Inline PL Imaging Techniques for Crystalline Silicon Cell Production. F. Korsós, Z. Kiss, Ch. Defranoux and S. Gaillard

Transcription

1 Inline PL Imaging Techniques for Crystalline Silicon Cell Production F. Korsós, Z. Kiss, Ch. Defranoux and S. Gaillard

2 OUTLINE I. Categorization of PL imaging techniques II. PL imaging setups III. Inline PLI for as-cut wafer sorting a) Band-to-band PLI correlation to µ-pcd b) Defect-band PLI IV. Inline PL imaging of passivated surface PV wafers a) Correlation between 2D and Inline PLI b) Correlation to QSS-µPCD J0 V. PL imaging on finished cells a) Defect visualization comparison to EL VI. Summary 2

3 PL imaging techniques The PL phenomenon hν < hν 0 Light source (hν 0 > E g ) Detector (Camera, diode, spectrograph) And the physics behind E g The PL result by different detector types hν 0 > E g E c hν 2 < E g hν 1 ~ E g hν 3 < E g E v Band-to-band luminescence (ν 1 ): I 2 PL, B B ~ n( Ndop n) ~ a eff b eff From Princeton instruments homepage Defect-band luminescence (ν 1,2 ): I ~ PL, D B N defect 3

4 Classification of PL measurements techniques Photoluminescence Spectroscopy (λ resolved) Imaging (Integrated in λ) Single point spectroscopy Camera based spectroscopy Scanning On the fly (Line camera based) Imaging by camera Stationary (2D camera) Band to Band PL Defect - Band PL Band to Band PL Defect - Band PL (not for Si CCD) 4

5 Inline PL imaging for as-cut wafers 2D camera based setup: Most common imaging setup Stationary wafer position Full wafer illumination steady state conditions Si CCD only B-B PL carrier lifetime imaging Low injection level on as-cut wafers (~1E13cm -3 ) Inline line camera based setup for as-cut wafes: Moving wafer IR-laser Illumination in a focused stripe InGaAs camera Capability for both B-B (lifetime) and Defect Band PL (depends on the applied filter set) Higher injection level (~1E16cm -3 ) comparable to n in solar cells in operation Sharp PL image of moving as-cut wafer within 1s Possible differences due to different injection level: Contrast at grain boundaries (for mc:si) Visualization of O rings due to different impact of thermal donors Related publication is in preparation 5

6 Inline PL imaging for as-cut wafers The primary inline PLI application is the inspection of as-cut mc:si wafers Usually B-B PL photons are captured by applying 1000nm-1300nm bandpass filters So, practically carrier lifetime is imaged F. Korsós, et al. 26th EU-PVSEC 2012 Parabolic relationship between PL signal and carrier lifetime. PL images can be converted into carrier lifetime maps by the application in-site µ-pcd carrier lifetme measurement 6

7 Inline PL imaging for as-cut wafers Material defects detected on the wafers Grain boundaries High recombination rate at grain boundaries Dislocations Contamination along the wafer edges Highly contaminated wafer (top / bottom wafers) 7

8 Inline PL imaging for as-cut wafers Theese defects can be identified and classified Too large defect area means efficiency loss So, maximum reachable efficiency of the cell can be predicted Relative efficiency [%] Efficiency comparison Sample No. Measured relative efficiency Measured Efficiency Efficiency Forecast Efficiency correlation Forecasted relative efficiency Measured Efficiency 8

9 Inline PL imaging for as-cut wafers By applying >1300nm long-pass filters, the defect-band luminescence can be captured. Dislocations emit significant flux of defect-band PL photons. So, they are directly visible in defect PL images. PL spectrum at the dislocation PL spectrum far from dislocation 9

10 Inline PL imaging after passivation process Production of high effiency solar cells requires the control of surface passivation by the measurment of: Carrier lifetime Emitter saturation current Implied Voc PLI is capable to image these quantities, but requires in-situ calibration. QSS-uPCD is applicable for this purpose* Correct measurement needs steady-state conditions. The focused beam inline PL does not suitable. Our patented solution** is the application of an extended steady-state zone around the line captured by the line camera. Inline QSS-PL *F. Korsós et al, 28th EU-PVSEC (2013) **J. Lagowski, et al. US Patent Application 10

11 Inline PL imaging after passivation process Experimental proof of the importance of the steady-state excitation zone Inline QSS-PL give the same PL image as the 2D camera PL which uses full wafer steady state illumination. 11

12 Inline PL imaging after passivation process Inline QSS-PL image can be converted into J oe or V oc map applying calibration by QSS-uPCD measurement 12

13 Inline PL imaging of finished cells The standard method to detect the metallization defects (shunts, bad fingers) is ElectroLuminscence (EL) We have found that inline PL is suitable for the same purpose but without requiring electrical contact to the wafer. Example PL image Example PL image raw material: contaminatio n wires: shunts, bad fingers 13

14 Inline PL imaging of finished cells Comparison of PL and EL images: Shunts EL : higher current flows through and near the shunts local warming brighter luminescence, PL : uniform illumination, higher current at shunt lower carrier density at the shunts fainter luminescence; inverse features in EL and PL with respect to each other. Inline- PL EL 14

15 Inline PL imaging of finished cells Comparison of PL and EL images: Bad Fingers EL : Higher local resistance lower local carrier injection lower local carrier density fainter luminescence; PL: Decreased conductance removes larger carrier density around the bad finger thinner darkened stripe around the wire and brighter luminescence in its surrounding inverse features in EL and PL with respect to each other. Comparison of PL and EL images: Contamination Lower local lifetime lower carrier concentration lower PL singal in fainter luminescence in both EL and PL resulting the same dark features on the images. Inline- PL EL 15

16 Summary Summary: PL imaging is applicable in each step of solar cell production to identify wafer quality and / or process problems We developed inline PL systems which is capable to do this job within 1s on moving wafers On as-cut wafers B-B PL signal correlates to upcd lifetime. It can be converted to a lifetime image. Crystal defects can be visulaized with high resolution. Based on this bad quality wafers can be sorted out from production The capturing of defect-band PL photons makes possible the direct imaging of specific defects (e.g. dislocations) Proper PL imaging of passivated PV wafers requires the application of steady-state illumination zone. By this accurate PL measurement can be realized in on-the-fly PL configuration as well. Based on the excellent correlation to QSS-uPCD maps, the PL image can be directly converted to Joe and implied Voc maps ON finished cells EL and PL images may differ but both indicates the same metallization failures. 16