1 00:00:05,509 --> 00:00:12,389 In the previous block we have discussed various technological aspects on crystalline silicon 2 00:00:12,389 --> 00:00:14,969 wafer based PV technology. 3 00:00:14,969 --> 00:00:21,610 In this block I will give you three examples of highly efficient solar cells based on crystalline 4 00:00:21,610 --> 00:00:24,250 silicon wafers. 5 00:00:24,250 --> 00:00:31,600 As discussed earlier this week, we have various types of wafers with different qualities of silicon. 6 00:00:32,029 --> 00:00:39,029 To achieve the highest efficiencies, the bulk recombination must be as low as possible. 7 00:00:39,420 --> 00:00:47,200 Therefore the high-efficiency crystalline silicon solar cells are based on monocrystalline wafers. 8 00:00:48,400 --> 00:00:54,379 Let's start with the first high-efficiency crystalline silicon solar cell developed by 9 00:00:54,379 --> 00:01:00,400 Martin Green's group at the University of New South Wales in the late 80s and the early 90s. 10 00:01:00,829 --> 00:01:07,829 Here you see an illustration of the PERL concept, which uses a p-type float-zone silicon wafer. 11 00:01:09,620 --> 00:01:17,700 This concept has approached a conversion efficiency of 25% and has been an example for various technology 12 00:01:17,700 --> 00:01:20,000 developed afterwards. 13 00:01:20,350 --> 00:01:27,350 PERL is an abbreviation for Passivated Emitter Rear Locally diffused. 14 00:01:27,910 --> 00:01:32,710 The name indicates the two important concepts integrated into the solar cell. 15 00:01:32,710 --> 00:01:38,640 The optical losses of the PERL solar cell at the front side are minimized using three 16 00:01:38,640 --> 00:01:41,340 important concepts. 17 00:01:41,340 --> 00:01:48,340 First, the top surface of the solar cell is textured using inverted pyramid structures. 18 00:01:49,340 --> 00:01:55,350 This microscopic texture allows a fraction of the reflected light to be incident on the 19 00:01:55,350 --> 00:01:58,140 front surface for a second time. 20 00:01:58,140 --> 00:02:03,409 This enhances the total amount of light coupled into the solar cell. 21 00:02:03,409 --> 00:02:10,800 Secondly, the inverted pyramid structures are covered by a double-layer anti-reflection 22 00:02:10,840 --> 00:02:17,469 coating (ARC) which results in an extremely low top surface reflection. 23 00:02:17,469 --> 00:02:26,400 Often a double-layer coating of magnesium fluoride and zinc sulfide is used as an anti-reflection coating. 24 00:02:26,849 --> 00:02:33,769 Thirdly, the contact area at the front side has to be as small as possible, to reduce 25 00:02:33,769 --> 00:02:35,370 the shading losses. 26 00:02:35,370 --> 00:02:44,900 In the PERL concept these very thin and fine metal fingers are processed using photolithography technology. 27 00:02:47,099 --> 00:02:50,069 The emitter layer is smartly designed. 28 00:02:50,069 --> 00:02:55,900 As discussed in the previous blocks, the emitter should be highly doped underneath the contacts, 29 00:02:55,900 --> 00:03:02,189 which in the PERL concept is achieved by heavily phosphorous diffused regions. 30 00:03:02,189 --> 00:03:08,799 The rest of the emitter is moderately doped, or in other words lightly diffused, to preserve 31 00:03:08,799 --> 00:03:12,189 an excellent "blue response". 32 00:03:12,189 --> 00:03:18,450 The emitter is passivated with an silicon oxide layer on top of the emitter to suppress the 33 00:03:18,450 --> 00:03:23,079 surface recombination velocity as much as possible. 34 00:03:23,079 --> 00:03:28,379 The surface recombination velocity has been suppressed to the level that the open-circuit 35 00:03:28,379 --> 00:03:35,379 voltages with values of above 700 mV have been obtained using the PERL concept. 36 00:03:37,790 --> 00:03:43,909 At the rear surface of the solar cell, point contacts have been used in combination with 37 00:03:43,909 --> 00:03:47,180 thermal oxide passivation layers. 38 00:03:47,180 --> 00:03:54,400 The oxide operates as a passivation layer of the non-contacted area, to reduce the unwelcome 39 00:03:54,569 --> 00:03:57,540 surface recombination. 40 00:03:57,540 --> 00:04:04,090 A highly doped boron region, created by local boron diffusion, operates as a local back 41 00:04:04,090 --> 00:04:11,000 surface field, to limit the recombination of the minority electrons at the metal contact. 42 00:04:13,000 --> 00:04:19,000 The PERL concept as we presented here includes some expensive processing steps. 43 00:04:19,500 --> 00:04:26,130 The Chinese Suntech company developed in collaboration with the University of New South Wales a more 44 00:04:26,130 --> 00:04:31,570 commercially viable crystalline silicon wafer technology, which is inspired on the PERL 45 00:04:31,570 --> 00:04:33,090 cell configuration. 46 00:04:33,090 --> 00:04:41,000 A second successful cell concept which is commercialized by SunPower is the interdigitated 47 00:04:41,000 --> 00:04:44,170 back contact (IBC) solar cell. 48 00:04:44,170 --> 00:04:50,820 The principle of interdigitated back contact concepts is that it does not suffer from shading 49 00:04:50,820 --> 00:04:53,990 losses of a front metal contact grid. 50 00:04:53,990 --> 00:05:00,260 All the contacts responsible for collecting of charge carriers at the n- and p-side are 51 00:05:00,260 --> 00:05:05,900 positioned at the back of the crystalline wafer solar cell. 52 00:05:05,900 --> 00:05:12,900 An advantage of these interdigitated concepts are that you are able to use monocrystalline 53 00:05:13,590 --> 00:05:15,480 float-zone n-type wafers. 54 00:05:15,480 --> 00:05:18,160 Why is that interesting? 55 00:05:18,160 --> 00:05:23,010 The n-type wafers have some advantages above p-type wafers. 56 00:05:23,010 --> 00:05:29,280 First, the n-type wafers do not suffer from light-induced degradation. 57 00:05:29,280 --> 00:05:36,280 In p-type wafers simultaneously boron and oxygen are present, which under light exposure 58 00:05:36,350 --> 00:05:40,960 start to make complexes that act like defects. 59 00:05:40,960 --> 00:05:48,600 The light-induced degradation causes a reduction of the power output with 2-3% 60 00:05:48,600 --> 00:05:51,400 after the first week of installation. 61 00:05:51,600 --> 00:05:55,390 This effect is not present in n-type wafers. 62 00:05:55,390 --> 00:06:01,110 The second advantage is that n-type silicon is not that sensitive for impurities like 63 00:06:01,110 --> 00:06:04,280 for instance iron impurities. 64 00:06:04,280 --> 00:06:11,280 As a result less efforts have to be made to make a high electronic quality of n-type silicon, 65 00:06:11,880 --> 00:06:18,880 meaning that high-quality n-type silicon can be processed cheaper than p-type. 66 00:06:19,220 --> 00:06:25,590 On the other hand, p-doped wafers have the advantage that the boron doping is more homogeneously 67 00:06:25,590 --> 00:06:29,690 distributed over the wafer as for n-type. 68 00:06:29,690 --> 00:06:36,690 This means that within one n-type wafer the electrical properties can vary within the 69 00:06:36,690 --> 00:06:38,700 same wafer. 70 00:06:38,700 --> 00:06:45,800 This effect lowers again the yield of solar cell production based on n-type monocrystalline wafers. 71 00:06:46,750 --> 00:06:54,300 Back contacted solar cells use, in contrast to the PERL concept, n-type float-zone monocrystalline 72 00:06:54,310 --> 00:06:56,890 silicon wafers. 73 00:06:56,890 --> 00:07:03,890 An interdigitated back contact is lacking one large p-n junction. 74 00:07:03,900 --> 00:07:07,380 Instead the cell has many localized junctions. 75 00:07:07,380 --> 00:07:13,880 The holes are separated at a junction of p+ and the n-type silicon, whereas the electrons 76 00:07:13,880 --> 00:07:18,680 are collected using a n+ type silicon. 77 00:07:18,680 --> 00:07:25,680 The semiconductor-metal interface is kept as small as possible to reduce the unwelcome 78 00:07:25,750 --> 00:07:30,310 recombination at this defect-rich interface. 79 00:07:30,310 --> 00:07:35,900 Another advantage is that the cross-section of the metal fingers can be made much larger 80 00:07:35,900 --> 00:07:41,300 to reduce the resistive losses of the contacts as much as possible. 81 00:07:41,300 --> 00:07:47,600 The fact that the contacts do not cause any shading losses at the back, allows them to become larger. 82 00:07:48,340 --> 00:07:54,700 The passivation layer can have a low refractive index such that it operates like a backside mirror. 83 00:07:54,780 --> 00:08:01,000 It will reflect the light above 900 nm, which is not absorbed during the first pass, 84 00:08:01,000 --> 00:08:06,150 back into the absorber layer, enhancing the absorption path length. 85 00:08:06,150 --> 00:08:10,280 An interdigitated back contact solar cell would look like this. 86 00:08:10,280 --> 00:08:13,300 At the back you have two metal grids. 87 00:08:13,300 --> 00:08:18,300 One collects the current of the n-type contacts and the other collects 88 00:08:18,700 --> 00:08:22,500 the current of the p-type contacts. 89 00:08:24,000 --> 00:08:30,000 At the front side the losses of the light excited charge carriers due to surface recombination 90 00:08:30,000 --> 00:08:35,800 is suppressed by using the same tricks like the back surface field as discussed for the 91 00:08:35,800 --> 00:08:39,500 rear surface for solar cells, based on p-type wafers. 92 00:08:39,500 --> 00:08:45,500 The surface recombination velocity of the front surface is determined by the minority 93 00:08:45,500 --> 00:08:48,200 charge carriers, in this case, the electrons. 94 00:08:48,200 --> 00:08:54,260 At the front side the losses of the light-excited charge carriers, due to the surface recombination 95 00:08:54,260 --> 00:09:00,840 is suppressed by using the same tricks like the back surface field as discussed for the 96 00:09:00,840 --> 00:09:06,630 rear surface for solar cells based on p-type wafers. 97 00:09:06,630 --> 00:09:11,580 The surface recombination velocity of the front surface is determined by the minority 98 00:09:11,580 --> 00:09:15,300 charge carriers, in this case, the holes. 99 00:09:15,300 --> 00:09:21,390 Consequently, we have to create a front surface field. 100 00:09:21,390 --> 00:09:27,710 A higher n-doped region is placed at the front surface indicated by n+. 101 00:09:27,710 --> 00:09:34,130 The interface between the higher-doped n-region and the lower-doped n-region acts again like 102 00:09:34,130 --> 00:09:35,490 a p-n junction. 103 00:09:35,490 --> 00:09:41,540 In this case it will act as a barrier for the light-excited minority holes in the lower 104 00:09:41,540 --> 00:09:45,820 doped region to diffuse to the front surface. 105 00:09:45,820 --> 00:09:52,400 The front surface field behaves like a passivation of the defects at the front interface and 106 00:09:52,400 --> 00:09:58,400 allows to have higher levels for the hole minority densities in the p-doped bulk. 107 00:09:59,690 --> 00:10:05,250 At the front side the reflective losses can be reduced using the same tricks as discussed 108 00:10:05,250 --> 00:10:07,880 for the PERL solar cell. 109 00:10:07,880 --> 00:10:13,310 Deposition of double-layered anti-reflection coatings and texturing of the front surfaces. 110 00:10:13,310 --> 00:10:17,730 SunPower is the company that has developed a cell technology based on interdigitated 111 00:10:17,730 --> 00:10:24,730 back contacts, and they have achieved high solar cell efficiencies of 24.2%. 112 00:10:27,860 --> 00:10:34,170 An alternative concept with high efficiencies is the so-called crystalline wafer based hetero-junction 113 00:10:34,170 --> 00:10:37,089 solar cells as you see in this illustration. 114 00:10:37,089 --> 00:10:42,570 First I have to answer the question: what is a heterojunction? 115 00:10:42,570 --> 00:10:49,570 So far I have introduced you to the concept of p-n junctions with a depletion zone. 116 00:10:50,000 --> 00:10:56,900 These junctions are fabricated by different doping types within the same semiconductor material. 117 00:10:56,990 --> 00:11:01,690 This means the band gap in the p- and n-doped material is the same. 118 00:11:01,690 --> 00:11:05,250 Such p-n junction is called a homojunction. 119 00:11:05,250 --> 00:11:11,370 However, you can also make a junction by two different semiconductor materials. 120 00:11:11,370 --> 00:11:18,080 For instance one semiconductor material that is p-doped and another type of semiconductor 121 00:11:18,080 --> 00:11:21,100 material that is n-doped. 122 00:11:21,100 --> 00:11:28,100 This is what we call a heterojunction. In the c-Si wafer based heterojunction 123 00:11:28,279 --> 00:11:32,580 we make use of two types of silicon based semiconductor materials. 124 00:11:32,580 --> 00:11:39,130 One is again a n-type float zone monocrystalline silicon wafer, the other material is hydrogenated 125 00:11:39,130 --> 00:11:41,650 amorphous silicon. 126 00:11:41,650 --> 00:11:47,180 This is a silicon material in which the atoms are not ordered in a crystalline lattice but 127 00:11:47,180 --> 00:11:49,610 in a disordered lattice. 128 00:11:49,610 --> 00:11:54,960 Next week I will come back to this material when we are discussing thin-film technologies. 129 00:11:54,960 --> 00:12:00,700 For the moment you only have to know that this material has a higher band gap than that 130 00:12:00,700 --> 00:12:02,700 of crystalline silicon. 131 00:12:02,700 --> 00:12:10,800 Secondly, amorphous silicon can be n-doped and p-doped as well, using phosphorous or boron. 132 00:12:11,200 --> 00:12:16,930 I won't go in detail, that is out of the scope of this course, but the electronic band diagram 133 00:12:16,930 --> 00:12:21,410 of a heterojunction of n-doped crystalline silicon and p-doped amorphous silicon in the 134 00:12:21,410 --> 00:12:26,500 dark and thermal equilibrium will look like this. 135 00:12:27,000 --> 00:12:34,029 You see that next to the induced field due to the space charge region, some local energy 136 00:12:34,029 --> 00:12:37,290 steps are introduced. 137 00:12:37,290 --> 00:12:42,870 These steps are caused by the fact that both band gaps are not the same. 138 00:12:42,870 --> 00:12:48,190 At this junction you see that the valence band is higher positioned in the p-type amorphous 139 00:12:48,190 --> 00:12:52,070 silicon in reference to the n-type crystalline silicon. 140 00:12:52,070 --> 00:12:59,070 This will allow the minority charge carriers, the holes, to drift to the p-type silicon. 141 00:12:59,990 --> 00:13:06,060 However, you see in this example that the holes experience a smaller barrier. 142 00:13:06,060 --> 00:13:12,980 In contrast to classical mechanics a particle cannot move through such barrier, in the case 143 00:13:12,980 --> 00:13:15,700 of quantum mechanics, this is still the case. 144 00:13:15,700 --> 00:13:22,400 A large fraction of the holes can move through this barrier and this phenomena is called tunneling. 145 00:13:22,860 --> 00:13:28,770 Now let's go to the crystalline silicon wafer based heterojunction solar cell. 146 00:13:28,770 --> 00:13:33,200 This is a concept which has been invented by the Japanese company Sanyo, 147 00:13:33,200 --> 00:13:36,700 which is currently part of Panasonic. 148 00:13:36,730 --> 00:13:43,500 The Panasonic cell is called the HIT cell, which stands for heterostructure with intrinsic thin film. 149 00:13:43,900 --> 00:13:47,480 The HIT cell configuration has two junctions. 150 00:13:47,480 --> 00:13:53,430 The junction at the front side is formed using a thin layer of only 5 nanometers of intrinsic 151 00:13:53,430 --> 00:13:58,940 amorphous silicon, which is indicated by the red color. 152 00:13:58,940 --> 00:14:05,420 A thin layer of p-doped amorphous silicon is deposited on top and here is indicated 153 00:14:05,420 --> 00:14:07,750 with the blue color. 154 00:14:07,750 --> 00:14:13,270 The heterojunction forces the holes to drift to the p-layer. 155 00:14:13,270 --> 00:14:15,950 At the rear surface a similar junction is made. 156 00:14:15,950 --> 00:14:21,510 First, a thin layer of intrinsic amorphous silicon is deposited on the wafer surface, 157 00:14:21,510 --> 00:14:23,570 indicated by the red. 158 00:14:23,570 --> 00:14:30,700 On top of the intrinsic layer an n-doped amorphous silicon is deposited, indicated by the yellow color. 159 00:14:31,580 --> 00:14:37,740 As discussed earlier, for high-quality wafers, like this n-type float-zone monocrystalline 160 00:14:37,740 --> 00:14:43,990 silicon wafer, the recombination of charge carriers at the surface determines the lifetime 161 00:14:43,990 --> 00:14:46,000 of the charge carriers. 162 00:14:46,990 --> 00:14:53,900 The advantage of the HIT concept is that the amorphous silicon acts like a very good passivation material. 163 00:14:54,200 --> 00:15:00,580 In this approach the highest possible lifetimes for charge carriers are accomplished. 164 00:15:00,580 --> 00:15:07,230 The c-Si wafer based heterojunction solar cell has the highest achieved open-circuit voltages 165 00:15:07,230 --> 00:15:09,779 among the crystalline silicon technologies. 166 00:15:09,779 --> 00:15:16,779 Panasonic achieved an open-circuit voltage of 750 mV. 167 00:15:17,790 --> 00:15:21,649 How do the charge carriers travel to the contact? 168 00:15:21,649 --> 00:15:27,190 The conductive properties of the p-doped amorphous silicon are relatively poor. 169 00:15:27,190 --> 00:15:32,900 In the homojunction solar cells the diffusion to the contacts takes place in the emitter layer. 170 00:15:32,920 --> 00:15:39,540 In contrast, in a HIT solar cell this occurs through the transparent conductive oxide material, 171 00:15:39,540 --> 00:15:45,870 like an ITO, which is deposited on top of the p-doped layer. 172 00:15:45,870 --> 00:15:51,269 The ITO is needed as the conductivity of the p-type layer is too poor. 173 00:15:51,269 --> 00:15:56,830 This results in such small diffusion lengths that a practical metal finger spacing can 174 00:15:56,830 --> 00:16:00,709 not be achieved using the p-type layers. 175 00:16:00,709 --> 00:16:04,589 Therefore ITO is used. 176 00:16:04,589 --> 00:16:10,260 An advantage of the HIT cell concept is that it allows to introduce the same contact scheme 177 00:16:10,260 --> 00:16:12,200 at the n-type back side. 178 00:16:12,200 --> 00:16:18,360 It means that this solar cell can be used in a bifacial configuration, it can collect 179 00:16:18,360 --> 00:16:24,600 light from the front, and scattered and diffuse light falling on the backside of the solar cell. 180 00:16:26,649 --> 00:16:30,790 Another important advantage of the HIT solar cell is that the amorphous silicon layers 181 00:16:30,790 --> 00:16:37,089 are deposited using cheap and straightforward plasma-enhanced chemical vapor deposition 182 00:16:37,089 --> 00:16:42,360 technology at low temperatures, not higher than 200 degrees Celsius. 183 00:16:42,360 --> 00:16:48,420 This means that making the front surface and back surface field in this type of solar cells 184 00:16:48,420 --> 00:16:50,500 is very cheap. 185 00:16:50,529 --> 00:16:55,500 Furthermore, this technology allows to use the n-type wafers. 186 00:16:55,600 --> 00:17:02,600 Summarized, the high-efficiency crystalline silicon wafer based solar cells are shown here. 187 00:17:02,870 --> 00:17:09,140 The record efficiency of a PERL solar cell was 25%, however, this was a lab-scale solar cell 188 00:17:09,140 --> 00:17:13,780 with an area of 4 cm^2. 189 00:17:13,780 --> 00:17:19,250 The record efficiency for an interdigitated crystalline silicon solar cell has been achieved 190 00:17:19,250 --> 00:17:20,400 by SunPower. 191 00:17:20,400 --> 00:17:29,400 They achieved an efficiency of 24.2 % on a wafer size of 155 cm^2. 192 00:17:29,650 --> 00:17:36,610 Finally, for the c-Si wafer based heterojunction solar cell, Panosonic achieved an efficiency 193 00:17:36,610 --> 00:17:43,400 of 24.7% on a wafer size of 102 cm^2. 194 00:17:43,400 --> 00:17:49,400 The efficiencies for multicrystalline silicon solar cells are lower as the wafer quality is lower. 195 00:17:49,760 --> 00:17:57,500 The best efficiency achieved is 19.5% by Q-cells on a wafer with a size of 243 cm^2. 196 00:17:57,600 --> 00:18:05,100 This is around 5% absolute below the record efficiencies based on monocrystalline silicon wafers. 197 00:18:05,470 --> 00:18:09,480 Now we know the efficiencies of solar cells. 198 00:18:09,480 --> 00:18:14,429 However, in practice we install panels on our roof. 199 00:18:14,429 --> 00:18:22,400 In the next block we will answer the question: How do we make solar modules out of solar cells?