Revealing fundamentals of charge extraction in photovoltaic devices through potentiostatic photoluminescence imaging

•A method for electrical-bias-dependent photoluminescence microscopy is introduced•The enables imaging of the local photocurrent in solar cells•Time-resolved charge-extraction demonstrated perovskite cells•Electrical coupling at interfaces and surface recombination revealed Spatial characterization methods are essential a deeper understanding physical phenomena quality control opto-electric devices such as cells. Here, to derive images by microcopy presented. As derived from basic principles, based on phenomenon quenching charge carriers externally extracted. It III-V cells that approach real-time assessment spatially resolved extraction. This allows us identify performance-loss mechanisms electrical interconnection or microscopic resolution. The density-voltage (J(V)) curve fundamental characteristic assess opto-electronic devices, particular However, it only yields information performance integrated over entire active device area. determine voltage-dependent principles. opportunities limitations studied investigation microscopically J(V) steady-state Jsc well transient effects. In addition, measurement contains extraction interfacial recombination. facilitates identification regions non-ideal linking these processing conditions. proposed technique highlights that, combined with potentiostatic measurements, luminescence can be powerful tool losses improvement current builds basis any further other devices. provides broad range charge-carrier generation, recombination, transport losses. Conventionally, measured illuminating full probing flowing between electrodes source meter. therefore A techniques have been established analysis photovoltaic devices1Schubert M.C. Mundt L.E. Walter D. Fell A. Glunz S.W. Spatially cells.Adv. Energy Mater. 2020; 10: 1904001https://doi.org/10.1002/aenm.201904001Crossref Scopus (14) Google Scholar,2Fertig F. Padilla M. Breitenstein O. Höffler H. Geisemeyer I. Schubert Rein S. Short-circuit density silicon cells.Energy Proc. 2015; 77: 43-56https://doi.org/10.1016/j.egypro.2015.07.008Crossref (5) Scholar light-beam-induced mapping (LBIC),3Mundt Kwapil W. Yakoob M.A. Herterich J.P. Kohlstädt Würfel U. Quantitative loss blade-coated cells.IEEE J. Photovolt. 2019; 9: 452-459https://doi.org/10.1109/JPHOTOV.2018.2888835Crossref (6) Scholar, 4Seager C.H. determination grain-boundary rates scanned spot excitation methods.J. Appl. Phys. 1982; 53: 5968-5971https://doi.org/10.1063/1.331389Crossref (63) 5Padilla Michl B. Thaidigsmann Warta cells.Sol. Sol. Cells. 2014; 120: 282-288https://doi.org/10.1016/j.solmat.2013.09.019Crossref (54) lock-in thermography (LIT),6Breitenstein Nondestructive current–voltage characteristics thermography.Sol. 2011; 95: 2933-2936https://doi.org/10.1016/j.solmat.2011.05.049Crossref (90) 7Breitenstein Fertig Bauer An empirical short circuit dark 143: 406-410https://doi.org/10.1016/j.solmat.2015.07.027Crossref (16) 8Fertig Greulich short-circuit via thermography.Appl. Lett. 104: 201111https://doi.org/10.1063/1.4876926Crossref (22) (PL) imaging,3Mundt Scholar,9Shen C. Kampwerth Green Thorsten Trupke T. Carstensen Schutt cell parameters comparison CELLO measurements.Sol. 2013; 109: 77-81https://doi.org/10.1016/j.solmat.2012.10.010Crossref (44) 10Trupke Pink E. Bardos R.A. Abbott M.D. series resistance obtained imaging.Appl. 2007; 90: 093506https://doi.org/10.1063/1.2709630Crossref (212) 11Kampwerth Weber J.W. Augarten Y. Advanced effective cells.Appl. 2008; 93: 202102https://doi.org/10.1063/1.2982588Crossref (121) 12Glatthaar Haunschild Zeidler R. Demant Preu Evaluating voltage cells.J. 2010; 108014501https://doi.org/10.1063/1.3443438Crossref (89) 13Glatthaar Kasemann Giesecke saturation cells.Phys. Status Solidi RRL. 4: 13-15https://doi.org/10.1002/pssr.200903290Crossref (100) 14Breitenstein Photoluminescence image evaluation implied distribution.Sol. 128: 296-299https://doi.org/10.1016/j.solmat.2014.05.040Crossref (19) 15Haunschild Glatthaar E.R. Fast using electroluminescence.Phys. 2009; 3: 227-229https://doi.org/10.1002/pssr.200903175Crossref (96) 16Hoffler PL distribution.IEEE 5: 613-618https://doi.org/10.1109/JPHOTOV.2014.2379097Crossref (10) 17Mastroianni Heinz F.D. Im J.-H. Veurman Grätzel Park N.-G. Hinsch Analysing effect crystal size structure highly efficient CH3NH3PbI3 photo- electroluminescence imaging.Nanoscale. 7: 19653-19662https://doi.org/10.1039/C5NR05308KCrossref PubMed 18Hameiri Z. Mahboubi Soufiani Juhl M.K. Jiang L. Huang Cheng Y.-B. cells.Prog. Res. 23: 1697-1705https://doi.org/10.1002/pip.2716Crossref (61) 19Okano Endo Wakamiya Yoshita Akiyama Kanemitsu Degradation mechanism diode spectroscopy.Appl. Express. 8: 102302https://doi.org/10.7567/APEX.8.102302Crossref (29) 20Soufiani A.M. Hameiri Meyer Lim Tayebjee M.J.Y. Yun J.S. Ho-Baillie Conibeer G.J. Spiccia Lessons learnt electro- delamination planar upon illumination.Adv. 2016; 1602111https://doi.org/10.1002/aenm.201602111Crossref (37) 21Stolterfoht Le Corre V.M. Feuerstein Caprioglio P. Koster L.J.A. Neher Voltage-dependent how correlates fill factor open-circuit cells.ACS 2887-2892https://doi.org/10.1021/acsenergylett.9b02262Crossref (46) 22Mundt Albrecht Mundus Saliba Correa-Baena Anaraki E.H. Korte Gratzel Hagfeldt et al.Nondestructive tandem multiwavelength mapping.IEEE 2017; 1081-1086https://doi.org/10.1109/JPHOTOV.2017.2688022Crossref (21) 23Walter Yiliang Wu Duong Peng Fong K.C. K. On use intensity quantified distribution resistance.Adv. 1701522https://doi.org/10.1002/aenm.201701522Crossref (24) 24Rietwyk K.J. Tan Surmiak Lu McMeekin D.P. Raga S.R. Duffy N. Bach Light modulated rapid cells.Nano Energy. 73: 104755https://doi.org/10.1016/j.nanoen.2020.104755Crossref (2) (EL) imaging.25Soufiani Kim Luminescence cells: note reporting results.Adv. 2018; 1702256https://doi.org/10.1002/aenm.201702256Crossref (13) discussed supplemental (section A), either probe portion require numerical fitting postulated model. Moreover, (PSCs), scanning has yield valuable below optical resolution limit.26Leblebici S.Y. Leppert Li Reyes-Lillo S.E. Wickenburg Wong Lee Melli Ziegler Angell D.K. al.Facet-dependent efficiency variations single grains hybrid halide perovskite.Nat. 1: 16093https://doi.org/10.1038/nenergy.2016.93Crossref (254) 27Weber S.A.L. Hermes I.M. Turren-Cruz S.-H. Gort Bergmann V.W. Gilson Graetzel Tress Berger How formation causes hysteresis Environ. Sci. 11: 2404-2413https://doi.org/10.1039/C8EE01447GCrossref 28Li Nasr Esfahani Wei Yao Zhao Chen Touching believing: interrogating nanoscale microscopy.npj Quant. 2: 56https://doi.org/10.1038/s41535-017-0061-4Crossref (32) 29Song Zhou Padture N.P. Huey B.D. Anomalous 3D photoconduction semiconductors tomographic atomic force microscopy.Nat. Commun. 3308https://doi.org/10.1038/s41467-020-17012-yCrossref (27) Thus, especially crucial emerging (PV) technologies like PSCs gain about macroscopic scale enhance power-conversion long-term stability Therefore, novel advanced necessary. physics optoelectronic divided into operation processes. For PV power photogeneration absorption photons, separation photogenerated electrons holes layers (CELs), outer terminals (cables). three processes described so-called reciprocity relations Donolato,30Donolato theorem collection.Ultramicroscopy. 1985; 46: 270-272https://doi.org/10.1063/1.95654Crossref (169) Rau,31Rau Reciprocity relation quantum electroluminescent emission Rev. 76: 085303-085308https://doi.org/10.1103/PhysRevB.76.085303Crossref (750) Green,32Wong From junction terminal: extended operation.Phys. 2012; 85: 235205https://doi.org/10.1103/PhysRevB.85.235205Crossref (38) respectively. To each processes, assigned, namely losses, radiative non-radiative mechanisms, this work, we introduce immediate carriers. makes observation difference two different bias voltages direct J(V). We propose refer “potentiostatic imaging” (PPI). outline theoretical principles PPI experimentally demonstrate validity means close-to-ideal GaInAsP cell. Next, show its potential characterize high (PCE >20%) terms their charge-extracting ability investigating both steady state curves. Based results, distinct signatures linked properties layer, CELs, respective interfaces. detailed-balance approach, considering state, contacts, J(V), expressed carriers, Jgen, (internal) layer interfaces, Jrec(V),J(V)=Jgen−Jrec(V).(Equation 1) assume considered (0 V ≤ Voc; [Voc]), internal generation Jgen not affected applied bias. without generality, photon flux illuminate sample constant time homogeneous. comprise components, Jn.r. Jrad, respectively: Jrec = + Jrad. They related proportionality k byJn.r.V=kV·JradV.(Equation 2) Equation generally valid if function bias, k(V). probed ideality close one displays negligible resistive (high parallel resistance, low resistance), then justified linear relationship Jn.r.(V) Jrad(V) and, hence, const., which will following. Empirical indications assumption already presented Stolterfoht coworkers.21Stolterfoht detailed discussion general expression (sections B, C, D). Now, (1) asJ(V)=Jgen−(1+k)·Jrad(V).(Equation 3) signal photodetector SPL(V) bySPL(V)=cJrad(V)e,(Equation 4) where e elementary c describes probability photons generated enter detector area translated signal. Using Equations (3) (4), now relate PL(V) open circuit, PL(Voc). By normalizing term, becomes independent setup-specific factors, elaborate calibration measures. find thatSPL(Voc)−SPL(V)SPL(Voc)=J(V)Jgen.(Equation 5a) shows used time-resolved device. Two applications interesting: first, recording images, derived. Second, various voltages, specific spots investigated. These approaches investigated Thereby, J(V)/Jgen results determined denoted |PL. samples display behavior, PSC below, assumptions made cannot strictly applied, ratio signals an approximation normalized (cf., information, sections B C). samples, suggest term “charge-extraction coefficient”, cex(V), withcex(V)=SPL(Voc)−SPL(V)SPL(Voc)≈J(V)Jgen.(Equation 5b) Many compound semiconductors, gallium-arsenide (GaAs), indium-phosphide (InP), combinations GaxIn1-xAsyP1-y, band gap. Hence, semiconductor crystals defect density, dominant mechanism. coefficient thin absorber order few micrometers. fact, along sufficiently mobilities lifetimes, promises PCEs material class applications.33Rau Blank Müller T.C.M. Kirchartz Efficiency materials unveiled analysis.Phys. 7044016https://doi.org/10.1103/PhysRevApplied.7.044016Crossref (154) Scholar,34Kirchartz Rau What good cell?.Adv. 1703385https://doi.org/10.1002/aenm.201703385Crossref grown epitaxially using, instance, metalorganic vapor-phase epitaxy (MOVPE). highest efficiencies single- multi-junction under AM1.5 g spectrum concentration were achieved reaching 47.1%.35Green Dunlop Hohl-Ebinger Kopidakis Hao X. Solar tables (version 57).Prog. 2021; 29: 3-15https://doi.org/10.1002/pip.3371Crossref (374) 36Kayes B.M. Nie Twist Spruytte S.G. Reinhardt Kizilyalli I.C. Higashi G.S. 27.6% conversion efficiency, new record single-junction 1 sun illumination.in: 37th IEEE Photovoltaic Specialists Conference. 2011: 4-8https://doi.org/10.1109/PVSC.2011.6185831Crossref (395) 37Geisz J.F. France R.M. Schulte K.L. Steiner Norman A.G. Guthrey H.L. Young M.R. Song Moriarty Six-junction III–V 47.1% 143 Suns concentration.Nat. 326-335https://doi.org/10.1038/s41560-020-0598-5Crossref (231) test (3), Ga0.91In0.09As0.83P0.17 almost ideal behavior one-diode model fit current-voltage characteristic, was Figure S5 section E). simultaneously recorded curve. remarkable overlap measurements above outlined considerations. applying baseline correction reflectance sample, (J(0) Jsc), Jsc/Jgen |PL assessed 0.87 ± 0.03 measurements. validate result, (IQE) external (EQE) wavelength 620 nm, setup 623 nm. IQE amounted 0.98 0.05 S6). discrepancy 11 %abs partially attributed errors occurred involvement multiple (details error estimation found I-3). Additionally, observed non-linear offset spatial inhomogeneities system induce additional uncertainties I-4). Due hardware constraints, no four-point configuration measurement, introducing resistances contact needles wiring. contacted micromanipulators, may introduced resistances. parasitic another reason lower value compared value. Overall, while there remain improve method, line predictions. Monocrystalline, wafer-based typically very homogeneous S7). contrast even significant heterogeneities.38Tennyson E.M. Doherty T.A.S. Stranks S.D. Heterogeneity length scales semiconductors.Nat. 573-587https://doi.org/10.1038/s41578-019-0125-0Crossref (135) carried out high-efficiency glass/fluorine-doped tin oxide (FTO)/compact titanium dioxide (TiO2)/mesoporous TiO2/perovskite/Spiro-OMeTAD/Au consisting FA0.92MA0.04Cs0.04PbI3 (PbI2)0.067 F, details). Stabilized J-V illumination yielded stabilized PCE 21.2%, 26.4 mA/cm2, Voc 1.03 S9). Despite performance, conditions (PL(Voc)) shown 2A inhomogeneous reveals morphological features assigned films formed spin coating stripes dots. PL(Voc), (i) (ii), exceeds surrounding more than five. 2B i.e., after stabilizing longer 60 s. seen cex(0) image, misleading directly conclude PL(Voc) performance. representation, comparatively reveal meaningful aforementioned patterns. Concerning distinguish features: feature cex(0), quenched when switching open- While indicator presence high-quality suggests poor connection CELs. similar recently also al., who employing (i.e., recombination) EL injection) indicative resistances.39Mahboubi Zhu Mussakhanuly Contactless illumination.Sol. 2100655https://doi.org/10.1002/solr.202100655Crossref contrast, (ii) due benign coupling. systematic LBIC, EL, PL, H), demonstrating adds insights complementary findings provided techniques. experimental strategies I, estimate Figures S18A–S18C. 3A PSC, displaying patterns above. grouped according pattern type image: dots (A) stripe (B) higher (C) 3B, magnified image). With retrieve cex(V) pixel microscope image. 3C corresponding curves A, C (cf. J, One see similarities lowest demonstrates At point, important implicitly assumes equal coupled sheet resistivity low. Finally, take unique account measurement.25Soufiani PSCs, changes illumination, atmospheric minutes hours.40Xu De Rosia Weeks Photoluminescence–voltage (PL–V) Chem. 121: 24389-24396https://doi.org/10.1021/acs.jpcc.7b06711Crossref (12) 41Chen Wen induced enhancement lead perovskite.Sol. 1600001https://doi.org/10.1002/solr.201600001Crossref (94) 42Howard J.M. Tennyson Barik Szostak Waks Toney M.F. Nogueira A.F. Neves B.R.A. Leite M.S. Humidity-induced variable Cs/Br perovskites.J. 3463-3469https://doi.org/10.1021/acs.jpclett.8b01357Crossref (35) non-reversible degradation place frames conditions.43Khenkin M.V. Visoly-Fisher Kolusheva Galagan Di Giacomo Vukovic Patil B.R. Sherafatipour G. Turkovic V. al.Dynamics photoinduced photovoltaics: reversible irreversible processes.ACS 799-806https://doi.org/10.1021/acsaem.7b00256Crossref (62) Scholar,44Gomez Sanchez Campoy-Quiles Abate Topological 45: 94-100https://doi.org/10.1016/j.nanoen.2017.12.040Crossref investigate spin-coated n-i-p studied, optimal glass/transparent conductive (TCO)/electron-extraction layer/perovskite/hole-extraction layer/Au intentionally altered, same 2. 4 respectively, reduced thickness (Figure 4A 4B), electron-extraction (EEL) 4C 4D), hole-extraction (HEL) 4E F). averaged 0.71, 0.24, 0.11, Note EEL-free displayed comparably resistances, whereas HEL-free showed L, S18). sketch stack 4G. interesting observations: 4B, wave recognized most probably caused spin-coating process. patterns, cex(0). Comparing furthermore some concealed, become visible cex(0)-representation. sense, (iii) (iv) bottom striking, indicates (iii), low, indicating severely degraded (or absent) spot. (iv), surroundings, present here. charges Looking 4D, observe absence pattern. process EEL. appear here: 2, (v) but Inversely, many small spots, represented circles (vi) prominent they can, however, distinguished regions. wave-pattern (Figures 4F). much (note color 10). Also, throughout There (vii) relatively (v), (viii) Further assessments metal back electrode I-V data M simple analyze representation PL(V 0), alongside, condition, unravel would possible LBIC alone H). able detect why Such solution-processed key challenge establishment CELs.45Wagner Mathiazhagan Mastroianni Distinguishing crystallization stages influence during real-time.Sci. Rep. 14899https://doi.org/10.1038/s41598-017-13855-6Crossref map 5A illustrates influenced 0) intensities. zero circuit. map, classify within set extreme cases.1In case, unity. minimized, implies maximized. Practically, situation reached coverage (corresponding maximal circuit), ideally extracted white ordinate graph 5B, schematic cross-section displayed. Features (vi), situation. occurs worst-case scenarios:2There functional case complete depicted 5B. identified zero, origin 5A. (iii).3A photoactive present, terminals. dramatic electrode, illustrated changed thus black bisector practice, likely those layers. (i), representative situation.4Finally, possibility selectivity. charge-selective front (case (4, 4′) 5B). Another e.g., selectivity EEL HEL. HEL 4D 4F represent photocurrent-voltage usually do contain improvements developed, Yet, none them so far offer combination all 0 VOC simultaneous fast acquisition microscopical