Rear interface engineering of hybrid organic-silicon nanowire solar cells via blade coating

In this work, we investigate blade-coated organic interlayers at the rear surface of hybrid organic-silicon photovoltaics based on two small molecules: Tris(8-hydroxyquinolinato) aluminium (Alq3) and 1,3-bis(2-(4tert-butylphenyl)-1,3,4-oxadiazol-5-yl) benzene (OXD-7). In particular, soluble Alq3 resulting in a uniform thin film with a root-mean-square roughness < 0.2nm is demonstrated for the first time. Both devices with the Alq3 and OXD-7 interlayers show notable enhancement in the open-circuit voltage and fill-factor, leading to a net efficiency increase by over 2% from the reference, up to 11.8% and 12.5% respectively. The capacitance-voltage characteristics confirm the role of the small-molecule interlayers resembling a thin interfacial oxide layer for the Al-Si Schottky barrier to enhance the built-in potential and facilitate charge transport. Moreover, the Alq3 interlayer in optimized devices exhibits isolated phases with a large surface roughness, in contrast to the OXD-7 which forms a continuous uniform thin film. The distinct morphological differences between the two interlayers further suggest different enhancement mechanisms and hence offer versatile functionalities to the advent of hybrid organic-silicon photovoltaics. ©2016 Optical Society of America OCIS codes: (160.4890) Organic materials; (310.1620) Interference coatings; (350.6050) Solar energy; (040.5350) Photovoltaic. References and links 1. International Technology Roadmap for Photovoltaic (ITRPV) 2014 Results; 2015. Available from: http://www.itrpv.net/Reports/Downloads/2015/ 2. P. V. Magazine, Photovoltaic Markets & Technology; 2012. Available from: http://www.pvmagazine.com/archive/articles/beitrag/switch-from-p-to-n-_100007072/501/#axzz3jI3NkDxM 3. Panasonic Press Release, 10 April 2014. Panasonic HIT® Solar Cell Achieves World’s Highest Energy Conversion Efficiency of 25.6%. Available from: http://news.panasonic.net/stories/2014/0416_26881.html 4. M. Taguchi, A. Yano, S. Tohoda, K. Matsuyama, Y. Nakamura, T. Nishiwaki, K. Fujita, and E. Maruyama, “24.7% record efficiency HIT solar cell on thin silicon wafer,” IEEE J. Photovolt. 4(1), 96–99 (2014). 5. S. C. Shiu, J. J. Chao, S. C. Hung, C. L. Yeh, and C. F. Lin, “Morphology dependence of silicon nanowire/poly (3, 4-ethylenedioxythiophene):poly(Styrenesulfonate) heterojunction solar cells,” Chem. Mater. 22(10), 3108– 3113 (2010). 6. J. C. Nolasco, R. Cabré, J. Ferré-Borrull, L. F. Marsal, M. Estrada, and J. Pallares, “Extraction of poly (3hexylthiophene)(P3HT) properties from dark current voltage characteristics in a P3HT/n-crystalline-silicon solar cell,” J. Appl. Phys. 107(4), 044505 (2010). 7. L. He, C. Rusli, C. Jiang, H. Wang, and D. Lai, “Simple approach of fabricating high efficiency Si nanowire/conductive polymer hybrid solar cells,” IEEE. Electron. Dev. Lett. 32(10), 1406–1408 (2011). 8. X. Shen, B. Sun, D. Liu, and S. T. Lee, “Hybrid heterojunction solar cell based on organic-inorganic silicon nanowire array architecture,” J. Am. Chem. Soc. 133(48), 19408–19415 (2011). #253576 Received 17 Nov 2015; revised 6 Jan 2016; accepted 7 Jan 2016; published 21 Jan 2016 © 2016 OSA 25 Jan 2016 | Vol. 24, No. 2 | DOI:10.1364/OE.24.00A414 | OPTICS EXPRESS A414 9. T. G. Chen, B. U. Huang, E. C. Chen, P. Yu, and H. F. Meng, “Micro-textured conductive polymer/silicon heterojunction photovoltaic devices with high efficiency,” Appl. Phys. Lett. 101(3), 033301 (2012). 10. C. Xie, X. Zhang, Y. Wu, X. Zhang, X. Zhang, Y. Wang, W. Zhang, P. Gao, Y. Han, and J. Jie, “Surface passivation and band engineering: a way toward high efficiency graphene–planar Si solar cells,” J. Mater. Chem. A Mater. Energy Sustain. 1(30), 8567–8574 (2013). 11. Y. Wu, X. Zhang, J. Jie, C. Xie, X. Zhang, B. Sun, Y. Wang, and P. Gao, “Graphene Transparent Conductive Electrodes for Highly Efficient Silicon Nanostructures-Based Hybrid Heterojunction Solar Cells,” J. Phys. Chem. C 117(23), 11968–11976 (2013). 12. Y. Zhang, F. Zu, S. T. Lee, N. Zhao, and B. Sun, “Heterojunction with organic thin layers on silicon for record efficiency hybrid solar cells,” Adv. Energy Mater. 4(2), 1300923 (2014). 13. Y. Zhang, W. Cui, Y. Zhu, F. Zu, L. Liao, S. T. Lee, and B. Sun, “High efficiency hybrid PEDOT:PSS/nanostructured silicon Schottky junction solar cells by doping-free rear contact,” Energy Environ. Sci. 8(1), 297–302 (2015). 14. Y. Zhang, R. Liu, S. T. Lee, and B. Sun, “The role of a LiF layer on the performance of poly(3,4ethylenedioxythiophene):poly(styrenesulfonate)/Si organic-inorganic hybrid solar cells,” Appl. Phys. Lett. 104(8), 083514 (2014). 15. D. Zielke, A. Pazidis, F. Werner, and J. Schmidt, “Organic-silicon heterojunction solar cells on n-type silicon wafers: the backpedot concept,” Sol. Energy Mater. Sol. Cells 131, 110–116 (2014). 16. D. Zielke, C. Niehaves, W. Lövenich, A. Elschner, M. Hörteis, and J. Schmidt, “Organic-silicon solar cells exceeding 20% efficiency,” Energy Procedia 77, 331–339 (2015). 17. Y. F. Chang, H. F. Meng, G. L. Fan, K. T. Wong, H. W. Zan, H. W. Lin, H. L. Huang, and S. F. Horng, “Blade coating of tris(8-hydroxyquinolinato)aluminum as the electron-transport layer for all-solution blue fluorescent organic light-emitting diodes,” Org. Electron. 29, 99–106 (2016). 18. J. Piao, S. Katori, T. Ikenoue, and S. Fujita, “Formation of aluminum tris(8‐hydroxyquinoline) solution in methanol and fabrication of thin films by ultrasonic spray‐assisted vapor deposition,” Phys. Status Solidi A 209(7), 1298–1301 (2012). 19. H. C. Card, “Aluminum—silicon Schottky barriers and ohmic contacts in integrated circuits,” IEEE Trans. Electron. Dev. 23(6), 538 (1976). 20. P. Yu, C. Y. Tsai, J. K. Chang, C. C. Lai, P. H. Chen, Y. C. Lai, P. T. Tsai, M. C. Li, H. T. Pan, Y. Y. Huang, C. I. Wu, Y. L. Chueh, S. W. Chen, C. H. Du, S. F. Horng, and H. F. Meng, “13% efficiency hybrid organic/siliconnanowire heterojunction solar cell via interface engineering,” ACS Nano 7(12), 10780–10787 (2013). 21. J. He, P. Gao, M. Liao, X. Yang, Z. Ying, S. Zhou, J. Ye, and Y. Cui, “Realization of 13.6% efficiency on 20 μm thick Si/organic hybrid heterojunction solar cells via advanced nanotexturing and surface recombination suppression,” ACS Nano 9(6), 6522–6531 (2015). 22. M. Ikai, S. Tokito, Y. Sakamoto, T. Suzuki, and Y. Taga, “Highly efficient phosphorescence from organic lightemitting devices with an exciton-block layer,” Appl. Phys. Lett. 79(2), 156 (2001). 23. L. Hou, L. Duan, J. Qiao, W. Li, D. Zhang, and Y. Qiu, “Efficient single layer solution-processed blue-emitting electrophosphorescent devices based on a small-molecule host,” Appl. Phys. Lett. 92(26), 263301 (2008). 24. T. Yasuda, Y. Yamaguchi, D. C. Zou, and T. Tsutsui, “Carrier mobilities in organic electron transport materials determined from space charge limited current,” Jpn. J. Appl. Phys. 41(9), 5626–5629 (2002). 25. K. L. Wang, B. Lai, M. Lu, X. Zhou, L. S. Liao, X. M. Ding, X. Y. Hou, and S. T. Lee, “Electronic structure and energy level alignment of Alq3/Al2O3/Al and Alq3/Al interfaces studied by ultraviolet photoemission spectroscopy,” Thin Solid Films 363(1-2), 178–181 (2000). 26. J. H. Lee, D. W. Moon, and Y. Yi, “The initial interface formation between Al and tris-(8-hydroquinoline) aluminum (Alq3) with LiF interlayer,” Org. Electron. 11(1), 164–168 (2010). 27. Y. T. Chang, J. K. Chang, Y. T. Lee, P. S. Wang, J. L. Wu, C. C. Hsu, I. W. Wu, W. H. Tseng, T. W. Pi, C. T. Chen, and C. I. Wu, “High-efficiency small-molecule-based organic light emitting devices with solution processes and oxadiazole-based electron transport materials,” ACS Appl. Mater. Interfaces 5(21), 10614–10622 (2013). 28. X. H. Yang, F. Jaiser, S. Klinger, and D. Neher, “Blue polymer electrophosphorescent devices with different electron-transporting oxadiazoles,” Appl. Phys. Lett. 88(2), 021107 (2006). 29. L. He, L. Duan, J. Qiao, D. Zhang, L. Wang, and Y. Qiu, “Highly efficient solution-processed blue-green to red and white light-emitting diodes using cationic iridium complexes as dopants,” Org. Electron. 11(7), 1185–1191 (2010). 30. B. V. Zehbroeck, Principles of Semiconductor Devices (Colarado University, 2004) Chap.3.7.


Introduction
According to the ITRPV roadmap, n-type monocrystalline silicon photovoltaics will gradually increase its market share due to the capability for high efficiency and reliability, particularly by excluding the boron-associated photon-induced degradation in p-type Si [1,2].Aligned with the reduced substrate thickness to save on the material cost, Panasonic Inc. has recently demonstrated a back-contact "HIT" cell with a record power conversion efficiency (PCE) of 25.6% on a 100 μm-thick n-type monocrystalline silicon wafer [3].The term "HIT" stands for "heterojunction with an intrinsic thin layer" of which the device structure is essentially an n-Si wafer sandwiched by wide-bandgap amorphous silicon (a-Si) materials for surface passivation and heterojunction formation.The processing temperature for HIT solar cells is less than 300 °C, which is advantageous to avoid wafer bowing and breakage issues commonly encountered for thin-silicon substrates.However, the cost of HIT solar cells is almost twice of the conventional devices due to the high-vacuum process of a-Si, which in turn limits the overall yield and throughput [4].In this regard, solution-processed hybrid organic-silicon heterojunction photovoltaics offers an even lower process temperature (<150 °C) and more cost-effective solutions than the inorganic counterpart for the realization of high-efficiency thin-film solar cells [5][6][7][8][9][10][11]. Since the absorption coefficient of silicon is an order of magnitude smaller than that of the direct-bandgap semiconductors, insufficient optical absorption in thin substrates gives rise to large rear surface recombination.Therefore, one of the key aspects to achieve high PCE relies on the rear surface processing.Over the past few years, several low-temperature approaches, including rear organic interlayers and back-PEDOT heterojunction concepts have been proposed [12][13][14][15][16].However, effective rear surface engineering via solution processes is still quite limited.Zhang et al. have demonstrated improved charge transport and PCE by introducing CsCO 3 (spin-coating), Liq (evaporation), or LiF (evaporation) between silicon and the aluminum cathode [12][13][14].These approaches essentially provide an n-type doping by releasing electrons from Cs or Li ions upon their chemical reaction with Al, giving rise to a weak back surface field at the rear surface.Despite the enhanced device characteristics, CsCO 3 can hardly be dispersed uniformly due to its large molecules, and small-molecule Liq or LiF cannot be easily dissolved in common solvents due to poor solubility, which also limits the scalability of hybrid organic-silicon devices.In this work, we investigate other organic interlayers based on two small molecules: Tris(8hydroxyquinolinato) aluminium (Alq 3 ) and 1,3-bis(2-(4-tert-butylphenyl)-1,3,4-oxadiazol-5yl)benzene (OXD-7) via a scalable blade coating technique.In particular, a stable and soluble Alq 3 process resulting in uniform and continuous thin film (the root-mean-square, RMS roughness < 0.2nm) is demonstrated for the first time [17,18].Either an Alq 3 or OXD-7 interlayer is next introduced into the rear surface of hybrid solar cells composed of poly (3,4ethylenedioxythiophene): poly(styrenesulfonate) (PEDOT:PSS) spun-casted on silicon nanowire (SiNW) templates.The photovoltaic characteristics versus the blade coating speed are systematically compared, where both approaches result in a >2% net efficiency gain, representing an enhancement factor of 21.6% and 20.2% from the reference devices, respectively.The capacitance-voltage characteristics reveal the role of the small-molecule interlayers similar to a thin interfacial oxide layer for the Al-Si Schottky barrier to increase the built-in potential and facilitate the charge transport [19].However, despite the notable characteristics improvement, the interfacial morphologies of the two materials in the best performing devices are very different.The optimized Alq 3 interlayer exhibits isolated phases with a large surface roughness, in contrast to the OXD-7, which forms a continuous uniform thin film.The distinct morphological differences between the two interlayers further suggest different enhancement mechanisms and may also be complementary to previously-reported interfacial doping materials.

Experimental
The general fabrication procedures of hybrid PEDOT:PSS-silicon solar cells have been described previously with slight variations on the preparation of silicon templates, organic surfactants, and metallization [7][8][9]20,21].Here, we started our process by preparing silicon nanowire templates using a metal-assisted electrode-less chemical etching method [20], where n-type monocrystalline Si (100) substrates with a thickness of 650 µm and a resistivity of 5 Ω-cm were employed.Next, the templates were dipped into a dilute hydrofluoric acid (DHF, HF:H 2 O = 1:10) solution for 15 sec to remove the native oxide and then moved into a nitrogen-purged glove box immediately.The rear organic interlayer was to be applied through a blade coating process.However, the small-molecule Alq 3 has been known to have very limited solubility due to its highly symmetric molecular structure; thus often requires thermal evaporation in high vacuum conditions at the cost of little material utilization [17].In our developed process, both Alq 3 and OXD-7 powders were purchased from Lumtec Corp..The powder was mixed with methanol as the organic solvent and then the solution was heated to 60°C for over 12 hours to completely dissolve the Alq 3 powder at the 0.1wt% and OXD-7 at the 0.2wt% concentration ratios.If a small portion of solids still precipitates, further heating to 80 °C and filtering through a Teflon filter help obtain a clear solution before the application.The silicon wafer undergone blade coating was first pre-heated at 60°C for 3 minutes to deplete the oxygen and water.Then, the blade with a groove spacing of 60 μm was mounted onto an automatically controlled stage, where the moving speed was adjusted to control the film thickness.At the beginning of the coating process, a 20 μL organic solution was first delivered from a micropipette and dropped underneath the blade, followed by blade movement at different speeds ranging from 25 to 250 mm/sec.Subsequently, hot wing was evenly blew over the surface for a quick dry off, resulting in a thin uniform film almost instantaneously.The film thickness was then measured by a microfigure measuring Instrument (ET200, Kosaka Laboratory Ltd.) with a resolution of 0.1nm in the vertical direction.After the blade-coating of the small-molecule interlayer on the rear surface, a 100-nmthick aluminum back electrode was deposited in a thermal evaporator.For the front-side process, the highly conductive aqueous PEDOS:PSS dispersion (Clevios PH1000) was prepared by mixing with 5 wt % dimethyl sulfoxide (DMSO) to enhance the conductivity and 0.3 wt % decaethylene glycol monododecyl ether (C12E10) as the surfactant.The PEDOT:PSS solution was spun-cast on top of the n-type silicon nanowire template at 8000 rpm for 100 sec, followed by hot-plate annealing at 115 °C for 10 min in ambient.Finally, 100-nm-thick silver grids as the frontal electrode were thermally evaporated through a shadow mask with a shading ratio of 18% and an opening area of 0.5 x 0.5 cm 2 .Figure 1(a) shows the schematic of a fabricated device structure.The upper inset on the right shows a cross-sectional scanning electron micrograph (SEM) of nanowire arrays covered with PEDOT:PSS.On the rear side of the device, a thin layer of small-molecule materials, either Alq 3 or OXD-7, is blade-coated to the back silicon surface prior to the deposition of the aluminum cathode.The chemical structures of Alq 3 and OXD-7 are provided aside, where elements are represented by different colors: C (gray) N (blue), and O (red). Figure 1(b) shows the flat energy bands of the device layers, where OXD-7 exhibits a relatively high lowest-unoccupied-molecular-orbital (LUMO, −2.6eV) and low highest-occupied-molecularorbital (HOMO, −6.3eV) levels, compared to those of Alq 3 (−3.0 and −5.7eV, respectively) [22,23].Since both Alq 3 and OXD-7 have also been widely employed as the electron transport materials (ETL) for organic light emitting and photovoltaic devices, the enabling solution processes of Alq 3 and OXD-7 can facilitate the scalability and deployment of organic optoelectronic devices considerably.Figure 1(c) shows the thickness dependence of the Alq 3 and OXD-7 thin films on the coating speed.It can be seen that a faster coating speed gives rise to a larger film thickness.Here, approximately a film thickness between 2 to 10 nm is achieved for the 0.1wt% Alq 3 and between 5 to 20nm for the 0.2wt% OXD-7 solutions.

Results and discussion
The statistical photovoltaic characteristics, including the short-circuit current density (J sc ), open-circuit voltage (V oc ), fill-factor (FF), and PCE, are plotted as a function of the blade coating speed for devices with the Alq 3 and with the OXD-7 interlayer as shown in Fig. 2(a) and 2(b), respectively.The reference counterpart without the rear interlayer is indicated at the zero speed.The full devices characteristics including the best and averaged values are provided Table 1 and Table 2, where the standard deviations are calculated by averaging four devices with a 95% confidence level.As shown in Fig. 2(a), the V oc , J sc , FF, and PCE, are all enhanced after the insertion of Alq 3 on the rear surface and peaked at the coating speed of 50 mm/sec, corresponding to an average layer thickness of 3nm in Fig. 1(c).As shown in Table 1, the averaged PCE of the hybrid solar cell with the Alq 3 interlayer is increased from the 9.7% without the interlayer, to 11.8%, mainly by boosting the averaged V oc from 533.0 to 553.2 mV and the FF from 63.3% to 67.8%.As the Alq 3 layer becomes thicker, the device characteristics are quickly deteriorated.Although in Fig. 2(b) the OXD-7 devices do not show as a strong dependence on the coating speed as the Alq 3 does, the interlayer also improves the device performance considerably.The best process condition occurred at 200 mm/sec, corresponding to a thickness of ~15nm in Fig. 1(c).The averaged PCE is increased from the 10.4% without the interlayer to 12.5% with the OXD-7 where the averaged V oc is improved from 536.4 to 557.2 mV and FF from 64.2% to 68.7%.In addition, the J sc of the OXD-7 devices is in general larger than the referenced devices possibly due to an increased built-in potential that improves the charge carrier collection, which will be discussed later.Similar to those with the Alq3 interlayer, the devices also start to degrade when the OXD-7 thin-film becomes thicker than 15nm.
We next examine the morphology of organic thin films using atomic force microscopy (AFM) to evaluate the film quality and investigate the enhancement mechanisms behind the two small-molecule materials.3(e)-3(h)) images of Alq 3 thin films resulting from a coating speed of 25, 50, 100, 200 mm/sec, respectively.The corresponding film thicknesses have been determined previously with a RMS roughness of 1.6, 2.5, 0.2, and 0.2nm.Based on the topography and phase diagrams, we indeed achieve a very continuous and uniform Alq 3 thin film using solution processes at a coating speed of 100 mm/sec, where the film thickness is as small as 5nm with a RMS roughness below 0.2nm.Such topographic specifications of the Alq 3 thin films can hardly be realized via thermal evaporation and hence is beneficial to the fabrication of organic devices at lower cost.However, by cross-referenced with the results in Fig. 2(a), the best Alq 3 devices occur at the 50 mm/sec coating speed, where the small-molecule material had yet formed a continuous network and still exhibits isolated phases.The depth differences between peaks and valleys can be as high as 25nm, suggesting that the rear aluminum electrode may still be partially in contact with the silicon, as well as the Alq 3 islands.On the other hand, Figs.The morphological discrepancy between the optimized Alq 3 and OXD-7 interlayers suggest different mechanisms in the FF and efficiency enhancement.First, the electron mobility of OXD-7 is approximately 2.1x10 −5 cm 2 /s•V, almost an order of magnitude larger than that of Alq 3 at 4.7x10 −6 cm 2 /s•V [24].Therefore, the OXD-7 can achieve optimized transport properties with a thicker film than the Alq 3 .The low electron mobility of Alq 3 also accounts for the low carrier conduction in the continuous thin film.In addition, it has been previously reported that thermally evaporated Al is diffusive into the Alq 3 , which results in an interface dipole about 0.8 eV [23,24].Giving the observation that in our devices Al may still partially contact with the Si surface and Alq 3 islands, the existence of the interfacial dipoles can locally lower the work function of the cathode without impeding charge transport.On the contrary, previous reports suggest that the chemical structure of OXD-7 has a lone pair of oxadiazole moiety which tends to grab hot metal ions upon thermal evaporation [27].The calculated Gibbs free energy is also negative, indicating that the chemical reaction between OXD-7 and Al is energetically favorable to generate metal complex [27].Therefor the formation of Al-oxadiazole network is capable to facilitate interfacial charge transfer by promoting carrier extraction through the reacted oxadiazole moiety.Compared to the reference, the OXD-7 devices show a much improved FF.Finally, the large ionization potential of OXD-7 also permits the interlayer to function as a hole-blocking layer (HBL) which could increase the minority carrier lifetime and boost the V oc for solar cells [28,29].The characteristics of the hybrid solar cell devices with the two rear interlayers can be further explored through a capacitance-voltage (C-V) profiling technique with a modulated frequency.Upon the voltage variation, carriers are swept in and out of the depletion regions to induce a junction capacitance change.Therefore, the C-V results are informative to understand mechanisms behind the physical property changes due to modifications of materials or device structures.First of all, a simplified hybrid solar cell structure contains two depletion regions.One is located at the PEDOT:PSS/Si interface and the other at the Si/Al rear interface.The junction capacitances associated with the two depletion regions are therefore in series connection.The hybrid solar cells presented in this work employ silicon nanowire templates, which are quite complex and have a indefinable junction area at the PEDOT:PSS/Si interface.As a result, the small junction capacitance would dominate the inverse square capacitance (1/C 2 )-voltage curve, compared to the Schottky junction capacitance [30].To avoid the complexity arising from the nanowire template and concentrate on the rear Al/Si Schottky barrier, we have prepared three planar devices without the rear interlayer (reference), as well as with the Alq 3 (50 mm/sec), and OXD-7 interlayers (200 mm/sec), as shown in the schematic inset of Fig. 5(a).3. It can be seen that the saturation current for both the Alq 3 and OXD-7 devices exhibit a ten-fold reduction, compared to the reference counterpart.Moreover, the measured V oc is gained by additional 33.2 mV (Alq 3 ) and 60.1 mV (OXD-7) compared to the reference without the interlayer, as shown in Fig. 5(b) and Table 3.The C-V measurement from −1V to a small positive voltage is next conducted with a modulation frequency of 10kHz for the three devices, which show the characteristics of Schottky diodes [19].The effective junction capacitance at the zero bias is about 10nF/cm 2 .Figure 5(c) plots the 1/C 2 versus the applied voltage for the three devices, which exhibits a liner dependence on the voltage at small reverse biases.The slope in a 1/C 2 -V plot for a Schottky diode is in general inversely proportional to the doping concentration of semiconductors and the intersection with the x-axis determines the built-in potential across the Schottky barrier.As shown in Fig. 5(c), the measure junction capacitance of the Alq 3 and OXD-7 devices are both smaller than the reference without significant slope changes, which indicates an additional capacitance in series with the junction capacitance [30].The results are also different from the previous approaches with a rear back surface field, where the rear interface doping increases the junction capacitance and reduces the slope in a 1/C 2 -V plot considerably and due to the inverse doping dependence [12,13,26].Therefore, our results suggest that the role of the Alq 3 or OXD-7 interlayer is analogous to a thin interfacial oxide layer for the Al/Si Schottky barrier, instead of the interfacial doping [30].In inorganic semiconductor devices, the existence of an interfacial oxide layer reduces the Schottky barrier height and increase the built-in potential, leading to a higher current at the turn-on condition than that without the interfacial oxide [19,30].Here, as shown in Fig. 5(c), the built-in potential, determined by the intersection at the x-axis, is indeed increased by 40 mV (Alq 3 ), and 60 mV (OXD-7), respectively, which agree with the enhancement of V oc shown in Fig. 5(b) very well.Finally, the incorporation of interfacial doping materials, such as LiF, Liq, and CsCO 3 , with either Alq 3 or OXD-7 have also been investigated for organic light emitting diodes and demonstrates improvements previously [21,22].We think that the proposed rear interfacial engineering via the solution process of Alq 3 and OXD-7 may also benefit from similar approaches.

Conclusion
In summary, we successfully demonstrate soluble processes for two small-molecules: Alq 3 and OXD-7 for the rear interfacial engineering of hybrid organic-silicon nanowire solar cells using a blade coating technique.The photovoltaic characteristics of the Alq 3 and OXD-7 devices exhibit enhanced open-circuit voltages and fill-factors, where the averaged power conversion efficiency achieves 11.8% and 12.5%, respectively, both corresponding to a net efficiency gain by over 2% compared to the reference devices.Through capacitance-voltage characterizations, the two rear interlayers play a role analogous to a thin interfacial oxide layer for the Al/Si Schottky barrier to increase the built-in potential and facilitate charge transport.Moreover, the distinct morphologies between the two interlayers further suggest different enhancement mechanisms, and hence can offer versatile functionalities to the advent of hybrid organic-silicon photovoltaics.

Fig. 1 .Fig. 2 .
Fig. 1.(a) Schematic of a fabricated hybrid organic-silicon nanowire solar cell incorporating a rear organic interlayer.The upper inset is a cross-sectional micrograph of the conductive polymer, PEDOT:PSS laying over the silicon nanowire template, where the scale bar is 100 nm.The lower inset shows the chemical structures of the two small-molecule materials Alq 3 and OXD-7, where elements are represented by different colors: C (black), N (blue), O (red).(b) Energy band diagram of a hybrid PEDOT:PSS silicon solar cell with a Alq 3 or OXD-7 rear interlayer.(c) The thickness of the blade-coated Alq 3 and OXD-7 layer as a function of the blade coating speed.

Figures 5 (
Figures 5(a) and 5(b) respectively plot the dark and light current density-voltage characteristics for the three planar devices.The photovoltaic characteristics under a simulated AM1.5g illumination condition is summarized in Table3.It can be seen that the saturation current for both the Alq 3 and OXD-7 devices exhibit a ten-fold reduction, compared to the reference counterpart.Moreover, the measured V oc is gained by additional 33.2 mV (Alq 3 ) and 60.1 mV (OXD-7) compared to the reference without the interlayer, as shown in Fig.5(b) and Table3.The C-V measurement from −1V to a small positive voltage is next conducted with a modulation frequency of 10kHz for the three devices, which show the characteristics of Schottky diodes[19].The effective junction capacitance at the zero bias is about 10nF/cm 2 .Figure5(c) plots the 1/C 2 versus the applied voltage for the three devices, which exhibits a liner dependence on the voltage at small reverse biases.The slope in a 1/C 2 -V plot for a Schottky diode is in general inversely proportional to the doping concentration of semiconductors and the intersection with the x-axis determines the built-in potential across the Schottky barrier.As shown in Fig.5(c), the measure junction capacitance of the Alq 3 and OXD-7 devices are both smaller than the reference without significant slope changes, which indicates an additional capacitance in series with the junction capacitance[30].The results are also different from the previous approaches with a rear back surface field, where the rear interface doping increases the junction capacitance and reduces the slope in a 1/C 2 -V plot considerably and due to the inverse doping dependence[12,13,26]. Therefore, our results suggest that the role of the Alq 3 or OXD-7 interlayer is analogous to a thin interfacial oxide layer for the Al/Si Schottky barrier, instead of the interfacial doping[30].In inorganic semiconductor devices, the existence of an interfacial oxide layer reduces the Schottky barrier height and increase the built-in potential, leading to a higher current at the turn-on condition than that without the interfacial oxide[19,30].Here, as shown in Fig.5(c), the built-in potential, determined by the intersection at the x-axis, is indeed increased by 40 mV (Alq 3 ), and 60 mV (OXD-7), respectively, which agree with the enhancement of V oc shown in Fig.5(b) very well.Finally, the incorporation of interfacial doping materials, such as LiF, Liq, and CsCO 3 , with either Alq 3 or OXD-7 have also been investigated for organic light emitting diodes and demonstrates improvements previously[21,22].We think that the proposed rear interfacial engineering via the solution process of Alq 3 and OXD-7 may also benefit from similar approaches.

Table 1 . Best and Averaged Photovoltaic Characteristics of Hybrid Heterojunction Solar Cells with and without the Rear Alq 3 Interlayer at Different Blade Coating Speeds a
The statistics are calculated by averaging four devices with a 95% confidence level. a

Table 2 . Best and Averaged Photovoltaic Characteristics of Hybrid Heterojunction Solar Cells with and without the Rear OXD-7 Interlayer at Different Blade Coating Speeds a
aThe statistics are calculated by averaging four devices with a 95% confidence level.