1 00:00:15,966 --> 00:00:16,700 An N type 2 00:00:16,700 --> 00:00:18,735 and I take a P type and I put them together, 3 00:00:18,735 --> 00:00:22,539 some really important things happen. 4 00:00:22,539 --> 00:00:25,241 So important that it led to this, which 5 00:00:25,241 --> 00:00:27,777 is the very first transistor. 6 00:00:27,777 --> 00:00:29,446 That's a transistor. 7 00:00:29,446 --> 00:00:32,983 It's a dope semiconductor, one type, next to a dope 8 00:00:32,983 --> 00:00:36,052 semiconductor, another type. 9 00:00:36,052 --> 00:00:37,821 Now, those are the three people-- 10 00:00:37,821 --> 00:00:39,856 Bardeen, Shockley, and Brattain-- 11 00:00:39,856 --> 00:00:42,459 who won the Nobel Prize for this work. 12 00:00:42,459 --> 00:00:44,861 And that's the first transistor. 13 00:00:44,861 --> 00:00:49,499 Does anybody know how many transistors we make today? 14 00:00:49,499 --> 00:00:52,402 Per second. 15 00:00:52,402 --> 00:00:55,605 That's the only time metric we can use. 16 00:00:55,605 --> 00:00:58,575 Yeah, it's something around 10 trillion. 17 00:00:58,575 --> 00:01:02,812 We make 10 trillion transistors per second today. 18 00:01:02,812 --> 00:01:03,680 This is a big deal. 19 00:01:03,680 --> 00:01:06,282 This started it all, and it's all about chemistry 20 00:01:06,282 --> 00:01:09,986 and the physics of this device. 21 00:01:09,986 --> 00:01:12,889 This led to diodes and photodiodes. 22 00:01:12,889 --> 00:01:16,658 This led to the whole revolution. 23 00:01:16,658 --> 00:01:24,134 And what I want to point out here is that this also led-- 24 00:01:24,134 --> 00:01:26,236 it called on chemistry. 25 00:01:26,236 --> 00:01:27,270 So I love this chart. 26 00:01:27,270 --> 00:01:28,805 Unfortunately, they didn't update it, 27 00:01:28,805 --> 00:01:33,143 but this is an Intel chart that they used to show. 28 00:01:33,143 --> 00:01:36,913 And what they're showing is how are they making chips. 29 00:01:36,913 --> 00:01:39,249 How much of the periodic table do they 30 00:01:39,249 --> 00:01:41,084 need to make the current chip? 31 00:01:41,084 --> 00:01:43,720 Well, in the '80s, there's almost nothing lit up here. 32 00:01:43,720 --> 00:01:46,156 There's a little bit over here that you can't see, 33 00:01:46,156 --> 00:01:47,289 stuff we just talked about. 34 00:01:47,289 --> 00:01:49,292 But look at the '90s and look at the 2000s. 35 00:01:49,292 --> 00:01:53,797 And today, it's 75% of this. 36 00:01:53,797 --> 00:01:56,866 It's literally 75% of this is in your phone. 37 00:01:56,866 --> 00:01:57,366 Yeah. 38 00:01:57,366 --> 00:01:58,935 Why? 39 00:01:58,935 --> 00:02:02,072 The reason is exactly what we've talked about today, 40 00:02:02,072 --> 00:02:04,974 because they need more and more and more 41 00:02:04,974 --> 00:02:08,845 flexibility and ways to tune the band gap 42 00:02:08,845 --> 00:02:11,948 and the semiconducting properties of these materials 43 00:02:11,948 --> 00:02:12,682 and the doping. 44 00:02:12,682 --> 00:02:13,750 Why do they need new ways? 45 00:02:13,750 --> 00:02:16,186 Because they're making them so darn small. 46 00:02:16,186 --> 00:02:18,454 They're so small, it gets harder to figure out 47 00:02:18,454 --> 00:02:20,657 how to do it right, how to do it in a way that's also 48 00:02:20,657 --> 00:02:23,927 stable over time, for example. 49 00:02:23,927 --> 00:02:25,929 So they need new chemistry. 50 00:02:25,929 --> 00:02:28,932 This is a call to action to the field of chemistry, 51 00:02:28,932 --> 00:02:32,769 and chemistry responded. 52 00:02:32,769 --> 00:02:34,838 And there's the data on the cost. 53 00:02:34,838 --> 00:02:36,305 I thought that was kind of cool. 54 00:02:36,305 --> 00:02:41,277 So these are orders of magnitude of cost and number 55 00:02:41,277 --> 00:02:42,978 of transistors. 56 00:02:42,978 --> 00:02:48,485 And I'll just say one more why this matters. 57 00:02:48,485 --> 00:02:51,387 Because the semiconductor is also the same thing 58 00:02:51,387 --> 00:02:55,325 we use to take electricity from the sun, 59 00:02:55,325 --> 00:02:57,260 to generate electricity from the sun. 60 00:02:57,260 --> 00:03:03,566 And you can imagine, you are doing that in your goody bag. 61 00:03:03,566 --> 00:03:08,037 You're literally using this semiconductor 62 00:03:08,037 --> 00:03:13,409 to generate electricity by shining light on it. 63 00:03:13,409 --> 00:03:14,344 Why is this important? 64 00:03:14,344 --> 00:03:15,745 Well, I like this comparison. 65 00:03:15,745 --> 00:03:18,648 All the coal, oil, and gas known to humans 66 00:03:18,648 --> 00:03:22,185 is what you get from the sun in about 20 days. 67 00:03:22,185 --> 00:03:22,952 This is the sun. 68 00:03:22,952 --> 00:03:25,388 Now, we've seen the sun before. 69 00:03:25,388 --> 00:03:27,891 This is how we see it in our class. 70 00:03:27,891 --> 00:03:31,728 The sun is a spectrum of different wavelengths 71 00:03:31,728 --> 00:03:32,829 and intensities. 72 00:03:32,829 --> 00:03:35,198 And this is already on planet Earth, 73 00:03:35,198 --> 00:03:37,333 because you can see those chemistries that 74 00:03:37,333 --> 00:03:39,969 are in the atmosphere have already absorbed. 75 00:03:39,969 --> 00:03:43,305 Remember ozone down here, helping us with UV radiation? 76 00:03:43,305 --> 00:03:46,276 And out here, you've got a lot of water and other things, CO2. 77 00:03:46,276 --> 00:03:47,110 Those are absorbing. 78 00:03:47,110 --> 00:03:50,446 That's why it doesn't look smooth. 79 00:03:50,446 --> 00:03:53,917 But the point is, I want to semiconductor to absorb as much 80 00:03:53,917 --> 00:03:54,984 of this as possible. 81 00:03:54,984 --> 00:03:58,188 If I want a good solar cell, I should absorb as much 82 00:03:58,188 --> 00:04:00,857 of this as possible. 83 00:04:00,857 --> 00:04:03,760 But the problem is that it's a constrained optimization 84 00:04:03,760 --> 00:04:04,861 problem. 85 00:04:04,861 --> 00:04:07,096 So let's see, if I-- 86 00:04:07,096 --> 00:04:07,897 I'll just use this. 87 00:04:12,368 --> 00:04:15,138 This goes back to Bohr, by the way. 88 00:04:15,138 --> 00:04:16,773 Remember, Bohr, I'm absorbing light. 89 00:04:16,773 --> 00:04:18,440 It's just now, I've got an actual solid. 90 00:04:18,440 --> 00:04:21,277 I don't have an atom. 91 00:04:21,277 --> 00:04:23,246 I've got a solid, which is what you need if you 92 00:04:23,246 --> 00:04:24,347 want to generate a current. 93 00:04:24,347 --> 00:04:27,150 I've got to hook leads up to it. 94 00:04:27,150 --> 00:04:29,352 But my solid is a semiconductor. 95 00:04:29,352 --> 00:04:33,623 And you can imagine, say, well, if my gap were really, really 96 00:04:33,623 --> 00:04:38,561 tiny, then any amount-- 97 00:04:38,561 --> 00:04:42,799 almost any amount of this spectrum 98 00:04:42,799 --> 00:04:46,736 would excite electrons. 99 00:04:46,736 --> 00:04:49,772 And I might grab most of that spectrum. 100 00:04:49,772 --> 00:04:51,641 I might absorb most of it. 101 00:04:51,641 --> 00:04:54,644 But the problem with that is that, then, all of these 102 00:04:54,644 --> 00:04:58,414 will thermalize, like I said in the beginning. 103 00:04:58,414 --> 00:04:59,249 Loss to heat. 104 00:05:02,485 --> 00:05:03,253 Loss to heat. 105 00:05:03,253 --> 00:05:05,321 So if my band gap is really small, 106 00:05:05,321 --> 00:05:11,261 I absorb a lot of the light, but I lose most of it as heat. 107 00:05:11,261 --> 00:05:15,531 If my band gap is really big, then there's so much of this 108 00:05:15,531 --> 00:05:16,933 light that I cannot absorb. 109 00:05:20,670 --> 00:05:22,605 By the way, also, the voltage that you get out 110 00:05:22,605 --> 00:05:25,207 of your solar cell is essentially this band gap. 111 00:05:25,207 --> 00:05:27,143 So if I get to really, really small band gaps, 112 00:05:27,143 --> 00:05:28,578 I have almost no voltage. 113 00:05:28,578 --> 00:05:30,446 That's bad, too. 114 00:05:30,446 --> 00:05:31,881 But if my band gap is too big, you 115 00:05:31,881 --> 00:05:32,949 think, oh, I'll get a high voltage. 116 00:05:32,949 --> 00:05:33,582 No. 117 00:05:33,582 --> 00:05:36,486 You won't absorb any photons. 118 00:05:36,486 --> 00:05:38,721 So it's a constraint problem. 119 00:05:38,721 --> 00:05:39,789 It's a constraint problem. 120 00:05:39,789 --> 00:05:43,593 You can actually solve this, and you can plot-- 121 00:05:43,593 --> 00:05:46,129 in fact, here's a little animation. 122 00:05:46,129 --> 00:05:48,031 Energy comes in from the sun. 123 00:05:48,031 --> 00:05:49,098 It excites an electron. 124 00:05:49,098 --> 00:05:50,600 This is how PowerPoint sees it. 125 00:05:50,600 --> 00:05:52,669 Leaves behind a hole, and you get them out. 126 00:05:52,669 --> 00:05:55,138 That's a solar cell. 127 00:05:55,138 --> 00:05:58,941 And this is the chart I want to show you. 128 00:05:58,941 --> 00:06:02,412 Because you see, if I take this constraint into account, 129 00:06:02,412 --> 00:06:04,714 that I don't absorb light if the gap is too big, 130 00:06:04,714 --> 00:06:06,883 but I do absorb light and it all goes 131 00:06:06,883 --> 00:06:08,685 to heat if the gap is too small, it 132 00:06:08,685 --> 00:06:10,753 means there's some sweet spot. 133 00:06:10,753 --> 00:06:12,455 There's some sweet spot. 134 00:06:12,455 --> 00:06:15,024 And that's what is plotted here. 135 00:06:15,024 --> 00:06:17,660 This is the band gap of a semiconductor, 136 00:06:17,660 --> 00:06:20,163 and this is the maximum that that solar cell 137 00:06:20,163 --> 00:06:21,964 efficiency could be. 138 00:06:21,964 --> 00:06:24,200 That's the maximum that it could be for that gap. 139 00:06:24,200 --> 00:06:27,470 So you can see the sweet spot is right there. 140 00:06:27,470 --> 00:06:29,372 That's a thermodynamic derivation 141 00:06:29,372 --> 00:06:31,574 called Shockley-Queisser of the maximum efficiency 142 00:06:31,574 --> 00:06:33,543 you could ever get out of a single material. 143 00:06:33,543 --> 00:06:35,278 And notice, if you put the materials here, 144 00:06:35,278 --> 00:06:38,214 silicon is 90% of the solar cells you buy today. 145 00:06:38,214 --> 00:06:41,784 And notice that it's not quite at its maximum potential. 146 00:06:41,784 --> 00:06:44,487 It's also not quite in the middle, where you'd want. 147 00:06:44,487 --> 00:06:46,422 It should have a slightly higher gap, 148 00:06:46,422 --> 00:06:48,825 and you'd get to higher efficiency. 149 00:06:48,825 --> 00:06:50,259 But these materials out of gallium 150 00:06:50,259 --> 00:06:51,728 arsenide, they do have the-- 151 00:06:51,728 --> 00:06:53,529 what is GS? 152 00:06:53,529 --> 00:06:55,898 That's not an element. 153 00:06:55,898 --> 00:06:59,569 Gallium arsenide, it does have a better gap, 154 00:06:59,569 --> 00:07:02,305 but it's much more expensive. 155 00:07:02,305 --> 00:07:03,072 And so we go. 156 00:07:03,072 --> 00:07:05,441 And a lot of solar cell research is 157 00:07:05,441 --> 00:07:08,211 on getting as close to this point as you can.