Supporting Information. Fully Solution-Processed Semitransparent Organic Solar Cells with a Silver Nanowire Cathode and a Conducting Polymer Anode

Transcription

1 Supporting Information Fully Solution-Processed Semitransparent Organic Solar Cells with a Silver Nanowire Cathode and a Conducting Polymer Anode Jong Hyuk Yim, Sung-yoon Joe, Christina Pang, Kyung Moon Lee, Huiseong Jeong, Ji-Yong Park, Yeong Hwan Ahn, John C. de Mello,,*, and Soonil Lee,* Department of Physics and Division of Energy Systems Research, Ajou University, Suwon, , Korea Centre for Plastic Electronics, Imperial College London, Exhibition Road, London, SW7 2AZ, United Kingdom Address correspondence to soonil@ajou.ac.kr, j.demello@imperial.ac.uk

2 Device architecture and electrode patterns Figure S1 shows a general schematic of the device architecture and the corresponding energy level diagrams for P3HT:PCBM bulk heterojunction (BHJ) solar cells with a bottom AgNW cathode. Two different electrode patterns were employed for the control and semitransparent devices as shown in Figure S2; also shown are the specific device architectures used for the two control devices and the semitransparent device. In all cases four identical pixels, each having an active area of 0.24 cm 2 were fabricated on each 3 3 cm 2 glass substrate. Accounting for the influence of illumination direction on the EQE The difference in transmittance between the cathode and anode windows alone is not sufficient to account for the difference in J sc with respect to the illumination directions. Figure S3 shows the transmittance spectra of Glass/AgNW/ZnO and PH 1000/GO/AI 4083, which determine the influx of solar spectrum photons to the P3HT:PCBM active layer when illuminated through the bottom (substrate) and top sides, respectively. The average transmittance between nm (the main spectral range for P3HT absorption) is 84 % for the cathode window and 83 % for the anode window. The measured transmittance of the P3HT:PCBM BHJ device matches closely with the product of the window transmittances and the transmittance through the active layer: Τ ( )= Τ ) Τ ) αλ ( ) P HT: PCBMd : [ ] 3 P3HT PCBM device λ ( λ Glass / AgNW/ ZnO ( λ PH / AI / GO e (S1) However, the external quantum efficiency (EQE) measurements show large differences at 530 nm, where the transmittances of cathode and anode windows are the same, whereas the EQE difference is reduced by a factor of ~2.7 at 600 nm. Figure S4 shows the respective EQE spectra corresponding to illumination through the PH 1000 and AgNW electrodes. These EQE spectra

3 were measured from a 9-month-old device, while the J-V curves in Figure 5(b) were measured directly after fabrication. This device maintained about half of its original PCE after 9 months. The reduction of J sc to half of its initial value is responsible for most of the PCE loss. The V oc and FF of this device by contrast suffered only very slight losses after 9-month storage: losses of 3 % for V oc and 2 % for FF. Because the operation of P3HT:PCBM BHJ OSCs can be modelled as sequential steps of exciton generation in response to light absorption, exciton diffusion to P3HT-PCBM interfaces, exciton dissociation into free charge carriers at the P3HT-PCBM junction, charge carrier transport in separate P3HT and PCBM domains, and extraction of charge carriers at the electrodes, 1 the EQE can be expressed as the product of efficiencies for each of these steps. We note that the efficiencies for the exciton dissociation, charge carrier transport, and charge carrier extraction processes, which are all electrical in nature, are insensitive to illumination direction. On the contrary, both exciton generation and exciton diffusion can be sensitive to the illumination direction. Figure S5 is the solar photon absorption spectra corresponding to the two illumination directions. These spectra have been estimated by multiplying the black-body radiation spectrum that mimics the solar photon emission with the transmittance of the respective window layers and the light harvesting efficiency of the 210-nm-thick P3HT:PCBM active layer: α( λ) P3HT: PCBMd P3HT: PCBM A( λ) Φ ( λ) ( λ) / / or ( λ) solar Τ PH1000 AI 4083 GO Τ Glass / AgNW/ ZnO 1 e. (S2) We note that the exciton generation spectra, which are directly proportional to the solar photon absorption spectra, must be very similar for the two illumination directions at wavelengths longer than 500 nm because there is no significant illumination-direction-dependent difference in solar photon absorption.

4 Figure S6 shows the normalized exciton generation rates with respect to the position in the P3HT:PCBM active layer for the two illumination directions: Gz (, ) Φ ( ) Τ( ) or Τ( ) ( ) αλ ( ) 3 λ solar λ λ PH1000/ AI 4083/ GO λ Glass / AgNW/ ZnO α λ e P3HT: PCBM P HT: PCBM. (S3) We note that the exciton generation rate is monotonically decreasing from its maximum value near the illumination-side electrodes. The rates of exciton generation by 530-nm photons for the two illumination directions are almost the same near the respective electrodes and at the middle of the ~210 nm P3HT:PCBM active layer. Therefore, we attribute the difference in J sc for anodeand cathode-side illumination to differences in exciton-diffusion efficiency due to the asymmetric location of P3HT-PCBM interfaces with respect to the two electrodes. Presumably, the majority of P3HT-PCBM junctions are not at the middle of the active layer, but located closer to the AgNW cathode. Because the diffusion length of excitons is extremely short, typically much less than 20 nm, a slight shift of P3HT-PCBM junctions towards the cathode would favour AgNWside illumination as this generates a larger number of excitons in the vicinity of P3HT-PCBM junctions, resulting in higher J sc. In other words, compared to the case of PH 1000 anode-side illumination, the exciton-diffusion efficiency becomes higher when illuminated from the AgNW cathode side. However, at a wavelength of 600 nm, the exciton generation rates are lower, but more uniform throughout the P3HT:PCBM active layer so that the difference in EQE between two illumination directions become much smaller. z 1. Blom, P. W. M.; Mihailetchi, V. D.; Koster, L. J. A.; Markov, D. E. Device Physics of Polymer:Fullerene Bulk Heterojunction Solar Cells. Adv. Mater. 2007, 19,

5 Figure S1. (a) Schematic device architecture of AgNW-based P3HT:PCBM bulk-hetero-junction solar cells with a top contact of PEDOT:PSS or evaporated Ag. (b) Energy level diagram of the materials used in device fabrication.

6 Figure S2. Schematic diagram showing electrode patterns for (a) control and (b) semi-transparent P3HT:PCBM bulk heterojunction solar cells. In the case of the control devices, the AgNW cathode was patterned in a single strip of area cm 2 and the evaporated silver anode was patterned in the shape of four fingers, each of area cm 2. In the case of the semitransparent devices, the AgNW cathode was patterned in the shape of four fingers, each of area cm 2, and the PH 1000 anode was patterned in a single strip of area cm 2. In both cases, four identical pixels were obtained per substrate, each having an active area of 0.24 cm 2 as defined by the overlap of the anode and cathode patterns. (C) Schematics showing the device structures for the three AgNW-based organic solar cells.

7 Figure S3. The transmittance spectra of the cathode-side (Glass/AgNW/ZnO) and anode-side (PH 1000/GO/AI 4083) windows, which determine the influx of solar spectrum photons to the active layer (P3HT:PCBM blend). Also shown is the measured transmittance of the complete semitransparent device (dotted black line). The measured transmittance of the complete device matches closely to the calculated transmittance as determined from the product of the transmittance spectra of the two window layers and the P3HT:PCBM active layer.

8 Figure S4. Comparison of the EQE spectra of a semitransparent P3HT:PCBM BHJ OSC under illumination through the AgNW cathode and the PH 1000 anode. The EQE difference at ~530 nm is about 2.7 times larger than that at ~600 nm. The EQE measurement was made using a 9- month-old semitransparent P3HT:PCBM device that had maintained about half of its original PCE.

9 Figure S5. Comparison of the calculated solar photon absorption spectra under illumination through the AgNW cathode and the PH 1000 anode. The spectra have been estimated as the product of the black-body radiation spectrum representing solar photons, the transmittance spectrum of the appropriate (anode-side or cathode-side) window, and the absorption spectrum of the 210 nm P3HT:PCBM blend layer. The exciton generation spectra are directly proportional to the respective solar photon absorption spectra.

10 Figure S6. Calculated exciton generation rates versus position in a 210-nm-thick P3HT:PCBM active layer for anode- and cathode-side illumination. Larger values of the absorption coefficient α(λ) lead to a steeper monotonic decrease in the exciton generation rates away from the window.