Characterization of Perovskite Solar Cells:
Towards a Reliable Measurement Protocol
Eugen Zimmermann1, Ka Kan Wong1, Michael Müller1, Hao Hu1, Philipp Ehrenreich1, Markus Kohlstädt2,3, Uli Würfel2,3, Simone Mastroianni2, Gayathri Mathiazhagan2, Andreas Hinsch2, TanajiP.Gujar4, Mukundan Thelakkat4, Thomas Pfadler1, Lukas Schmidt-Mende1
1Department of Physics, University of Konstanz, 78457 Konstanz, Germany
2Fraunhofer Institute for Solar Energy Systems ISE, Heidenhofstr. 2, 79110 Freiburg, Germany
3Freiburg Materials Research Center FMF, University of Freiburg, Stefan-Meier-Str. 21, 79104 Freiburg, Germany
4Applied Functional Polymers, Department of Macromolecular Chemistry I, University of Bayreuth, 95440 Bayreuth, Germany
Methods
Device Type A: ITO/PEDOT:PSS/CH3NH3PbI3–xClx/C60/LiF/Ag (University of Konstanz)
Devices were fabricated as described in [1]. In short, ITO-coated glass substrates (15Ω per sq) were ultrasonic cleaned for 5 min by detergent, acetone and isopropanol successively, and treated with oxygen plasma for 7 min. A PEDOT:PSS (Clevios™ VP AI 4083) film (30 nm) was spin-coated on cleaned substrates and subsequently annealed at 180 °C for 5 min. Afterwards, substrates were transferred into a glovebox with moisture and oxygen levels both below 5 ppm. The prepared precursor solution in anhydrous DMF (Sigma Aldrich 99.9%) of synthesized methylammonium iodide, PbI2 (Sigma Aldrich 99 %), and PbCl2 (Sigma Aldrich 98 %) in a ratio of 4:1:1 [2] was spin-coated onto the substrates with 3000 rpm for 20 s. Directly following, the films were transferred in a vacuum chamber on a hotplate at 90 °C. After evacuating, the films turned red-brown and were then further annealed without vacuum at 100 °C for full conversion. Finally, C60 (20 nm), LiF (1 nm), and Ag (100 nm) were thermally evaporated on top consecutively.
Device Type B: FTO/c-TiO2/PCBM/CH3NH3PbI3/P3HT/WO3/Ag (University of Konstanz)
All chemicals were purchased without further purification. FTO coated glass substrates (10Ω per sq) were ultrasonic cleaned for 5 min by detergent, acetone, and isopropanol consecutively, followed by N2 drying. After oxygen plasma cleaning for 7 min, a TiO2 blocking layer (70 nm) was deposited on the substrates by sputtering, and sintered at 450 °C for 1 hr. 1 w/v % PC60BM in chlorobenzene was spin-coated on top, and dried at 70 °C. Then, substrates were transferred to a glove-box where moisture and oxygen level were kept below 5ppm. Both PbI2 (0.75 M in anhydrous DMF) and substrates were maintained at 60 °C before spin-coating. Hot PbI2 solution was spin-coated on the substrates at 5000 rpm for 15 s, and then quickly transferred on a hotplate at 70 °C for drying. Subsequently, substrates were drop-casted with 4w/v % MAI (Solaronix) in anhydrous isopropanol for 20 s, spun at 4000 rpm for 20 s, and annealed at 70 °C for 30 min. P3HT (25 mg/ml, 51 kD, Rieke Metals Inc.) was spin-coated at 2000 rpm for 120 s, and WO3 (3 nm)/Ag (100 nm) were finally deposited on top by thermal evaporation.
Device Type C: ITO/PEDOT:PSS/CH3NH3PbI3/PCBM/Ba/Ag (ISE/University of Freiburg)
Devices were fabricated employing a one-step process based on lead acetate trihydrate (Pb(CH3CO2)2 · 3H2O) for the formation of perovskite layers, following a process developed by Forgács et al.[3]. Pb(CH3CO2)2 · 3H2O and all solvents were purchased from Sigma-Aldrich, methylammonium iodide (MAI) was purchased from Dyesol. All materials were used as received. Prepatterned ITO substrates were cleaned twice and subsequently in acetone, isopropanol and deionized water under ultrasonic agitation for 5 min each, blow dried with nitrogen and UV–ozone treated for 20 min. A 40 nm thick PEDOT:PSS (Clevios™ P VP AI 4083, Heraeus) layer was spin-coated onto the substrates which were then annealed at 130 °C for 10 min. All subsequent steps have been performed under nitrogen atmosphere. MAI (351 mg/mL) and Pb(CH3CO2)2 · 3H2O (279 mg/mL) were co-dissolved in anhydrous DMF, filtered and deposited on top of the substrates by spin-coating at 1750 rpm. The substrates were annealed at 85 °C for 25 min, during which perovskite was formed. Phenyl–C61–butyric acid methyl ester (PCBM) was deposited on top via spin-coating from a filtered 10 mg/mL solution in chloroform. Metal back electrodes (10 nm Ba/100 nm Ag) were thermally evaporated on top.
Device Type D: FTO/c-TiO2/mp-TiO2/mp-Al2O2/CH3NH3PbI3/Graphite (ISE Freiburg)
FTO coated glasses were cut (25 x 25 mm2), and etched by a laser. The substrates are cleaned ultrasonically for 2 min in Mucasol and 1 min in deionized water, both at 50 °C, rinsed in isopropanol, and finally dried. A compact TiO2 layer (12 nm) was sputtered on cleaned substrates. Afterwards, 600 nm of mesoporous TiO2 (mp-TiO2) was screen printed with a Dyesol DSL-18NRT paste mixed in a ratio of 1:2 with terpineol and sintered at 300 °C for 15 min. A 600 nm thick space layer consisting of 20 nm particles of Al2O3 (mp-Al2O2) was screen printed on top and sintered likewise. Subsequently, a 10 µm thick hole conductive and p-selective layer of porous carbon was formed by screen printing a Solaronix Elcocarb B/SP graphite paste. After processing the selective and porous layers, the substrates were sintered at 500 °C for 30 min and annealed at 120 °C for 20 min inside a N2 glovebox, to remove the remaining moisture. A 1-step perovskite solution was formed with 0.7225 g of PbI2 (TCI) and 0.25 g of MAI (Dyesol) in 1 ml of Dimethylsulfoxide. The solution was stirred at 60 °C overnight, loaded (10 µl) for 1 min and spin coated at 2500 rpm for 10 sec. Subsequently, substrates were annealed at 100 °C for 1 hour.
For recrystallization of the perovskite, the cells were kept under methyl amine gas for a few minutes and then introduced to normal atmosphere.
Device Type E: FTO/c-TiO2/mp-TiO2/CH3NH3PbI3/spiro-OMeTAD/WO3/Ag (University of Bayreuth)
FTO-coated glass sheets (17 Ω per sq) were etched with zinc powder and HCl (2M) to obtain the required electrode pattern. The substrates were then cleaned with detergent diluted in deionized water, rinsed with deionized water, acetone and ethanol, and dried with clean dry air. A compact TiO2 (50 nm) blocking layer was deposited by spray pyrolysis of titanium (IV) bis(acetoacetonato)-di(isopropanoxylate) diluted in ethanol at 450 °C on FTO coated glass substrates and annealed at 450 °C. After cooling, the substrates were transferred in a glovebox under nitrogen atmosphere. For perovskite formation, we adapted a published procedure [4, 5] and optimized it as follows. PbI2 (1 M) was dissolved in N, N-dimethyl formamide overnight under stirring conditions at 100 °C and 80 µl solution was spin coated on the FTO/c-TiO2/mp-TiO2 substrates at 2000 rpm for 50 s, and dried at 100 °C for 5 min. 100 mg MAI powder was spread out around the PbI2 coated substrates with a petri dish covering on the top and heated at 165 °C for 13 h for full conversion. For the optimization of conversion time, we monitored absorbance and found that the majority of PbI2 conversion is possible after 3h. However, the film shows a yellow spot at the centre of the substrate which indicates the presence of residual PbI2. Apparently the film becomes completely black after 12 h. The hole transporting material (HTM) deposition solution comprised of 2,2’,7,7’-tetrakis-(N,N-di-p-methoxyphenylamine) 9, 9’-spirobifluorene (spiro-OMeTAD) in chlorobenzene (102 mg/ml), 40 µl of 4-tert-butylpyridine (TBP) and 37 µl (520 mg/ml) solution of Li-TFSI in acetonitrile. No oxidants were used. HTM layer was deposited under ambient conditions. This was followed by thermal deposition of WO3 (3 nm), and Ag (100 nm) to form the back contact.
Device Type F: ITO/TiO2/BCP/PTB7:PCBM/PEDOT:PSS/Ag (University of Konstanz)
Devices were fabricated as described in [6]. ITO glass substrates (15Ω per sq, provided by Lumtec) were subsequently cleaned in acetone and isopropanol (IPA) in an ultrasonic bath for 5 min. After an oxygen plasma treatment for 7 min, a compact 40 nm thick TiO2 layer was sputtered and subsequently sintered at 450 °C for 30 min. Afterwards a 6.5 nm thick bathocupreine (BCP, provided by Lumtec) layer has been deposited via thermal evaporation (P 2*10-6 mbar) prior the spin-coating of the active layer. The active layer was deposited from a blended solution consisting of PTB7 (Poly[[4,8-bis[(2-ethylhexyl)oxy]benzo[1,2-b:4,5-b']dithiophene-2,6-diyl][3-fluoro-2-[(2-ethylhexyl)carbonyl]thieno[3,4-b]thiophenediyl]], 115.000 kD (Mw, PS Standard), Polydispersity(PDI): 2.5, provided by 1-Material) and [70]PCBM (Solenne BV) with a weight ratio of 1:1.5 in chlorbenzene. The solvent additive 1,8-di-iodooctane was added (3% by volume) to this blended solution. The film was spin-cast in a N2 glovebox at 1000 rpm for 60 s, directly followed by a 8 nm PEDOT:PSS (dissolved in IPA in a 1:10 volume ratio) film at 5000 rpm for 60 s. Finally, the 150 nm silver electrodes were deposited via thermal evaporation (P 5*10-6 mbar).
J-V measurements
Solar cells devices A, B, C, E, and F were measured at the University of Konstanz. Devices were illuminated with an AM 1.5G solar simulator. For measurements in inert atmosphere and ambient air the light intensity was adjusted to about 85 mW cm-2 and 110 mW cm-2, respectively, using a Fraunhofer ISE certified Si reference diode with an attached KG5 filter. J-V curves were acquired with a Keithley 2410 source meter controlled by a Matlab program. Solar cells were placed in a light tight sample holder (in order to avoid additional excitation of the active layer due to scattered light) and illuminated through a shadow mask defining three times an active area of 0.125 cm2 for the three pixels on each substrate.
Device C, and D were measured at the University of Freiburg. Current–voltage curves were measured under nitrogen atmosphere using a Keithley 2400 source meter and a KHS Steuernagel (Solar Cell Test 575) solar simulator providing simulated AM 1.5G illumination (100 mW/cm²) corrected for spectral mismatch. The active area of the cell (3x3 mm²) was illuminated by using metal masks of corresponding size.
Matlab code for controlling the Keithley sourcemeter can be found at GitHub [7].
Tracking algorithm
Initial values for the tracking of maximum power point and the VOC were extracted from a standard J-V curve with arbitrary parameters. Subsequently, the tracking algorithm is measuring five points with the initial value as centre point and searching for the voltage with maximum power or lowest current for MPP or VOC tracking, respectively. The found point is taken as new centre point for the next 5 point sweep. In case the new point lies on the boarder of the scan range (Position 1 or 5), the scan range will be increased, otherwise decreased. For tracking of the short circuit current density, the current is simply monitored as a function of time at 0 V.
Additional Measurements
J-V measurements in ambiant air
Figure S1: J-V measurements of Device A measured in ambient atmosphere and N2 (left) and corresponding steady state measurements.
Table S1: Comparison of J-V measurements in different atmospheres. Device A was measured in the glovebox with a light intensity of about 85 mW/cm², whereas Device A’ was measured in ambient with a light intensity of about 110 mW/cm².
MeasurementI.a / I.b / II.b
à / ß / ●
Device A
(N2) / PCE / 13.52 / 13.68 / 13.56
Ag
LiF
C60
Perovskite
PEDOT:PSS
ITO / FF / 75.62 / 76.17 / 77.25
JSC / 15.34 / 15.36 / 15.26
VOC / 0.96 / 0.97 / 0.95
Device A’
(air) / PCE / 12.93 / 12.71 / 13.16
Ag
LiF
C60
Perovskite
PEDOT:PSS
ITO / FF / 72.21 / 70.63 / 73.04
JSC / 20.44 / 20.60 / 20.37
VOC / 0.91 / 0.91 / 0.92
Extended tracking of the maximum power point
Figure S2: Extended tracking of the maximum power point for (left) 600 s, and (right) 1800 s. For device of Type B (right), incidental outliers have been removed.
Cyclic J-V measurements
Figure S3: Cyclic J-V measurements for all investigated device types. Symbols in are introduced for easier assignment and just represent the respective device. Device types are: A) ITO/PEDOT:PSS/CH3NH3PbI3–xClx/C60/LiF/Ag, B) FTO/c-TiO2/PCBM/CH3NH3PbI3/P3HT/WO3/Ag, C) ITO/PEDOT:PSS/CH3NH3PbI3/PCBM/Ba/Ag, D) FTO/c-TiO2/mp-TiO2/mp-Al2O2/CH3NH3PbI3/Graphite, E) FTO/c-TiO2/mp-TiO2/CH3NH3PbI3/spiro-OMeTAD/WO3/Ag, F) ITO/TiO2/BCP/PTB7:PCBM/PEDOT:PSS/Ag.
Time resolved J-V measurements
Figure S4: Time resolved J-V measurements (left) and corresponding reconstructed J-V curves (right) of investigated device types A to F. Device of type E was measured with 1 s resolution instead 10 s and is not shown in left. J-V curves in (right) were reconstructed by extracting the last measurement point of each individual voltage step in (left).
References
1. Hu, H., et al., Highly Efficient Reproducible Perovskite Solar Cells Prepared by Low-Temperature Processing. Molecules, 2016. 21(4): p. 542-552.
2. Wang, D., et al., Reproducible One-Step Fabrication of Compact MAPbI3–xClxThin Films Derived from Mixed-Lead-Halide Precursors. Chem. Mat., 2014. 26(24): p. 7145-7150.
3. Forgács, D., M. Sessolo, and H.J. Bolink, Lead acetate precursor based p-i-n perovskite solar cells with enhanced reproducibility and low hysteresis. J. Mater. Chem. A, 2015. 3(27): p. 14121-14125.
4. Chen, Q., et al., Controllable self-induced passivation of hybrid lead iodide perovskites toward high performance solar cells. Nano Lett., 2014. 14(7): p. 4158-63.
5. Chen, Q., et al., Planar Heterojunction Perovskite Solar Cells via Vapor-Assisted Solution Process. J. Am. Chem. Soc., 2014. 136(2): p. 622-625.
6. Kim, K.-D., et al., Decoupling optical and electronic optimization of organic solar cells using high-performance temperature-stable TiO2/Ag/TiO2 electrodes. APL Materials, 2015. 3(10): p. 106105.
7. GitHub Repository of Eugen Zimmermann. 2016 [cited 2016 27.06.2016]; Available from: https://github.com/EugenZimmermann/matlab-keithley-jv.
8