Efficiency enhancement of organic solar cells using ...

RSC Advances View Article Online

Published on 04 February 2015. Downloaded by National Taiwan University on 12/04/2016 05:43:15.

PAPER

Cite this: RSC Adv., 2015, 5, 18990

View Journal | View Issue

Efficiency enhancement of organic solar cells using peroxo-polytitanic acid coated silver nanowires as transparent electrodes Wei-Feng Xu,†ab Ming-Chih Tsai,†ac Po-Han Fu,b Tzu-Yen Huang,ad Shang-Jung Yang,af Wei-Cheng Tian,e Chih-Wei Chu,a Ding-Wei Huangb and Pei-Kuen Wei*af Solution-processed transparent electrodes made from silver-nanowires and peroxo-polytitanic (PPT) acid gel were developed for enhancing efficiency of organic solar cells. The electronic and optical properties of multilayer silver nanowires were significantly improved through the interconnection layers of amorphous titanium oxide (TiOx) from PPT acid gel. The AgNW–TiOx composite film showed averaged 92.14%

Received 7th December 2014 Accepted 2nd February 2015

optical transmittance with only 16.01 U sq

1

sheet resistance. Combining the transparent electrodes with

poly(3-hexylthiophene):[6,6]-phenyl-C61-butyric acid methyl ester (P3HT:PC61BM) blend, the efficiency of organic solar cells was 1.45 times that of devices using an ITO electrode. Optical simulations verified

DOI: 10.1039/c4ra15915b

that the improvement was attributed to the enhanced near-field absorption and substantial scattering of

www.rsc.org/advances

incident light resulted from the random nature of the AgNW–TiOx core–shell nanostructure.

Introduction Polymer solar cells have attracted much attention due to the advantages of easy fabrication, larger-scale production, and good compatibility with exible substrates.1–3 For the exible devices, one of the most important issues is the exploitation of a transparent conducting electrode (TCE) with high transparency and low sheet resistance to replace the extensively used indium-tin oxide (ITO), which its fragility, inexibility, and scarcity of indium limit the applications for exible organic solar cells.4,5 Many possible candidate electrode materials, such as carbon nanotubes,6–8 graphene,9–11 conductive polymers,12–14 metal grids,15 and metal thin lms16,17 have been investigated. However, the power conversion efficiencies (PCE) of devices with these alternative electrodes still cannot compete or even outstrip that of solar cells using an ITO electrode. The major factor for the lower PCE arises from the compromise between optical transparency and electric conductivity of the transparent electrodes. a

Research Center for Applied Sciences, Academia Sinica, Taipei 11529, Taiwan. E-mail: [email protected]; Fax: +886-2-2787-3146; Tel: +886-2-2787-3122

b

Graduate Institute of Photonics and Optoelectronics, National Taiwan University, Taipei 10617, Taiwan

c Graduate Institute of Biomedical Electronics and Bioinformatics, National Taiwan University, Taipei 10617, Taiwan d

Department of Chemical Engineer, National Taiwan University, Taipei 10617, Taiwan

e

Graduate Institute of Electronics Engineering, National Taiwan University, Taipei 10617, Taiwan

f

Institute of Optoelectronics Sciences, National Taiwan Ocean University, Keelung 20224, Taiwan † These authors contributed equally to this work.

18990 | RSC Adv., 2015, 5, 18990–18996

Recently, a solution-processed silver nanowires (AgNWs) mesh electrode was developed for bilayer solar cells with a comparable efficiency to that of cells with a commonly used metal-oxide electrode.18 However, the higher surface roughness of the AgNWs electrode leads to the possibility of inter-electrode shorting and thus poor device efficiency of less than 0.5% due to a higher leakage current. To overcome this problem, a thick layer (200 nm) of nano-particulate titania (TiOx) buffer layer was coated on AgNWs to reduce the nanowire rough surface.19 An improved PCE up to 3.5% was achieved due to a benecial reduction of dark leakage current and a signicantly increased photocurrent. The AgNWs with TiOx nanoparticles layer achieved 83% transparency at 550 nm and 15 U sq 1 sheet resistance. Its performance is comparable to their ITO counterpart.20 In this work, we developed a low cost-effective and convenient approach to fabricate ITO-comparable transparent electrodes by using a solution process of silver nanowires mixed with peroxopolytitanic (PPT) acid gel. Different from the TiOx nanoparticles coating, the AgNW–PPT form core–shell nanostructures. The electrode has a higher transmittance of 93.7% at 550 nm, while remaining a low sheet resistance of 16.01 U sq 1. With the advantage of enhanced near-eld absorption and substantial scattering of incident light resulted from the random nature of the AgNW–TiOx core–shell nanostructure, we achieved an average PCE of 4.69% in inverted polymer solar cells with this type of electrode. The efficiency is 36.7% improvement relative to that of reference devices with ITO electrode. Moreover, the composite electrode under mild thermal annealing can further reach a highest PCE of up to 5.05% due to an enhanced photocurrent generation resulting from the increased effective

This journal is © The Royal Society of Chemistry 2015

View Article Online

Paper

optical path length by the scattering from the annealinginduced Ag clusters.

Experimental section


Device fabrication Inverted polymer solar cells were fabricated in a layer structure of glass/cathode/zinc oxide (ZnO)/poly(3-hexylthiophene):[6,6]phenyl-C61-butyric acid methyl ester (P3HT:PC61BM)/ molybdenum oxide (MoOx)/Ag (cathode: AgNWs with or without PPT, ITO (Merck Inc., 24 U sq 1) as the reference device). Silver nanowires (1.0 wt% in ethanol) with a diameter of 35 nm and length of 15–25 mm were purchased from Blue Nano Inc. Aer dilution to 0.2 wt%, the AgNW dispersion was spin-coated onto pre-cleaned glass substrates at 2000 rpm for 30 s. A yellow transparent PPT acid gel was prepared by mixing 60 ml of titanium isopropoxide (TTIP) (Sigma-Aldrich, 97%), 430 ml of hydrogen peroxide (H2O2) (Acros Organic, 35 wt%), and 9.51 ml of H2O, followed by continuously stirring for 48 hours at room temperature (RT). The resulting PPT acid gel was diluted in isopropyl alcohol to obtain the PPT coating solution. The PPT coating solution was then spin-cast onto the AgNW-coated glass substrates at 2000 rpm for 60 s, followed by air drying for 2 min. To match with the lowest unoccupied molecular orbital (LUMO) of PC61BM and obtain a smooth the electrode surface, a ZnO layer was coated on the AgNW layer. The layer was made by a 0.46 M of ZnO precursor solution dissolving equimolar zinc acetate dehydrate (Alfa Aesar, 98%) and monoethanolamine (Sigma-Aldrich, 99%) in 2-methoxy ethanol (Sigma-Aldrich, 99.8%). The prepared ZnO precursor solution was spin-coated onto the as-deposited AgNW–TiOx composite lm at 5000 rpm for 60 s. The amorphous ZnO lms were then air annealed at 200 C for 1 h. Subsequently, the active layer was cast from a solution containing 1.75 wt% P3HT:PC61BM (Uni-Onward Corp., >99.5%) (1 : 1, w/w) in 1,2-dichlorobenzene at 800 rpm for 60 s inside N2 lled glove box. The photovoltaic layer was le in a glass Petri dish at RT for 60 min for slow solvent evaporation, and then thermally annealed at 170 C for 10 min to obtain a 220 nm-thick active layer. The devices were completed aer thermal evaporation of MoOx (10 nm) and Ag (100 nm) through a shadow mask under pressure of 6 10 6 Torr, giving an active area of 0.04 cm2 dened by the overlap between the patterned AgNW (or ITO) and top Ag contact. We dened the pattern of silver nanowire electrode by rst sticking Scotch tape onto the AgNWs coated glass substrate and then removing the redundant part by tearing away the tape. The ITO electrode was made by chemically etching ITO glass through a tape mask with acid and alkali resistance. The AgNW electrode and ITO had the same 2 mm-wide conductive strip. For the ITO reference device, all processing conditions, except for not using TiOx, were kept the same as for the AgNW-based one. Optical and electrical characterization The current density versus voltage (J–V) characteristics of devices were performed within a N2-lled glove box through a semiconductor parameter analyzer (HP 4145C) in the dark and under


RSC Advances

a simulated AM 1.5G spectrum provided by a Xe lamp-based solar simulator (Thermal Oriel 1000W) with a light intensity of 100 mW cm 2, calibrated through a silicon photodiode. The transmittance and reectance spectra were recorded using an ultraviolet-visible spectrophotometer (Jasco V-550) with an integrating sphere reectance accessory. The sheet resistance was measured using the four-point probe system and the average value was taken from at least 5 different locations. The lm thickness was evaluated using a V-VASE variable-angle spectroscopic ellipsometer (J. A. Woollam Inc.). The optical modeling of the devices was performed using the nite element method (FEM) with the use of the soware Comsol Multiphysics. The surface roughness and topography of electrodes were imaged by trapping mode atomic force microscopy (AFM) (Veeco di Innva SPM), and the surface morphology images were investigated by scanning electron microscopy (SEM) (LEO-1530 FEG SEM).

Results and discussion Optical, electrical properties and surface morphology Fig. 1(a) shows a schematic diagram for the fabrication of AgNW–TiOx composite lms. First, the AgNWs were spin-cast on the pre-cleaned glass substrate and presents a loosepacked structure. Aer spin-coating of the PPT coating solution onto the AgNWs and subsequently drying to remove the residual solvent at RT, the peroxo-polytitanic acid gel covers the surface of nanowires and distributes randomly in the openings of the crossed network. This gives rise to meniscus nanostructures around crossed AgNW positions and between the substrate and AgNWs due to capillary forces from solvent evaporation. The resulting AgNW–TiOx core–shell nanostructures facilitate the enhancement of contacts between AgNWs and the adhesion of AgNWs to the underlying glass substrate. The thickness of TiOx lm was analyzed using ImageJ soware from a TEM image of the selected area (Fig. 1(b)) and determined to be 1.82 0.25 nm. In Fig. 2, the sheet resistance and optical transmittance was investigated by measuring four different transparent conducting lms: bare AgNW mesh, AgNW–TiOx composite lm without thermal annealing, AgNW–TiOx (*) with thermal annealing at 150 C under vacuum for 10 min, and ITO. The bare AgNW mesh showed 22.99 2.02 U sq 1 sheet resistance and 91.77% transmission at 550 nm, as shown in Fig. 2(b). Upon lamination of TiOx thin lm onto AgNWs, the sheet resistance was substantially reduced to 16.01 0.81 U sq 1 due to fusion at the contact of the nanowires which led to the reduction of contact resistance between the nanowires, and the transmittance was increased to 93.7% at 550 nm. This is attributed to the plasmonic effect of high refractive index (>2) TiOx ller placed in the empty space in the crossed AgNW network. This extraordinary optical transmission improvement is seen near the 350 nm wavelength corresponding to the plasmonic absorption of the AgNWs, as can be seen in Fig. 2(a). Aer being treated with TiOx sol–gel solution and further annealing, the resulting annealed AgNW composite lms have an average transmittance of 89.21% from 350 to 800 nm with sheet resistance of 15.79 0.15 U sq 1, which met the RSC Adv., 2015, 5, 18990–18996 | 18991

View Article Online


RSC Advances

Paper

Fig. 1 (a) Schematic diagram of the procedures for fabricating AgNW–TiOx composite film. The AgNW–TiOx form a core–shell nanostructure. (b) TEM image of AgNW–TiOx composite film.

requirements for transparent conducting electrode of OPVs. The additional thermal annealing would further result in the formation of silver fusion at the contact of nanowires. The silver fusion formed larger Ag clusters and thus increased the

(a) Transmittance spectra of bare AgNW mesh, AgNW–TiOx composite film without and with thermal annealing, and ITO film (not including glass substrate losses). The inset shows a photograph of transparent conducting films on glass substrate. (b) The relationship between sheet resistance and transmittance at 550 nm. Fig. 2

18992 | RSC Adv., 2015, 5, 18990–18996

scattering and absorbance of the incoming light. Therefore, the optical transmittance was decreased aer thermal treatment. The sheet resistance of the ITO lm was 24.05 0.04 U sq 1 and the transmittance was 89.11% at 550 nm. In order to verify the composite lm suitable for transparent conducting electrode of device, the surface morphology and topography of AgNW-based lm was investigated by SEM and AFM. Fig. 3 shows the SEM images of a pristine AgNW network and AgNW–TiOx composite lm with thermal annealing. The SEM image of Fig. 3(a) indicates that the AgNWs formed a uniform mesh with 28% cover ratio area analyzed by ImageJ soware across the substrate. Upon lamination of TiOx onto nanowires, most of the TiOx lm is centered largely in the crossnanowire junctions which help strengthen the electrical contact between the nanowires and along edges of the nanowires, however, a few are randomly distributed in the openings between nanowires, as seen in the Fig. 3(b). Due to the lack of a supporting layer to ll the voids between neighboring nanowires, a further coating layer of amorphous ZnO onto AgNW– TiOx composite lm is necessary for charge extraction and transportation and surface smoothing. Fig. 3(c) and (d) show that the AgNW and composite networks are completely buried in the ZnO material and the sharp edge of the original AgNW image appeared blurred. We also examined the surface topography using AFM because the surface roughness of the electrodes is closely interrelated with the cell-performance. Fig. 4 shows an AFM topographical image of bare AgNW mesh and AgNW–TiOx composite lm with thermal annealing before and aer ZnO coating. The fabrication process of bare AgNWs lm was conducted by repeatedly spin-coating the AgNW solution on the glass substrate followed by drying at RT. Due to the incomplete removal of solvent at RT, the bare AgNWs lm had a loose-packed structure. Aer thermal annealing, the lm volume shrunk due to solvent evaporation and eventually formed a compact lm structure. The annealing process would result in Ag clusters at the contact of nanowires, which increased the roughness at localized area. For the case of


View Article Online


Paper

RSC Advances

Fig. 3 SEM images for (a) uncoated and (c) ZnO-coated bare AgNW mesh and for (b) uncoated and (d) ZnO-coated AgNW–TiOx composite film with thermal annealing. A 52 tilted SEM image was also shown in the inset.

AgNW–TiOx composite, the TiOx lled at the openings between the crossed networks of AgNWs. It substantially decreased the surface roughness. Therefore, the root mean square (RMS) roughness of AgNW–TiOx composite lm was signicantly reduced to 23.52 nm relative to the surface roughness of the

bare AgNW network (36.29 nm). The average height of the stacked nanowires (64.6 nm) decreased to half of that of the bare nanowire mesh (121.44 nm), which conrms the functionality of surface smoothing of TiOx, as seen in Fig. 4(a) and (b). However, there still are sharpe edges and ridges observed in

Fig. 4 AFM topography images for (a) uncoated and (c) ZnO-coated bare AgNW mesh and for (b) uncoated and (d) ZnO-coated AgNW–TiOx

composite film with thermal annealing. The height profiles shown in insets are taken along the white double-headed arrows.


RSC Adv., 2015, 5, 18990–18996 | 18993

View Article Online

RSC Advances


the height prole of insets. Aer being covered by a ZnO layer, the RMS and the average height of composite lms was further reduced to 15.42 nm and 52 nm (Fig. 4(d)). Though the RMS value of AgNW-based lm is relatively high compared to the ITO surface, the use of 220 nm thick active layer can effectively decrease the possibility of inter-electrode shorting and thus increase the device yield. Device performance To demonstrate the application of AgNW-based composite lm in polymer solar cells, they were incorporated into devices with an inverted device conguration, as seen in the inset of Fig. 5(a). The reference device was made with the commercially available ITO-coated (100 nm) glass substrate as the transparent electrodes. The ZnO layer inserted in between the cathode and the active layer can serve as a buffer layer to smooth out the roughness, to match the work function and provide efficient charge extraction and transportation. The J– V characteristics of the solar cells are shown in Fig. 5(a). The device with AgNW–TiOx composite electrode exhibited an open circuit voltage (Voc) of 0.58 0.01 V, short circuit current

Paper

density (Jsc) of 13.32 0.04 mA cm 2, ll factor (FF) of 61.23 0.41%, and power conversion efficiency (PCE) of 4.69 0.05%, while the ITO reference device showed an Voc of 0.58 0.01 V, Jsc of 10.95 0.06 mA cm 2, FF of 54.41 1.00%, and PCE of 3.43 0.07%. The 36.7% improvement of PCE relative to an ITO device is due to the increasing Jsc and FF. The increasing Jsc can be attributed to the light scattering effect and plasmon-enhanced absorption by highly localized neareld of randomly arranged silver nanostructure. The enhancement of FF comes from the reduced series resistance due to the increased contact area between the AgNW network and ZnO. In addition, the device with AgNW–TiOx core–shell electrode and additional thermal annealing further showed an increase of the PCE to 4.98% 0.11% as a result of Jsc enhancement from 13.32 to 13.91 0.93 mA cm 2. The Voc remains the same for AgNW-based and ITO device, which suggest that this newly developed AgNW/ZnO behaves as a good contact to the photoactive layer as that for ITO/ZnO, resulting in no contact-related photovoltage loss.21 The efficiency of the most efficient device is 5.05% with an Voc of 0.58 V, Jsc of 14.31 mA cm 2, and FF of 60.83%, which is the highest efficiency value reported for OPVs with P3HT:PC61BM active layer using AgNW electrodes to our knowledge. The obtained photovoltaic performance data are summarized in Table 1. In order to investigate the origin of the Jsc enhancement, we measured the total reectance spectra of devices with AgNWbased and ITO electrode illuminated from the substrate side, and then calculated the amount of light absorption in the cell, as seen in Fig. 5(b). It can be seen that the devices with AgNWbased electrodes can capture more light than that of ITO devices at l > 450 nm, reasonably deducing the efficient conversion of photocurrent at these wavelength range.

Simulations

Fig. 5 (a) J–V characteristics of the inverted OPV devices with different electrode configuration as cathode. The device configuration is also shown in the inset. (b) Measured total absorption spectra of the OPV devices using 1 T R, where T is zero in the devices. The absorbance spectra of 220 nm thick P3HT:PC61BM film on the glass was also shown.

18994 | RSC Adv., 2015, 5, 18990–18996

To clarify the enhanced device absorption behavior originating from the combined effect of light scattering and plasmonic absorption of AgNW–TiOx core–shell nanostructure, we also conducted numerical simulations of the external quantum efficiency (EQE) spectra of the devices with AgNW–TiOx and ITO electrode (reasonably assuming 70% of internal quantum efficiency). In Fig. 6(a), the simulated EQE of a device with a composite electrode was signicantly enhanced over a wide spectral region from 450 to 650 nm, and the maximum EQE value increased from 52% to 66.87% at 570 nm relative to that of an ITO device, which is consistent with the measured result in the inset of Fig. 6(a). For the transverse magnetic wave (TM) illumination, there is an observed dip ascribed to the absorption of Ag near 400 nm due to the excitation of localized surface plasmon resonance (LSPR); for the transverse electric wave (TE) illumination, there is no LSPR effect and the EQE spectrum shows a broad, rather at shape of the spectral response at visible light range. Finally, the simulated power absorbed per unit volume of the overall device with AgNW–TiOx and ITO electrode at 570 nm for TM polarization is shown in Fig. 6(b). A comparison between these two contour plots clearly


View Article Online

Paper Table 1

Device performance parameters of the OPVs under 100 mW cm

Device

Voc

AgNW AgNW–TiOx AgNW–TiOx (*)b ITO

0.58 0.58 0.58 0.58

a


RSC Advances

The best-performing cell.

b

0.01 (0.58)a 0.01 (0.57) 0.01 (0.58) 0.01 (0.57)

2

AM 1.5G illumination

Jsc (mA cm 2)

PCE (%)

FF (%)

12.47 0.24 (12.46) 13.32 0.04 (13.6) 13.91 0.93 (14.31) 10.95 0.06 (11.01)

4.45 4.69 4.98 3.43

61.86 2.17 (64.05) 61.23 0.41 (60.94) 61.49 2.28 (60.83) 54.41 1.00 (55.71)

0.14 (4.63) 0.05 (4.72) 0.11 (5.05) 0.07 (3.5)

Composite lm under thermal annealing at 150 C for 10 min.

demonstrates stronger power absorbed in the photoactive layer was obtained in the nanowire device rather than in the ITO device. Moreover, for the nanowire device, the maximum absorption concentrate underneath the protrusion of Ag nanowire, which can be attributed to the LSPR effect and thus lead to enhancement of the photocurrent. In contrast, the referenced ITO device did not produce a signicant plasmonic eld enhancement.

Conclusions In summary, we have successfully improved the optical and electrical performance of AgNW-based transparent electrodes by using an interconnection layer of TiOx from PPT acid gel. With a further coating layer of amorphous ZnO onto AgNW– TiOx composite lm, signicant enhancement of the solar cell performance with a PCE of 4.69% was demonstrated as compared with 3.43% of an ITO reference device. The efficiency enhancement is attributed to the increased FF and Jsc originating from surface plasmon resonances that lead to a strong near eld enhancement as well as highly efficient light scattering as veried by the optical simulation. The applications of this newly developed TCE on OPVs offer great potential for the fabrication of high performing plastic solar cells.

Acknowledgements This work was supported by the National Science Council, Taipei, Taiwan, under Contact no. 100-2221-E-001-010-MY3 and nanoscience program of Academia Sinica. We acknowledge the optoelectronic device fabrication and testing by the scientic instrument facility at Institute of Chemistry. The tilted SEM image was supported by dual-beam focus ion beam (DBFIB) core facility in Research Center for Applied Sciences, Academia Sinica.

Notes and references

Fig. 6 Numerical simulation results showing (a) the simulated external quantum efficiency (EQE) spectra of the devices with AgNW–TiOx and ITO electrode and (b) the power absorbed per unit volume in the cells at normal incident wavelength of 570 nm for the TM polarization. The measured EQE spectra are shown in the inset of (a).


1 P. Peumans and S. R. Forrest, Appl. Phys. Lett., 2001, 79, 126– 128. 2 H. F. Dam and F. C. Krebs, Sol. Energy Mater. Sol. Cells, 2012, 97, 191–196. 3 R. Roesch, F. C. Krebs, D. M. Tanenbaum and H. Hoppe, Sol. Energy Mater. Sol. Cells, 2012, 97, 176–180. 4 D. R. Cairns, R. P. Witte, D. K. Sparacin, S. M. Sachsman, D. C. Paine, G. P. Crawford and R. R. Newton, Appl. Phys. Lett., 2000, 76, 1425–1427. 5 G. F. Wang, X. M. Tao and R. X. Wang, Nanotechnology, 2008, 19, 145201–145205. 6 E. M. Doherty, S. De, P. E. Lyons, A. Shmeliov, P. N. Nirmalraj, V. Scardaci, J. Joimel, W. J. Blau, J. J. Boland and J. N. Coleman, Carbon, 2009, 47, 2466–2473. 7 D. S. Leem, S. Kim, J. W. Kim, J. I. Sohn, A. Edwards, J. S. Huang, X. H. Wang, J. J. Kim, D. D. C. Bradley and J. C. deMello, Small, 2010, 6, 2530–2534.

RSC Adv., 2015, 5, 18990–18996 | 18995

View Article Online


RSC Advances

8 S. Kim, J. Yim, X. Wang, D. D. C. Bradley, S. Lee and J. C. Demello, Adv. Funct. Mater., 2010, 20, 2310–2316. 9 K. S. Kim, Y. Zhao, H. Jang, S. Y. Lee, J. M. Kim, K. S. Kim, J. H. Ahn, P. Kim, J. Y. Choi and B. H. Hong, Nature, 2009, 457, 706–710. 10 G. Eda and M. Chhowalla, Adv. Mater., 2010, 22, 2392–2415. 11 P. H. Wobkenberg, G. Eda, D. S. Leem, J. C. de Mello, D. D. C. Bradley, M. Chhowalla and T. D. Anthopoulos, Adv. Mater., 2011, 23, 1558–1562. 12 P. A. Levermore, R. Jin, X. H. Wang, L. C. Chen, D. D. C. Bradley and J. C. de Mello, J. Mater. Chem., 2008, 18, 4414–4420. 13 S. I. Na, S. S. Kim, J. Jo and D. Y. Kim, Adv. Mater., 2008, 20, 4061–4067. 14 J. Huang, X. Wang, A. J. deMello, J. C. deMello and D. D. C. Bradley, J. Mater. Chem., 2007, 17, 3551–3554.

18996 | RSC Adv., 2015, 5, 18990–18996

Paper

15 B. Y. Ahn, D. J. Lorang and J. A. Lewis, Nanoscale, 2011, 3, 2700–2702. 16 H. M. Stec, R. J. Williams, T. S. Jones and R. A. Hatton, Adv. Funct. Mater., 2011, 21, 1709–1716. 17 W. F. Xu, C. C. Chin, D. W. Hung and P. K. Wei, Sol. Energy Mater. Sol. Cells, 2013, 118, 81–89. 18 J. Y. Lee, S. T. Connor, Y. Cui and P. Peumans, Nano Lett., 2008, 8, 689–692. 19 D. S. Leem, A. Edwards, M. Faist, J. Nelson, D. D. Bradley and J. C. de Mello, Adv. Mater., 2011, 23, 4371–4375. 20 R. Zhu, C. H. Chung, K. C. Cha, W. Yang, Y. B. Zheng, H. Zhou, T. B. Song, C. C. Chen, P. S. Weiss, G. Li and Y. Yang, ACS Nano, 2011, 5, 9877–9882. 21 P. Sehati, S. Braun, L. Lindell, X. J. Liu, L. M. Andersson and M. Fahlman, IEEE J. Sel. Top. Quantum Electron., 2010, 16, 1718–1724.
