50.259 --> 00:02:52.680 shove it into an aluminum can, and seal it with 00:02:52.680 --> 00:02:54.860 a rubber plug. Which, I mean, sounds like a basic 00:02:54.860 --> 00:02:56.479 high school science project. It really does. 00:02:56.620 --> 00:02:58.780 But the magic, and honestly, the engineering 00:02:58.780 --> 00:03:01.240 nightmare here, comes down to a really stubborn 00:03:01.240 --> 00:03:04.900 mathematical reality. The capacity to store energy, 00:03:05.180 --> 00:03:08.280 which we call capacitance, is dictated by a strict 00:03:08.280 --> 00:03:10.180 formula. All right. It's basically area divided 00:03:10.180 --> 00:03:12.860 by the thickness of the insulator. Exactly. It's 00:03:12.860 --> 00:03:15.340 the surface area of your metal plates divided 00:03:15.340 --> 00:03:17.699 by the distance between them. So if you want 00:03:17.699 --> 00:03:19.560 to store a massive amount of energy, you need 00:03:19.560 --> 00:03:21.879 massive metal plates. Yeah. But you also need 00:03:21.879 --> 00:03:24.460 this thing to fit onto a tiny computer motherboard. 00:03:25.180 --> 00:03:28.319 Like, you can't just put a kitchen table -sized 00:03:28.319 --> 00:03:31.120 sheet of aluminum inside a smartphone. You really 00:03:31.120 --> 00:03:34.240 can't. You need massive surface area in a tiny 00:03:34.240 --> 00:03:37.560 footprint. And the industry's solution to this 00:03:37.560 --> 00:03:40.199 is an electrochemical process called etching. 00:03:40.379 --> 00:03:43.139 Etching. OK, how does that work? Manufacturers 00:03:43.139 --> 00:03:46.060 don't use smooth aluminum foil for the positive 00:03:46.060 --> 00:03:48.659 anode. Before they ever roll it up, they run 00:03:48.659 --> 00:03:51.740 that foil through a bath of high aggressive acids 00:03:51.740 --> 00:03:54.460 and hit it with high electrical currents. Which 00:03:54.460 --> 00:03:56.620 basically eats away at the metal, right? It does. 00:03:56.900 --> 00:03:59.259 It chemically blasts these microscopic tunnels 00:03:59.259 --> 00:04:01.919 and pits directly into the aluminum. They are 00:04:01.919 --> 00:04:04.360 roughening it up on a deeply microscopic level. 00:04:04.479 --> 00:04:06.599 The source material has a staggering statistic 00:04:06.599 --> 00:04:09.280 on this, by the way. This etching process can 00:04:09.280 --> 00:04:11.419 increase the surface area of the fold by a factor 00:04:11.419 --> 00:04:14.759 of up to 200 compared to a smooth surface. 200 00:04:14.759 --> 00:04:18.160 times. Yeah, 200 times the area in the exact 00:04:18.160 --> 00:04:20.480 same physical footprint. It's like taking a perfectly 00:04:20.480 --> 00:04:23.060 smooth piece of paper and crumpling it up into 00:04:23.060 --> 00:04:25.899 a tiny dense ball. Or a mountain range. Yes. 00:04:26.420 --> 00:04:29.560 Imagine a perfectly flat plot of land. If you 00:04:29.560 --> 00:04:32.420 suddenly raise a jagged microscopic mountain 00:04:32.420 --> 00:04:35.279 range all over that land, the actual surface 00:04:35.279 --> 00:04:37.779 area of all those peaks and valleys combined 00:04:37.779 --> 00:04:41.800 is astronomical compared to the flat dirt. You're 00:04:41.800 --> 00:04:44.480 packing the surface area of a sprawling estate 00:04:44.480 --> 00:04:46.839 into the size of a thimble. But here's where 00:04:46.839 --> 00:04:49.240 the engineering gets incredibly tricky. You've 00:04:49.240 --> 00:04:52.879 just created this jagged microscopic mountain 00:04:52.879 --> 00:04:55.399 range of aluminum. Right. Now, to make a working 00:04:55.399 --> 00:04:57.699 capacitor, you have to cover that entire mountain 00:04:57.699 --> 00:05:00.000 range with a perfectly even electrical insulator. 00:05:00.180 --> 00:05:03.139 OK. And then you need a second metallic electrode 00:05:03.139 --> 00:05:05.699 to perfectly touch that insulator absolutely 00:05:05.699 --> 00:05:08.500 everywhere. Wait, if the oxide layer is the actual 00:05:08.500 --> 00:05:10.889 insulator, why do we even need the liquid? Can 00:05:10.889 --> 00:05:13.430 we just press the second piece of metal, like 00:05:13.430 --> 00:05:16.029 the cathode foil, right against the oxide -coated 00:05:16.029 --> 00:05:17.930 mountain range? You'd think so, but remember 00:05:17.930 --> 00:05:20.069 the scale of what we're talking about here. Metal 00:05:20.069 --> 00:05:23.250 is rigid. If you take a flat, solid piece of 00:05:23.250 --> 00:05:25.930 aluminum foil and press it against a microscopic 00:05:25.930 --> 00:05:29.639 mountain range, what happens? Oh. Ah. I see. 00:05:29.720 --> 00:05:31.980 It's only going to physically touch the very 00:05:31.980 --> 00:05:33.980 highest peaks of the mountains. Precisely. It 00:05:33.980 --> 00:05:36.100 wouldn't conform to the valleys at all. All of 00:05:36.100 --> 00:05:38.259 that massive surface area, you just spent so 00:05:38.259 --> 00:05:40.319 much time and money etching into the first foil. 00:05:40.500 --> 00:05:42.660 It would be completely wasted. Completely wasted 00:05:42.660 --> 00:05:44.560 because the second foil just can't reach down 00:05:44.560 --> 00:05:47.019 into the microscopic pits. OK. That makes perfect 00:05:47.019 --> 00:05:50.019 sense. And that is why these things cannot be 00:05:50.019 --> 00:05:53.060 completely solid. Exactly. The actual insulator, 00:05:53.240 --> 00:05:56.259 the dielectric, is a layer of aluminum oxide. 00:05:56.560 --> 00:06:00.300 grown directly onto that jagged anode foil. And 00:06:00.300 --> 00:06:02.699 the force points out that this layer is unbelievably 00:06:02.699 --> 00:06:06.699 thin. We are talking about 1 .4 nanometers thick 00:06:06.699 --> 00:06:09.899 for every volt of electricity it's designed to 00:06:09.899 --> 00:06:12.279 hold back. It's a number that's honestly hard 00:06:12.279 --> 00:06:15.120 to even conceptualize. Well, let's put it in 00:06:15.120 --> 00:06:17.259 perspective for a second. A single human hair 00:06:17.259 --> 00:06:20.259 is roughly 80 ,000 nanometers thick. Wow. So 00:06:20.259 --> 00:06:22.060 we are talking about an electrical insulator 00:06:22.060 --> 00:06:24.220 that is tens of thousands of times thinner than 00:06:24.220 --> 00:06:27.329 a human hair. Right. So you have an unimaginably 00:06:27.329 --> 00:06:31.370 thin insulator draped over a wildly jagged microscopic 00:06:31.370 --> 00:06:34.689 mountain range. To perfectly match that incredibly 00:06:34.689 --> 00:06:38.149 complex surface, you just cannot use a solid 00:06:38.149 --> 00:06:41.189 metal plate. You need a liquid. Yes. The liquid 00:06:41.189 --> 00:06:43.430 electrolyte, that stuff soaking the paper spacer 00:06:43.430 --> 00:06:46.810 we mentioned earlier, that acts as the true second 00:06:46.810 --> 00:06:49.990 electrode. Because it's a liquid. It flows effortlessly 00:06:49.990 --> 00:06:52.550 into every single microscopic valley, every pit, 00:06:52.889 --> 00:06:55.329 every tunnel. Yeah. It perfectly conforms to 00:06:55.329 --> 00:06:58.009 the exact shape of the incredibly thin oxide 00:06:58.009 --> 00:07:00.310 layer. Exactly. So the second aluminum foil, 00:07:00.389 --> 00:07:02.430 the one we normally call the negative cathode, 00:07:02.930 --> 00:07:05.310 is really just a metal bridge to carry the electricity 00:07:05.310 --> 00:07:07.310 from the circuit board into that conductive liquid. 00:07:07.470 --> 00:07:10.290 That's exactly it. The liquid is the real star 00:07:10.290 --> 00:07:12.769 of the show here. But the source material explains 00:07:12.769 --> 00:07:15.269 that this liquid electrolyte isn't just a passive 00:07:15.269 --> 00:07:17.889 conductor filling in the gaps. It actually has 00:07:17.889 --> 00:07:21.350 a vital active secondary job. And this, honestly, 00:07:21.529 --> 00:07:23.709 this blew my mind when I read the research. Because 00:07:23.709 --> 00:07:26.589 you have this 1 .4 nanometer thick insulator 00:07:26.589 --> 00:07:29.509 under immense electrical stress. Oh yeah, constantly. 00:07:29.709 --> 00:07:32.410 If a microscopic flaw develops in that oxide 00:07:32.410 --> 00:07:35.509 layer, or if it cracks from heat or vibration, 00:07:36.069 --> 00:07:37.829 the electricity would just punch right through 00:07:37.829 --> 00:07:40.170 it. The capacitor would short circuit, and your 00:07:40.170 --> 00:07:43.370 device would just die. Instantly. So are you 00:07:43.370 --> 00:07:45.410 telling me that the liquid electrolyte actually 00:07:45.410 --> 00:07:48.009 acts like a biological immune system to prevent 00:07:48.009 --> 00:07:50.170 this? I mean, that's not even an exaggeration. 00:07:50.209 --> 00:07:52.829 The immune system analogy is remarkably accurate, 00:07:52.930 --> 00:07:55.170 because the liquid electrolyte actually contains 00:07:55.170 --> 00:07:58.569 oxygen. So while your computer or TV is running, 00:07:58.949 --> 00:08:01.569 if a microscopic crack appears in the oxide layer 00:08:01.569 --> 00:08:04.029 and current starts to leak through, the electrolyte 00:08:04.029 --> 00:08:07.170 immediately reacts. Oh, so? The leaking electrical 00:08:07.170 --> 00:08:09.569 current hits the electrolyte. generates a tiny 00:08:09.569 --> 00:08:12.430 bit of heat, and causes the oxygen in the liquid 00:08:12.430 --> 00:08:15.529 to chemically bond with the newly exposed raw 00:08:15.529 --> 00:08:18.129 aluminum in the crack. It literally builds a 00:08:18.129 --> 00:08:20.529 new scab of aluminum oxide right over the flaw. 00:08:21.110 --> 00:08:23.810 Exactly. It chemically self -heals in real time. 00:08:24.009 --> 00:08:26.569 It patches the hole while your device is actively 00:08:26.569 --> 00:08:28.490 running without you ever knowing it even happened. 00:08:28.550 --> 00:08:31.810 That is just brilliant. It's a dynamic, living 00:08:31.810 --> 00:08:34.289 chemical reaction sitting right next to your 00:08:34.289 --> 00:08:37.039 computer's processor. Yeah, it really is. And 00:08:37.039 --> 00:08:40.179 this specific combination, the massively etched 00:08:40.179 --> 00:08:43.340 surface area, the ultra -thin oxide layer, and 00:08:43.340 --> 00:08:46.139 the self -healing liquid filling the gaps, that 00:08:46.139 --> 00:08:48.039 is what allows these components to be incredibly 00:08:48.039 --> 00:08:51.440 cheap while holding huge amounts of energy. The 00:08:51.440 --> 00:08:53.559 source notes they have a capacitance range from 00:08:53.559 --> 00:08:55.960 a fraction of a microfarad all the way up to 00:08:55.960 --> 00:08:59.450 a massive 2 .7 farads. What's fascinating here 00:08:59.450 --> 00:09:02.649 is how gorgeous of a synthesis this is between 00:09:02.649 --> 00:09:06.049 chemistry and electrical engineering. But relying 00:09:06.049 --> 00:09:09.110 on an active liquid chemical reaction inside 00:09:09.110 --> 00:09:11.610 a sealed metal can... There's a catch, right? 00:09:11.750 --> 00:09:13.649 Oh, there's a serious double -edged sword here. 00:09:14.009 --> 00:09:16.269 It introduces vulnerabilities that purely solid 00:09:16.269 --> 00:09:18.710 -state components simply don't have. Right. Because 00:09:18.710 --> 00:09:20.870 if it's so great at healing itself, what's the 00:09:20.870 --> 00:09:23.269 catch? Why do they have such a notoriously bad 00:09:23.269 --> 00:09:25.169 reputation in the repair community? Well, the 00:09:25.169 --> 00:09:27.389 biggest catch is that they are strictly polarized. 00:09:27.399 --> 00:09:30.080 They have a designated positive and negative 00:09:30.080 --> 00:09:33.179 terminal and the direct current, the DC voltage, 00:09:33.500 --> 00:09:35.720 must flow in the correct direction. Or else. 00:09:35.919 --> 00:09:38.539 Or else. The entire system is carefully designed 00:09:38.539 --> 00:09:41.559 so that the thick insulating oxide layer is heavily 00:09:41.559 --> 00:09:43.740 grown on the positive anode. And the negative 00:09:43.740 --> 00:09:46.100 cathode side. The negative cathode foil is essentially 00:09:46.100 --> 00:09:49.279 just bare aluminum. It only has a very thin, 00:09:49.580 --> 00:09:52.100 naturally occurring oxide layer just from sitting 00:09:52.100 --> 00:09:54.980 out in the air in the factory. Right. The encyclopedic 00:09:54.980 --> 00:09:58.080 source notes that this natural thin layer can 00:09:58.080 --> 00:10:01.019 only withstand a reverse voltage of about 1 .5 00:10:01.019 --> 00:10:03.879 volts. Barely anything. So what actually happens 00:10:03.879 --> 00:10:06.320 if someone at the factory wires this thing backward. 00:10:06.879 --> 00:10:09.700 Or if a power surge reverses the flow beyond 00:10:09.700 --> 00:10:13.159 negative 1 .5 volts, why does it explode? Because 00:10:13.159 --> 00:10:15.100 you're forcing electricity to flow the wrong 00:10:15.100 --> 00:10:18.259 way through a chemical bath. Instead of the current 00:10:18.259 --> 00:10:20.440 triggering the electrolyte to heal the oxide 00:10:20.440 --> 00:10:23.159 layer, the reverse current forcefully breaks 00:10:23.159 --> 00:10:25.159 down the water and chemicals in the electrolyte. 00:10:25.379 --> 00:10:27.720 And a major byproduct of breaking down those 00:10:27.720 --> 00:10:31.889 chemicals is hydrogen gas. Oh, wow. And gas expands. 00:10:32.629 --> 00:10:35.289 Inside a completely sealed aluminum can, gas 00:10:35.289 --> 00:10:37.610 means massive pressure. And that pressure builds 00:10:37.610 --> 00:10:39.809 rapidly. If you keep applying reverse voltage, 00:10:39.970 --> 00:10:42.330 the liquid boils, the gas expands, and the metal 00:10:42.330 --> 00:10:45.450 can bursts, often in a spectacular dramatic fashion. 00:10:45.669 --> 00:10:49.269 It explodes. It will literally spray boiling 00:10:49.269 --> 00:10:52.429 hot electrolyte and shredded electrically charged 00:10:52.429 --> 00:10:55.370 paper everywhere. Which is essentially a tiny 00:10:55.370 --> 00:10:58.129 fragmentation grenade going off inside your power 00:10:58.129 --> 00:11:00.600 supply. Basically. This is why if you look at 00:11:00.600 --> 00:11:02.840 the top of a modern cylindrical capacitor on 00:11:02.840 --> 00:11:05.720 a circuit board, you'll almost always see a little 00:11:05.720 --> 00:11:08.519 cross or a K shape stamped directly into the 00:11:08.519 --> 00:11:10.980 aluminum lid. Right, the vents. I always wondered 00:11:10.980 --> 00:11:13.200 what those were for. They are predetermined breaking 00:11:13.200 --> 00:11:16.460 points or safety vents. They are intentionally 00:11:16.460 --> 00:11:20.139 weakened areas of the metal designed to rupture 00:11:20.139 --> 00:11:23.460 and quietly vent the hydrogen gas before the 00:11:23.460 --> 00:11:26.200 entire can turns into shrapnel. It's a completely 00:11:26.200 --> 00:11:30.049 necessary safety valve, but sometimes Even correct 00:11:30.049 --> 00:11:32.250 wiring and safety vents aren't enough to prevent 00:11:32.250 --> 00:11:34.710 a disaster. Especially if the chemistry is wrong. 00:11:35.049 --> 00:11:37.509 Exactly. Especially if the fundamental chemistry 00:11:37.509 --> 00:11:39.610 of the electrolyte itself is compromised. Which 00:11:39.610 --> 00:11:41.429 brings us to the industrial espionage part of 00:11:41.429 --> 00:11:43.950 our deep dive, the infamous capacitor plague. 00:11:44.169 --> 00:11:47.169 Oh yeah, the plague. The research details this 00:11:47.169 --> 00:11:50.789 period roughly between the years 2000 and 2005. 00:11:51.590 --> 00:11:53.889 The industry was aggressively shifting toward 00:11:53.889 --> 00:11:57.850 new, highly conductive, water -based electrolytes. 00:11:58.159 --> 00:12:01.240 Why the sudden shift to water? It was all about 00:12:01.240 --> 00:12:04.360 efficiency. Water is an incredibly good conductor 00:12:04.360 --> 00:12:07.259 of electricity. If you use a high water content 00:12:07.259 --> 00:12:10.259 electrolyte, the internal electrical resistance 00:12:10.259 --> 00:12:13.559 of the capacitor drops significantly. This makes 00:12:13.559 --> 00:12:15.759 the component vastly more efficient at handling 00:12:15.759 --> 00:12:18.600 power and crucially much cheaper to manufacture. 00:12:18.840 --> 00:12:20.740 But there's a glaring chemical problem there. 00:12:21.240 --> 00:12:23.059 Anyone who has ever left a bicycle out in the 00:12:23.059 --> 00:12:25.620 rain knows that water naturally oxidizes bare 00:12:25.620 --> 00:12:28.679 metal. It rusts it. Exactly. Water reacts very 00:12:28.679 --> 00:12:31.220 aggressively with bare aluminum. So to safely 00:12:31.220 --> 00:12:33.700 use water inside a capacitor, manufacturers had 00:12:33.700 --> 00:12:36.659 to develop highly guarded proprietary chemical 00:12:36.659 --> 00:12:39.159 inhibitors. Like stabilizers. Right, stabilizers 00:12:39.159 --> 00:12:41.159 that essentially stop the water from naturally 00:12:41.159 --> 00:12:43.639 attacking the aluminum. And here is where the 00:12:43.639 --> 00:12:46.259 thriller movie plot kicks in. A recipe for one 00:12:46.259 --> 00:12:48.940 of these advanced, highly efficient water -based 00:12:48.940 --> 00:12:52.279 electrolytes was stolen. It was taken by a rogue 00:12:52.279 --> 00:12:55.779 scientist, allegedly, and sold to various materials 00:12:55.779 --> 00:12:58.799 manufacturers in Asia who then mass produced 00:12:58.799 --> 00:13:02.500 it and sold it to huge capacitor brands worldwide. 00:13:02.899 --> 00:13:05.000 But the stolen recipe they were working from 00:13:05.000 --> 00:13:08.500 was incomplete? Yes. It was missing the vital 00:13:08.500 --> 00:13:11.500 chemical stabilizing substances. the inhibitors. 00:13:11.620 --> 00:13:14.379 So you have millions of components rolling off 00:13:14.379 --> 00:13:16.919 the assembly lines, completely lacking the chemical 00:13:16.919 --> 00:13:18.820 breaks needed to stop the water from attacking 00:13:18.820 --> 00:13:21.159 the aluminum. What a disaster. It really was. 00:13:21.480 --> 00:13:23.779 Without those stabilizers, the water inside these 00:13:23.779 --> 00:13:25.799 capacitors began reacting violently with the 00:13:25.799 --> 00:13:27.960 aluminum housing, creating aluminum hydroxide 00:13:27.960 --> 00:13:30.440 and, crucially, large amounts of hydrogen gas. 00:13:30.679 --> 00:13:33.120 And this wasn't from being wired backward? No, 00:13:33.120 --> 00:13:35.000 not at all. This was happening under normal, 00:13:35.220 --> 00:13:37.759 everyday operating conditions. Just a runaway 00:13:37.759 --> 00:13:39.940 exothermic reaction sitting right there on your 00:13:39.940 --> 00:13:43.159 motherboard. The result was mass bursting capacitors 00:13:43.159 --> 00:13:46.860 worldwide. Desktop computers, server power supplies, 00:13:47.220 --> 00:13:50.039 early flat screen televisions, internet routers, 00:13:50.360 --> 00:13:53.720 all failing prematurely as their capacitors bulged, 00:13:54.200 --> 00:13:56.200 leaked this brownish fluid and just popped open. 00:13:57.019 --> 00:14:00.200 So what does this all mean? A botched industrial 00:14:00.200 --> 00:14:02.860 heist literally caused computers around the globe 00:14:02.860 --> 00:14:05.320 to explode like tiny popping fireworks because 00:14:05.320 --> 00:14:08.000 they forgot the chemical stabilizer. It really 00:14:08.000 --> 00:14:10.460 is a stark reminder of supply chain vulnerability. 00:14:10.799 --> 00:14:12.679 We tend to think of modern electronics as being 00:14:12.679 --> 00:14:15.460 defined by their incredibly complex silicon microchips 00:14:15.460 --> 00:14:18.179 and processors. Sure. But during the capacitor 00:14:18.179 --> 00:14:20.980 plague, a single missing chemical additive in 00:14:20.980 --> 00:14:23.940 a microscopic layer of a basic inexpensive component 00:14:23.940 --> 00:14:26.840 brought down a massive swath of the global consumer 00:14:26.840 --> 00:14:29.559 electronics market. It totally recontextualizes 00:14:29.559 --> 00:14:31.919 how fragile the ecosystem inside our devices 00:14:31.919 --> 00:14:34.299 truly is. Yeah. But here's the truly sobering 00:14:34.299 --> 00:14:36.000 part for anyone listening to us right now. Oh, 00:14:36.000 --> 00:14:39.299 the wearout. Yeah. Even if you don't have a stolen 00:14:39.299 --> 00:14:42.399 unstable recipe, and even if you never wire a 00:14:42.399 --> 00:14:45.360 circuit backward, these liquid -filled cans still 00:14:45.360 --> 00:14:48.220 have a fatal flaw. From the very second you turn 00:14:48.220 --> 00:14:51.539 your device on, they are slowly, inevitably dying. 00:14:51.779 --> 00:14:54.220 Yes. In engineering terms, this is known as wear 00:14:54.220 --> 00:14:57.379 -out failure. Because they fundamentally rely 00:14:57.379 --> 00:15:00.559 on a liquid electrolyte, it's physically impossible 00:15:00.559 --> 00:15:03.379 for them to last forever. Because liquids evaporate. 00:15:03.600 --> 00:15:06.500 Over time, that liquid slowly evaporates. No 00:15:06.500 --> 00:15:09.460 seal is perfect. The vapor diffuses right through 00:15:09.460 --> 00:15:11.379 the rubber plug at the bottom of the can. And 00:15:11.379 --> 00:15:13.820 as it dries out, what physically happens to the 00:15:13.820 --> 00:15:16.080 electrical performance? Well, as you lose liquid, 00:15:16.179 --> 00:15:18.399 you lose your connection to the microscopic values 00:15:18.399 --> 00:15:20.879 of the aluminum. The overall capacitance, the 00:15:20.879 --> 00:15:23.460 amount of energy it can store, slowly drops. 00:15:23.539 --> 00:15:26.620 OK. But more dangerously, the internal resistance 00:15:26.620 --> 00:15:30.080 starts to climb. Engineers call this ESR, or 00:15:30.080 --> 00:15:32.450 equivalent series resistance. Let's break down 00:15:32.450 --> 00:15:35.370 why ESR matters. The source explains that these 00:15:35.370 --> 00:15:37.629 capacitors are constantly dealing with ripple 00:15:37.629 --> 00:15:40.649 currents. What exactly is a ripple current? Imagine 00:15:40.649 --> 00:15:43.909 a river trying to flow smoothly, but there are 00:15:43.909 --> 00:15:46.110 constant surges and drops in the water pressure. 00:15:46.809 --> 00:15:49.789 The capacitor's job in a power supply is to absorb 00:15:49.789 --> 00:15:52.309 those surges and release water during the drops, 00:15:52.830 --> 00:15:54.929 keeping the overall flow perfectly steady to 00:15:54.929 --> 00:15:57.980 your processor. Right. Those surges are the ripple 00:15:57.980 --> 00:16:01.120 currents, superimposed alternating currents that 00:16:01.120 --> 00:16:04.279 ride on top of the steady DC power. Okay, so 00:16:04.279 --> 00:16:06.620 the capacitor is working hard to filter out those 00:16:06.620 --> 00:16:09.100 surges. But as the liquid evaporates and the 00:16:09.100 --> 00:16:11.679 ESR goes up, it's like the capacitor's internal 00:16:11.679 --> 00:16:13.559 plugs are getting clogged. Exactly. When those 00:16:13.559 --> 00:16:16.039 electrical surges, the ripple currents hit that 00:16:16.039 --> 00:16:18.740 high resistance. they create electrical friction. 00:16:19.320 --> 00:16:21.419 And electrical friction means heat. Which creates 00:16:21.419 --> 00:16:24.600 a vicious self -accelerating cycle. Oh no. The 00:16:24.600 --> 00:16:27.100 ripple currents hit the high resistance, which 00:16:27.100 --> 00:16:30.259 creates internal heat. That heat cooks the liquid, 00:16:30.679 --> 00:16:32.860 speeding up the evaporation. Which increases 00:16:32.860 --> 00:16:35.559 the resistance again. The loss of liquid increases 00:16:35.559 --> 00:16:38.320 the resistance even further, which then generates 00:16:38.320 --> 00:16:40.919 even more heat. It's a literal ticking clock. 00:16:41.399 --> 00:16:45.259 And managing that heat is the single most important 00:16:45.259 --> 00:16:47.360 factor in how long your devices will actually 00:16:47.360 --> 00:16:51.200 survive. In the engineering world, the degradation 00:16:51.200 --> 00:16:53.539 of these chemical components is governed by something 00:16:53.539 --> 00:16:57.039 called the Arrhenius rule, which is widely known 00:16:57.039 --> 00:16:59.179 in the electronics industry as the 10 degree 00:16:59.179 --> 00:17:02.259 rule. The Arrhenius rule dictates how chemical 00:17:02.259 --> 00:17:05.460 reaction rates change with temperature. In the 00:17:05.460 --> 00:17:07.480 context of our capacitors drying out, it gives 00:17:07.480 --> 00:17:10.140 us a remarkably simple mathematical reality. 00:17:10.660 --> 00:17:13.720 For roughly every 10 degrees Celsius, you lower 00:17:13.720 --> 00:17:15.700 the ambient temperature around the capacitor. 00:17:15.799 --> 00:17:17.799 You have the rate of the chemical evaporation. 00:17:17.839 --> 00:17:20.779 Wow. You literally double the component's lifespan. 00:17:21.140 --> 00:17:23.740 The 10 degree rule. This is a massive practical 00:17:23.740 --> 00:17:25.460 takeaway for you, the listener. Let's put some 00:17:25.460 --> 00:17:28.359 real numbers to that. If a manufacturer specifies 00:17:28.359 --> 00:17:31.579 a capacitor to last 2 ,000 hours at its maximum 00:17:31.579 --> 00:17:34.430 rated temperature of, say, 105 degrees. Celsius. 00:17:34.589 --> 00:17:37.029 2 ,000 hours sounds incredibly short. It is. 00:17:37.549 --> 00:17:40.190 That's less than three months of continuous use. 00:17:40.690 --> 00:17:43.490 But if you can design the airflow in your device 00:17:43.490 --> 00:17:46.150 to keep the operating temperature down to 45 00:17:46.150 --> 00:17:48.630 degrees Celsius, you are dropping the temperature 00:17:48.630 --> 00:17:51.650 by six increments of 10 degrees. Okay, so you 00:17:51.650 --> 00:17:54.789 double that 2 ,000 hours six times over. Yes. 00:17:55.009 --> 00:17:57.829 That same capacitor will theoretically last over 00:17:57.829 --> 00:18:02.589 128 ,000 hours. That is roughly 15 years of continuous 00:18:02.589 --> 00:18:05.769 operation. This is exactly why you need to care 00:18:05.769 --> 00:18:08.670 about the cooling fans in your PC or the ventilation 00:18:08.670 --> 00:18:11.049 space around your TV or your gaming console. 00:18:11.490 --> 00:18:13.529 Exactly. When you shove your electronics into 00:18:13.529 --> 00:18:17.069 a tight unventilated media cabinet, you are literally 00:18:17.069 --> 00:18:20.230 cooking the liquid inside these capacitors. If 00:18:20.230 --> 00:18:22.490 you clean the dust out of a fan or just give 00:18:22.490 --> 00:18:24.230 the console some breathing room and you drop 00:18:24.230 --> 00:18:26.390 the internal temperature by just 10 degrees, 00:18:26.700 --> 00:18:29.180 You aren't just making it run quieter. No, you're 00:18:29.180 --> 00:18:31.619 doing so much more. You are mathematically doubling 00:18:31.619 --> 00:18:34.339 the life of the capacitors inside. You are buying 00:18:34.339 --> 00:18:37.839 yourself years of extra device life. It's a direct, 00:18:38.119 --> 00:18:40.099 practical application of the chemistry we've 00:18:40.099 --> 00:18:43.339 been unpacking. Dealing with this heat and evaporation 00:18:43.339 --> 00:18:46.019 trade -off has been an ongoing engineering headache 00:18:46.019 --> 00:18:48.799 for a century. Which really begs the question, 00:18:49.069 --> 00:18:52.529 If heat and liquid evaporation are such massive 00:18:52.529 --> 00:18:56.470 fundamental flaws, whose idea was it to put a 00:18:56.470 --> 00:18:59.089 liquid chemical bath inside our electronics in 00:18:59.089 --> 00:19:00.789 the first place? That's a great question. To 00:19:00.789 --> 00:19:03.470 understand why we tolerate this ticking clock, 00:19:03.789 --> 00:19:06.269 we had to look back at the massive engineering 00:19:06.269 --> 00:19:08.069 problem they were trying to solve historically. 00:19:08.230 --> 00:19:10.549 We have to trace it back to the late 19th century. 00:19:11.190 --> 00:19:14.910 In 1896, a researcher named Carol Pollack actually 00:19:14.910 --> 00:19:18.150 patented the underlying idea of using an oxide 00:19:18.150 --> 00:19:21.799 layer aluminum as a capacitor in a liquid electrolyte. 00:19:21.980 --> 00:19:23.819 And the source notes that the early industrial 00:19:23.819 --> 00:19:26.339 versions of these were incredibly literal. They 00:19:26.339 --> 00:19:29.200 were called wet electrolytic capacitors because 00:19:29.200 --> 00:19:31.839 they were essentially massive metallic boxes 00:19:31.839 --> 00:19:34.440 completely filled with water and borax with a 00:19:34.440 --> 00:19:36.799 folded aluminum plate fully submerged inside. 00:19:36.940 --> 00:19:40.200 They were basically giant sloshing jars of conductive 00:19:40.200 --> 00:19:42.240 fluid. That sounds dangerous. They were bulky 00:19:42.240 --> 00:19:44.980 and messy, but they were absolutely crucial for 00:19:44.980 --> 00:19:47.759 early telecommunications. They were primarily 00:19:47.759 --> 00:19:50.619 used in large telephone exchanges to filter out 00:19:50.619 --> 00:19:54.160 the background noise or hash on the 48 volt DC 00:19:54.160 --> 00:19:56.359 power lines so people could actually hear each 00:19:56.359 --> 00:19:59.380 other over the phone. But a giant jar of Borax 00:19:59.380 --> 00:20:03.859 water is entirely impractical for anything. portable 00:20:03.859 --> 00:20:05.640 or meant for the living room. Yeah. Couldn't 00:20:05.640 --> 00:20:07.880 exactly put one of those on a laptop. No, definitely 00:20:07.880 --> 00:20:11.180 not. Which brings us to the real commercial breakthrough 00:20:11.180 --> 00:20:15.990 in 1925. A man named Samuel Rubin teamed up with 00:20:15.990 --> 00:20:18.849 Philip Mallory, whose company, by the way, eventually 00:20:18.849 --> 00:20:22.430 evolved into Duracell. Oh, wow. Yeah. Rubin invented 00:20:22.430 --> 00:20:24.809 what became known as the dry type of electrolytic 00:20:24.809 --> 00:20:27.029 capacitor. Though dry is obviously a relative 00:20:27.029 --> 00:20:29.049 term here. Right. It's not totally dry. It's 00:20:29.049 --> 00:20:31.410 just not a sloshing bucket of water anymore. 00:20:32.170 --> 00:20:34.029 Rubin figured out that you didn't need the whole 00:20:34.029 --> 00:20:36.869 metallic box to hold the liquid. Instead, he 00:20:36.869 --> 00:20:39.589 soaked a thin paper spacer with the liquid electrolyte 00:20:39.589 --> 00:20:42.289 and sandwiched it tightly between two wound up 00:20:42.220 --> 00:20:44.579 aluminum foils. Which is the modern sandwich 00:20:44.579 --> 00:20:46.640 design we talked about earlier. This meant the 00:20:46.640 --> 00:20:48.720 outer container was suddenly just for physical 00:20:48.720 --> 00:20:51.420 protection, not an active sloshing part of the 00:20:51.420 --> 00:20:54.619 electrical circuit. Exactly. This single innovation 00:20:54.619 --> 00:20:58.059 drastically reduced both the physical size and 00:20:58.059 --> 00:21:00.779 the cost of the capacitor. It suddenly made it 00:21:00.779 --> 00:21:03.160 possible to fit high capacity energy storage 00:21:03.160 --> 00:21:06.339 into small spaces. And the timing of this invention 00:21:06.339 --> 00:21:09.420 drove an entire technological revolution. Because 00:21:09.420 --> 00:21:11.880 it coincided directly with the home radio boom 00:21:11.880 --> 00:21:15.140 of the late 1920s and 30s. Yes. Before this, 00:21:15.240 --> 00:21:18.240 building a radio that ran on standard noisy household 00:21:18.240 --> 00:21:21.559 AC power required massive incredibly expensive 00:21:21.559 --> 00:21:24.380 oiled paper capacitors to smooth out the power 00:21:24.380 --> 00:21:27.069 enough to hear the audio. Rubin's invention of 00:21:27.069 --> 00:21:30.029 the tightly wound paper space capacitor made 00:21:30.029 --> 00:21:32.430 home radios small enough and affordable enough 00:21:32.430 --> 00:21:34.869 for the masses. It was a fundamental enabler 00:21:34.869 --> 00:21:37.609 of modern mass consumer electronics. Over the 00:21:37.609 --> 00:21:39.910 decades, engineers have heavily refined the etching 00:21:39.910 --> 00:21:42.250 processes to increase that microscopic surface 00:21:42.250 --> 00:21:45.190 area, and they've developed complex organic solvents 00:21:45.190 --> 00:21:47.509 to replace the plain water, which reduces the 00:21:47.509 --> 00:21:49.549 evaporation rate and extends the shelf life. 00:21:49.759 --> 00:21:52.920 But despite all those incredible chemical advancements, 00:21:53.480 --> 00:21:56.500 the core concept remains exactly the same as 00:21:56.500 --> 00:22:00.720 Rubin's 1925 design. Two foils, a paper spacer, 00:22:00.859 --> 00:22:03.680 and a liquid. It really is. Which leaves me with 00:22:03.680 --> 00:22:07.359 one final massive question. We have solid polymer 00:22:07.359 --> 00:22:09.920 capacitors now. We have advanced multi -layer 00:22:09.920 --> 00:22:13.599 ceramics. If these wet aluminum capacitors eventually 00:22:13.599 --> 00:22:16.180 dry out and die, taking our devices with them, 00:22:16.420 --> 00:22:18.759 why haven't we completely replaced them with 00:22:18.759 --> 00:22:22.240 solid state tech? Why invite this ticking liquid 00:22:22.240 --> 00:22:25.259 clock into our modern gadgets when we have alternatives? 00:22:25.480 --> 00:22:27.319 Well, if we connect this to the bigger picture, 00:22:27.680 --> 00:22:29.680 engineering is always a strict discipline of 00:22:29.680 --> 00:22:32.539 trade -offs. Always. Yes. Solid polymer electrolytes 00:22:32.539 --> 00:22:34.859 exist today. They have incredibly low internal 00:22:34.859 --> 00:22:37.079 resistance. And because they are solid, they 00:22:37.079 --> 00:22:39.299 don't evaporate and dry out in the same way. 00:22:40.040 --> 00:22:42.380 significantly more expensive to manufacture and 00:22:42.380 --> 00:22:44.880 they simply cannot achieve the same massive capacitance 00:22:44.880 --> 00:22:47.519 values or handle the same high voltage ranges 00:22:47.519 --> 00:22:50.119 in the same tiny physical footprint. So it all 00:22:50.119 --> 00:22:52.440 comes down to harsh economics and sheer energy 00:22:52.440 --> 00:22:56.220 density. It does. The non -solid wet aluminum 00:22:56.220 --> 00:22:59.420 electrolytic capacitor remains absolutely unparalleled 00:22:59.420 --> 00:23:01.559 when a designer needs to filter out heavy low 00:23:01.559 --> 00:23:04.200 -frequency electrical noise, store a massively 00:23:04.200 --> 00:23:06.579 large amount of energy, and keep the manufacturing 00:23:06.579 --> 00:23:09.980 costs down to pennies per unit. Wow. Without 00:23:09.980 --> 00:23:12.759 them, the modern mass market for consumer electronics, 00:23:13.220 --> 00:23:15.160 from the cheapest smartphone charger to high 00:23:15.160 --> 00:23:18.119 -end home appliances, would simply not be economically 00:23:18.119 --> 00:23:21.420 viable. We accept the finite lifespan because 00:23:21.420 --> 00:23:23.859 it's the only way to make the technology affordable 00:23:23.859 --> 00:23:27.480 and accessible to everyone. It's just a fascinating, 00:23:27.700 --> 00:23:29.500 profound compromise we've gone on such a journey 00:23:29.500 --> 00:23:33.900 today. From literal jars of Borax water smoothing 00:23:33.900 --> 00:23:36.740 out 19th century telephone lines to the realization 00:23:36.740 --> 00:23:39.220 that microscopic mountain ranges of aluminum 00:23:39.220 --> 00:23:41.519 can hold massive amounts of power. Yeah, it's 00:23:41.519 --> 00:23:43.579 wild. We've unpacked how that incredibly thin 00:23:43.579 --> 00:23:46.819 1 .4 nanometer oxide layer relies on a liquid 00:23:46.819 --> 00:23:49.500 immune system to constantly heal itself. And 00:23:49.500 --> 00:23:51.519 we've seen how intensely vulnerable that chemical 00:23:51.519 --> 00:23:53.839 system is, whether it's blowing up from a stolen 00:23:53.839 --> 00:23:56.579 unstable recipe during the capacitor plague or 00:23:56.579 --> 00:23:59.200 just the slow inevitable evaporation driven by 00:23:59.200 --> 00:24:01.789 the 10 degree rule. It really changes how you 00:24:01.789 --> 00:24:03.650 view the hardware surrounding you every day. 00:24:04.009 --> 00:24:06.730 You look at your sleek, modern devices and think 00:24:06.730 --> 00:24:09.609 of them as permanent, static objects of silicon 00:24:09.609 --> 00:24:13.359 and metal. But remember... the aluminum electrolytic 00:24:13.359 --> 00:24:15.859 capacitor. Inside your devices right now, there 00:24:15.859 --> 00:24:18.759 is a microscopic liquid chemical ecosystem that 00:24:18.759 --> 00:24:21.819 is slowly breathing, healing itself, and inevitably 00:24:21.819 --> 00:24:24.440 drying out. Your technology isn't just a machine. 00:24:24.880 --> 00:24:27.400 In a chemical sense, it is aging, just like you. 00:24:27.700 --> 00:24:29.980 What a profound way to look at it. Thank you 00:24:29.980 --> 00:24:32.000 for joining us on this deep dive into the hidden 00:24:32.000 --> 00:24:34.480 world of your electronics. Keep your devices 00:24:34.480 --> 00:24:37.420 cool, stay curious, and we'll see you next time.