4.9
Conclusions on optimized a-Si:H/c-Si in-
terface passivation
In conclusion, intrinsic a-Si:H acts as a high performance passivation layer on all kind of c-Si substrates. This simple, low temperature passivation process does not need to fear comparison to other high performing passi-
vation schemes such as SiO2 and SiNx. The i a-Si:H’s excellent passivation
quality is identified in this work to be related to a marked decrease of the interface dangling bond density. This conclusion is reached by modeling in- terface recombination by means of amphoteric dangling bonds, as observed in bulk a-Si:H. The a-Si:H/c-Si interface recombination limit imposed on
VOC is pushed to record values well over 700 mV on all substrates studied
experimentally, that include FZ and CZ of various doping.
Field-effect passivation can be added by microdoping i a-Si:H layers or fixing their outer surface potential by an overlying doped a-Si:H layer.
In contrast to SiO2 and SiNx, the lowest effective surface recombination
velocity (1 cm/s corresponding to a lifetime of 7.5 ms) is measured on <111> oriented c-Si. Because textured, complete high performance solar cells rely on the passivation of <111>-oriented pyramidal facets, this is an important finding.
The abruptness of the crystallographic nature of the a-Si:H/c-Si het- erointerface plays an important role in determining the interface passiva- tion quality. Abrupt interfaces are observed when passivating flat c-Si by VHF-PECVD grown thin-film Si layers (and layer stacks), provided that epitaxial growth is avoided by choosing reasonable PECVD process param- eters. This optimal crystallographic interface is much more challenging to obtain in textured c-Si substrates. Lifetime measurements on differently textured c-Si substrates emphasize the importance of the texture’s size and quality. With the appropriate surface preconditioning, the same high
implVOCs of well over 700 mV, as on flat c-Si, are reached by i a-Si:H pas-
sivation. TEM micrographs of i a-Si:H plus doped a-Si:H/µc-Si:H layer stacks, as used as emitter and BSF in Si HJ solar cells, reveal the pres- ence of epitaxial growth at the bottom of the pyramidal valleys. These features are identified here as the cause of the poorer performances of i a- Si:H passivation on textured c-Si. Therefore, contrariwise to flat c-Si, the substrate’s crystalline nature is not lost through the i a-Si:H interlayer for the subsequent growth of the doped a-Si:H/µc-Si:H emitter and BSF layer. In contrast to the flat doped a-Si:H/µc-Si:H transition layers’ preliminary development on glass, layer characterization tools do not work any longer
4.9. Conclusions on optimized a-Si:H/c-Si interface passivation
that the doped a-Si:H/µc-Si:H layers’ growth conditions can be adapted to the textured surface’s nature and optimized passivation layer stacks are developed. Finally, using fully amorphous doped Si layers (still in stack with i a-Si:H) instead of a-Si:H/µc-Si:H transition layers for emitter and BSF formation in Si HJ solar cells, this detrimental epitaxial growth could possibly be prevented.
Chapter 5
Amorphous/crystalline
silicon heterojunction
solar cells
First of all, the carrier transport through the interfaces in Si HJ solar cells is briefly addressed in this chapter (Sec. 5.2). Lifetime measurement guided Si HJ solar cell optimization by device diagnostics is then presented, where some redundancy to Chap. 4 (a-Si:H/c-Si interface passivation) is included in this new context of Si HJ solar cells (Sec. 5.3). From there on, the achievement of well performing flat Si HJ solar cells is straight-forward (Sec. 5.4). On textured c-Si, TEM micrographs greatly assist lifetime mea- surements by identifying local epitaxial growth in pyramidal valleys as the main concern for textured Si HJ solar cell optimization (Sec. 5.5). This chapter is supposed to be self-consistent.
5.1
Introduction
Silicon heterojunction solar cells made by Sanyo, called HIT (heterojunc-
tion with intrinsic thin-layer) [WTS+91], achieve record efficiencies of 22.3%
with VOC = 725 mV, JSC = 38.9 mA/cm2 and F F = 79.1% on a total area
of 100.5 cm2 as confirmed by AIST [TYT+09]. The fabrication of a-Si:H/c-
Si heterojunction solar cells is of great interest for several reasons:
• A fabrication process at low temperature (200 ◦C), which reduces
the wafer breakage (in particular when Al layers are used) and the energy necessary to invest in the fabrication process.
5.1. Introduction
• As ultra-low surface recombination is achieved and the low fabrica- tion temperature reduces material breakage, thinner wafers can be used, resulting in a strong reduction in the consumption of silicon (with a long-term potential of 4 g/Wp against 10 − 11 g/Wp of stan- dard c-Si solar cells, including material losses).
• The possibility to achieve ultra high solar cell efficiency > 20%. • A better temperature coefficient than standard c-Si solar cells, i.e. a
reduction from −0.5%/◦C to −0.25%/◦C.
• As the solar cell fabrication process is now based on thin-film sil- icon, it offers the prospect of applying the cost-effective industrial deposition techniques used for thin Si layers.
Therefore, Si HJ solar cells combine the best of crystalline silicon solar cells on one hand and amorphous Si solar cells on the other hand, illustrated in Fig. 5.1. That is why this kind of ultra-high efficient solar cell is one of the most promising candidates to reach grid parity in the segment of highly efficient c-Si cells.
80 nm SiN 15 μm p+c-Si 400 nm n+c-Si 200 μm p c-Si 30 μm Al 250 μm metal grid contact
(a) 85 nm ITO 15 nm p+a-Si:H/μc-Si:H 5 nm i a-Si:H 15 nm n+a-Si:H/μc-Si:H 5 nm i a-Si:H 100 μm n c-Si 100 nm Al or Ag 110 nm ITO 100 μm metal grid contact
(b) 1-2 μm ZnO 15 nm p+a-Si:H/μc-Si:H glass 15 nm n a-Si:H/μc-Si:H 250 nm i a-Si:H 0.2-2 μm back contact 2 μm (c)
Figure 5.1: a) Conventional crystalline Si solar cell, b) Si HJ solar cell and c) amorphous Si solar cell.
The interpretation of lifetime measurements is a powerful tool for the characterization of the performances of a-Si:H layers and layer stacks with