High deposition rate nanocrystalline silicon with enhanced homogeneity - PowerPoint PPT Presentation

1 / 16
About This Presentation
Title:

High deposition rate nanocrystalline silicon with enhanced homogeneity

Description:

High deposition rate nanocrystalline silicon with enhanced homogeneity. A. D. Verkerk, J.K. Rath and R.E.I. Schropp. Utrecht University, The Netherlands ... – PowerPoint PPT presentation

Number of Views:63
Avg rating:3.0/5.0
Slides: 17
Provided by: Bla74
Category:

less

Transcript and Presenter's Notes

Title: High deposition rate nanocrystalline silicon with enhanced homogeneity


1
High deposition rate nanocrystalline silicon with
enhanced homogeneity
  • A. D. Verkerk, J.K. Rath and R.E.I.
    SchroppUtrecht University, The
    NetherlandsNanophotonics Physics of Devices
  • ICANS23 August 25, 2009

2
Introduction
  • Hydrogenated nanocrystalline silicon (nc-SiH)
  • VHF PECVD - 60 MHz
  • High deposition rate (rd) 4.5 nm/s
  • Solar cells with 6.4 efficiency
  • Comparison between high and low rd
  • Low rd High rd
  • rd 0.5 nm/s 4.5 nm/s
  • pressure 5 mbar 9 mbar
  • SiH4 flow 3 sccm 20 sccm
  • H2 flow 84 sccm 300 sccm
  • VHF power 20 W 350 W

3
Outline
  • Goal
  • Crystalline fraction in low and high deposition
    rate
  • Mechanisms for instability
  • Length scales and time scales for back diffusion
  • Stabilisation time for back diffusion effects
  • Methods of homogeneity improvement
  • Results

4
Goal
  • High efficiency solar cells need homogeneous
    phase composition
  • The growth-direction homogeneity of high
    deposition rate (rd) nc-SiH needs to be improved

5
Development of crystalline fraction
Intrinsic films deposited onnc-SiH seed layer
6
Possible explanations
  • Surface diffusion length of silicon shorter at
    high rd
  • Very high depletion gives stronger start-up
    effects
  • Back diffusion of silane from outside plasma
    volume

7
Diffusion length
  • Calculate diffusion length of Ar in H2
  • Diffusion coefficient D 1/p
  • Residence time in plasma tres,p p and
    1/flowtotal
  • Diffusion length is
  • Low rd High rd
  • rd 0.5 nm/s 4.5 nm/s
  • pressure 5 mbar 9 mbar
  • total flow 87 sccm 320 sccm
  • Residence time 0.39 s 0.19 s
  • Diffusion constant 240 cm2s-1 133 cm2s-1
  • Diffusion length 19 cm 10 cm

8
Possible explanations
  • Surface diffusion length of silicon shorter at
    high rd
  • Very high depletion gives stronger start-up
    effects
  • Back diffusion of silane from outside plasma
    volume

POSSIBLE
9
Stabilisation time for silane back diffusion
  • Depends on residence time in inactive region
  • Depends on amount of unused silane depletion
  • Solution can be written
  • Background initial mixture is important

N SiH4 concentration S Change due to source
flow G Change due to deposition tres,r Reactor
residence time
10
Calculation of SiH4 concentration (high rd)
11
Compare with low rd regime
  • Stabilisation time depends on flow and pressure
  • High rd (4.5 nm/s) stabilizes in 7 seconds
  • Compare Low rd (0.5 nm/s) stabilizes in 15
    seconds
  • Diffusion length and stabilisation time
    comparable
  • Stabilisation affects thicker section of the film
    in high rd

12
Methods of homogeneity improvement
  • After deposition start
  • Profiling of hydrogen flow
  • Profiling of silane flow
  • Profiling of VHF power
  • Before deposition start
  • Tuning of background mixture
  • Start with hydrogen plasma
  • Features of hydrogen plasma start
  • Hydrogen background gas mixture
  • Increased surface mobility by atomic H abundance

13
Effect of 5-second hydrogen plasma start
14
Further tuning of the crystalline development
15
Discussion and Conclusions
  • Possible increasing back diffusion at lower
    depletion
  • Stabilisation time comparable between high and
    low rd
  • Diffusion length comparable between high and low
    rd
  • Still incubation layer is much thicker in high rd
  • Hydrogen start effectively reduces incubation
    layer
  • Depletion profiling can be used for fine-tuning

16
Acknowledgement
  • Thanks to
  • Martin Huijzer for depositions
  • Wim Goedheer for fruitful discussions
  • This project is undertaken with financial support
    from the Ministry of Economic Affairs of The
    Netherlands program EOS (Subsidy Energy
    Research)
  • Thank you for your attention!
Write a Comment
User Comments (0)
About PowerShow.com