Insights in Perovskite Solar Cell Fabrication: Unraveling the Hidden Challenges of Each Layer

IEEE P oo IEEE JOURNAL OF PHOTOVOLTAICS 1 Insigh s in Pe o ski e Sola Cell Fab ica ion: Un a eling he Hidden Challenges o Each Laye 1 2 Ve ena S ockhausen, Isabel Mesqui a, Lu´ ısa And ade, and Ad´ elio Mendes3 Abs ac —Pe o ski e sola cells (PSC) a e undoub edly he4 mos ac i e esea ch a ea in pho o ol aics a his momen .5 Ac ually, since 2009 his eme ging echnology passed om 3.8%6 o he p esen >22% o ene gy con e sion e iciency. Along wi h7 ha , a huge amoun o some imes con adic ing and incomple e8 in o ma ion abou how o p epa e and cha ac e ize PSC is9 p o ided, which makes i di icul o no ge los . This pape is10 mainly di ec ed owa d newcome s in his a ea, wi h he goal o11 gi e o ien a ion o PSC ab ica ion p o ocols ha a e quickly12 implemen able and ha lead o eliable and accep able e iciencies.13 The e o e, a s ep-by-s ep analysis o each laye is p o ided and,14 wi hin his scope, se e al ab ica ion echniques a e compa ed in15 e ms o e iciency op imiza ion. Fu he mo e, a new and e sa ile16 al e na i e o lase -assis ed sc ibing o subs a e pa e ning is17 p esen ed. Elec ochemical cha ac e iza ion o dummy cells as an18 easy and e sa ile ool o isola ed laye cha ac e iza ion is demon-19 s a ed o TiO2blocking laye s. A e op imiza ion o each laye ,20 PSC wi h an a e age e iciency o (14.8 ±1.0)% was ob ained.21 Index Te ms—Blocking laye , ab ica ion de ails, pe o mance,22 pe o ski e sola cells.23 I. INTRODUCTION24 GLOBAL ene gy consump ion is p ojec ed o aise by 48%25 om 2012 o 2040 [1], which makes he in ensi ica ion o 26 enewable ene gy implemen a ion una oidable. Among hem,27 sola ene gy p oduc ion has been he as es g owing sec o 28 wi h he bigges sha e in newly c ea ed jobs in he pas ew29 Manusc ip ecei ed Decembe 22, 2017; e ised Ap il 4, 2018; accep ed Ap il 6, 2018. This wo k was suppo ed in pa by he Eu opean Union’s Ho i- Q1 zon 2020 P og amme, h ough a FET Open esea ch and inno a ion ac ion unde G an 687008, in pa by he P ojec POCI-01-0145-FEDER-006939 (LEP- ABE - Labo a o y o P ocess Enginee ing, En i onmen , Bio echnology and Ene gy – UID/EQU/00511/2013), unded by he Eu opean Regional De elop- men Fund, h ough COMPETE2020 – P og ama Ope acional Compe i i idade e In e nacionalizac¸˜ ao (POCI) and by na ionals unds h ough FCT (Fundac¸˜ ao pa a a Ciˆ encia e a Tecnologia), and in pa by NORTE-01-0145-FEDER-000005 – LEPABE-2-ECO-INNOVATION, suppo ed by No h Po ugal Regional Ope a ional P og amme (No e 2020), unde he Po ugal 2020 Pa ne ship Ag eemen , h ough he Eu opean Regional De elopmen Fund. The wo k o V. S ockhausen was suppo ed by he Eu opean Commission h ough he Se - en h F amewo k P og am, he Speci ic P og am “Ideas” o he Eu opean Re- sea ch Council o esea ch and echnological de elopmen as pa o an Ad- anced G an unde G an 321315 (BI-DSC). The wo k o I. Mesqui a was suppo ed by he FCT o he Ph.D. ellow ( e .: PD/PB/105985/2014). The wo k o L. And ade was suppo ed by he FCT (IF/01331/2015). (Co espond- ing au ho : Ad´ elio Mendes.) The au ho s a e wi h he Faculdade de Engenha ia, Uni e sidade do Po o, Po o 4200-465, Po ugal (e-mail:,[email p o ec ed]; [email p o ec ed]; [email p o ec ed]; [email p o ec ed]). Colo e sions o one o mo e o he igu es in his pape a e a ailable online a h p://ieeexplo e.ieee.o g. Digi al Objec Iden i ie 10.1109/JPHOTOV.2018.2826055 yea s [2], [3]. Besides he ma u e silicon echnology, ad anced 30 coppe indium gallium selenide and CdTe sola cells la ely en e - 31 ing he ma ke , pe o ski e sola cells (PSC) ha e been ea ning 32 a lo o a en ion due o hei s iking pe o mance e olu ion 33 since 2012 [4]. Since hen, PSC e iciencies ha e been amp- 34 ing up quickly, eaching ce i ied eco d e iciencies o 22.7% 35 o labo a o y de ices [5]; mo e ecen ly, in Feb ua y 2018, 36 G ¨ a zel epo ed 23.3% a ABXPV con e ence, Rennes [6]. De- 37 spi e he as p og ess in ab ica ing PSC wi h high e iciency, 38 s abili y has been a limi ing ac o so a . Thus, a emp ing o 39 add ess e iciency and s abili y, a huge a ie y o o mula ions 40 and cell a chi ec u es has been published. I includes plana de- 41 ices using an in e ed p-i-n a chi ec u e and PSC employing 42 a mesopo ous s uc u e ha can ei he ac i ely pa icipa e in 43 he elec on ans e (ac i e mesopo ous laye ) o me ely se e 44 as sca old s uc u e (passi e mesopo ous laye ). Wi hin pe - 45 o ski es, chemical enginee ing has o igina ed a huge quan i y 46 o mixed s uc u es, employing mixed ca ions and anions. Many 47 labo a o ies ha e been deciding o di ec esea ch e o s owa d 48 his “shoo ing s a ,” bu no all o hem we e capable o ep o- 49 duce he ou s anding esul s published in he li e a u e. E en 50 wi hou ega ding long- e m s abili y, e iciencies o en emain 51 below expec a ions because usually, c ucial echnical de ails e- 52 main ba ely explained o e en unmen ioned in esea ch a icles. 53 Thus, li le ab ica ion e o s wi hin each laye o he PSC will 54 sum up and lead o an o e all e iciency d op. The e o e, me ely 55 conside ing e iciencies o en i e de ices makes ep oduc ion o 56 published esul s a ha d ask. 57 In his pape , a s ep-by-s ep analysis o he echnical p oblems 58 o each laye is p o ided and a possible impac o hei modi i- 59 ca ion on he cell pe o mance will be assessed. In he end, he 60 cha ac e iza ion o he en i e de ice is discussed. As eco d e i- 61 ciency PSC’s gene ally possess a cell a chi ec u e wi h an ac i e 62 mesopo ous laye [7], ocus will lie on his PSC s uc u e. Fo a 63 deepe discussion abou al e na i e cell a chi ec u es, in e es ed 64 eade s a e e e ed o he in o ma i e e iew o Salim e al.[8], 65 Mesqui a e al. [9] o he ecen book w i en by Pa k e al. [10]. 66 II. MATERIALS AND METHODS 67 Fig. 1 shows a schema ic ep esen a ion o he PSC. On op o 68 a anspa en conduc i e oxide (TCO) subs a e ha was sc ibed 69 in o de o impede sho ci cui ing (g ey line), a dense TiO2laye 70 is deposi ed, ollowed by a mesopo ous laye . The adjacen pe - 71 o ski e laye pa ially in il a es in o he mesopo ous s uc u e 72 and o ms a capping laye . I is ollowed by a laye o hole 73 2156-3381 © 2018 IEEE. Pe sonal use is pe mi ed, bu epublica ion/ edis ibu ion equi es IEEE pe mission. See h p://www.ieee.o g/publica ions s anda ds/publica ions/ igh s/index.h ml o mo e in o ma ion. IEEE P oo 2IEEE JOURNAL OF PHOTOVOLTAICS Fig. 1. Schema ic ep esen a ion o a mesoscopic PSC in c oss sec ion (le ) and op iew ( igh ). anspo ma e ial (HTM); inally, a nanome ic me allic laye 74 se es as a cu en collec o .75 In he ollowing, laye ab ica ion de ails and, when applica-76 ble, al e na i e ab ica ion me hods a e p esen ed. Fu he mo e,77 i will be discussed wha equipmen is equi ed o cell ab ica-78 ion and which equipmen acquisi ion can be pos poned, hanks79 o al e na i e ab ica ion p o ocols.80 A. Subs a e P epa a ion81 Fluo ide-doped in oxide (FTO) subs a es (2.2 mm hick-82 ness, TEC7, Sola onix) we e pa e ned ia Ve saLase (VLS83 2.30, Uni e sal Lase Sys ems, USA) o c ea e wo sepa a e84 cha ge collec ion a eas on he FTO subs a e—me hod (A). As85 an al e na i e o lase sc ibing, which equi es he a ailabili y86 o such an equipmen , an elec ochemical educ i e ea men 87 can be pe o med o emo e selec i ely he conduc i e laye —88 me hod (B). Keep in o conside a ion ha he chemicals used89 o ha pu pose a e highly co osi e, which equi es adequa e90 p o ec ion and ca e. In o de o do so, he subs a e a ea o TCO91 emo al was delimi ed by Kap on ape and exposed oa3M92 HCl solu ion. A cons an po en ial o −2.4 V was applied un il93 ca hodic cu en dec ease s a ed o la en. Meanwhile, he in94 oxide o he FTO u ned g ey and s a ed o peel o ; samples95 we e emo ed om he solu ion and insed wi h wa e . Wi h96 a co on swab dipped in a dilu ed ni ic acid solu ion (0.5 M),97 emaining in esidues we e cleaned o . Then, Kap on ape was98 emo ed and samples we e abundan ly insed wi h wa e .99 In a nex s ep, samples we e mechanically cleaned, using a100 oo hb ush and a 10 % Hellmanex III (Hellma GmbH, Ge many)101 solu ion. Subsequen ly, subs a es we e abundan ly insed wi h102 wa e and sonica ed in e hanolic KOH solu ion o 5 min. The103 subs a es we e again abundan ly insed wi h wa e and son-104 ica ed in wa e o 5 min, be o e being insed wi h ace one105 and d ied in ni ogen lux. P io o blocking laye deposi ion,106 subs a es we e addi ionally cleaned o 20 min by an ozone107 cleane (UVO-Cleane , Jeligh Company Inc., USA). Al e na-108 i ely o an ozone cleane , plasma ea men can be applied [11],109 among o he e icien me hods.110 B. Elec on Blocking (BL) and Mesopo ous Laye P epa a ion111 TiO2blocking laye was deposi ed by wo di e en me hods.112 Me hod (A) was done by spincoa ing o a comme cial solu ion113 (Ti-Nanoxide BL/SC, Sola onix, Swi ze land) (5000 /min, 30 s, 114 2000 ( /min)/s). Be o e ilm deposi ion, he a ea o pho oan- 115 ode con ac was p o ec ed by adhesi e s ip (Sco ch Magic 116 Tape, 3M) and he ilms we e subsequen ly calcined a 550 °C117 o 1 h, unde applica ion o a s epwise empe a u e inc ease 118 o 100 °C each 10 min. Me hod (B) employed sp ay py ol- 119 ysis o a p ecu so solu ion con aining 0.56 M ace ylace one 120 (Sigma-Ald ich, 99.6%) and 0.18 M i anium diisop opoxide 121 bis(ace ylace ona e) (Sigma-Ald ich, 75 w .% in isop opanol) 122 in 7 mL isop opanol (Sigma-Ald ich, anhyd ous, 99.5%) ha 123 was su icien o 64 samples. He e, subs a es we e p ehea ed 124 a 450 °C and he pho oanode a ea was p o ec ed wi h a glass 125 s ipe be o e applying he sp ay ia an a omize , using ei he ai 126 o oxygen as ca ie gas. A e wa d, samples we e le o 45 min 127 mo e a ha empe a u e. Fo applica ion o mesopo ous TiO2,128 a comme cial pas e (gene ally 30-NR-D, Dyesol, Aus alia, un- 129 less o he wise s a ed) was dilu ed in pu e e hanol (1:6 w/w) and 130 applied on he subs a es ia spincoa ing (5000 /min, 10 s, 2000 131 ( /min)/s). P io o deposi ion, pho oanode con ac had been 132 p o ec ed by adhesi e s ipes. Samples we e hen immedia ely 133 ans e ed on a hea pla e a 100 °C o p ed ying be o e being 134 calcined in a u nace a 500 °C o 30 min. Subsequen ly, sam- 135 ples we e ans e ed o oxygen- ee and d y condi ions (glo e 136 box) be o e allowing o cool below 100 °C. 137 C. Pe o ski e Ac i e Laye P epa a ion 138 The pe o ski e p ecu so solu ion was p epa ed acco ding o 139 he ollowing condi ions published by Saliba e al. [12]: 1.1 M 140 PbI2(Sigma-Ald ich, 99.999% ace me al basis), 0.2 M PbB 2141 (Sigma-Ald ich, 99.999% ace me al basis), 0.2 M me hylam- 142 monium b omide (Dyesol), and 1.0 M o mamidinium iodide 143 (Dyesol) we e dissol ed in 1 mL o a DMF/DMSO mix u e (8:2 144 / , bo h Sigma-Ald ich, 99.8 and ࣙ99.9%, espec i ely). F om Q2145 his solu ion, 0.95 mL we e added o 0.05 mL o a 1.5 M CsI 146 s ock solu ion in DMSO (Sigma-Ald ich, 99.999% ace me als 147 basis). This inal solu ion was deposi ed on he subs a es by 148 applying a wo-s ep spincoa ing p og am (s ep 1: 1000 /min, 149 10 s, 200 ( /min)/s, s ep 2: 6000 /min, 30 s, 2000 ( /min)/s). 150 A e 25 s, 100 µL chlo obenzene was pou ed on o he spinning 151 subs a e, a p ocedu e which is known as an isol en echnique. 152 Ca e ul adjus men o d ipping speed and ip- o-sample dis ance 153 had o be ained o ab ica e samples in a ep oducible manne . 154 IEEE P oo STOCKHAUSEN e al.: INSIGHTS IN PSC FABRICATION: UNRAVELING THE HIDDEN CHALLENGES OF EACH LAYER 3 Samples appea ed b own immedia ely a e spincoa ing and155 we e subsequen ly sin e ed a 100 °C o 40 min be o e be-156 ing allowed o cool down. A e each deposi ion, he in e io o 157 he spin coa e was cleaned wi h a clo h o emo e he condensed158 chemicals.159 D. Hole Conduc ing Laye and Cu en Collec o 160 Two di e en hole conduc o s we e es ed: me hod161 (A) spi o-OMeTAD solu ion con ained 75 mM spi o-162 OMeTAD (Chembo un, 99.7% sublimed g ade), 0.24 M163 4- e -bu ylpy idine (Sigma-Ald ich, 96 %), 41 mM li hium164 bis i luo ome hanesul onimida e (Li-TFSI, Ac os O ganics)165 ha was ob ained om a 1.8 M s ock solu ion in ace oni ile166 (Sigma-Ald ich, 99.999 % elec onic g ade), and 27 mM FK167 209 Co(III) TFSI sal (Dyesol) ha was ob ained om a 0.27168 M s ock solu ion in ace oni ile. The solu ion was deposi ed169 ia spincoa ing (4000 /min, 20 s, 2000 ( /min)/s). Me hod (B):170 P3HT solu ion was ab ica ed om 15 mg/mL P3HT (Chem-171 bo un China), 23 mM 4- e -bu ylpy idine, and 0.7 mM Li-172 TFSI. I was deposi ed by spincoa ing (3000 /min, 30 s, 2000173 ( /min)/s). A e wa d, pho oanode con ac s co e ed wi h pe -174 o ski e and hole conduc o we e mechanically cleaned wi h175 a scalpel and co on swabs dipped in ace oni ile. Finally, a176 60-nm- hick gold laye as cu en collec o was applied h ough177 a s ainless s eel mask by wo di e en me hods: 1) by he -178 mal e apo a ion on a Vapo S a ion 4 (Ox o d Vacuum Science,179 U.K.), applying a deposi ion a e o 0.01 nm/s o he i s 4 nm,180 ollowed by 0.1 nm/s o he emaining hickness; and 2) by181 spu e ing using a Leica EM ACE200 (Leica Mic osys ems,182 Ge many) and applying a cu en o 60 mA and a deposi ion183 du a ion o 360 s. A mask o adhesi e black ape wi h an ac i e184 a ea o 0.2 cm2was applied on he glass side o he cell p io o185 pho oelec ochemical cha ac e iza ion.186 E. Dummy Cell P epa a ion187 The p epa a ion was analogous o PSC, howe e , applying188 me ely blocking laye , hole anspo laye , and gold laye by189 he mal e apo a ion.190 F. Cha ac e iza ion191 Fo pho oelec ochemical cha ac e iza ion, a 150-W sola 192 simula o O iel class A sola simula o , (Newpo , USA) using193 a 1.5 ai mass il e (Newpo , USA) was employed. The e ec-194 i e i adia ion in ensi y was measu ed wi h a single c ys al Si195 pho odiode (Newpo , USA). I–V cu es we e eco ded wi h a196 po en ios a (Zennium, Zahne -Elek ik GmbH, Ge many) a a197 scan a e o 10 mV/s, sweeping om open-ci cui o sho -ci cui 198 po en ial (backwa d scan). Be o e each measu emen , he open-199 ci cui po en ial VOC was allowed o s abilize unde i adia ion,200 which gene ally ook less han a minu e. Ca e was aken ha 201 s a ing po en ials we e chosen o be no mo e han 20 mV su-202 pe io o VOC in o de o p o ec he cell [13]. A leas h ee203 cells o each ype we e es ed o a e aged e iciencies. SEM204 images we e eco ded wi h a Quan a 400 FEG (FEI, USA) a 205 he CEMUP ma e ials analysis cen e o he Uni e si y o Po o.206 III. RESULTS AND DISCUSSION 207 FTO on glass is usually employed as anspa en conduc i e 208 subs a e, due o i s s abili y owa d ele a ed empe a u es. Se - 209 e al shee esis ances a e a ailable on he ma ke and gene ally 210 FTO wi h a shee esis ance o 7–10 Ω/sq is chosen, as i is a 211 good comp omise in e ms o conduc i i y e sus anspa ency. 212 Tho ough subs a e cleaning is an essen ial s ep and o en unde - 213 a ed; howe e , i plays a pi o al ole as he pe o ski e sola cell 214 is cons i u ed by se e al laye s ha a e all wi hin he nanome e 215 scale and any con amina ion o he subs a e will hus lead o 216 ilm de ec s ha lowe o e all cell e iciency. The sc ibing o he 217 subs a e locally emo es he TCO laye and impedes he elec- 218 ic sho ci cui h ough he subs a e o he pho oanode and he 219 ca hode. I is o en ob ained by lase abla ion o he conduc i e 220 laye bu no e e y labo a o y possesses a sui able equipmen . 221 A low-cos al e na i e is chemical e ching o he conduc i e 222 laye [14], hough i leads o a he inhomogeneous FTO e- 223 mo al. A e y e sa ile and inno a i e, ye low-cos s a egy is 224 he elec ochemical educ i e ea men o he FTO, which leads 225 o a clean and comple e FTO emo al on he exposed a eas [15]. 226 The compac n- ype i anium dioxide ilm ac s as an elec on- 227 selec i e laye and hus p e en s he ecombina ion o exci ons 228 a he TCO su ace. I his laye is absen , no dense enough 229 o possesses pinholes, he ab ica ed cells will show dec eased 230 e iciencies due o ecombina ion e en s. A he same ime, i 231 has o be hin enough o p o ide e icien elec on anspo by 232 minimizing cha ge accumula ion and he e o e ecombina ion. 233 The so-called blocking laye can be ab ica ed by se e al ways, 234 including, bu no es ic ing o chemical ba h deposi ion [16], 235 spincoa ing [4], [17], sp ay py olysis [12], [18], [19], spu e ing 236 [20], [21], elec on-beam e apo a ion [22], and a omic laye de- 237 posi ion [23], [24]. We decided o compa e TiO2blocking laye s 238 ob ained by sp ay py olysis and spincoa ing o a comme cial so- 239 lu ion (Ti-Nanoxide BL/SC, Sola onix, Swi ze land). An easy 240 means o check i he elec ochemical beha io o he blocking 241 laye ollows a diode-like beha io is o ab ica e dummy cells. 242 Such cells a e composed o he compac TiO2laye on op o he 243 TCO subs a e, a hole- anspo laye like spi o-OMeTAD and 244 a gold con ac , hus simila o a pe o ski e cell, howe e wi h- 245 ou any pho oac i e laye . Cyclic ol amme y (CV) has been 246 pe o med on dummy cells wi h di e en BL and he esul s a e 247 shown in Fig. 2(a). In he case o a BL made by sp ay py olysis, 248 he CV shows ze o anodic cu en and a s eep inc ease o he 249 ca hodic cu en which sugges s a dense and pinhole- ee laye 250 wi h high elec onic conduc i i y. In case o he BL o med by 251 spin coa ing, he CV shows, in addi ion o he ca hodic cu en 252 inc ease a lowe po en ial, a sluggish ca hodic and anodic cu - 253 en e olu ion ac oss he en i e po en ial window, which is an 254 indica ion o pinholes. Fo compa ison, he CV o a dummy 255 cell wi hou any blocking laye demons a es a ypical ohmic 256 beha io , p o ing he absence o any blocking e ec a posi i e 257 po en ial. The PSC co esponding o he BL ab ica ion me hods 258 show I–V cu es ha unde line he ex emely impo an ole o 259 he blocking laye . The cell wi h he BL made by sp ay py oly- 260 sis shows bes e iciencies, whe eas ha made wi h spincoa ed 261 blocking laye pe o ms wo se. The cell wi hou any BL shows 262 IEEE P oo 4IEEE JOURNAL OF PHOTOVOLTAICS Fig. 2. (a) CV o dummy cells wi hou blocking laye (g ey diamonds), blocking laye made by spincoa ing ( ed ci cles) and by sp ay py olysis (black squa es). (b) I–V cu es o PSC wi h a blocking laye made by sp ay py olysis (black squa es), spincoa ing o a comme cial solu ion ( ed ci cles) and wi hou any blocking laye (g ey diamonds) a 0.95 sun. Fig. 3. (a) Scanning elec on mic oscopy images ( op- iew) o a ba e TCO subs a e. (b) TiO2compac laye deposi ed by sp ay py olysis. (c) Spincoa ing o a comme cial solu ion. Black ba s co espond o 2 µm. TABLE I EFFICIENCIES AND I–V CHARACTERISTICS OF PSC WITH COMPACT TIO2MADE BY SPRAY PYROLYSIS USING DIFFERENT CARRIER GASES he lowes e iciencies ha , howe e , a e no ze o. The eason263 is ha he pe o ski e laye i sel is an elec on anspo e [17]264 as well as i is capable o anspo holes [25], [26]. This en-265 de s he TCO-pe o ski e in e ace in o a nonselec i e con ac 266 ha p omo es ecombina ion and he e o e leads o dec eased267 e iciencies.268 Images o he di e en blocking laye s eco ded by scanning269 elec on mic oscopy show some undamen al di e ences, see270 Fig. 3. The laye deposi ed by sp ay py olysis is a he hin and271 homogeneous, whe eas he laye deposi ed by spin coa ing is272 hicke and shows c acks, see Fig. 3(c) (uppe igh co ne ). I 273 can be concluded ha PSC wi h a BL ab ica ed by sp ay py-274 olysis show supe io e iciencies and he e o e, his ab ica ion275 me hod migh be ecommended. The in luence o he ca ie gas276 on cell e iciencies was u he es ed bu i came ou ha pu e277 oxygen did no imp o e cell e iciencies, see Table I. Due o 278 lack o deposi ion con ol, BL hickness may a y be ween 30 279 and 80 nm, as occasional SEM c oss sec ions showed. Howe e , 280 no co ela ed impac on PSC e iciency could be s a ed. 281 Mesopo ous i ania laye has been employed in dye-sensi ized 282 sola cells (DSSC), wi h he unc ion o inc ease he ac i e su - 283 ace a ea and anspo elec ons unde ligh exci a ion [27]. As 284 he ini ial pe o ski e sola cells we e hough as a con inui y o 285 DSSC, a mesopo ous i ania ilm also was applied he e, e en i 286 he ex inc ion coe icien o pe o ski es such as (CH3NH3)PbI3287 is abou en imes highe han ha o N719 dye [28]. As 288 a consequence, he necessi y o ac i e su ace inc ease is 289 u ned obsole e. Thus, PSC wi hou mesopo ous laye , so-called 290 plana de ices, ha e been de eloped, hough hei e iciencies 291 we e lagging behind hose employing a mesopo ous laye o 292 IEEE P oo STOCKHAUSEN e al.: INSIGHTS IN PSC FABRICATION: UNRAVELING THE HIDDEN CHALLENGES OF EACH LAYER 5 Fig. 4. I–V cu e o cells wi h a mesopo ous TiO2 laye made om 30NR- D pas e (a e age pa icle size o 30 nm, black squa es) and om 18-NR-T pas e (a e age pa icle size o 20 nm, ed ci cles) a 0.94 sun. Bo h pas es we e pu chased om Dyesol L d. a long ime [7]. Only ecen ly, he e iciency gap has become293 a he small, which is due o imp o ed in e ace enginee ing [7],294 [18], [29]. Beside highe e iciencies, PSC wi h a mesopo ous295 laye show educed e iciency de ia ion be ween o wa d and296 backwa d scan, a phenomenon desc ibed as hys e esis [19], [30],297 [31]. When Snai h e al. demons a ed ha e en wi h meso-298 po ous laye s o alumina, a ma e ial ha canno pa icipa e in299 elec on ans e due o band ene gy misma ch, high e iciencies300 o 15.9% could be ob ained [32], i was deduced ha he meso-301 po ous laye mainly ul ills a s uc u al ole o c ys al g ow h302 and geome y de e mina ion, e en i e icien elec on injec ion303 om he pe o ski e in o he TiO2mesopo ous laye was epo ed304 [33]. Howe e , la es esul s demons a e ha ionic mig a ion a 305 he pe o ski e/TiO2in e ace, which is esponsible o cha ge306 accumula ion and he e o e ecombina ion e en s, is educed o 307 e en supp essed in he p esence o he TiO2mesopo ous laye 308 [34]. As hys e esis also depends on ionic mig a ion [35]–[38],309 i s educ ion in he p esence o mesopo ous TiO2is he conse-310 quence. Pu ing all oge he , bes esul s ha e been achie ed so311 a using a hin TiO2mesopo ous laye and a pe o ski e cap-312 ping laye ha p e en s ecombina ion be ween TiO2and he313 hole anspo laye [33], [34].314 The e exis se e al comme cial pas es wi h di e en sizes o 315 TiO2pa icles and he e o e, i was decided o compa e wo316 di e en pa icle sizes, namely one possessing 20 nm and one317 wi h 30 nm a e age pa icle diame e . The same pas e dilu ion318 a io in pu e e hanol (1:6) as well as he same deposi ion and319 sin e ing condi ions we e applied. Fig. 4 shows ha bo h meso-320 po ous laye s lead o e y compa able cell e iciencies ha a e321 wi hin he e o scale. This poin s owa d a highe ole ance and322 owa d a mesopo ous laye a chi ec u e, as long as he pa icle323 size emains simila .324 Wi hin pe o ski e ma e ials, he e exis s a huge a ie y o 325 ecipes and deposi ion echniques, which a e well summa ized326 in he book by Pa k e al. [10] and in he e iew by Song e al.327 [39]. I migh be no an easy ask o decide o he sui able328 pe o ski e ype and ab ica ion p ocess. The name pe o ski e329 e e s o a c ys alline s uc u e o he ype ABX3, A and B be-330 ing ca ions and X an anion. The pe o ski e class sui able o 331 sola cells is an o ganic lead halide, wi h A being gene ally an 332 o ganic ca ion, B being he lead ion, and X being a halogen, 333 usually b omine, iodide, chlo ine, and mix u es he eo . Lead 334 subs i u ion by in and ge manium analogs leads o pe o ski es 335 wi h se e e s abili y p oblems [32], [40], [41] and he e o e will 336 no be add essed he e. Ou ocus was o de e mine a pe o ski e 337 o mula ion easy o implemen and ha esul s in ep oducible 338 pe o ski e laye s wi h enhanced s abili y. Many esul s ha e 339 been published wi h monoca ionic pe o ski es, howe e , wi h 340 some inhe en limi a ions ha a e b ie ly exposed he e: MAPbI3341 has been in ensi ely s udied [42] bu has some d awbacks such 342 as weak s abili y owa d mois u e [43], [44] and empe a u e 343 [45]. Fo mamidinium (FA) was p oposed as al e na i e ca ion; 344 howe e , i s pe o ski e analog FAPbI3c ys allizes in he pho- 345 oinac i e phase below 60 °C [46], such as he ino ganic ca ion 346 analog CsPbI3[47]. Whe eas se e al g oups obse ed imp o ed 347 s abili y o he pho oac i e phase upon using bina y mixed ca ion 348 pe o ski es [48]–[52], Saliba e al. decided o combine he h ee 349 ca ions in a pe o ski e and achie ed high e iciencies (>20%) 350 on a e y ep oducible basis [12]. 351 Se e al me hods exis o solu ion-p ocessed ilm ab ica ion, 352 he mos common being simple sp eading o he pe o ski e p e- 353 cu so solu ion on he subs a e, also known as one-s ep depo- 354 si ion. Howe e , ilms wi h poo su ace con ol and he e o e 355 huge e iciency a ia ions gene ally eme ge [42]. A mo e so- 356 phis ica ed app oach is he sequen ial s ep deposi ion, whe e 357 he me al halide is i s deposi ed and annealed be o e being 358 b ough in con ac wi h he ammonium sal as apo o in solu- 359 ion [42], [53]. Ne e heless, se e al d awbacks o his deposi- 360 ion me hod we e expe ienced in ou g oup, such as incomple e 361 con e sion o he me al halide o pa ial dissolu ion o he pe - 362 o ski e du ing he subsequen washing s ep. Fu he mo e, i is 363 mo e ime-consuming as i equi es wo sin e ing s eps. The 364 an isol en echnique was in oduced in 2014 by he g oup o 365 Seok [19] and since hen, i has been he me hod o choice o 366 subsequen ly published eco d e iciencies [7]. I is qui e simple 367 o implemen and equi es only one p ecu so solu ion, whe eas 368 he c ys alliza ion o he pe o ski e is ini ia ed by adding a so- 369 called an isol en . This an isol en is chosen no o dissol e he 370 pe o ski e on one side and o displace he sol en o he la e on 371 he o he side. The main d awback o his deposi ion me hod is 372 i s a isanal aspec , equi ing a ce ain deg ee o aining be o e 373 eaching enhanced ep oducibili y. Howe e , smoo h pe o ski e 374 ilms wi h homogeneous composi ion and la ge g ain bounda ies 375 a e ob ained a e a sho aining ime. Fig. 5(a) shows a c oss 376 sec ion o a PSC wi h he monoli hic pe o ski e capping laye 377 on op o he mesopo ous TiO2laye wi h g ains g owing om 378 he bo om o he op and which a e hough o enhance cha ge 379 anspo , acco ding o Saliba e al. [12]. The op iew o he 380 pe o ski e laye [see Fig. 5(b)]shows g ains possessing diame- 381 e s be ween 200 and 500 nm, which is in good ag eemen wi h 382 he o iginal epo [12]. 383 One o he bigges de imen al ac o s o pe o ski e ab- 384 ica ion and s abili y is a mosphe ic humidi y, oge he wi h 385 oxygen [54]. PSC ha a e mean o exhibi p olonged s abili y 386 equi e ab ica ion and s o age in ine a mosphe e o de ice 387 encapsula ion a e ab ica ion [55], [56]. The e o e, PSC a e 388 IEEE P oo 6IEEE JOURNAL OF PHOTOVOLTAICS TABLE II BEST AND AVERAGE EFFICIENCIES FOR PSC, WHILE THE PEROVSKITE LAYER WAS FABRICATED WITHIN DIFFERENT CONDITIONS Fig. 5. SEM images. (a) C oss sec ion o he en i e PSC de ice showing he compac TiO2laye (ca. 80 nm), ollowed by he mesopo ous laye o 150– 200 nm. The adjacen pe o ski e laye pa ially in il a es in o he mesopo ous s uc u e and i s capping laye has a hickness o 200–300 nm. The spi o- MeOTAD hole anspo laye (da k g ey) has a hickness o 130–190 nm, ollowed by a 60-nm- hick gold laye as cu en collec o (ligh g ey). (b) Top iew o he pe o ski e laye , showing g ains wi h ca. 200–500 nm diame e . The issu es e ol ed du ing image cap u e and he e o e a e belie ed o be due o imaging. Ba s co espond o 1 µm. gene ally ab ica ed in glo e boxes wi h d y and oxygen- ee389 a mosphe e. And no only ha , also he subs a es should be390 absolu ely mois u e- ee. A e sin e ing he mesopo ous laye ,391 subs a es should hus be handled only in d y a mosphe e o 392 ans e ed o d y a mosphe e such as a glo e box be o e cooling393 below 150 °C. Wi hin he deg ada ion mechanism o pe o ski eQ3 394 s uc u es, oxygen only in e e es subsequen ly o hyd a a ion395 [56] and can he e o e be conside ed less c i ical i mois u e is396 absen o e y low. Labo a o ies ha a e newcome s in his a ea397 o esea ch migh no possess a glo e box in as uc u e. Low398 a mosphe ic humidi y le els a e assumed o be less c i ical o399 pe o ski e ab ica ion bu hese condi ions depend s ongly on400 he geog aphic localiza ion, he season, and some mo e inhe -401 en ac o s. To demons a e he e ec o mois u e and oxygen402 a mosphe e on pe o ski e o ma ion, a compa a i e es o PSC403 de ices ha we e ab ica ed inside and ou side he glo e box404 ( ela i e humidi y ou side he glo e box: 58%) was pe o med.405 The esul s a e displayed in Fig. 6 and Table II and i can be406 obse ed ha in case o pe o ski e being ab ica ed in ambi-407 en a mosphe e, he ac i e laye showed a ligh e colo and he408 co esponding PSC showed a bo h lowe VOC and JSC, whe eas409 he ill ac o emained a he unin luenced. As a s a egy o410 minimize wa e up ake by he subs a e, samples we e hea ed411 o 100 °C immedia ely be o e he pe o ski e p ecu so solu-412 ion was deposi ed. Co esponding PSC showed an imp o ed413 VOC and JSC, howe e he ill ac o dec eased. This is likely414 due o inhomogeneous c ys al g ow h, induced by he ele a ed415 empe a u e o he subs a e. Fo he bes cells ob ained, a li -416 Fig. 6. I–V cu es o PSC ab ica ed in a glo e box wi h 0% ela i e humidi y a 25 °C (black squa es), a ambien humidi y (58% ela i e humidi y) a 25 °C (g ey diamonds), and a ambien humidi y wi h subs a e p ehea ing a 100 °C ( ed ci cles) wi h an inciden ligh in ensi y o 0.94 sun. Image: Sample wi h pe o ski e p oduced inside (le ) and ou side he glo e box ( igh ) a 25 °C. le imp o emen can be s a ed when ho subs a es we e used, 417 hough a e age e iciencies came ou o be e y simila o hose 418 wi hou hea ea men . 419 This s udy shows ha i is highly ecommended o wo k 420 wi h a glo e box, p o iding e y low humidi y (<0.002% el. 421 humidi y) and oxygen le els. Ano he possible s a egy migh 422 be he use o a pe o ski e o mula ion ha is op imized owa d 423 enhanced esis ance a ele a ed humidi y le els [57]. 424 A op he pho oac i e laye , he hole conduc ing laye selec- 425 i ely anspo s he holes o he cu en collec o and he e o e 426 ul ills he complemen a y ole o he TiO2laye . I has o be 427 pinhole- ee o inhibi con ac o he cu en collec o wi h he 428 pe o ski e laye , o he same easons ha we e al eady s essed 429 ou conce ning he elec on conduc ing laye . Gene ally, a o - 430 mula ion using spi o-OMeTAD is used ha con ains, among o h- 431 e s, he ionic liquid LiTFSI o inc ease hole conduc i i y. How- 432 e e , bo h LiTFSI and spi o-OMeTAD ha e hyd ophilic p ope - 433 ies and p omo e humidi y inges ion, leading o poo humidi y 434 s abili y o he en i e de ice. A empe a u es abo e 55 °C, he 435 molecula hole anspo e c ys allizes, which se e ely a ec s 436 cell e iciencies. Two di e en hole conduc o laye s we e com- 437 pa ed owa d hei s abili y, namely spi o-oMeTAD as molecula 438 HTL and poly(3-hexyl hiophene) (P3HT) as polyme ic HTL. 439 IEEE P oo STOCKHAUSEN e al.: INSIGHTS IN PSC FABRICATION: UNRAVELING THE HIDDEN CHALLENGES OF EACH LAYER 7 Fig. 7. (a) I–V cu es o PSC employing spi o-OMeTAD (black squa es) and P3HT ( ed ci cles) as HTL a 0.95 sun. (b) No malized e iciency s abili y o PSC using wo PSC wi h spi o-OMeTAD (black squa es and g ey diamonds) and wo wi h P3HT ( ed ci cles and blue iangles). The ha ched g een a ea indica es a de ia ion less han 5% om he ini ial e iciency. Fig. 8. I–V cu es o PSC possessing a gold cu en collec o made by he mal e apo a ion (black line) and by spu e ing ( ed line) a 0.93 sun. I u ns ou ha spi o-OMeTAD leads o be e e iciencies,440 see Fig. 7(a), showing be e cu en densi y, open-ci cui po-441 en ial, and ill ac o al oge he . A po en ial ad an age could442 lie in an enhanced s abili y despi e lowe e iciency when P3HT443 is employed, bu he expe imen al da a could no con i m his444 assump ion, see Fig. 7(b), wi hin he limi ed ime ame.445 Gene ally, cu en collec o s a e made o gold despi e i s446 highe cos , as al e na i e me als such as Ag and Al ha e been447 demons a ing weak s abili ies [58]–[60]. Among gold depo-448 si ion me hods, one o he mos common ones a e spu e ing449 and he mal e apo a ion. Howe e , almos all published wo ks450 use he mal e apo a ion. We decided he e o e o compa e gold451 ilms wi h simila hickness ha we e ab ica ed by hese wo452 echniques. Indeed, he mal e apo a ion leads o a be e o e all453 cell pe o mance, see Fig. 8.454 The eason o he wo se pe o mance o PSC wi h spu e ed455 cu en collec o was e idenced by doing a sco ch es . While456 in case o he mal e apo a ion, he gold laye could be easily 457 s ipped o , in case o spu e ing deposi ion, he gold emained 458 s uck in o HTM. E en a e dissol ing he HTM laye , gold 459 aces we e s ill de ec ed wi h he naked eye inside he pe - 460 o ski e laye . This means ha du ing he gold laye deposi ion, 461 su ace bomba dmen p o okes pene a ion o gold deep in o 462 he de ice s uc u e, c ea ing ecombina ion cen e s. Thus, a 463 he mal e apo a o is needed o e icien pe o ski e sola cell 464 ab ica ion, e en i his s ep conside ably inc eases he ene gy 465 payback ime o PSC [61]. 466 Following all he laye ab ica ion s eps men ioned be o e, 467 i was possible o ab ica e PSC wi h an a e age e iciency o 468 (14.8 ±1.0)% o a se o 49 cells, see Fig. 9 le . 469 In labo a o y condi ions, cell e iciencies a e gene ally mea- 470 su ed o cell ac i e a eas in e io o 1 cm2. The cell a ea de- 471 limi ed by he deposi ion o he cu en collec o should be only 472 sligh ly supe io o he ac i e cell a ea (delimi ed by a mask) o 473 a oid ecombina ion e en s. Ins ead o using c ocodile clamps 474 a bi a ily connec ed o he cell, a sui able sample holde is 475 p e e able, see Fig. 9 igh , o maximal ep oducibili y. Among 476 all ac o s ha desc ibe he cell’s pe o mance, he maximum 477 powe poin (MPP) is he mos aluable in o ma ion in e ms o 478 applicabili y in sola de ices as i desc ibes bes he ope a ing 479 pa ame e s o he cell [7], [13], [62]. The MPP is ob ained ia 480 ma hema ic ex ac ion om I–Vcu es and su p isingly has no 481 ye gained big a en ion in published scien i ic wo ks. I–Vcu es 482 a e gene ally ob ained by dynamic scanning o ex e nal loads 483 hough ca e mus be aken ha he scan a e does no o e pass 484 he dynamic elec ochemical e en s inside he pe o ski e cell. 485 An example is gi en in Fig. 10, whe e a PSC was measu ed 486 a se e al scan a es. I me ely he I–V cu e is conside ed o 487 e iciency de e mina ion, bes esul s a e ob ained wi h a scan 488 a e o 1 V/s. Bu i CV o he same cell a e eco ded a 10 mV/s 489 and a 1 V/s, a s iking di e ence is obse ed conce ning he 490 hys e esis, see Fig. 11. Whe eas hys e esis is a he low in he 491 o me case, i conside ably inc eases in he la e case. This 492 IEEE P oo 8IEEE JOURNAL OF PHOTOVOLTAICS Fig. 9. E iciency dis ibu ion o all cells ab ica ed in s anda d condi ions (le ) and cell a chi ec u e wi h es ing de ice o ep oducible es ing condi ions ( igh ). Fig. 10. I–V cu es and e iciencies o a PSC eco ded a di e en scan a es a 0.98 sun: 10 mV/s (black squa es), 100 mV/s ( ed ci cles), 1 V/s (g ey diamonds). Fig. 11. Hys e esis o a PSC eco ded a a scan a e o 10 mV/s (black squa es) and 1 V/s ( ed ci cles) a 0.98 sun. means ha i he I–V cu e is eco ded in backwa d scan in such 493 condi ions, he ob ained esul s do no e lec he cell’s eal be- 494 ha io and he e o e lead o o e es ima ion o cell pe o mance 495 [7], [13]. Lacking so a easily implemen able measu emen 496 p o ocols o MPP acking, condi ions o I–V cu e eco ding 497 should be ca e ully chosen, wi h he goal o no o e es ima e 498 eal cell cha ac e is ics. 499 IV. CONCLUSION 500 PSC ep esen a e y a ac i e pho o ol aic echnology, as 501 innume ous publica ions ha e demons a ed, howe e ini ia ing 502 in his a ea may be a ha d ask. PSC a e made o se e al hin 503 laye s and no only each laye , bu also each in e ace plays 504 an impo an ole o he manu ac u ing o e icien de ices. 505 This pape is mainly di ec ed owa d esea ch g oups and sci- 506 en is s ha a e beginne s in his ac i e ield o esea ch and as 507 such, i was in ended o poin ou ab ica ion de ails ha emain 508 ba ely discussed in mos publica ions, bu ha a e signi ican 509 o he p epa a ion o e icien cells. I was e idenced ha be- 510 sides he usual equipmen o hin- ilm p epa a ion (ho pla e, 511 spin coa e , p og ammable u nace, e c.) and pho oelec ochem- 512 ical cha ac e iza ion (po en ios a o a iable ex e nal load, sola 513 simula o ), a glo e box and a he mal e apo a o o he depo- 514 si ion o he gold cu en collec o a e s ongly ad ised. The 515 goal o his pape is o analyze and op imize each laye and, as 516 a consequence, demons a e hei in luence on he en i e PSC 517 de ice. 518 I was e idenced by expe imen s ha o PSC pe o ming 519 bes , blocking laye has o be ab ica ed ia sp ay py olysis, 520 whe eas compa able esul s we e ob ained when pu e oxygen o 521 ai was used as ca ie gas. Fo he mesopo ous laye , no di e - 522 ence could be s a ed o bo h pa icle sizes used (20 and 30 nm); 523 howe e , i is likely ha a bigge di e ence in size may be o 524 ma e . Ou expe ience showed ha bes esul s o pe o ski e 525 ilms we e ob ained by applying he an isol en echnique wi h 526 a iple ca ion o mula ion. I was demons a ed ha conce ning 527 he adjacen hole anspo laye , spi o-OMeTAD was esul ing 528 IEEE P oo STOCKHAUSEN e al.: INSIGHTS IN PSC FABRICATION: UNRAVELING THE HIDDEN CHALLENGES OF EACH LAYER 9 in cells wi h supe io e iciency compa ed wi h he polyme ic529 analog P3HT. Finally, wo echniques o he ab ica ion o gold530 cu en collec o we e p esen ed and i was demons a ed ha 531 he mal e apo a ion leads o be e PSC han spu e ing, as in532 case o he la e he ene ge ic su ace bomba dmen p o oked533 gold pa icle pene a ion un il he pe o ski e ac i e laye , hus534 c ea ing ecombina ion cen e s. Inco po a ing all he discussed535 op imiza ions, PSC wi h an a e age e iciency o (14.8 ±1.0)%536 we e ab ica ed. Fu he mo e, an inno a i e, e sa ile, and quick537 me hod o elec ochemical subs a e e ching was applied in538 PSC, making lase -assis ed sc ibing and he e o e he neces-539 si y o such an equipmen obsole e. Dummy cells as selec i e540 elec ochemical cha ac e iza ion me hod o single laye s we e541 in oduced and yielded e sa ile esul s o he quali a i e com-542 pa ison o TiO2blocking laye s. I is belie ed ha his pape 543 con ibu es o a as e implemen a ion o PSC ab ica ion in544 esea ch g oups wi h ew expe ience in his a ea, hanks o a545 deepe unde s anding o ab ica ion de ails and use ul analysis546 ools ha a e easily a ailable.547 ACKNOWLEDGMENT548 The au ho s would like o hank M. G ¨ a zel o hos ing549 I. Mesqui a a EPFL o deepen he knowledge abou PSC ab-550 ica ion and O. Bellon om G ea cellsola L d. o ui ul551 discussions.552 REFERENCES553 [1] U.S.E.I. Adminis a ion, Washing on, DC, USA, Annual Ene gy Ou look,554 2017.555 [2] I.I.R.E. Agency, Abu Dhabi, UAE, RE hinking Ene gy 2017: Accele a ing556 he global ene gy ans o ma ion, 2017.557 [3] I.I.R.E. Agency, Abu Dhabi, UAE, Renewable Ene gy and Jobs Annual558 Re iew, 2016.559 [4] H.-S. Kim e al., “Lead iodide pe o ski e sensi ized all-solid-s a e submi-560 c on hin ilm mesoscopic sola cell wi h e iciency exceeding 9%,” Sci.561 Rep., ol. 2, 2012, A . no. 591.562 [5] NREL e iciency cha . [Online.]A ailable: h ps://www.n el.go /p /563 asse s/images/e iciency-cha .png. Accessed on: Ma . 26, 2018. Q4 564 [6] [Online.]A ailable: h p://www.g ea cellsola .com/wp-con en /uploads/565 2018/03/Au o a-Newsle e -Ma ch-2018-Final-ENG-RC.pd 566 [7] J.-P. Co ea-Baena e al., “The apid e olu ion o highly e icien pe -567 o ski e sola cells,” Ene gy En i on. Sci., ol. 10, pp. 710–727, 2017.568 [8] T. Salim, S. Sun, Y. Abe, A. K ishna, A. C. G imsdale, and Y. M. Lam,569 “Pe o ski e-based sola cells: Impac o mo phology and de ice a chi ec-570 u e on de ice pe o mance,” J. Ma e . Chem. A, ol. 3 pp. 8943–8969,571 2015.572 [9] I. Mesqui a, L. And ade, and A. Mendes, “Pe o ski e sola cells: Ma e ials,573 con igu a ions and s abili y,” Renewable Sus ain. Ene gy Re ., ol. 82,574 pp. 2471–2489, 2018.575 [10] N. G. Pa k, M. G ¨ a zel, and T. Miyasaka, O ganic-Ino ganic Halide Pe -576 o ski e Pho o ol aics: F om Fundamen als o De ice A chi ec u es.New577 Yo k, NY, USA: Sp inge , 2016.578 [11] C. C. Wu, C. I. Wu, J. C. S u m, and A. Kahn, “Su ace modi ica ion o 579 indium in oxide by plasma ea men : An e ec i e me hod o imp o e he580 e iciency, b igh ness, and eliabili y o o ganic ligh emi ing de ices,”581 Appl. Phys. Le ., ol. 70, pp. 1348–1350, 1997.582 [12] M. Saliba e al., “Cesium-con aining iple ca ion pe o ski e sola cells:583 Imp o ed s abili y, ep oducibili y and high e iciency,” Ene gy En i on.584 Sci., ol. 9, pp. 1989–1997, 2016.585 [13] J. A. Ch is ians, J. S. Manse , and P. V. Kama , “Bes p ac ices in pe o ski e586 sola cell e iciency measu emen s. A oiding he e o o making bad cells587 look good,” J. Phys. Chem. Le ., ol. 6, , pp. 852–857, 2015.588 [14] H.-R. Xia, J. Li, W.-T. Sun, and L.-M. Peng, “O ganohalide lead pe -589 o ski e based pho ode ec o s wi h much enhanced pe o mance,” Chem.590 Commun., ol. 50, pp. 13695–13697, 2014.591 [15] S. P. Koi y e al., “An elec ochemical me hod o as and con olled e ch- 592 ing o luo ine-doped in oxide coa ed glass subs a es,” J. Elec ochem. 593 Soc., ol. 164, pp. E1–E4, 2017. 594 [16] G. Yin e al., “Enhancing e iciency and s abili y o pe o ski e sola 595 cells h ough Nb-doping o TiO2 a low empe a u e,” ACS Appl. Ma e . 596 In e aces, ol. 9, pp. 10752–10758, 2017. 597 [17] M. Liu, M. B. Johns on, and H. J. Snai h, “E icien plana he e ojunc ion 598 pe o ski e sola cells by apou deposi ion,” Na u e, 501, pp. 395–398, 599 2013. 600 [18] D. Bi e al., “Polyme - empla ed nuclea ion and c ys al g ow h o pe - 601 o ski e ilms o sola cells wi h e iciency g ea e han 21%,” Na u e 602 Ene gy, ol. 1, 2016, A . no. 16142. 603 [19] N. J. Jeon, J. H. Noh, Y. C. Kim, W. S. Yang, S. Ryu, and S. I. Seok, 604 “Sol en enginee ing o high-pe o mance ino ganic–o ganic hyb id pe - 605 o ski e sola cells,” Na u e Ma e , ol. 13, pp. 897–903, 2014. 606 [20] D. Yang, Z. Yang, W. Qin, Y. Zhang, S. Liu, and C. Li, “Al e na ing 607 p ecu so laye deposi ion o highly s able pe o ski e ilms owa ds e - 608 icien sola cells using acuum deposi ion,” J. Ma e . Chem. A, ol.3, 609 pp. 9401–9405, 2015. 610 [21] D. Yang e al., “Su ace op imiza ion o elimina e hys e esis o eco d 611 e iciency plana pe o ski e sola cells,” Ene gy En i on. Sci., ol.9, 612 pp. 3071–3078, 2016. 613 [22] K. Wang e al., “CO2 Plasma- ea ed TiO2 ilm as an e ec i e elec on 614 anspo laye o high-pe o mance plana pe o ski e sola cells,” ACS 615 Appl. Ma e . In e aces, ol. 9, pp. 33989–33996, 2017. 616 [23] W. Yongzhen e al.,“Highly compac TiO 2 laye o e icien hole- 617 blocking in pe o ski e sola cells,” Appl. Phys. Exp ess, ol. 7, 2014, 618 A . no. 052301. 619 [24] H. Hu e al., “A omic laye deposi ion o TiO2 o a high-e iciency hole- 620 blocking laye in hole-conduc o - ee pe o ski e sola cells p ocessed in 621 ambien ai ,” ACS Appl. Ma e . In e aces, ol. 8, pp. 17999–18007, 2016. 622 [25] L. E ga , Hole T anspo Ma e ial (HTM) F ee Pe o ski e Sola Cell, 623 Hole Conduc o F ee Pe o ski e-Based Sola Cells. Cham, Swi ze land: 624 Sp inge , 2016, pp. 9–24. 625 [26] L. E ga e al., “Mesoscopic CH3NH3PbI3/TiO2 he e ojunc ion sola 626 cells,” J. Ame . Chem. Soc., ol. 134, pp. 17396–17399, 2012. 627 [27] A. Hag eld , G. Boschloo, L. Sun, L. Kloo, and H. Pe e sson, “Dye- 628 sensi ized sola cells,” Chem. Re ., ol. 110, pp. 6595–6663, 2010. 629 [28] J.-H. Im, C.-R. Lee, J.-W. Lee, S.-W. Pa k, and N.-G. Pa k, “6.5% e - 630 icien pe o ski e quan um-do -sensi ized sola cell,” Nanoscale, ol.3, 631 pp. 4088–4093, 2011. 632 [29] E. H. Ana aki e al., “Highly e icien and s able plana pe o ski e so- 633 la cells by solu ion-p ocessed in oxide,” Ene gy En i on. Sci., ol.9, 634 pp. 3128–3134, 2016. 635 [30] H. J. Snai h e al., “Anomalous hys e esis in pe o ski e sola cells,” J. 636 Phys. Chem. Le ., ol. 5, pp. 1511–1515, 2014. 637 [31] B. Chen, M. Yang, S. P iya, and K. Zhu, “O igin o J–V hys e esis in 638 pe o ski e sola cells,” J. Phys. Chem. Le ., ol. 7, pp. 905–917, 2016. 639 [32] K. Wojciechowski, M. Saliba, T. Leij ens, A. Aba e, and H. J. Snai h, “Sub- 640 150 [deg ee]C p ocessed meso-supe s uc u ed pe o ski e sola cells wi h 641 enhanced e iciency,” Ene gy En i on. Sci., ol. 7, pp. 1142–1147, 2014. 642 [33] T. Leij ens, B. Laube , G. E. Epe on, S. D. S anks, and H. J. Snai h, 643 “The impo ance o pe o ski e po e illing in o ganome al mixed halide 644 sensi ized TiO2-based sola cells,” J. Phys. Chem. Le ., ol. 5, pp. 1096– 645 1102, 2014. 646 [34] M. Anaya e al., “Elec on injec ion and sca old e ec s in pe o ski e 647 sola cells,” J. Ma e . Chem. C, ol. 5, pp. 634–644, 2017. 648 [35] E. L. Unge e al., “Hys e esis and ansien beha io in cu en - ol age 649 measu emen s o hyb id-pe o ski e abso be sola cells,” Ene gy En i on. 650 Sci., ol. 7 pp. 3690–3698, 2014. 651 [36] W. T ess, N. Ma ino a, T. Moehl, S. M. Zakee uddin, M. K. Nazee uddin, 652 and M. G a zel, “Unde s anding he a e-dependen J-V hys e esis, slow 653 ime componen , and aging in CH3NH3PbI3 pe o ski e sola cells: The 654 ole o a compensa ed elec ic ield,” Ene gy En i on. Sci., ol. 8, pp. 995– 655 1004, 2015. 656 [37] S. Meloni e al., “Ionic pola iza ion-induced cu en – ol age hys e esis 657 in CH3NH3PbX3 pe o ski e sola cells,” Na u e Commun., ol. 7, 2016, 658 A . no. 10334. 659 [38] G. Richa dson e al., “Can slow-mo ing ions explain hys e esis in he 660 cu en - ol age cu es o pe o ski e sola cells?” Ene gy En i on. Sci.,661 ol. 9, pp. 1476–1485, 2016. 662 [39] Z. Song, S. C. Wa hage, A. B. Phillips, and M. J. Heben, “Pa hways owa d 663 high-pe o mance pe o ski e sola cells: Re iew o ecen ad ances in 664 o gano-me al halide pe o ski es o pho o ol aic applica ions,” J. Pho on. 665 Ene gy, ol. 6, 2016, A . no. 022001. 666 IEEE P oo STOCKHAUSEN e al.: INSIGHTS IN PSC FABRICATION: UNRAVELING THE HIDDEN CHALLENGES OF EACH LAYER 5 Fig. 4. I–V cu e o cells wi h a mesopo ous TiO2 laye made om 30NR- D pas e (a e age pa icle size o 30 nm, black squa es) and om 18-NR-T pas e (a e age pa icle size o 20 nm, ed ci cles) a 0.94 sun. Bo h pas es we e pu chased om Dyesol L d. a long ime [7]. Only ecen ly, he e iciency gap has become293 a he small, which is due o imp o ed in e ace enginee ing [7],294 [18], [29]. Beside highe e iciencies, PSC wi h a mesopo ous295 laye show educed e iciency de ia ion be ween o wa d and296 backwa d scan, a phenomenon desc ibed as hys e esis [19], [30],297 [31]. When Snai h e al. demons a ed ha e en wi h meso-298 po ous laye s o alumina, a ma e ial ha canno pa icipa e in299 elec on ans e due o band ene gy misma ch, high e iciencies300 o 15.9% could be ob ained [32], i was deduced ha he meso-301 po ous laye mainly ul ills a s uc u al ole o c ys al g ow h302 and geome y de e mina ion, e en i e icien elec on injec ion303 om he pe o ski e in o he TiO2mesopo ous laye was epo ed304 [33]. Howe e , la es esul s demons a e ha ionic mig a ion a 305 he pe o ski e/TiO2in e ace, which is esponsible o cha ge306 accumula ion and he e o e ecombina ion e en s, is educed o 307 e en supp essed in he p esence o he TiO2mesopo ous laye 308 [34]. As hys e esis also depends on ionic mig a ion [35]–[38],309 i s educ ion in he p esence o mesopo ous TiO2is he conse-310 quence. Pu ing all oge he , bes esul s ha e been achie ed so311 a using a hin TiO2mesopo ous laye and a pe o ski e cap-312 ping laye ha p e en s ecombina ion be ween TiO2and he313 hole anspo laye [33], [34].314 The e exis se e al comme cial pas es wi h di e en sizes o 315 TiO2pa icles and he e o e, i was decided o compa e wo316 di e en pa icle sizes, namely one possessing 20 nm and one317 wi h 30 nm a e age pa icle diame e . The same pas e dilu ion318 a io in pu e e hanol (1:6) as well as he same deposi ion and319 sin e ing condi ions we e applied. Fig. 4 shows ha bo h meso-320 po ous laye s lead o e y compa able cell e iciencies ha a e321 wi hin he e o scale. This poin s owa d a highe ole ance and322 owa d a mesopo ous laye a chi ec u e, as long as he pa icle323 size emains simila .324 Wi hin pe o ski e ma e ials, he e exis s a huge a ie y o 325 ecipes and deposi ion echniques, which a e well summa ized326 in he book by Pa k e al. [10] and in he e iew by Song e al.327 [39]. I migh be no an easy ask o decide o he sui able328 pe o ski e ype and ab ica ion p ocess. The name pe o ski e329 e e s o a c ys alline s uc u e o he ype ABX3, A and B be-330 ing ca ions and X an anion. The pe o ski e class sui able o 331 sola cells is an o ganic lead halide, wi h A being gene ally an 332 o ganic ca ion, B being he lead ion, and X being a halogen, 333 usually b omine, iodide, chlo ine, and mix u es he eo . Lead 334 subs i u ion by in and ge manium analogs leads o pe o ski es 335 wi h se e e s abili y p oblems [32], [40], [41] and he e o e will 336 no be add essed he e. Ou ocus was o de e mine a pe o ski e 337 o mula ion easy o implemen and ha esul s in ep oducible 338 pe o ski e laye s wi h enhanced s abili y. Many esul s ha e 339 been published wi h monoca ionic pe o ski es, howe e , wi h 340 some inhe en limi a ions ha a e b ie ly exposed he e: MAPbI3341 has been in ensi ely s udied [42] bu has some d awbacks such 342 as weak s abili y owa d mois u e [43], [44] and empe a u e 343 [45]. Fo mamidinium (FA) was p oposed as al e na i e ca ion; 344 howe e , i s pe o ski e analog FAPbI3c ys allizes in he pho- 345 oinac i e phase below 60 °C [46], such as he ino ganic ca ion 346 analog CsPbI3[47]. Whe eas se e al g oups obse ed imp o ed 347 s abili y o he pho oac i e phase upon using bina y mixed ca ion 348 pe o ski es [48]–[52], Saliba e al. decided o combine he h ee 349 ca ions in a pe o ski e and achie ed high e iciencies (>20%) 350 on a e y ep oducible basis [12]. 351 Se e al me hods exis o solu ion-p ocessed ilm ab ica ion, 352 he mos common being simple sp eading o he pe o ski e p e- 353 cu so solu ion on he subs a e, also known as one-s ep depo- 354 si ion. Howe e , ilms wi h poo su ace con ol and he e o e 355 huge e iciency a ia ions gene ally eme ge [42]. A mo e so- 356 phis ica ed app oach is he sequen ial s ep deposi ion, whe e 357 he me al halide is i s deposi ed and annealed be o e being 358 b ough in con ac wi h he ammonium sal as apo o in solu- 359 ion [42], [53]. Ne e heless, se e al d awbacks o his deposi- 360 ion me hod we e expe ienced in ou g oup, such as incomple e 361 con e sion o he me al halide o pa ial dissolu ion o he pe - 362 o ski e du ing he subsequen washing s ep. Fu he mo e, i is 363 mo e ime-consuming as i equi es wo sin e ing s eps. The 364 an isol en echnique was in oduced in 2014 by he g oup o 365 Seok [19] and since hen, i has been he me hod o choice o 366 subsequen ly published eco d e iciencies [7]. I is qui e simple 367 o implemen and equi es only one p ecu so solu ion, whe eas 368 he c ys alliza ion o he pe o ski e is ini ia ed by adding a so- 369 called an isol en . This an isol en is chosen no o dissol e he 370 pe o ski e on one side and o displace he sol en o he la e on 371 he o he side. The main d awback o his deposi ion me hod is 372 i s a isanal aspec , equi ing a ce ain deg ee o aining be o e 373 eaching enhanced ep oducibili y. Howe e , smoo h pe o ski e 374 ilms wi h homogeneous composi ion and la ge g ain bounda ies 375 a e ob ained a e a sho aining ime. Fig. 5(a) shows a c oss 376 sec ion o a PSC wi h he monoli hic pe o ski e capping laye 377 on op o he mesopo ous TiO2laye wi h g ains g owing om 378 he bo om o he op and which a e hough o enhance cha ge 379 anspo , acco ding o Saliba e al. [12]. The op iew o he 380 pe o ski e laye [see Fig. 5(b)]shows g ains possessing diame- 381 e s be ween 200 and 500 nm, which is in good ag eemen wi h 382 he o iginal epo [12]. 383 One o he bigges de imen al ac o s o pe o ski e ab- 384 ica ion and s abili y is a mosphe ic humidi y, oge he wi h 385 oxygen [54]. PSC ha a e mean o exhibi p olonged s abili y 386 equi e ab ica ion and s o age in ine a mosphe e o de ice 387 encapsula ion a e ab ica ion [55], [56]. The e o e, PSC a e 388 IEEE P oo 6IEEE JOURNAL OF PHOTOVOLTAICS TABLE II BEST AND AVERAGE EFFICIENCIES FOR PSC, WHILE THE PEROVSKITE LAYER WAS FABRICATED WITHIN DIFFERENT CONDITIONS Fig. 5. SEM images. (a) C oss sec ion o he en i e PSC de ice showing he compac TiO2laye (ca. 80 nm), ollowed by he mesopo ous laye o 150– 200 nm. The adjacen pe o ski e laye pa ially in il a es in o he mesopo ous s uc u e and i s capping laye has a hickness o 200–300 nm. The spi o- MeOTAD hole anspo laye (da k g ey) has a hickness o 130–190 nm, ollowed by a 60-nm- hick gold laye as cu en collec o (ligh g ey). (b) Top iew o he pe o ski e laye , showing g ains wi h ca. 200–500 nm diame e . The issu es e ol ed du ing image cap u e and he e o e a e belie ed o be due o imaging. Ba s co espond o 1 µm. gene ally ab ica ed in glo e boxes wi h d y and oxygen- ee389 a mosphe e. And no only ha , also he subs a es should be390 absolu ely mois u e- ee. A e sin e ing he mesopo ous laye ,391 subs a es should hus be handled only in d y a mosphe e o 392 ans e ed o d y a mosphe e such as a glo e box be o e cooling393 below 150 °C. Wi hin he deg ada ion mechanism o pe o ski eQ3 394 s uc u es, oxygen only in e e es subsequen ly o hyd a a ion395 [56] and can he e o e be conside ed less c i ical i mois u e is396 absen o e y low. Labo a o ies ha a e newcome s in his a ea397 o esea ch migh no possess a glo e box in as uc u e. Low398 a mosphe ic humidi y le els a e assumed o be less c i ical o399 pe o ski e ab ica ion bu hese condi ions depend s ongly on400 he geog aphic localiza ion, he season, and some mo e inhe -401 en ac o s. To demons a e he e ec o mois u e and oxygen402 a mosphe e on pe o ski e o ma ion, a compa a i e es o PSC403 de ices ha we e ab ica ed inside and ou side he glo e box404 ( ela i e humidi y ou side he glo e box: 58%) was pe o med.405 The esul s a e displayed in Fig. 6 and Table II and i can be406 obse ed ha in case o pe o ski e being ab ica ed in ambi-407 en a mosphe e, he ac i e laye showed a ligh e colo and he408 co esponding PSC showed a bo h lowe VOC and JSC, whe eas409 he ill ac o emained a he unin luenced. As a s a egy o410 minimize wa e up ake by he subs a e, samples we e hea ed411 o 100 °C immedia ely be o e he pe o ski e p ecu so solu-412 ion was deposi ed. Co esponding PSC showed an imp o ed413 VOC and JSC, howe e he ill ac o dec eased. This is likely414 due o inhomogeneous c ys al g ow h, induced by he ele a ed415 empe a u e o he subs a e. Fo he bes cells ob ained, a li -416 Fig. 6. I–V cu es o PSC ab ica ed in a glo e box wi h 0% ela i e humidi y a 25 °C (black squa es), a ambien humidi y (58% ela i e humidi y) a 25 °C (g ey diamonds), and a ambien humidi y wi h subs a e p ehea ing a 100 °C ( ed ci cles) wi h an inciden ligh in ensi y o 0.94 sun. Image: Sample wi h pe o ski e p oduced inside (le ) and ou side he glo e box ( igh ) a 25 °C. le imp o emen can be s a ed when ho subs a es we e used, 417 hough a e age e iciencies came ou o be e y simila o hose 418 wi hou hea ea men . 419 This s udy shows ha i is highly ecommended o wo k 420 wi h a glo e box, p o iding e y low humidi y (<0.002% el. 421 humidi y) and oxygen le els. Ano he possible s a egy migh 422 be he use o a pe o ski e o mula ion ha is op imized owa d 423 enhanced esis ance a ele a ed humidi y le els [57]. 424 A op he pho oac i e laye , he hole conduc ing laye selec- 425 i ely anspo s he holes o he cu en collec o and he e o e 426 ul ills he complemen a y ole o he TiO2laye . I has o be 427 pinhole- ee o inhibi con ac o he cu en collec o wi h he 428 pe o ski e laye , o he same easons ha we e al eady s essed 429 ou conce ning he elec on conduc ing laye . Gene ally, a o - 430 mula ion using spi o-OMeTAD is used ha con ains, among o h- 431 e s, he ionic liquid LiTFSI o inc ease hole conduc i i y. How- 432 e e , bo h LiTFSI and spi o-OMeTAD ha e hyd ophilic p ope - 433 ies and p omo e humidi y inges ion, leading o poo humidi y 434 s abili y o he en i e de ice. A empe a u es abo e 55 °C, he 435 molecula hole anspo e c ys allizes, which se e ely a ec s 436 cell e iciencies. Two di e en hole conduc o laye s we e com- 437 pa ed owa d hei s abili y, namely spi o-oMeTAD as molecula 438 HTL and poly(3-hexyl hiophene) (P3HT) as polyme ic HTL. 439 IEEE P oo STOCKHAUSEN e al.: INSIGHTS IN PSC FABRICATION: UNRAVELING THE HIDDEN CHALLENGES OF EACH LAYER 7 Fig. 7. (a) I–V cu es o PSC employing spi o-OMeTAD (black squa es) and P3HT ( ed ci cles) as HTL a 0.95 sun. (b) No malized e iciency s abili y o PSC using wo PSC wi h spi o-OMeTAD (black squa es and g ey diamonds) and wo wi h P3HT ( ed ci cles and blue iangles). The ha ched g een a ea indica es a de ia ion less han 5% om he ini ial e iciency. Fig. 8. I–V cu es o PSC possessing a gold cu en collec o made by he mal e apo a ion (black line) and by spu e ing ( ed line) a 0.93 sun. I u ns ou ha spi o-OMeTAD leads o be e e iciencies,440 see Fig. 7(a), showing be e cu en densi y, open-ci cui po-441 en ial, and ill ac o al oge he . A po en ial ad an age could442 lie in an enhanced s abili y despi e lowe e iciency when P3HT443 is employed, bu he expe imen al da a could no con i m his444 assump ion, see Fig. 7(b), wi hin he limi ed ime ame.445 Gene ally, cu en collec o s a e made o gold despi e i s446 highe cos , as al e na i e me als such as Ag and Al ha e been447 demons a ing weak s abili ies [58]–[60]. Among gold depo-448 si ion me hods, one o he mos common ones a e spu e ing449 and he mal e apo a ion. Howe e , almos all published wo ks450 use he mal e apo a ion. We decided he e o e o compa e gold451 ilms wi h simila hickness ha we e ab ica ed by hese wo452 echniques. Indeed, he mal e apo a ion leads o a be e o e all453 cell pe o mance, see Fig. 8.454 The eason o he wo se pe o mance o PSC wi h spu e ed455 cu en collec o was e idenced by doing a sco ch es . While456 in case o he mal e apo a ion, he gold laye could be easily 457 s ipped o , in case o spu e ing deposi ion, he gold emained 458 s uck in o HTM. E en a e dissol ing he HTM laye , gold 459 aces we e s ill de ec ed wi h he naked eye inside he pe - 460 o ski e laye . This means ha du ing he gold laye deposi ion, 461 su ace bomba dmen p o okes pene a ion o gold deep in o 462 he de ice s uc u e, c ea ing ecombina ion cen e s. Thus, a 463 he mal e apo a o is needed o e icien pe o ski e sola cell 464 ab ica ion, e en i his s ep conside ably inc eases he ene gy 465 payback ime o PSC [61]. 466 Following all he laye ab ica ion s eps men ioned be o e, 467 i was possible o ab ica e PSC wi h an a e age e iciency o 468 (14.8 ±1.0)% o a se o 49 cells, see Fig. 9 le . 469 In labo a o y condi ions, cell e iciencies a e gene ally mea- 470 su ed o cell ac i e a eas in e io o 1 cm2. The cell a ea de- 471 limi ed by he deposi ion o he cu en collec o should be only 472 sligh ly supe io o he ac i e cell a ea (delimi ed by a mask) o 473 a oid ecombina ion e en s. Ins ead o using c ocodile clamps 474 a bi a ily connec ed o he cell, a sui able sample holde is 475 p e e able, see Fig. 9 igh , o maximal ep oducibili y. Among 476 all ac o s ha desc ibe he cell’s pe o mance, he maximum 477 powe poin (MPP) is he mos aluable in o ma ion in e ms o 478 applicabili y in sola de ices as i desc ibes bes he ope a ing 479 pa ame e s o he cell [7], [13], [62]. The MPP is ob ained ia 480 ma hema ic ex ac ion om I–Vcu es and su p isingly has no 481 ye gained big a en ion in published scien i ic wo ks. I–Vcu es 482 a e gene ally ob ained by dynamic scanning o ex e nal loads 483 hough ca e mus be aken ha he scan a e does no o e pass 484 he dynamic elec ochemical e en s inside he pe o ski e cell. 485 An example is gi en in Fig. 10, whe e a PSC was measu ed 486 a se e al scan a es. I me ely he I–V cu e is conside ed o 487 e iciency de e mina ion, bes esul s a e ob ained wi h a scan 488 a e o 1 V/s. Bu i CV o he same cell a e eco ded a 10 mV/s 489 and a 1 V/s, a s iking di e ence is obse ed conce ning he 490 hys e esis, see Fig. 11. Whe eas hys e esis is a he low in he 491 o me case, i conside ably inc eases in he la e case. This 492 IEEE P oo 8IEEE JOURNAL OF PHOTOVOLTAICS Fig. 9. E iciency dis ibu ion o all cells ab ica ed in s anda d condi ions (le ) and cell a chi ec u e wi h es ing de ice o ep oducible es ing condi ions ( igh ). Fig. 10. I–V cu es and e iciencies o a PSC eco ded a di e en scan a es a 0.98 sun: 10 mV/s (black squa es), 100 mV/s ( ed ci cles), 1 V/s (g ey diamonds). Fig. 11. Hys e esis o a PSC eco ded a a scan a e o 10 mV/s (black squa es) and 1 V/s ( ed ci cles) a 0.98 sun. means ha i he I–V cu e is eco ded in backwa d scan in such 493 condi ions, he ob ained esul s do no e lec he cell’s eal be- 494 ha io and he e o e lead o o e es ima ion o cell pe o mance 495 [7], [13]. Lacking so a easily implemen able measu emen 496 p o ocols o MPP acking, condi ions o I–V cu e eco ding 497 should be ca e ully chosen, wi h he goal o no o e es ima e 498 eal cell cha ac e is ics. 499 IV. CONCLUSION 500 PSC ep esen a e y a ac i e pho o ol aic echnology, as 501 innume ous publica ions ha e demons a ed, howe e ini ia ing 502 in his a ea may be a ha d ask. PSC a e made o se e al hin 503 laye s and no only each laye , bu also each in e ace plays 504 an impo an ole o he manu ac u ing o e icien de ices. 505 This pape is mainly di ec ed owa d esea ch g oups and sci- 506 en is s ha a e beginne s in his ac i e ield o esea ch and as 507 such, i was in ended o poin ou ab ica ion de ails ha emain 508 ba ely discussed in mos publica ions, bu ha a e signi ican 509 o he p epa a ion o e icien cells. I was e idenced ha be- 510 sides he usual equipmen o hin- ilm p epa a ion (ho pla e, 511 spin coa e , p og ammable u nace, e c.) and pho oelec ochem- 512 ical cha ac e iza ion (po en ios a o a iable ex e nal load, sola 513 simula o ), a glo e box and a he mal e apo a o o he depo- 514 si ion o he gold cu en collec o a e s ongly ad ised. The 515 goal o his pape is o analyze and op imize each laye and, as 516 a consequence, demons a e hei in luence on he en i e PSC 517 de ice. 518 I was e idenced by expe imen s ha o PSC pe o ming 519 bes , blocking laye has o be ab ica ed ia sp ay py olysis, 520 whe eas compa able esul s we e ob ained when pu e oxygen o 521 ai was used as ca ie gas. Fo he mesopo ous laye , no di e - 522 ence could be s a ed o bo h pa icle sizes used (20 and 30 nm); 523 howe e , i is likely ha a bigge di e ence in size may be o 524 ma e . Ou expe ience showed ha bes esul s o pe o ski e 525 ilms we e ob ained by applying he an isol en echnique wi h 526 a iple ca ion o mula ion. I was demons a ed ha conce ning 527 he adjacen hole anspo laye , spi o-OMeTAD was esul ing 528 IEEE P oo STOCKHAUSEN e al.: INSIGHTS IN PSC FABRICATION: UNRAVELING THE HIDDEN CHALLENGES OF EACH LAYER 9 in cells wi h supe io e iciency compa ed wi h he polyme ic529 analog P3HT. Finally, wo echniques o he ab ica ion o gold530 cu en collec o we e p esen ed and i was demons a ed ha 531 he mal e apo a ion leads o be e PSC han spu e ing, as in532 case o he la e he ene ge ic su ace bomba dmen p o oked533 gold pa icle pene a ion un il he pe o ski e ac i e laye , hus534 c ea ing ecombina ion cen e s. Inco po a ing all he discussed535 op imiza ions, PSC wi h an a e age e iciency o (14.8 ±1.0)%536 we e ab ica ed. Fu he mo e, an inno a i e, e sa ile, and quick537 me hod o elec ochemical subs a e e ching was applied in538 PSC, making lase -assis ed sc ibing and he e o e he neces-539 si y o such an equipmen obsole e. Dummy cells as selec i e540 elec ochemical cha ac e iza ion me hod o single laye s we e541 in oduced and yielded e sa ile esul s o he quali a i e com-542 pa ison o TiO2blocking laye s. I is belie ed ha his pape 543 con ibu es o a as e implemen a ion o PSC ab ica ion in544 esea ch g oups wi h ew expe ience in his a ea, hanks o a545 deepe unde s anding o ab ica ion de ails and use ul analysis546 ools ha a e easily a ailable.547 ACKNOWLEDGMENT548 The au ho s would like o hank M. G ¨ a zel o hos ing549 I. Mesqui a a EPFL o deepen he knowledge abou PSC ab-550 ica ion and O. Bellon om G ea cellsola L d. o ui ul551 discussions.552 REFERENCES553 [1] U.S.E.I. Adminis a ion, Washing on, DC, USA, Annual Ene gy Ou look,554 2017.555 [2] I.I.R.E. Agency, Abu Dhabi, UAE, RE hinking Ene gy 2017: Accele a ing556 he global ene gy ans o ma ion, 2017.557 [3] I.I.R.E. Agency, Abu Dhabi, UAE, Renewable Ene gy and Jobs Annual558 Re iew, 2016.559 [4] H.-S. Kim e al., “Lead iodide pe o ski e sensi ized all-solid-s a e submi-560 c on hin ilm mesoscopic sola cell wi h e iciency exceeding 9%,” Sci.561 Rep., ol. 2, 2012, A . no. 591.562 [5] NREL e iciency cha . [Online.]A ailable: h ps://www.n el.go /p /563 asse s/images/e iciency-cha .png. Accessed on: Ma . 26, 2018. Q4 564 [6] [Online.]A ailable: h p://www.g ea cellsola .com/wp-con en /uploads/565 2018/03/Au o a-Newsle e -Ma ch-2018-Final-ENG-RC.pd 566 [7] J.-P. Co ea-Baena e al., “The apid e olu ion o highly e icien pe -567 o ski e sola cells,” Ene gy En i on. Sci., ol. 10, pp. 710–727, 2017.568 [8] T. Salim, S. Sun, Y. Abe, A. K ishna, A. C. G imsdale, and Y. M. Lam,569 “Pe o ski e-based sola cells: Impac o mo phology and de ice a chi ec-570 u e on de ice pe o mance,” J. Ma e . Chem. A, ol. 3 pp. 8943–8969,571 2015.572 [9] I. Mesqui a, L. And ade, and A. Mendes, “Pe o ski e sola cells: Ma e ials,573 con igu a ions and s abili y,” Renewable Sus ain. Ene gy Re ., ol. 82,574 pp. 2471–2489, 2018.575 [10] N. G. Pa k, M. G ¨ a zel, and T. Miyasaka, O ganic-Ino ganic Halide Pe -576 o ski e Pho o ol aics: F om Fundamen als o De ice A chi ec u es.New577 Yo k, NY, USA: Sp inge , 2016.578 [11] C. C. Wu, C. I. Wu, J. C. S u m, and A. Kahn, “Su ace modi ica ion o 579 indium in oxide by plasma ea men : An e ec i e me hod o imp o e he580 e iciency, b igh ness, and eliabili y o o ganic ligh emi ing de ices,”581 Appl. Phys. Le ., ol. 70, pp. 1348–1350, 1997.582 [12] M. Saliba e al., “Cesium-con aining iple ca ion pe o ski e sola cells:583 Imp o ed s abili y, ep oducibili y and high e iciency,” Ene gy En i on.584 Sci., ol. 9, pp. 1989–1997, 2016.585 [13] J. A. Ch is ians, J. S. Manse , and P. V. Kama , “Bes p ac ices in pe o ski e586 sola cell e iciency measu emen s. A oiding he e o o making bad cells587 look good,” J. Phys. Chem. Le ., ol. 6, , pp. 852–857, 2015.588 [14] H.-R. Xia, J. Li, W.-T. Sun, and L.-M. Peng, “O ganohalide lead pe -589 o ski e based pho ode ec o s wi h much enhanced pe o mance,” Chem.590 Commun., ol. 50, pp. 13695–13697, 2014.591 [15] S. P. Koi y e al., “An elec ochemical me hod o as and con olled e ch- 592 ing o luo ine-doped in oxide coa ed glass subs a es,” J. Elec ochem. 593 Soc., ol. 164, pp. E1–E4, 2017. 594 [16] G. Yin e al., “Enhancing e iciency and s abili y o pe o ski e sola 595 cells h ough Nb-doping o TiO2 a low empe a u e,” ACS Appl. Ma e . 596 In e aces, ol. 9, pp. 10752–10758, 2017. 597 [17] M. Liu, M. B. Johns on, and H. J. Snai h, “E icien plana he e ojunc ion 598 pe o ski e sola cells by apou deposi ion,” Na u e, 501, pp. 395–398, 599 2013. 600 [18] D. Bi e al., “Polyme - empla ed nuclea ion and c ys al g ow h o pe - 601 o ski e ilms o sola cells wi h e iciency g ea e han 21%,” Na u e 602 Ene gy, ol. 1, 2016, A . no. 16142. 603 [19] N. J. Jeon, J. H. Noh, Y. C. Kim, W. S. Yang, S. Ryu, and S. I. Seok, 604 “Sol en enginee ing o high-pe o mance ino ganic–o ganic hyb id pe - 605 o ski e sola cells,” Na u e Ma e , ol. 13, pp. 897–903, 2014. 606 [20] D. Yang, Z. Yang, W. Qin, Y. Zhang, S. Liu, and C. Li, “Al e na ing 607 p ecu so laye deposi ion o highly s able pe o ski e ilms owa ds e - 608 icien sola cells using acuum deposi ion,” J. Ma e . Chem. A, ol.3, 609 pp. 9401–9405, 2015. 610 [21] D. Yang e al., “Su ace op imiza ion o elimina e hys e esis o eco d 611 e iciency plana pe o ski e sola cells,” Ene gy En i on. Sci., ol.9, 612 pp. 3071–3078, 2016. 613 [22] K. Wang e al., “CO2 Plasma- ea ed TiO2 ilm as an e ec i e elec on 614 anspo laye o high-pe o mance plana pe o ski e sola cells,” ACS 615 Appl. Ma e . In e aces, ol. 9, pp. 33989–33996, 2017. 616 [23] W. Yongzhen e al.,“Highly compac TiO 2 laye o e icien hole- 617 blocking in pe o ski e sola cells,” Appl. Phys. Exp ess, ol. 7, 2014, 618 A . no. 052301. 619 [24] H. Hu e al., “A omic laye deposi ion o TiO2 o a high-e iciency hole- 620 blocking laye in hole-conduc o - ee pe o ski e sola cells p ocessed in 621 ambien ai ,” ACS Appl. Ma e . In e aces, ol. 8, pp. 17999–18007, 2016. 622 [25] L. E ga , Hole T anspo Ma e ial (HTM) F ee Pe o ski e Sola Cell, 623 Hole Conduc o F ee Pe o ski e-Based Sola Cells. Cham, Swi ze land: 624 Sp inge , 2016, pp. 9–24. 625 [26] L. E ga e al., “Mesoscopic CH3NH3PbI3/TiO2 he e ojunc ion sola 626 cells,” J. Ame . Chem. Soc., ol. 134, pp. 17396–17399, 2012. 627 [27] A. Hag eld , G. Boschloo, L. Sun, L. Kloo, and H. Pe e sson, “Dye- 628 sensi ized sola cells,” Chem. Re ., ol. 110, pp. 6595–6663, 2010. 629 [28] J.-H. Im, C.-R. Lee, J.-W. Lee, S.-W. Pa k, and N.-G. Pa k, “6.5% e - 630 icien pe o ski e quan um-do -sensi ized sola cell,” Nanoscale, ol.3, 631 pp. 4088–4093, 2011. 632 [29] E. H. Ana aki e al., “Highly e icien and s able plana pe o ski e so- 633 la cells by solu ion-p ocessed in oxide,” Ene gy En i on. Sci., ol.9, 634 pp. 3128–3134, 2016. 635 [30] H. J. Snai h e al., “Anomalous hys e esis in pe o ski e sola cells,” J. 636 Phys. Chem. Le ., ol. 5, pp. 1511–1515, 2014. 637 [31] B. Chen, M. Yang, S. P iya, and K. Zhu, “O igin o J–V hys e esis in 638 pe o ski e sola cells,” J. Phys. Chem. Le ., ol. 7, pp. 905–917, 2016. 639 [32] K. Wojciechowski, M. Saliba, T. Leij ens, A. Aba e, and H. J. Snai h, “Sub- 640 150 [deg ee]C p ocessed meso-supe s uc u ed pe o ski e sola cells wi h 641 enhanced e iciency,” Ene gy En i on. Sci., ol. 7, pp. 1142–1147, 2014. 642 [33] T. Leij ens, B. Laube , G. E. Epe on, S. D. S anks, and H. J. Snai h, 643 “The impo ance o pe o ski e po e illing in o ganome al mixed halide 644 sensi ized TiO2-based sola cells,” J. Phys. Chem. Le ., ol. 5, pp. 1096– 645 1102, 2014. 646 [34] M. Anaya e al., “Elec on injec ion and sca old e ec s in pe o ski e 647 sola cells,” J. Ma e . Chem. C, ol. 5, pp. 634–644, 2017. 648 [35] E. L. Unge e al., “Hys e esis and ansien beha io in cu en - ol age 649 measu emen s o hyb id-pe o ski e abso be sola cells,” Ene gy En i on. 650 Sci., ol. 7 pp. 3690–3698, 2014. 651 [36] W. T ess, N. Ma ino a, T. Moehl, S. M. Zakee uddin, M. K. Nazee uddin, 652 and M. G a zel, “Unde s anding he a e-dependen J-V hys e esis, slow 653 ime componen , and aging in CH3NH3PbI3 pe o ski e sola cells: The 654 ole o a compensa ed elec ic ield,” Ene gy En i on. Sci., ol. 8, pp. 995– 655 1004, 2015. 656 [37] S. Meloni e al., “Ionic pola iza ion-induced cu en – ol age hys e esis 657 in CH3NH3PbX3 pe o ski e sola cells,” Na u e Commun., ol. 7, 2016, 658 A . no. 10334. 659 [38] G. Richa dson e al., “Can slow-mo ing ions explain hys e esis in he 660 cu en - ol age cu es o pe o ski e sola cells?” Ene gy En i on. Sci.,661 ol. 9, pp. 1476–1485, 2016. 662 [39] Z. Song, S. C. Wa hage, A. B. Phillips, and M. J. Heben, “Pa hways owa d 663 high-pe o mance pe o ski e sola cells: Re iew o ecen ad ances in 664 o gano-me al halide pe o ski es o pho o ol aic applica ions,” J. Pho on. 665 Ene gy, ol. 6, 2016, A . no. 022001. 666 IEEE P oo 10 IEEE JOURNAL OF PHOTOVOLTAICS [40] F. Hao, C. C. S oumpos, D. H. Cao, R. P. H. Chang, and M. G. Kana zidis,667 “Lead- ee solid-s a e o ganic-ino ganic halide pe o ski e sola cells,”668 Na u e Pho on., ol. 8, pp. 489–494, 2014.669 [41] T. K ishnamoo hy e al.,“Lead- ee ge manium iodide pe o ski e ma e-670 ials o pho o ol aic applica ions,” J. Ma e . Chem. A, ol. 3, pp. 23829–671 23832, 2015.672 [42] J. Bu schka e al., “Sequen ial deposi ion as a ou e o high-pe o mance673 pe o ski e-sensi ized sola cells,” Na u e, ol. 499 pp. 316–319, 2013.674 [43] M. A. G een, A. Ho-Baillie, and H. J. Snai h, “The eme gence o pe o ski e675 sola cells,” Na u e Pho on., ol. 8pp. 506–514, 2014.676 [44] M. G a zel, “The ligh and shade o pe o ski e sola cells,” Na u e Ma e .,677 ol. 13, pp. 838–842, 2014.678 [45] B. Conings e al., “In insic he mal ins abili y o me hylammonium lead679 ihalide pe o ski e,” Ad . Ene gy Ma e ., ol. 5, 2015, A . no. 1500477.680 [46] C. C. S oumpos, C. D. Malliakas, and M. G. Kana zidis, “Semiconduc -681 ing in and lead iodide pe o ski es wi h o ganic ca ions: Phase ansi-682 ions, high mobili ies, and nea -in a ed pho oluminescen p ope ies,”683 Ino ganic Chem., ol. 52, pp. 9019–9038, 2013.684 [47] C. K. Molle , “C ys al s uc u e and pho oconduc i i y o caesium685 plumbohalides,” Na u e, ol. 182, pp. 1436–1436, 1958.686 [48] N. J. Jeon e al., “Composi ional enginee ing o pe o ski e ma e ials o 687 high-pe o mance sola cells,” Na u e, ol. 517, pp. 476–480, 2015.688 [49] J. P. Co ea Baena e al., “Highly e icien plana pe o ski e sola 689 cells h ough band alignmen enginee ing,” Ene gy En i on. Sci., ol.8,690 pp. 2928–2934, 2015.691 [50] H. Choi e al., “Cesium-doped me hylammonium lead iodide pe o ski e692 ligh abso be o hyb id sola cells,” Nano Ene gy, ol. 7, pp. 80–85,693 2014.694 [51] J.-W. Lee, D.-H. Kim, H.-S. Kim, S.-W. Seo, S. M. Cho, and N.-G. Pa k,695 “Fo mamidinium and cesium hyb idiza ion o pho o- and mois u e-s able696 pe o ski e sola cell,” Ad . Ene gy Ma e ., ol. 5, 2015, A . no. 1501310.697 [52] X. Li e al., “A acuum lash–assis ed solu ion p ocess o high-e iciency698 la ge-a ea pe o ski e sola cells,” Science, ol. 353, pp. 58–62, 2016.699 [53] Q. Chen e al., “Plana he e ojunc ion pe o ski e sola cells ia apo -700 assis ed solu ion p ocess,” J. Ame . Chem. Soc., ol. 136, pp. 622–625,701 2014.702 [54] T. Leij ens, G. E. Epe on, N. K. Noel, S. N. Habis eu inge , A. Pe ozza,703 and H. J. Snai h, “S abili y o me al halide pe o ski e sola cells,” Ad .704 Ene gy Ma e ., ol. 5, 2015.Q5 705 [55] F. Ma eocci e al., “Encapsula ion o long- e m s abili y enhancemen 706 o pe o ski e sola cells,” Nano Ene gy, ol. 30, pp. 162–172, 2016.707 [56] I. Mesqui a, L. And ade, and A. Mendes, “Pe o ski e sola cells: Ma e ials,708 con igu a ions and s abili y,” Renewable Sus ain. Ene gy Re ., ol. 82,709 pp. 2471–2489, 2017.710 [57] Q. Tai e al., “E icien and s able pe o ski e sola cells p epa ed in711 ambien ai i espec i e o he humidi y,” Na u e Commun., ol. 7, 2016,712 A . no. 11105.713 [58] T. Leij ens, G. E. Epe on, S. Pa hak, A. Aba e, M. M. Lee, and H. J.714 Snai h, “O e coming ul a iole ligh ins abili y o sensi ized TiO2 wi h715 meso-supe s uc u ed o ganome al i-halide pe o ski e sola cells,” Na-716 u e Commun., ol. 4, 2013, A . no. 2885.717 [59] Y. Han e al., “Deg ada ion obse a ions o encapsula ed plana 718 CH3NH3PbI3 pe o ski e sola cells a high empe a u es and humidi y,”719 J. Ma e . Chem. A, ol. 3, pp. 8139–8147, 2015.720 [60] J. You e al., “Imp o ed ai s abili y o pe o ski e sola cells ia solu ion-721 p ocessed me al oxide anspo laye s,” Na u e Nano echnol., ol. 11,722 pp. 75–81, 2016.723 [61] J. Gong, S. B. Da ling, and F. You, “Pe o ski e pho o ol aics: Li e-cycle724 assessmen o ene gy and en i onmen al impac s,” Ene gy En i on. Sci.,725 ol. 8, pp. 1953–1968, 2015.726 [62] E. Zimme mann e al., “Cha ac e iza ion o pe o ski e sola cells:727 Towa ds a eliable measu emen p o ocol,” APL Ma e ., ol. 4, 2016,728 A . no. 091901.729 Ve ena S ockhausen ecei ed he B.S. deg ee in 730 chemis y and biochemis y and he M.S. deg ee 731 in chemis y om Ludwig Maximilians Uni e si y, 732 Munich, Ge many, in 2005 and 2007, espec i ely, 733 and he Ph.D. deg ee in analy ical physical chem- 734 is y om Denis Dide o Uni e si y, Pa is, F ance, 735 in 2011. 736 In 2013 and om 2015 o 2017, she was a Pos - 737 doc Fellow wi hin he g oup o P o . Ad´ elio Mendes. 738 He esea ch in e es s include pho o ol aic sys ems 739 such as dye-sensi ized sola cells and pe o ski e so- 740 la cells, as well as pho oelec ochemical sys ems o wa e spli ing. 741 742 Isabel Mesqui a was bo n in Po o, Po ugal, in 1988. 743 She ecei ed he M.S. deg ee in chemical enginee - 744 ing om he Uni e si y o Po o, Po o, in 2011. She 745 is cu en ly wo king owa d he Ph.D. deg ee a he 746 Facul y o Enginee ing, Uni e si y o Po o, ocusing 747 in he pe o ski e sola cells and hei s abili y. 748 F om Sep embe o 2011 o Ma ch 2015, she 749 was a Resea ch Fellow wi h LEPABE—Labo a o y 750 o P ocess Enginee ing, En i onmen , Bio echnol- 751 ogy and Ene gy wo king in me hanol uel cells and 752 dye-sensi ized sola cells. . 753 754 Lu´ ısa And ade ecei ed he Ph.D. deg ee om he 755 Uni e si y o Po o, Po o, Po ugal, in 2010, wi h 756 he hesis “S udy and Cha ac e iza ion o G ¨ a zel 757 Sola Cells.” 758 She is cu en ly an Assis an Resea che wi h LEP- 759 ABE, Facul y o Enginee ing, Uni e si y o Po o, 760 wo king wi h dye-sensi ized sola cells and pe - 761 o ski e sola cells. She wo ked di ec ly wi h P o . 762 G ¨ a zel a he Labo a o y o Pho onic and In e aces, 763 Lausanne, Swi ze land. 764 D . And ade was he ecipien o he ACP Diogo 765 Vasconcelos awa d in 2011, he Sol ay&Ho ione Inno a ion Challenge awa d 766 in 2011 and 2012, and he Ramos Ca a ino awa d in 2012. 767 768 Ad´ elio Mendes was bo n in 1964. He ecei ed he 769 Ph.D. deg ee om he Uni e si y o Po o, Po o, Po - 770 ugal, in 1993. 771 He is a Full P o esso wi h he Chemical Enginee - 772 ing Depa men , Facul y o Enginee ing, Uni e si y 773 o Po o. He coo dina es a la ge esea ch eam wi h e- 774 sea ch in e es s mainly including dye-sensi ized sola 775 cells and pe o ski e sola cells, pho oelec ochemi- 776 cal cells, including wa e spli ing and sola edox 777 low cells, edox low ba e ies, PEMFC, me hanol 778 s eam e o ming, memb ane, and adso ben -based 779 gas sepa a ions. He ecei ed an Ad anced Resea ch G an om he ERC on 780 dye-sensi ized sola cells o building in eg a ed o ca. 2 MEu os. He is cu - 781 en ly he Coo dina o o CEne -FEUP, he Compe ence Cen e o Ene gy o 782 he Facul y o Enginee ing a he Uni e si y o Po o. 783 784 IEEE P oo GENERAL INSTRUCTION 785 Au ho s: Please no e ha we canno accep new sou ce iles as co ec ions o you pape . I possible, please anno a e he PDF 786 p oo we ha e sen you wi h you co ec ions, using Adobe Ac oba edi ing so wa e, and upload i ia he Au ho Ga eway. 787 Al e na i ely, you may send us you co ec ions in a simple . x ile, u ilizing he line numbe s in he ma gins o he p oo o 788 indica e exac ly whe e you would like o us o make co ec ions. You may, howe e , upload e ised g aphics ia he Au ho 789 Ga eway. 790 QUERIES 791 Q1. Au ho : Please check whe he he unding in o ma ion is co ec . 792 Q2. Au ho : Please p o ide he expansion o DMF, DMSO, and HTL. 793 Q3. Au ho : Please check he usage o he e m “hyd a a ion” in he sen ence “Wi hin he deg ada ion mechanism o pe o ski e 794 s uc u es . . . ”. 795 Q4. Au ho : Please p o ide missing yea o Re s. [5]and [6].796 Q5. Au ho : Please p o ide page age o Re . [54].797