1 00:00:05,750 --> 00:00:11,370 Let's discuss the operating principle of the crystalline silicon solar cell. 2 00:00:11,370 --> 00:00:17,580 In this block we will discuss several technical aspects which play an important role in the 3 00:00:17,580 --> 00:00:24,580 collection of the light excited charge carriers and the reduction of the optical losses. 4 00:00:26,150 --> 00:00:28,970 Let's look at the solar cell. 5 00:00:28,970 --> 00:00:35,970 In week 2, we have discussed how an illuminated p-n junction can operate as a solar cell. 6 00:00:36,360 --> 00:00:43,360 In the illustrations we have used in week 2, both the p-doped and n-doped region have 7 00:00:43,370 --> 00:00:45,750 the same thickness. 8 00:00:45,750 --> 00:00:51,910 If we look at the c-Si technology, this is not the case in real devices. 9 00:00:51,910 --> 00:00:56,879 Let's look at the most conventional crystalline silicon solar cell. 10 00:00:56,879 --> 00:01:01,879 The solar cell is based on a p-type silicon wafer. 11 00:01:01,879 --> 00:01:08,060 The n-layer, in this illustration colored yellow, is the n-doped layer in the crystalline 12 00:01:08,060 --> 00:01:10,049 silicon PV technology. 13 00:01:10,049 --> 00:01:14,049 This layer is often called the emitter layer. 14 00:01:14,049 --> 00:01:21,800 This n-type emitter layer is much thinner than the wafer, the emitter has a thickness 15 00:01:21,850 --> 00:01:24,869 in the order of 1 micron. 16 00:01:24,869 --> 00:01:31,869 The rest p-doped wafer can be in the order of 100 up to 300 microns. 17 00:01:33,070 --> 00:01:39,110 As discussed in week 3, a large fraction of light is absorbed close to the front surface 18 00:01:39,110 --> 00:01:41,479 of the solar cell. 19 00:01:41,479 --> 00:01:46,200 This is the generation profile as we have discussed in week 3. 20 00:01:46,200 --> 00:01:53,170 You see that in the first 10 microns most charge carriers are excited. 21 00:01:53,170 --> 00:02:01,500 By making the front emitter layer very thin, a large fraction of the light excited charge carriers 22 00:02:01,600 --> 00:02:10,100 generated by the incoming light are created within the diffusion length of the p-n junction. 23 00:02:10,239 --> 00:02:16,950 One of the processing methods to make the emitter layer is by solid state diffusion. 24 00:02:16,950 --> 00:02:23,819 The wafers are placed in a furnace, where the dopant atoms are present in atmosphere 25 00:02:23,819 --> 00:02:26,849 in the form of phosphine. 26 00:02:26,849 --> 00:02:30,660 These atoms react at these high temperatures with the surface. 27 00:02:30,660 --> 00:02:37,660 At these high temperatures the dopant atoms are mobile in the solid and based on Fick's 28 00:02:37,739 --> 00:02:42,450 law the dopant atoms will diffuse into the wafer. 29 00:02:42,450 --> 00:02:48,569 Similar as we have discussed in week 2 for charge carrier diffusion, net diffusion 30 00:02:48,569 --> 00:02:51,819 occurs when there is a density gradient. 31 00:02:51,819 --> 00:02:58,440 The diffusion process is controlled such that the dopants penetrate into the solid to establish 32 00:02:58,440 --> 00:03:02,890 the desired emitter thickness. 33 00:03:02,890 --> 00:03:09,800 Now we are going to look at how the charge carriers in crystalline silicon solar cells are collected. 34 00:03:09,920 --> 00:03:16,159 We will discuss the crucial issues which play an important role in this charge collection: 35 00:03:16,159 --> 00:03:23,159 this includes the emitter layer, the metal contacts and the back contact. 36 00:03:25,230 --> 00:03:27,269 Let's start with the emitter layer: 37 00:03:27,269 --> 00:03:33,760 At the p-n junction the light excited minority charge carriers are separated at the p-n junction. 38 00:03:33,760 --> 00:03:39,220 The minority electrons in the p-layer drift to the n-layer. 39 00:03:39,220 --> 00:03:42,709 These electrons have to be collected. 40 00:03:42,709 --> 00:03:48,670 Since the silicon n-emitter is not conductive enough we have to use the more conductive 41 00:03:48,670 --> 00:03:51,170 metal contacts. 42 00:03:51,170 --> 00:03:55,720 The metal contacts are placed on top of the emitter layer. 43 00:03:55,720 --> 00:04:00,720 Here the metal contacts are made of the cheap metal aluminum. 44 00:04:00,720 --> 00:04:05,700 This means that the electrons have to diffuse laterally through the emitter layer 45 00:04:05,700 --> 00:04:09,090 to the contact to be collected. 46 00:04:09,090 --> 00:04:15,909 Let's see what the challenges for this transport of electrons to the contacts are . 47 00:04:15,909 --> 00:04:22,590 First, last week we have discussed the importance of the high lifetimes of charge carriers. 48 00:04:22,590 --> 00:04:29,400 High lifetime guarantees large open-circuit voltages, or in other words the best utilization 49 00:04:29,400 --> 00:04:31,680 of the band gap energy. 50 00:04:31,680 --> 00:04:38,680 This means that we have to reduce the recombination losses as much as possible. 51 00:04:39,190 --> 00:04:45,000 In addition recombination limits the collected current as well. 52 00:04:45,000 --> 00:04:51,900 Recombination in silicon occurs as discussed earlier through Shockley-Read-Hall recombination 53 00:04:51,900 --> 00:04:55,990 and Auger recombination. 54 00:04:55,990 --> 00:04:58,410 Let's consider Shockley-Read-Hall recombination. 55 00:04:58,410 --> 00:05:03,800 A bare c-Si surface contains many defects. 56 00:05:03,800 --> 00:05:10,060 The surface silicon atoms have some valence electrons which cannot make a molecular orbital 57 00:05:10,060 --> 00:05:13,660 with the absent neighboring atoms. 58 00:05:13,660 --> 00:05:20,360 These valence orbitals containing only one electron at the surface act like defects, also 59 00:05:20,360 --> 00:05:23,930 called dangling bonds. 60 00:05:23,930 --> 00:05:29,690 At these defects the charge carriers can recombine through the SRH process. 61 00:05:29,690 --> 00:05:35,280 The probability and speed at which charge carriers can recombine is usually expressed 62 00:05:35,280 --> 00:05:39,770 in terms of the surface recombination velocity. 63 00:05:39,770 --> 00:05:44,160 Since a large fraction of the charge carriers are generated close to the front surface, 64 00:05:44,160 --> 00:05:51,600 a high surface recombination velocity at the emitter front surface will lead to charge carriers losses 65 00:05:51,600 --> 00:05:57,650 and consequently lower short-circuit current densities as well. 66 00:05:58,620 --> 00:06:06,600 For example, high-quality monocrystalline silicon wafers have no defect rich grain boundaries in the bulk. 67 00:06:06,600 --> 00:06:13,580 The lifetime of the charge carriers is limited by the recombination processes at the wafer surface. 68 00:06:14,520 --> 00:06:21,520 In order to reduce the surface recombination, two approaches are used. 69 00:06:21,940 --> 00:06:27,900 The first is based on reducing the number of defects at the surface. 70 00:06:27,900 --> 00:06:36,100 This can be achieved by the deposition of a thin layer of a different material on top of the surface. 71 00:06:36,160 --> 00:06:41,400 This material partially restores the bonding environment of the silicon atoms. 72 00:06:41,400 --> 00:06:48,900 In addition, the material must be an insulator, it must force the electrons to remain in and 73 00:06:49,009 --> 00:06:55,300 move through the emitter layer. Typical chemical passivation layers being 74 00:06:55,300 --> 00:07:01,669 used on emitter layers are silicon oxide and silicon nitride. 75 00:07:01,669 --> 00:07:08,490 Silicon oxide layers are formed by heating up the silicon surface in an oxygen-rich atmosphere. 76 00:07:08,490 --> 00:07:14,080 The surface oxidizes resulting in a thin layer of silicon oxide. 77 00:07:14,080 --> 00:07:19,919 SiN can be deposited using plasma-enhanced chemical vapor deposition. 78 00:07:19,919 --> 00:07:26,919 We will come back to that process technology when we talk about thin-film PV technologies. 79 00:07:28,220 --> 00:07:35,900 A second approach to reduce the surface recombination velocity is to reduce the minority charge 80 00:07:36,000 --> 00:07:38,460 carrier density near the surface. 81 00:07:38,460 --> 00:07:44,910 As the surface recombination velocity is limited by the minority charge carriers, a low as 82 00:07:44,910 --> 00:07:50,620 possible minority charge carrier density at the surface region would be beneficial. 83 00:07:50,620 --> 00:07:56,449 It means that with increasing the doping of the emitter layer, we decrease the density 84 00:07:56,449 --> 00:07:59,150 of the minority charge carriers. 85 00:07:59,150 --> 00:08:03,789 This results in a lower recombination velocity at the surface. 86 00:08:03,789 --> 00:08:10,000 However, this is again in competition with the diffusion length of the minority charge carriers. 87 00:08:10,200 --> 00:08:15,199 In this case the hole is indicated by the blue dot. 88 00:08:15,199 --> 00:08:21,800 Note, that many charge carriers are generated by the blue-ish photons close to the surface. 89 00:08:21,800 --> 00:08:27,330 To utilize these light excited minority charge carriers, the diffusion length of the holes 90 00:08:27,330 --> 00:08:36,100 has to be large enough to reach the depletion zone at the p-n junctions as indicated by the blue arrow. 91 00:08:36,100 --> 00:08:42,200 Increasing doping levels, decreases the diffusion length of the minority holes in the emitter layer. 92 00:08:42,200 --> 00:08:47,209 Here you see an example in which the diffusion length has become so small that the minority 93 00:08:47,209 --> 00:08:51,649 charge carriers are unable to reach the p-n junction. 94 00:08:51,649 --> 00:08:58,149 Therefore, too high doping levels or too thick emitter layers would result in a poor blue 95 00:08:58,149 --> 00:09:05,149 response or in other words low EQE values in the blue part of the solar spectrum. 96 00:09:05,260 --> 00:09:11,300 Such layer could be referred to as a "dead layer" as the light excited minority charge 97 00:09:11,300 --> 00:09:12,890 carriers can not be collected. 98 00:09:12,890 --> 00:09:18,300 Now, we look in more detail to the metal-emitter interface. 99 00:09:18,300 --> 00:09:25,200 A problem at this metal-emitter interface is that we cannot use an insulating passivation layer, 100 00:09:25,200 --> 00:09:33,800 like SiOx or SiN anymore as we need electrons to conduct from the semiconductor to the metal. 101 00:09:34,220 --> 00:09:41,220 This means that the metal-semiconductor interface has more defects and therefore an unwelcome 102 00:09:41,420 --> 00:09:44,510 high interface recombination velocity. 103 00:09:44,510 --> 00:09:52,000 In addition a metal-semiconductor material induces a barrier for the majority charge carriers. 104 00:09:52,300 --> 00:09:58,200 In view of the scope of this course, it is too much detail to discuss this barrier in great detail. 105 00:09:58,400 --> 00:10:04,990 Keep in mind that this high barrier will give rise to a higher contact resistance. 106 00:10:04,990 --> 00:10:11,000 Again high doping levels can reduce the recombination velocity at the metal-semiconductor interface 107 00:10:11,000 --> 00:10:14,140 and reduce the contact resistance. 108 00:10:14,140 --> 00:10:20,050 As a consequence it is preferred to reduce the area of the interface between the metal 109 00:10:20,050 --> 00:10:27,050 and the semiconductor to minimize the recombination at the interface defects as much as possible, 110 00:10:27,660 --> 00:10:34,660 and to have the emitter directly under the contact as heavily doped as possible. 111 00:10:36,070 --> 00:10:42,550 This higher doped region in is indicated by N++. 112 00:10:42,550 --> 00:10:48,400 In the figure you see that the contact area between metal and interface is kept as small as possible. 113 00:10:48,570 --> 00:10:55,570 The sides of the metal contacts are buried in the insulating passivation layer. 114 00:10:57,360 --> 00:11:01,910 The area below the contact has been heavily doped. 115 00:11:01,910 --> 00:11:08,910 These two approaches reduce the recombination and collection losses at the metal contact. 116 00:11:09,130 --> 00:11:13,300 The classic metal grid pattern on top of a solar cell looks like this. 117 00:11:13,300 --> 00:11:17,000 It is the road map for the electrons. 118 00:11:17,000 --> 00:11:25,600 Indicated in red, we see the highway for the electrons in the middle of the top surface of the solar cell. 119 00:11:25,680 --> 00:11:29,649 These are called the busbars. 120 00:11:29,649 --> 00:11:36,600 The smaller country roads going from the busbar to the edge of the solar cell, are called 'fingers'. 121 00:11:37,180 --> 00:11:43,329 First, the fingers have a resistance R. 122 00:11:43,329 --> 00:11:50,329 If L is the length of the finger, W the width of the finger and H is the height of the finger, 123 00:11:50,380 --> 00:11:57,380 and rho is the electrical resistivity of the metal, we can describe the resistance R as 124 00:11:57,829 --> 00:12:02,130 rho times length divided by the height and the width. 125 00:12:02,130 --> 00:12:07,790 The equation shows that the longer the path length for an electron will be, the larger 126 00:12:07,790 --> 00:12:10,760 the resistance the electron experiences. 127 00:12:10,760 --> 00:12:17,760 In addition, the smaller the cross-section (width times height) of the finger, 128 00:12:17,760 --> 00:12:20,600 the larger the resistance will be. 129 00:12:20,600 --> 00:12:29,000 Note, that the resistance of the contacts will act as a series resistance in the equivalent electric circuit. 130 00:12:29,170 --> 00:12:34,400 Larger series resistance will result in lower fill factors of the solar cell as discussed in week 3. 131 00:12:34,589 --> 00:12:41,589 This means that you would like to have the finger cross-section as large as possible. 132 00:12:42,019 --> 00:12:46,860 In this figure we see that the electrons in the emitter have to travel in the lateral 133 00:12:46,860 --> 00:12:50,089 direction to arrive at the contact. 134 00:12:50,089 --> 00:12:57,089 Since the n-type silicon has a higher resistivity as the metal, the charge carriers in the emitter 135 00:12:57,490 --> 00:13:01,050 layer also experience a resistance. 136 00:13:01,050 --> 00:13:06,860 In one of the exercises in this week it is demonstrated that the power loss due to the 137 00:13:06,860 --> 00:13:13,860 resistivity of the emitter layer, scales with the finger spacing to the power 3. 138 00:13:14,959 --> 00:13:21,959 As the metal contacts are at the front surface, they act as an unwelcome shading object, 139 00:13:21,970 --> 00:13:29,500 or in other words light incident on the front contact area cannot be absorbed in the PV active layers. 140 00:13:29,660 --> 00:13:35,300 So you would like to keep the contact area as small as possible, which is in competition 141 00:13:35,300 --> 00:13:40,769 with the fact that you would like to have a large as possible cross-section. 142 00:13:40,769 --> 00:13:47,019 So basically you would like to have a small as possible width and a high as possible height 143 00:13:47,019 --> 00:13:53,190 to comply with these requirements as demonstrated by the left contact. 144 00:13:53,190 --> 00:13:58,190 As you see we have several effects, which are in competition. 145 00:13:58,190 --> 00:14:03,720 Let's look for a moment at the relation between losses and the finger spacing. 146 00:14:03,720 --> 00:14:10,800 With increasing finger spacing the power losses of the solar cell decreases due to shading. 147 00:14:12,019 --> 00:14:20,000 In contrast, the losses due to the resistivity in the emitter layer and the metal fingers increases. 148 00:14:20,000 --> 00:14:24,600 This means that the total loss has a minimum. 149 00:14:25,200 --> 00:14:29,389 Similar plots can be made for the width of the contact. 150 00:14:29,389 --> 00:14:33,910 The larger the width, the larger the losses due to shading. 151 00:14:33,910 --> 00:14:38,820 The resistive losses in the metal fingers decrease with increasing finger width. 152 00:14:38,820 --> 00:14:45,820 Again there exists a width, which has a minimum loss. 153 00:14:46,199 --> 00:14:53,800 Optimizing the front contact pattern is therefore a complex interplay between the finger width and spacing. 154 00:14:55,279 --> 00:14:59,880 At the back surface of the solar cell we have similar issues playing a role. 155 00:14:59,880 --> 00:15:06,000 As the electrons have to be collected in the n-type layer, the holes are collected at the back contact. 156 00:15:06,490 --> 00:15:11,630 As electrons are the only charge carriers that exist in the metal, you have to realize that 157 00:15:11,630 --> 00:15:17,920 the holes recombine with the electrons at the contact interface. 158 00:15:17,920 --> 00:15:23,449 If the distance between the p-n interface and the back contact is smaller than the typical 159 00:15:23,449 --> 00:15:28,850 diffusion length of the minority electrons, the latter mentioned minority electrons can 160 00:15:28,850 --> 00:15:35,850 be lost at the defects at the back contact interface due to the SRH recombination. 161 00:15:36,490 --> 00:15:40,889 How can we reduce this loss mechanism? 162 00:15:40,889 --> 00:15:46,649 First, we can reduce again the area between the metal contact and the semiconductor. 163 00:15:46,649 --> 00:15:49,420 We can make point contacts. 164 00:15:49,420 --> 00:15:54,700 The rest of the rear surface is then passivated by an insulating passivation layer, 165 00:15:54,700 --> 00:16:02,029 similar as we have discussed for the cases of the emitter front surface. 166 00:16:02,029 --> 00:16:07,279 The loss mechanism of recombination of electrons at the back contact can be further reduced 167 00:16:07,279 --> 00:16:09,790 by a so-called back surface field. 168 00:16:09,790 --> 00:16:18,700 A higher p-doped region is placed above the point contacts at the rear surface, indicated by P++. 169 00:16:19,700 --> 00:16:23,000 How does the back surface field work? 170 00:16:23,000 --> 00:16:27,860 For the moment we have put the solar cell on its side. 171 00:16:27,860 --> 00:16:34,779 Below the solar cell you see the band diagram indicating the various interfaces. 172 00:16:34,779 --> 00:16:42,000 The interface between the lower-doped p-region and the higher-doped p-region acts like a p-n junction. 173 00:16:42,329 --> 00:16:48,139 In this case it will act as a barrier for the light excited minority electrons in the 174 00:16:48,139 --> 00:16:53,250 lower-doped region to diffuse to the back surface. 175 00:16:53,250 --> 00:16:59,040 The space charge field behaves like a passivation of the defects at the back contact interface 176 00:16:59,040 --> 00:17:04,800 and allows to have higher levels for the electron minority density in the p-doped bulk. 177 00:17:04,800 --> 00:17:10,000 Up to now we have talked about managing the charge carriers. 178 00:17:10,200 --> 00:17:17,100 Now we are going to discuss the management of the photons in a crystalline silicon solar cell device. 179 00:17:17,539 --> 00:17:23,720 The optical loss mechanisms are shading, reflection losses, parasitic absorption losses in the 180 00:17:23,720 --> 00:17:29,700 non-PV active layers and light that is not absorbed in the silicon layers and is lost 181 00:17:29,700 --> 00:17:31,490 at the back contact. 182 00:17:31,490 --> 00:17:38,490 Shading losses are caused by the front contact grid as discussed earlier. 183 00:17:38,640 --> 00:17:43,590 The optimum contact grid at the front is a competition between reduction of the shading 184 00:17:43,590 --> 00:17:50,299 losses and a reduction of the resistivity losses in the metal contact and the emitter layer. 185 00:17:51,440 --> 00:17:57,360 Secondly, the reflection at the front surface can be considered as a loss mechanism. 186 00:17:57,360 --> 00:18:01,870 Last week we have discussed two approaches to reduce the losses. 187 00:18:01,870 --> 00:18:05,780 The first one was based on the Rayleigh film principle. 188 00:18:05,780 --> 00:18:10,100 By putting a film with a refractive index smaller than that of silicon wafer between 189 00:18:10,190 --> 00:18:14,470 air and the wafer, you can reduce the losses. 190 00:18:14,470 --> 00:18:19,450 The optimum value for the refractive index of the intermediate layer equals the square 191 00:18:19,450 --> 00:18:25,460 root of the product of refractive indices of the two other media. 192 00:18:25,460 --> 00:18:31,539 For silicon and air at 500 nm this leads to an optimum refractive index of 2.1. 193 00:18:31,539 --> 00:18:37,280 Note, that in practice a solar cell is encapsulated under a glass plate. 194 00:18:37,280 --> 00:18:42,000 The glass plate will be beneficial to refractive index grading as well, 195 00:18:42,000 --> 00:18:45,539 reducing the reflection losses further. 196 00:18:45,539 --> 00:18:51,640 Secondly, using the concept of destructive interference we can choose the thickness and 197 00:18:51,640 --> 00:18:58,470 refractive index of an anti-reflection coating such that in a certain wavelength the irradiance 198 00:18:58,470 --> 00:19:01,900 coupled out of solar cell is minimized. 199 00:19:01,900 --> 00:19:06,669 The light reflected from the front surface is in anti-phase with the light reflected 200 00:19:06,669 --> 00:19:10,890 from the back surface, as discussed last week. 201 00:19:10,890 --> 00:19:15,900 The thickness of such layer should be the wavelength divided by the 4 times the refractive index. 202 00:19:16,240 --> 00:19:24,700 If we take a welcome reflective index of 2.1 we arrive for light of 500 nm at an optimal 203 00:19:24,700 --> 00:19:31,700 thickness of 60 nm for an anti-reflection based on destructive interference. 204 00:19:33,140 --> 00:19:37,400 As we have discussed earlier, one of the typical passivation layers of standard crystalline 205 00:19:37,400 --> 00:19:39,730 silicon is silicon nitride. 206 00:19:39,730 --> 00:19:44,559 Here we see a multicrystalline silicon wafer without the SiNx layer. 207 00:19:44,559 --> 00:19:49,760 We see that a lot of light is reflected by the wafer. 208 00:19:49,760 --> 00:19:54,679 Next to it, we see a multicrystalline silicon with a silicon nitride passivation layer. 209 00:19:54,679 --> 00:19:58,980 As you can see much less light is being reflected. 210 00:19:58,980 --> 00:20:05,980 Interestingly, the refractive index of SiN is in the range of 2 up to 2.2 at 500 nm, 211 00:20:06,390 --> 00:20:10,380 close to the optimum for refractive index grading. 212 00:20:10,380 --> 00:20:15,630 Typical thicknesses are in the same range as the 60 nm discussed earlier. 213 00:20:15,630 --> 00:20:21,250 Note, that the surface of the solar cell looks a bit blue-ish. 214 00:20:21,250 --> 00:20:28,250 This shows that the reflection of blue light by the SiNx is larger than for the other wavelengths. 215 00:20:28,650 --> 00:20:35,240 As discussed last week, texturing of the surface can improve the light incoupling as well. 216 00:20:35,240 --> 00:20:42,120 Light that is reflected at the textured surface, can be reflected at angles in which the trajectory 217 00:20:42,120 --> 00:20:46,470 of the light ray is incident somewhere else on the surface. 218 00:20:46,470 --> 00:20:49,350 Here the light can be still coupled into the silicon. 219 00:20:49,350 --> 00:20:55,400 Secondly, the scattering at textured surfaces will couple the light under angles into the wafer. 220 00:20:55,500 --> 00:21:00,000 This means that the absorption path length will be enhanced. 221 00:21:00,300 --> 00:21:07,640 This certainly enhances the welcome absorption of the light with a wavelength above 900 nm. 222 00:21:08,570 --> 00:21:14,250 The texture on the wafers can be realized with wet-etching approaches. 223 00:21:14,250 --> 00:21:18,000 Here you can use an interesting property of the 100 surface. 224 00:21:18,000 --> 00:21:23,110 This anisotropic etching can be used when etching a crystalline silicon wafer with an 225 00:21:23,110 --> 00:21:26,460 initial 100 surface orientation. 226 00:21:26,460 --> 00:21:35,400 Etching of such wafer leads to textured surfaces with pyramid structures having only 111 oriented surfaces. 227 00:21:35,700 --> 00:21:40,780 In this picture you can see such textured surface of a monocrystalline silicon wafer 228 00:21:40,780 --> 00:21:44,000 as processed at the DIMES lab in Delft. 229 00:21:44,000 --> 00:21:50,500 Using such etching procedures it is even possible to have a silicon wafer looking black, 230 00:21:50,500 --> 00:21:54,270 as shown here in the second picture. 231 00:21:54,270 --> 00:22:00,720 Here we see a flat multicrystalline silicon layer with and without silicon nitride. 232 00:22:00,720 --> 00:22:06,330 Under these wafers, we see two wafers with and without silicon nitride layers, as well. 233 00:22:06,330 --> 00:22:12,450 The only difference is that the two wafers at the bottom have textured surfaces. 234 00:22:12,450 --> 00:22:17,730 The wafers are placed in the same illumination conditions as the flat wafers. 235 00:22:17,730 --> 00:22:23,500 By eye you can already see that these surfaces are darker and therefore reflect less light. 236 00:22:23,600 --> 00:22:29,659 So, we have discussed the general design rules for crystalline silicon solar cells. 237 00:22:29,659 --> 00:22:37,000 In the next block I will show you three different concepts of high-efficiency crystalline silicon solar cells. 238 00:22:37,100 --> 00:22:43,000 As you will see, they are based on the design rules we have discussed in this block. 239 00:22:43,000 --> 00:22:44,250 See you in the next block!