Comparison of RF-PCD and MW-PCD Lifetime Measurements on Silicon Wafers and Bricks

Transcription

1 Comparison of RF-PCD and MW-PCD Lifetime Measurements on Silicon Wafers and Bricks Dr. Kevin Lauer CiS Forschungsinstitut für Mikrosensorik und Photovoltaik GmbH An-Institut der TU Ilmenau, Konrad-Zuse-Str. 14, D Erfurt Workshop on Test Methods for Silicon Feedstock Materials, Bricks and Wafers, München,

2 Overview Something about CiS Lifetime measurement basics Round-robin experiment 2

3 Something about CiS 3

4 History 1992 Formation of CiS e.v Formation of CiS Institute as a limited company 1997 First prototyping of silicon radiation detectors for HEP 1999 Change of corporate form to a nonprofit limited company 2003 New location: Technology Center AZM Formation of the Solar Center 2004 New clean room 2008 Best Supplier Award (CERN) 2010 AMA Innovation Award for the In-Ear Sensor system 4

5 Business Unit Photovoltaic Focus 1: Silicon wafer analysis Analysis of slurry and sawing wire FTIR analysis of wafers Breakage tests Lifetime measurement Focus 3: Module testing Climate chamber Flasher UV chamber / light soaking test High voltage test Focus 2: Solar cell processing Texturing, diffusion, anti-reflective coating, metallization up to 210x210 mm 2 Development of improved process steps for industrial solar cell manufactures Development of new cell designs Analysis of cell parameters (IV, EQE, IQE, thermography, electroluminescence, LBIC ) 5

6 Silicon wafer analysis R&D mainly focusing on impact of silicon quality on solar cell performance Development of characterization methods 6

7 Correlation of lifetime and cell efficiency relative solar cell efficiency Block I Block II lifetime measured at passivated raw wafer τ [µs] correlation visible limiting defects? 7

8 Determination of Interstitial Iron Concentration E12 1/cm³ E12 1/cm³ 2.6E12 1/cm³ E11 1/cm³ 10 y [mm] lifetime τ [µs] 1.5E12 1/cm³ iron boron pairs interstitial iron E11 1/cm³ 10-3 excess charge carrier density n [cm ] x [mm] Lifetime measurement in both states of the meta-stable iron boron defect complex Map of the interstitial iron concentration K. Lauer et al., J. Appl. Phys. 104 (10), (2008) 8

9 Low temperature photoluminescence spectroscopy 0.05 B 0.04 Al PL intensity [a. u.] P wavelength λ [nm]] analysis of shallow impurities 9

10 FTIR (low temperature) Al C B P oxygen, carbon, nitrogen shallow impurities 10

11 Defect etching methods helical dislocation etch pits of dislocation line driving through silicon surface 11

12 Silicon wafer analysis Characterization methods Electrical quality of silicon (excess charge carrier lifetime, resistivity,...) Mapping of electrical active species (e.g. interstitial iron) Oxygen, carbon, nitrogen analysis using room temperature FTIR Low temperature FTIR Low temperature photoluminescence spectroscopy DLTS SIMS REM/EDX Defect etching methods Sawing slurry analysis Wafer breakage tests Analysis of saw damage 12

13 Lifetime measurement basics 13

14 Definition of the carrier recombination lifetime τ n U n n U recombination rate n excess charge carrier density SEMI, PV

15 Recombination mechanism in silicon 1 b 1 Auger 1 SRH,1 1 Rad... excess charge carrier lifetime τ [µs] 10 4 τ rad τ Auger τ SRH, Cr p 0 = 1x10 16 cm -3 [Fe i ] = 1x10 12 cm -3 [Cr i ] = 1x10 11 cm -3 τ SRH, Fe τ b Working area of Si-solar cells excess charge carrier density n [cm -3 ] Lifetime is a strong function of the excess carrier density. Lifetime should be measured where solar cells are operating. radiation: W. Gerlach, phys. stat. sol.(a) 13 (1972) 277 Auger: M. Kerr et al., J. Appl. Phys. 91 (2002) 2473 Fe: A. Istratov et al., Applied Physics A 69 (1999) 13 Cr: J. Schmidt et al. J. Appl. Phys. 102 (2007)

16 Recombination at the silicon surface as-cut surface: S eff = 10 4 cm s -1 τ b τ eff 100µs 2.19µs 200µs 2.21µs 1% of τ eff contains information about silicon quality surface passivation needed Simulation, thickness of wafer: w = 200 µm 16

17 How to measure carrier recombination lifetime τ? Measuring of excess carrier density Excess carrier density during measurement process is described by diffusion equation: n t G n n D 2 n see e. g. J. S. Blakemore, Semiconductor Statistics 17

18 How to measure carrier recombination lifetime τ? Relationship between excess charge carrier density within the diffusion equation and the actual measured signal has to be known for each type of measurement. measured signal calibration? excess charge carrier density diffusion equation? carrier recombination lifetime 18

19 MW-PCD 19

20 Microwave-detected Photoconductance Decay (MW-PCD) measurement setup Microwave generator Detector Computer Laser diode/ bias Microwave antenna Wafer 20

21 Linearity of microwave reflection signal P r t n av t e.g. deviation of about 10% from linearity for n = cm -3 p 0 = cm -3 K. Lauer et al., J. Appl. Phys 104 (2008)

22 Standard analysis of microwave reflection signal using SEMI PV exponential fit of microwave reflection signal reveals decay time Decay time is lifetime if conditions for linearity are given and lifetime does not depend on excess carrier density. 22

23 Standard analysis of microwave reflection signal MW-PCD measurement of SiNx-passivated mc-silicon wafer 100 MWPCD Signal U [mv] ? Lifetime depends on excess carrier density. Monoexponential fit is not appropriate! Result depends on fitting algorithm Time t [µs] 23

24 Standard analysis of microwave reflection signal Map of the decay time constant of a multicrystalline silicon wafer high resolution images are possible 24

25 RF-PCD 25

26 RF-PCD measurement setup Conductivity measured by radio frequency (RF) electro-magnetic fields (eddy-current) Illumination intensity determined by reference solar cell SEMI, PV

27 RF-PCD Calibration of RF sensor 5 voltage U [V] U = A + B σ s + C σ s sheet conductance σ s [ Ω -1 ] calibration by differently doped reference wafers Resistivity of reference wafers measured by 4-point-probe measurement signal linked to excess carrier density (mobility model has to be used) Sinton et al., Appl. Phys. Lett. 69 (1996)

28 RF-PCD Assumption: steady state (Generation = Recombination) RF-coil signal U [V] time t [ms] I (~ G) U (~ n) illumination intensity I [suns] lifetime τ b [µs] excess carrier density n [cm -3 ] Diffusion equation reduces to: b n n G n from intensity, using optical constant 28

29 RF-PCD RF-QSSPC measurement of SiNx-passivated mc-silicon wafer 100 carrier lifetime τ [µs] excess carrier density n [cm -3 ] Augerrecombination Trapping SRH-recombination 29

30 ROUND-ROBIN OF EXCESS CHARGE CARRIER LIFETIME AND DECAY TIME MEASUREMENTS ON SILICON BLOCKS AND ADJACENT SURFACE PASSIVATED WAFERS K. Lauer 1, A. Lawerenz 1, S. Walter 2, A. Sidelnicov 3, M. Herms 3, S. Diez 4, Y. Ludwig 4, M. Turek 5, S. Reißenweber 6 1 CiS Forschungsinstitut für Mikrosensorik und Photovoltaik GmbH, SolarZentrum Erfurt, Konrad-Zuse-Str. 14, Erfurt, Germany 2 Schott Solar Wafer GmbH, Werk Jena 1, Ilmstraße 8, Jena, Germany 3 Bosch Solar Energy AG, Solar Energy, Wilhelm-Wolff-Str. 23, Erfurt, Germany 4 Q-Cells SE, Sonnenallee 17-21, Bitterfeld-Wolfen, Germany 5 Fraunhofer CSP, Walter-Hülse-Str. 1, Halle, Germany 6 Institut für Experimentelle Physik, TU Bergakademie Freiberg, Leipziger Str. 23, Freiberg, Germany 30

31 Aims Investigate reproducibility and comparability of standard lifetime measurement methods Verify consistency of lifetime measurements on silicon blocks and wafers 31

32 Samples adjacent surface passivated wafers one mono (CZ, ρ=5.1 Ωcm) and one multicrystalline (ρ=2.3 Ωcm) silicon sample set ingot adjacent wafers are cut from block => silicon in wafer comparable to silicon in block 32

33 Preparation Si block lifetime [µs] time [h] saw damage etched off from wafers surface recombination at the wafer level minimized using silicon nitride layers (enables measurement of bulk lifetime) lifetime stabilized by light soaking of all samples 1sun) verified that no meta-stable defects exist (e.g. iron-boron-pairs) as-cut block surface 33

34 Measurement positions block wafer 10.4 cm R Beschriftung 15.6 cm one block and three passivated wafers for multi-crystalline and mono-crystalline silicon sample set measurement at 3 positions per sample average of 3 measurement points for the block and 9 for the wafers is reported 34

35 MW-PCD settings MW-PCD device: WT2000 from Semilab impact of surface recombination visible for block measurements block autosetting microwave frequency laser power laser pulse width bias light head height MW-PCD no 10.2 GHz photons/puls 200 ns no 2 mm passivated wafer 35

36 RF-QSSPC settings RF-PCD device: WCT100/120 and BCT100 from SintonInstruments optical constant wafer thickness analysis type excess charge carrier density RF-PCD 0.7/0.85 (wafer/ingot) 150/180µm (mono/multi) generalized/qss (wafer/ingot) cm -3 lifetime τ [µs] RF-PCD QSS multicrystalline silicon ingot monocrystalline silicon ingot excess charge carrier density n [cm -3 ] strong dependence of the lifetime on the excess carrier density visible 36

37 Round-Robin results lifetime / decay time τ [µs] ingot monocrystalline silicon multicrystalline silicon surface passivated wafers different colors represent different partners average of 3 measurement points for the block and 9 for the wafers is displayed small deviation between partners standard deviation for mono-si smaller than multi-si 10 MW-PCD RF-PCD QSS RF-PCD QSS estimated RF-PCD Gen measurement methods MW-PCD 37

38 Round-Robin results lifetime / decay time τ [µs] Ingot Ingot Wafer Ingot Wafer Ingot Wafer (estimated) RF-PCD quasi steady-state multicrystalline silicon monocrystalline silicon MW-PCD short pulse excitation measurement methods MW-PCD long pulse excitation mono ingot mono wafer multi ingot multi wafer life time [µs] RF-PCD dev [µs] dev [%] decay time [µs] M W-PCD dev [µs] dev [%] MW-PC decay time measured on block dominated by surface recombination excess carrier dependence of lifetime leads to deviation of decay time and lifetime on passivated wafer deviation in lifetime measurements of 8% for passivated mono-si wafers lifetime measured on the wafer lies within the error of the estimated lifetime measured on the ingot 38

39 Comparison to MWPCD with long pulse excitation multicrystalline silicon monocrystalline silicon 200 lifetime / decay time τ [µs] long pulse excitation reveals carrier profile ranging deeper in the silicon carriers diffuse during recording the microwave reflection decay from the bulk into the region sensed by the microwaves higher decay times compared to short pulse excitation are measured 10 Ingot Ingot Wafer Ingot Wafer Ingot Wafer (estimated) RF-PCD quasi steady-state MW-PCD short pulse excitation measurement methods MW-PCD long pulse excitation 39

40 Summary Good reproducibility for each measurement method (deviation ranges from 3% to 19%) Bad comparability between decay time (MWPCD) and lifetime (RF-PCD) (different sensing depths, unknown excess carrier density for MW-PCD, erroneous evaluation of MW-PCD signal) Bad consistency between block and passivated wafer for decay time (MWPCD, surface recombination) Good consistency between block and passivated wafer for estimated lifetime RF-PCD 40

41 Thank you for your kind attention! 41

42 Thanks! This work was supported by the German Ministry for Education and Research in the project xµ-material in the framework of the Excellence Cluster Solar Valley Central Germany. 42

43 Mikrowellen-detektiertes Photoleitfähigkeitsabklingen (MWPCD) Auswertung bei Verwendung eines Bias-Lichts konstante Hintergrundbeleuchtung während MWPCD Messung Laserintensität klein gegen Hintergrundintensität => Messung einer differentiellen Lebensdauer => Bestimmung der Lebensdauer durch Integration über verschiedene Hintergrundbeleuchtungsintensitäten 43

44 Mikrowellen-detektiertes Photoleitfähigkeitsabklingen (MWPCD) Auswertung bei Verwendung eines Bias-Lichts je stärker die Injektionsabhängigkeit der Lebensdauer desto größer ist die Abweichung der differentiellen Lebensdauer 44

45 How to measure carrier recombination lifetime τ? Diffusionsgleichung für Überschussladungsträger U Rekombinationsrate n allgemein als Überschussladungsträgerdichte bezeichnet n p Kein Ladungsträgertrapping! n t G n n D 2 n SEMI, Document Number: 4738A 45

46 verschiedene Messmodi transient n t G n n D 2 n stationär generalisiert Annahme: homogene Verteilung der Ladungsträger im Wafer! 46

47 Messeinstellungen der WT2000 für Wafer und Block kein Autosetting durchführen! wichtige Parameter sind rot eingekreist Skalierung des Messfensters der Transiente auf die jeweilige Probe anpassen. Messung des gesamten Blocks/Wafers mit 0.5 mm Raster kein Bias-Licht 2 mm Messkopfhöhe (Abstand Probe zu Messkopfverkleidung) => 3 mm Abstand Probe zu Sensor 47

48 Messeinstellungen WCT100/120 (QSSPC für Wafer) Waferdicke => Mono: 150µm, Multi: 180µm Messeinstellungen BCT210 (QSSPC für Blöcke) 48