1. Research Opportunities in Laser Surface Texturing/Crystallization of Thin-Film Solar Cells Y. Lawrence Yao Columbia University January 4 th, 2011 Research Opportunities in Energy Manufacturing 2011 CMMI Grantees Conference 1

2. Outline Overview of Photovoltaic (PV) Technology Optical Confinement Methods Laser Surface Texturing (LST) Applications Simultaneous texturing/crystallization of a- Si:H thin films Research Opportunities Research Opportunities in Energy Manufacturing 2011 CMMI Grantees Conference 2

3. Comparison of PV Absorbers AbsorberProsCons Bulk crystalline silicon Stable, high efficiency High cost, low absorption coefficient (indirect band gap) Cost is $2.50/watt a-Si:H Low cost, has potential Lowest cost can be $0.5/watt Unstable, low efficiency Nanocrystalline silicon Stable, large-area deposition Thicker than a-Si:H nc-Si:H/a-Si:H (stable) III-V (GaAs, InP) High efficiency and absorption coefficient High cost of producing devices, easily cleaved and weak, crystal imperfection, cannot use lower-cost deposition method. For space application, multi-junction devices. Cost of electricity is ~1000 times of silicon cells. CdTe/CdS High absorption coefficient, a few micron thick cell Complex deposition process, efficiency is not very high, cost is $0.98/watt Chalcopyrite compounds (I,II,VI) CuInSe 2, CuInS 2, CuGa 1-x In x Se 2 High absorption coefficient, a few micron thick cell higher cost of electricity than a-Si:H cells Dye sensitized and organic Low cost of both material and substrate Low efficiency, still under development Thin Films Research Opportunities in Energy Manufacturing 2011 CMMI Grantees Conference 3

4. Performance gaps between best device efficiencies in the lab and attainable efficiencies for several solar cell technologies (Kazmerski, 2005) Research Opportunities in Energy Manufacturing 2011 CMMI Grantees Conference 4 Performance Gaps in Efficiency At ~1.4eV highest attainable  - III-V (GaAs) Si: 1.12eV a-Si:H – 1.7eV a-Si:H – largest potential gain in 

5. Research Opportunities in Energy Manufacturing 2011 CMMI Grantees Conference 5

6. Overview of Solar Cells Thin films of more interest due to the large-area manufacturing feasibility a-Si:H has the lowest cost, however, it also suffers low efficiency and instability (the Staebler- Wronski Effect) GaAs has the highest efficiency, however, it costs 1000 times to make as other thin film absorbers III-V compound based multi-junction + concentrator can achieve the best efficiency (42.4%) Research Opportunities in Energy Manufacturing 2011 CMMI Grantees Conference 6

7. Optical Confinement Methods Anti-reflection coating (ARC)  Universally used Chemical etching/texturing  Anisotropic alkaline and isotropic acid  Not applicable for amorphous and thin films Mechanical texturing  Use mechanical dicing saws and blades - damage KOH (c-Si) (D. Heslinga, 2008) HF and HNO 3 (polyc-Si) (D. Heslinga, 2008) Research Opportunities in Energy Manufacturing 2011 CMMI Grantees Conference 7

8. Optical Confinement Methods Reactive ion etching Low throughput Laser surface texturing  Sharper surface features  Better absorption  More uniform absorption  Low throughput not easy for scaling up Plasma (c-Si) (D. Heslinga, 2008) Acid 1 Acid 2 Research Opportunities in Energy Manufacturing 2011 CMMI Grantees Conference 8

9. LST Applications Tribology Biological Other applications in PV Research Opportunities in Energy Manufacturing 2011 CMMI Grantees Conference 9

10. Beyond Light Trapping (c-Si) LST of c-Si in different atmosphere Below-band-gap abs. (a)SF 6, (b) N 2, (c) Cl 2, (d) air, (e) vacuum all used fs laser Carey, PhD Thesis, Harvard, 2004 Research Opportunities in Energy Manufacturing 2011 CMMI Grantees Conference 10 Energy Valence band Conduction band Sub dopant band Band gap

11. Beyond Light Trapping (c-Si) c-Si, SF 6, Crouch et al,2004  fs laser: recessed surface, smaller pitch (2 to 3 times of, interference), ns laser: protruded surfaced, larger pitch (capillary wave generation)  Below-band-gap absorption: ns-laser allows higher doping concentration; annealing diffuses out dopants Research Opportunities in Energy Manufacturing 2011 CMMI Grantees Conference 11

12. 800nm, 130fs, 0.4J/cm 2 Film thickness 1.6 µm Feasible for thin films Below-band-gap absorption enhancement without dopant nc-Si layer (1,100 nm) Increased defects H. Wang, et al, 2009 a-Si:H Thin Films 248nm, 30ns, 0.4J/cm 2 Film thickness 1.6 µm Research Opportunities in Energy Manufacturing 2011 CMMI Grantees Conference 12

13. nc-Si layer in a-Si:H film Texturing and surface crystallization in one step (XRD, TEM, EBSD) ns laser induces more crystallinity Potential for stability improvement Crystalline structure to be further studied Cavities in ns laser to be studied H. Wang, et al, 2009 ns laser sample Cross-section TEM ns laser sample Cross-section TEM fs laser sample Research Opportunities in Energy Manufacturing 2011 CMMI Grantees Conference 13

14. Research Needs The one-step surface texturing/crystallization of thin film  Understand laser type and process conditions on resultant crystalline structures  Understand how the partial crystallization affecting stability of a-Si:H cells  Simultaneous doping (e.g., sulfur) –how does doping affect a-Si:H (minority carrier mobility and lifetime) Research Opportunities in Energy Manufacturing 2011 CMMI Grantees Conference 14

15. Research Needs How to apply LST on III-V (e.g., GaAs) and multijunction cells  MOCVD for crystalline GaAs thin films is very expensive  Low-cost MBD for amorphous GaAs is much cheap –LST to surface texturing and crystallization  To address the high sensitivity to impurities introduced during the process Research Opportunities in Energy Manufacturing 2011 CMMI Grantees Conference 15

16. Research Needs How to apply LST on III-V (e.g., GaAs) and multi-junction cells (cont.)  LST can potentially be used for texturing+crystallization+junction doping as a one-step process for each junction  Issues associated with complete crystallization throughout film thickness instead of partial crystallization  Effects of the tunnel junctions Research Opportunities in Energy Manufacturing 2011 CMMI Grantees Conference 16

17. Research Needs Large-area, high-throughput LST  Effects of spatial and temporal characteristics of laser irradiation  Spatial: Homogeneous intensity/mask projection (R. Delmdahl, et al,2010) Research Opportunities in Energy Manufacturing 2011 CMMI Grantees Conference 17

18. Research Needs  Temporal:  longer laser pulse-width for crystallization  Double-peak pulse for high- throughput crystallization  But to address issues associate with increased HAZ and hydrogen explosion Research Opportunities in Energy Manufacturing 2011 CMMI Grantees Conference 18