Surface and bulk passivation of multicrystalline silicon solar cells by ...
Surface and bulk passivation of multicrystalline silicon solar cells by ... Surface and bulk passivation of multicrystalline silicon solar cells by ...
57 seen that, initially Ss decreases with an increase in the injection level and then increases Figure 3.10 Measured effective SRV showing dependence on excess carrier density [89]. According to this model calculation, very high Q f (1-3 x 10 12 cm-2) will lead to no injection level dependence of the effective surface recombination velocity for p-type Si wafer, which is in contradiction to experimental results as shown in Figure 3.10 [89]. This discrepancy is assumed to be caused by carrier generation and recombination in the depletion region [87]. It is implied that a complete model of surface recombination at SiΝ -Si interface should also consider other recombination mechanisms. 3.2.3 A Modified Model including Recombination in Damaged Region As mentioned above, the existing SRH formalism fails to obtain a reasonable agreement between experiment and theory in terms of injection level dependence. It is due to the
58 more complicated properties of SiΝ :H-Si interface compared to SiO 2-Si interface. In order to solve this problem, a "deeper" insight of SiΝ :H-Si interface is needed. It has been found that, during the nitridation procedure, a damaged layer is formed [90 - 92], which is not only critical for H storage and subsequent bulk passivation after firing, but also is crucial for surface passivation. Figure 3.11 is an XTEM image which shows the existence of the surface damage between PECVD SiΝ :H layer and Si wafer. Figure 3.11 High-resolution cross-sectional TEM of Si-SiΝ :H interface before firing showing process-induced damage by plasma [92]. Α large distribution of traps or recombination centers results in a large carrier density across the damaged layer. This suggests that a complete surface recombination model should go beyond the current SRH formalism to a deeper extent. The recombination is evoked in the damaged region that is formed by processinduced defects in the vicinity of the surface. Hence, there is an increased minority-
- Page 25 and 26: 6 bond is called a hole. It too can
- Page 27 and 28: 8 Figure 1.4 The I-V characteristic
- Page 29 and 30: 10 First generation cells consist o
- Page 31 and 32: 12 Basically, materials for manufac
- Page 33 and 34: 14 Defects are generally categorize
- Page 35 and 36: 16 copper, or nickel in concentrati
- Page 37 and 38: 18 the SiNx:H layer during the ther
- Page 39 and 40: CHAPTER 2 SILICON NITRIDE LAYER FOR
- Page 41 and 42: 22 reflectance of polished Si can b
- Page 43 and 44: 24 information is application-orien
- Page 45 and 46: 26 film fed growth (EFG) ribbon sil
- Page 47 and 48: 28 Figure 2.5 Deposition of SiΝ :
- Page 49 and 50: 30 Figure 2.6 shows the dependence
- Page 51 and 52: 32 atoms, the interface states are
- Page 53 and 54: 34 2.5 Bulk Passivation of Si by Si
- Page 55 and 56: 36 It was found that the bulk lifet
- Page 57 and 58: CHAPTER 3 MODELING OF SURFACE RECOM
- Page 59 and 60: 40 Figure 3.2 Schematic diagram of
- Page 61 and 62: 42 σ and σp are the capture cross
- Page 63 and 64: 44 Qsi — charge density induced i
- Page 66 and 67: 47 Figure 3.5 The calculated depend
- Page 68 and 69: 49 * 10 Λ m; m is in a range from
- Page 70 and 71: 51 Na, sigma_n, sigma_p: enter x.xx
- Page 72 and 73: 53 Figure 3.7 Measured Seff(Δn) de
- Page 74 and 75: 55 curves converge to a single valu
- Page 78 and 79: 59 carrier recombination within the
- Page 80 and 81: 61 recombination in the SCR influen
- Page 82 and 83: 63 Figure 3.13 shows that: 1) after
- Page 84 and 85: CHAPTER 4 MINORITY-CARRIER LIFETIME
- Page 86 and 87: 67 Figure 4.1 Α photograph of QSSP
- Page 88 and 89: 69 work. The most convenient is 1 m
- Page 90 and 91: 7Ι dependence of the minority carr
- Page 92 and 93: 73 It was tempting to assume that l
- Page 94 and 95: 75 resistivities and lifetime) do n
- Page 96 and 97: 77 5.2 Objective An electronic mode
- Page 98 and 99: 79 Figure 5.2 is a photograph of a
- Page 100 and 101: 81 impurity-gettering methods which
- Page 102 and 103: 83 distribution of local currents a
- Page 104 and 105: 85 modeling. Wafers were selected f
- Page 106 and 107: 87 Figure 5.5 A comparison of (a) d
- Page 108 and 109: 89 alloying results in metallizatio
- Page 110 and 111: 91 (i) Defect clusters are the prim
- Page 112 and 113: 93 SiNX induced charge density on t
- Page 114 and 115: APPENDIX I PROGRAMS TO CALCULATE SR
- Page 116 and 117: 97 phin = -ΕΙ - 1 / beta * log(nd
- Page 118 and 119: 99 ίter3 = 0 for xi=1 to nmax/2-1
- Page 120 and 121: 101 input "output file name {XXXXXX
- Page 122 and 123: 103 F (i) = (exp (beta * (phip - ph
- Page 124 and 125: APPENDIX III COMPUTATIONAL METHOD F
57<br />
seen that, initially Ss decreases with an increase in the injection level <strong>and</strong> then increases<br />
Figure 3.10 Measured effective SRV showing dependence on excess carrier density [89].<br />
According to this model calculation, very high Q f (1-3 x 10 12 cm-2) will lead to<br />
no injection level dependence <strong>of</strong> the effective surface recombination velocity for p-type<br />
Si wafer, which is in contradiction to experimental results as shown in Figure 3.10 [89].<br />
This discrepancy is assumed to be caused <strong>by</strong> carrier generation <strong>and</strong> recombination in the<br />
depletion region [87]. It is implied that a complete model <strong>of</strong> surface recombination at<br />
SiΝ -Si interface should also consider other recombination mechanisms.<br />
3.2.3 A Modified Model including Recombination in Damaged Region<br />
As mentioned above, the existing SRH formalism fails to obtain a reasonable agreement<br />
between experiment <strong>and</strong> theory in terms <strong>of</strong> injection level dependence. It is due to the