Hydrogen Redistribution and Surface Effects in Silicon Solar Cells - PowerPoint PPT Presentation

Faculty of Engineering School of Photovoltaic and Renewable Energy Engineering Hydrogen Redistribution and Surface Effects in Silicon Solar Cells Dr. Phillip Hamer, ACAP Postdoctoral Fellow 27 th March 2019

Outline • Shameless Self Indulgence Hydrogen in Silicon • Transport of Hydrogen in Silicon • A number of surface effects important for solar cell structures reported recently • Contact Resistance • Surface Degradation • Nature and location of surface defects

Hydrogen – an unrequited love • Necessary for highest efficiencies in nearly every commercial cell architecture • Surface Passivation • Mitigation of B-O defect • Passivation of other impurities and crystallographic defects • Can’t get rid of it • Enormous impact on device performance while remaining below detection limit • Interacts with vacancies, interstitials, dangling bonds, dopants, dislocations, grain boundaries and other impurities • Even the most basic properties are not well established • Has been identified as playing a role in LeTID

Simulation of Hydrogen Transport • Simulate general behaviour of hydrogen in solar cell structures during thermal processing • Immobile • Hydrogen-boron (HB) and hydrogen-phosphorous (HP) pairs [1] • Largely Immobile • Hydrogen dimers (H 2 A and H 2 C) [2] • Mobile • Interstitial hydrogen (H + ,H 0 ,H - ) • Concentrations dependent upon fermi/quasi-fermi levels [3,4] • Interaction with electric fields critical • Model will be updated throughout ACAP fellowship [1] Zundel, T. and Weber, J. (1989) Phys. Rev. B, 39(8), 13549, [2] Voronkov, V.V. and Falster, R. (2017) Phys. Stat. Sol. (B), 254(6), 1600779 [3] Herring, C. et al. (2001) Phys. Rev. B, 64(12), 125209, [4] Sun, C. et al. (2015) J. Appl. Phys. 117(4), 45702

GADEST 2017 300 o C Hamer, P. et al. (2018) J. Appl. Phys., 123(4), 043108

GADEST 2017 • Most hydrogen in the structure 300 bound to phosphorous o C • Majority in the bulk exists as dimers • However these forms are not responsible for hydrogen transport Hamer, P. et al. (2018) J. Appl. Phys., 123(4), 043108

GADEST 2017 300 o C Hamer, P. et al. (2018) J. Appl. Phys., 123(4), 043108

GADEST 2017 300 o C “Cross -over Point” Hamer, P et al. (2018) J. Appl. Phys., 123(4), 043108

GADEST 2017 300 o C “Cross -over Point” Hamer, P. et al. (2018) J. Appl. Phys., 123(4), 043108

GADEST 2017 • Hydrogen profiles are 300 o C H - H near steady state + • Balance of drift and diffusion • Cross-over point critical • Trapping complicates raising interstitial hydrogen concentrations Hamer, P. et al. (2018) J. Appl. Phys., 123(4), 043108

GADEST 2017 300 o C Hamer, P. et al. (2018) J. Appl. Phys., 123(4), 043108

GADEST 2017 300 500 o C 700 o C o C 12 Hamer, P. et al. (2018) J. Appl. Phys., 123(4), 043108

GADEST 2017 • 700 o C Hydrogen trapped as HP is reduced with temperature 300 o C o Greater dissociation o Lower H - concentrations • Inflexion point shifts towards surface • Much greater H transport Hamer, P et al. (2018) J. Appl. Phys., 123(4), 043108

ANU 2017 Surface Effects - Contact Resistance Fill factor (a) and pseudo fill factor (b) as a function of belt furnace annealing (BFA) set temperature before (black squares) and after (red circles) annealing for p-type mc-SI PERC cells [1] . • Extended thermal processes post-firing lead to drop in fill-factors • Investigation reveals this is due to an increase in front contact resistance [1] Chan, C. et al. (2017) Solar RRL, 1(11), 1700129

ANU 2017 In-Situ Monitoring of R S • Series resistance completely 4 terminal I-V overwhelms diode measurements @ 350 o C characteristics Total time: 4 hours • At lower temperature the time taken for R S to increase goes up • However maximum R S achievable also increases • Effect is unstable and reversible • Re-distribution of Hydrogen

ANU 2017 In-Situ Monitoring of R S • Series resistance completely 4 terminal I-V overwhelms diode measurements @ 350 o C characteristics Total time: 4 hours • At lower temperature the time taken for R S to increase goes up • However maximum R S achievable also increases • Effect is unstable and reversible • Re-distribution of Hydrogen

ANU 2017 350 o C Total H 25 20 0 V 2 ) 15  R S (  cm 10 Experimental 5 Simulation – Interstitial H only, neglecting dimers 0 0 2000 4000 6000 8000 10000 12000 14000 time (s) Hamer, P. et al. (2018) Sol. Energy Mat. Sol. Cells, 184, pp. 91-97

ANU 2017 350 o C 25 25 25 20 20 20 0 V 0 V 0 V +0.1 V +0.1 V +0.2 V 2 ) 2 ) 2 ) 15 15 15  R S (  cm  R S (  cm  R S (  cm 10 10 10 Experimental 5 5 5 0 0 0 0 0 0 2000 2000 2000 4000 4000 4000 6000 6000 6000 8000 8000 8000 10000 10000 10000 12000 12000 12000 14000 14000 14000 time (s) time (s) time (s) Hamer, P. et al. (2018) Sol. Energy Mat. Sol. Cells, 184, pp. 91-97

ANU 2017 350 o C 25 25 25 20 20 20 0 V 0 V 0 V +0.1 V +0.1 V +0.1 V +0.2 V +0.2 V +0.2 V 2 ) 2 ) 2 ) 15 15 15  R S (  cm  R S (  cm  R S (  cm -0.1 V -0.1 V -0.2 V 10 10 10 Experimental -0.2 V 5 5 5 +0.2V 0 0 0 0 0 0 2000 2000 2000 4000 4000 4000 6000 6000 6000 8000 8000 8000 10000 10000 10000 12000 12000 12000 14000 14000 14000 time (s) time (s) time (s) Hamer, P. et al. (2018) Sol. Energy Mat. Sol. Cells, 184, pp. 91-97

ANU 2017 After Process • Increase in R S observed in-situ corresponds to increase at room temperature • Reverse biased samples show negligible increase over original R S • The majority of the increase is unstable, and can be temporarily reversed by applying a large forward current at room temperature. FB: 1A forward current passed at room temperature for 120 s to reduce contact res. Hamer, P. et al. (2018) Sol. Energy Mat. Sol. Cells, 184, pp. 91-97

ANU 2017 Reversibility • Previous thermal treatment plays a role • Possible to (almost) completely reverse change in R S • Then stable at room temperature • Implies long range redistribution Final R S 1.06 Ω .cm 2 Final R S 0.93 Ω .cm 2 Final R S 0.75 Ω .cm 2 Hamer, P. et al. (2018) Sol. Energy Mat. Sol. Cells, 184, pp. 91-97

Kinetics of Hydrogen Redistribution • Transport of hydrogen dominated by interstitial form • Overall rate likely determined by release from bound forms • Expect hydrogen dimers in monocrystalline silicon • Multi? Hamer, P. et al. (2018) Proceedings of the 7 th WCPEC, pp. 1682-1686

Dependence upon bulk material and cell structure PERC Cells o C 450 100 o C 400 10 2 )  R S.TEMP (  .cm 1 o C 350 0.1 • Best fit to data with quadratic 0.01 Mono relation 0.2V forward bias Multi • Mono shows more rapid increase 1E-3 0 20 40 60 80 100 120 in series resistance, with reduced Time (min) temperature dependence In Situ measurements of change in series resistance for Mono and Multi PERC cells annealed under forward bias at temperatures between 350-450 o C Hamer, P. et al. (2018) Proceedings of the 7 th WCPEC, pp. 1682-1686

Dependence upon bulk material and cell structure 0.2 V Forward Bias -10 • No significant difference in 2 /s) activation energy observed 2.34 eV between PERC and Al-BSF ln[k] ( cm -15 • Mono Al-BSF approximately 2 3.25 eV orders of magnitude slower at all temps multi Al-BSF • Less Hydrogen -20 multi PERC • 2.09 eV Multi devices show a significantly mono Al-BSF higher activation energy mono PERC • Different bound form -25 0.0014 0.0015 0.0016 -1 ) 1/T (K Arrhenius plot of fitted quadratic rate constant for Mono and Multi PERC and Al-BSF cells at temperatures between 350 and 450 o C. Hamer, P. et al. (2018) Proceedings of the 7 th WCPEC, pp. 1682-1686

Reversibility 3 10 o C p-type mono PERC, 400 • Contact Resistance Increase has a reversible and 0.2 V Process 2 10 non-reversible part • 2 ) Changes persist to room temperature  R S (  cm 1 10 measurements • Still shows a strong dependence on material 0 10 • ΔR S a somewhat questionable measure 0.2 V Process -1 10 • In order to extract physically meaningful 0.2 V Forward Bias 0.4 V Reverse Bias information we require a more detailed model of -2 10 0 20 40 60 80 100 120 140 the contact Time (min) 2 10 o C p-type multi PERC, 400 1 10 0.2 V Process 2 )  R S (  cm 0 10 -1 10 0.2 V Process -2 10 0.2 V Forward Bias In Situ measurement of change in series resistance 0.4 V Reverse Bias for Mono and Multi PERC cells with fixed applied -3 10 bias (open symbols) and with switched bias (closed 0 20 40 60 80 100 120 140 symbols). Time (min) Hamer, P. et al. (2018) Proceedings of the 7 th WCPEC, pp. 1682-1686