Supplementary Information A hybrid CMOS-imager with a solution-processable polymer as photoactive layer Daniela Baierl*, Lucio Pancheri, Morten Schmidt, David Stoppa, Gian-Franco Dalla Betta, Giuseppe Scarpa, Paolo Lugli Affiliations: Institute for Nanoelectronics, Technische Universität München, Munich, Germany Daniela Baierl*, Morten Schmidt, Giuseppe Scarpa, Paolo Lugli Fondazione Bruno Kessler, Trento, Italy Lucio Pancheri, David Stoppa Department of Information Engineering and Computer Science, University of Trento, Trento, Italy Gian-Franco Dalla Betta *Corresponding author. Email: baierl@nano.ei.tum.de 1
Supplementary Figures Supplementary Figure S1: Large-scale sprayed organic photodiodes. (a) The currentvoltage behavior under dark and illuminated (100 mw/cm²) conditions for different mean blend layer thicknesses. (b) The spectral EQE at -4 V external bias voltage. 2
Supplementary Figure S2: Spectral EQE of panchromatic OPDs. Different blend mixing ratios between squaraine, PCBM and P3HT are applied. The external bias is -4 V. 3
Supplementary Figure S3: Influence of silicon photogeneration in a PCBM:P3HT detector. Images, showing a horse, acquired by the hybrid CMOS-detector with a P3HT:PCBM bulk heterojunction. The left one is acquired by illumination with green light (535 nm), the other with a NIR-LED at 850 nm. 4
Supplementary Discussion Influence of the photoactive layer thickness on the external quantum efficiency (EQE) and the dark current. The EQE of an OPD is set by the fraction of photons reaching the photoactive blend and the number of collected carriers relative to the absorbed photons (internal quantum efficiency). A thicker blend layer absorbs more photons, but also the serial resistance is increased considerably due to the general low carrier mobility. To define an optimal mean blend thickness for the rough layers, large-scale OPDs with a 7 mm² big active area were sprayed on glass substrates and the impact on dark current and EQE was investigated. Theses OPDs were fabricated like for the hybrid-imager but with a simpler build-up (non-inverted, with the sprayed conductive PEDOT:PSS as bottom electrode and a thermal evaporated aluminum layer as top contact). Supplementary Figure S1b shows the change of EQE with varying mean photoactive layer thickness for OPDs which were apart from this variance identically fabricated. The mean layer thickness was determined with a profilometer on sprayed reference layers on glass substrates. For increasing the layer thickness, starting at about 300 nm, the maximum EQE at a wavelength at about 500 nm is decreasing with an about 4 % loss of EQE for a 100 nm thickness increase. Regarding the integrated EQE over the entire absorption range, the maximum EQE can be found around 500 nm layer thickness, as also reported by Rauch et al. 48. A thinner sprayed layer is prone to pinholes since the local thickness variation can be in the order of some hundreds nanometers due to film build-up of single sprayed blend droplets. These pinholes, arising from conductive bridges between the two electrodes, through a canal of the donator respective acceptor material, decrease the shunt resistance of the device which is related to an increase of the dark current. This behavior is demonstrated in Supplementary Figure S1a, showing the decrease 5
and saturation of the dark current with increasing layer thickness. Generally, we observed the dark current saturation, which occurs for shunt resistance >> serial resistance, for a mean photoactive layer thicknesses above 400-450 nm which is a quite high value compared to typical spin-coated layer thicknesses (around 100-300 nm) and considering the root-meansquare roughness which is considerably smaller, about 100 nm. This means that for thinner layers, in spite of the low mean roughness, there are still pinholes present through the organic blend layer which could be particular detrimental to the performance of diodes with dimensions comparable to single droplets (5-10 µm). In this case, a broad distribution of the dark currents for different pixels would be expected. To avoid the resulting high fixed pattern noise, the photoactive layer thickness of the hybrid imager was chosen to a slightly higher value (650 nm) as the optimum for high EQE. The downscaling of the diode area from 7 mm² to pixel pad sizes of 50 µm and smaller led to a significant reduction of the saturation level of the dark current. In Supplementary Figure S1a, it is more than one order of magnitude higher than the mean dark level of the hybrid imager. Adding squaraine to the PCBM:P3HT bulk heterojunction. To exploit also a part of the NIR spectrum and to demonstrate the portability of the hybrid CMOS imager concept, a squaraine (SQ) dye 49-52 was added to the PCBM:P3HT bulk heterojunction. To determine an optimal mixing ratio for an uniform EQE distribution over the wavelength range between 300 nm and 900 nm, large scale (7mm² active area) spin-coated OPDs were fabricated and characterized. About a photodetector with a bulk-heterojunction of PCBM:SQ with an optimum mixing ratio of 3:1, it was already reported 51. This detector exhibits a low EQE in the green and red region of the spectrum, as shown in Supplementary Figure S2, since neither PCBM nor SQ absorb in this region considerably. 6
We increased the amount of P3HT in this blend and observed that for SQ:P3HT:PCBM=1:1.5:3 the dip in the EQE spectrum vanished. For changing the amount of PCBM, neither an increase (EQE decrease in the green region) nor a decrease (EQE decrease in the NIR region) improved the performance of this panchromatic detector. Influence of silicon absorbance on the hybrid imager performance. In Fig. 6 of the article we showed an image acquired with the panchromatic CMOS-imager at 850 nm. At this wavelength, also the photogeneration in the silicon substrate can contribute to the signal since at this wavelength, the EQE of the hybrid imager is quite low, as shown in Supplementary Figure S2. The photons, which are not absorbed in the organic layer, generate photocarriers in the silicon which are partly collected in the MOS switches of the chip. To determine the influence of this effect, two pictures at different wavelengths, as in Fig.6, were acquired with the hybrid imager containing only the PCBM:P3HT bulk heterojunction. By omitting the squaraine, almost all photons pass the organic layer (which is transparent at 850 nm) and the resulting blurred image with very weak contrast, shown in Supplementary Figure S3, accounts for the photocarriers generated in silicon. This image proves that the chip is almost blind to NIR light, and the image at 850 nm shown in Fig.6 can be attributed to the photoactive squaraine. Supplementary References 48 Tedde, S. F. et al. Fully Spray Coated Organic Photodiodes. Nano Lett. 9, 980-983 (2009). 49 Beverina, L. & Salice, P. Squaraine Compounds: Tailored Design and Synthesis towards a Variety of Material Science Applications. Eur. J. Org. Chem. 2010, 1207-1225 (2010). 50 Binda, M. et al. Fast and air stable near-infrared organic detector based on squaraine dyes. Org. Electron. 10, 1314-1319 (2009). 51 Binda, M. et al. High detectivity squaraine-based near infrared photodetector with na/cm² dark current. App. Phys. Lett. 98, 073303 (2011). 52 Wei, G., Wang, S., Renshaw, K., Thompson, M. E. & Forrest, S. R. Solution-Processed Squaraine Bulk Heterojunction Photovoltaic Cells. ACS Nano 4, 1927-1934 (2010). 7