We show experimentally and theoretically enhancement of external quantum efficiency in the green-NIR spectrum for organic photovoltaic device, by the incorporation of patterned Au nano-disk arrays that extend from the front electrode into the active layer. Enhancement mechanisms and design rules are extracted by comprehensive simulations which match the experimental findings. The enhanced efficiency is shown to stem from field enhancement originating from both localized plasmonic resonances and periodic nano patch antennas configuration.
© 2010 OSA
Organic photovoltaic devices (OPV), especially bulk hetero-junction (BHJ) cells, have gained much attention due to their potential for low-cost and high throughput production [1–3]. However, the efficiency of OPV cells is still low, up to 5-6% . In order to achieve high efficiencies the device is kept very thin to reduce recombination losses, but must be optically thick to absorb most of the incoming light. Several methods have been suggested to overcome this limitation: tandem cells , folded structures , and new active materials [7,8].
An alternative approach is the exploitation of localized plasmon resonance of metallic nano-particle (MNP) arrays . Utilizing MNP to increase solar cell efficiency has been demonstrated in various thin film technologies [10–12] and in silicon [13,14]. In organic cells an increase in photocurrent was exhibited by incorporating either a solution of randomly distributed small MNP in the buffer layer of P3HT:PCBM cells [15,16], or as a thin layer of Ag in the buffer layer , or as a replacement of the ITO electrode by perforated Ag layer . Colloidal metal particles were reported also inside the active layer of a MEH-PPV:PCBM cell . However, while in inorganic cells good correlation between experimental results and theoretical-simulation models has been shown, investigations in OPV cells typically involve randomly distributed particles with size typically below 30nm which also exhibit random clustering rates, making it difficult to trace and optimize enhancement mechanisms using appropriate simulation models. The employment of small particles presents an additional drawback: allowing maximum resonance wavelength of ~450nm and relatively large absorption in the metal. The benefit of using larger sized silver prisms was recently shown using photo-induced absorption spectroscopy . In order to maximize the effect of the MNP on the cell efficiency, tuning the properties of the plasmon resonances and interactions is desirable. Thus the strategy of our study is based on a precisely controlled fabrication method using electron beam lithography (EBL) enabling a tight control of MNP array parameters such as particle size, shape, aspect ratio, and array period, all are affecting the resonance properties [21–23], allowing us to study the enhancement mechanisms and develop design guidelines. Upon success, mass fabrication techniques such as nano-imprinting can be used for commercial cell fabrication. Nanoimprint lithography, based on stamps prepared by EBL, preserves the fine accuracy of EBL, but can be used by a step and repeat procedure to cover the large areas of solar cells in a massive, cheap and high throughput process. As a first step, 3D Finite difference time domain (FDTD) simulations of the entire cell were performed (discussed later) to analyze and optimize the MNP array parameters; subsequently the optimized parameters were implemented experimentally with good correlation to the predicted performance. A schematic illustration of the typical cell is shown in Fig. 1 .
2. Sample fabrication
The fabrication of the solar cell starts by patterning the Au nano particle array on the front electrode of the device. The array is produced on a glass substrate covered with a 100nm indium tin-oxide (ITO) layer by lift-off technique, employing EBL patterning (Raith E-Line). The EBL was performed by exposing a 200nm thick PMMA layer using with a pattern of an array of circles. Subsequently a 5nm thick Cr and 100nm Au layers were deposited and then lift-off procedure, using acetone, took place to leave array of Au/Cr disks with height of 100nm, a diameter of 100nm (with variance of ~10nm) and center to center separation of 400nm. The inset of Fig. 1 displays a SEM image of the resulting MNP array. A solution of PEDOT: PSS (poly(3,4-ethylenedioxythiophene):poly(styrene sulfonate) is then spin coated and dried at 110°C to form a 50nm buffer layer partially covering the Au particles. Subsequently a 1:1 solution of P3HT (poly(3-hexylthiophene)) and PCBM ([6,6]-phenyl C61-butyric acid methyl ester) is spin coated to form a 170nm active layer. Finally, Calcium (10nm) and Aluminum (120nm) are thermally evaporated using a contact mask.
The MNP array parameters were carefully chosen according to simulation results, as will be further discussed, such that the peak of the plasmon resonance is located at wavelength close to the P3HT:PCBM band-gap where it is less absorbing . The MNP height was chosen such that the particles will significantly protrude into the active layer to maximize the field enhancement effect. The array period is a tradeoff between minimizing metal absorption at shorter wavelength and increasing plasmon resonance strength. A detailed examination of the influence of each MNP array’s parameter is discussed in the followings.
3. Results and discussion
Figure 2a shows the measured external quantum efficiency (EQE) of the cell with the Au MNP array compared to that of a reference cell. The cell under study and reference cell are located only few hundred microns apart to rule out variations due to slightly different processing of the organic material. Examining the enhancement ratio - Fig. 2b, a wide enhancement peak is observed for wavelength between 550 and 750nm with a major peak of 53% enhancement at 660nm and a minor peak of 33% at 710nm. Another small peak is registered between 780 and 820nm. An average reduction of ~10% in efficiency is observed at wavelength below 550nm. In Fig. 2c we depict the J-V characteristics, measured under 100 mW/cm2 illumination by a solar simulator with an AM 1.5G filter, of the cell with the Au nano particles compared to a reference cell without. An increase of ~3.5% in short circuit current is observed, while open circuit voltage is unchanged.
To better understand the nature of the enhancement mechanisms we performed a FDTD simulation using Lumerical software  for a cell of the same layer structure as the experimental one. The MNP array was modeled by Au cylindrical particles with height of 100nm and radius of 50nm embedded in the PEDOT:PSS and P3HT:PCBM layers with an array period of 400nm. A square unit cell and periodic boundary conditions with a normal incidence plane wave excitation from the ITO side was used. The dielectric functions for the different constituents were taken from the literature [25,26]. The calculation mesh resolution was 5nm and it was verified that mesh of 2nms a 1nm gave the same results, while the time step was 0.1fsec to assure convergence. The input plane wave was a short pulse (2.65fsec) to cover simultaneously all the range of the sun spectrum.
Figures 3a shows the fraction of power absorbed in the active P3HT:PCBM layer and in the metal for the cell with the MNP compared to the power absorbed in the active layer of a reference cell. The absorbed power at each wavelength is calculated in a standard procedure by integrating the (light intensity (E2) multiplied by the effective conductivity of the absorber) over its volume. Although the exact field distribution may be not so accurate due to the lack of knowledge of the exact nature of the interfaces, the absorbed power, as an integral measure, is highly accurate. The extracted spectral enhancement factor is depicted in Fig. 3b, exhibiting two enhanced absorption peaks at 680nm and 785nm while a reduction of ~10% in absorption is seen at shorter wavelengths. Clearly there is a very good correlation both in the enhancement peaks locations between simulation and measurement results, with the peak-width in the measurements slightly wider due to manufacturing inhomogeneity. The ~10% efficiency reduction at shorter wavelength also correlates well with the experimental results. The difference in enhancement magnitudes between simulations and the experiments as well as additional details such as the splitting of the main experimentally observed peak will be discussed later.
To further understand the enhancement mechanisms we examine the electric field intensity distribution at the two absorption enhancement peaks at 680nm and 785nm – Figs. 3c and 3d respectively. The resonance at 680nm exhibits a very strong local field enhancement in the vicinity of the particle, typical of the quadruple plasmon resonance, with lobes both in the PEDOT:PSS and P3HT:PCBM leading to enhanced absorption. The field penetrating into the metal results in undesired metal absorption, as seen in Fig. 3a, however, this loss is minimized by using larger particles which have a higher scattering vs. absorption cross section . It should be noted that such larger particles with regular spacing can be easily fabricated by nano lithography patterning but are difficult to achieve when deposited from stable liquid dispersion – that was used in previous studies. The resonance at λ = 785nm, Fig. 3d, has a different origin. The electric field intensity distribution indicates that this resonance mode is the TM110 mode of a circular nano-patch cavity  generated between the Au nanodisk and the Aluminum back contact. This low radiation-Q optical antenna mode is beneficial for solar cell application – the energy is trapped mainly in the P3HT:PCBM layer leading to enhanced absorption in the active layer with less metal loss compared to the plasmon resonance mode case. Similar resonance was observed in OPV cells with a gratings based Ag anode . The patch resonance wavelength can be controlled by changing cavity height (i.e active layer thickness, or particle height) or size (particle diameter). The periodicity of the nano patch antenna array should be approximately a multiple of half the desired resonance wavelength (~390nm) to avoid destructive interference, thus the structure acts at this resonance as diffraction gratings coupled with patch antenna array (also discussion related to Fig. 4 below).
Now we look into optimizing the plasmonic particles density (fill factor). The efficiency reduction at short wavelength, evident both in experiment and simulation results (Figs. 2 & 3), is caused by the Au particles, absorbing and backscattering some of the light. This phenomenon may look as posing an inherent tradeoff when determining the optimal density (or array period) of the MNP: lower MNP density will show less loss at shorter wavelength, but also reduced efficiency enhancement at the plasmon resonances. However, here the antenna effect of the plasmonic nano particles comes about to resolve this conflict. The influence of MNP density on the fraction of power absorbed in the cell is shown in Fig. 4. Decreasing MNP area density from 20% to 10% to 5% (array periods of 200nm, 280nm, 400nm respectively) decreases the short wavelength loss linearly, while the enhanced cell absorbance in plasmon resonance wavelength (660-720nm) is only slightly affected (the excursions of the resonance shape and splitting as a function of periodicity will be resolved in connection with Fig. 7). Only when increasing the period to 450nm (density ~4%) the absorbance at plasmon resonance wavelength starts to decrease. The cause for this interesting phenomenon is the fact that at the plasmon resonance the effective particle cross section is much larger than its geometrical cross section , yielding a much larger effective density, while away from this resonance the cross section is the actual particle size. The desired particle density should be roughly inversely proportional to the effective particle cross section at the plasmon resonance, yielding both maximum absorption enhancement at plasmon resonance wavelength and minimum reduction outside the resonance. The dependence of patch resonance on periodicity discussed above is also evident in Fig. 4 where the patch cavity resonance is exhibited only for periods of 400-450nm (densities of 4-5%).
At this stage, and using Figs. 5 and 6 , we look into the influence of the proximity of the metal structure to the active layer and also interpret by this effect the reduced experimental enhancement compared to the ideal simulations. First, we analyze a cell with shorter particles (height of 60nm), such that they are only slightly protruding the active layer. Absorption enhancement in the cell with the shorter particles is significantly reduced compared to the taller 100nm particles – Fig. 5a. The electric field intensity cross section at the resonance for particle height of 60nm – Fig. 5b, shows increased intensity mainly in the PEDOT:PSS and ITO layers, thus not contributing to enhanced absorption, leading to the conclusion that particles significantly embedded in the active layer are the better design choice. The small blue shift in the resonance stems from a decrease in the effective refractive index of the dielectric medium embedding the particle, now closer to the smaller PEDOT index (~1.45) than the higher P3HT:PCBM index (~2.1).
Similar effect - remoteness of the metal particle from the active layer can explain the lower enhancement value in measurements compared to simulations. Although partially attributed to the inhomogeneity in manufacturing , most plausibly the dominant effect is that during the spin coating of the 50nm PEDOT buffer a residual very thin layer coated also the top 50nm part of the Au disks extending above this layer. This residual coating is reducing the absorption enhancement in the active layer, as verified by the simulation of such structure shown in Fig. 6. It should be noted that this thin PEDOT coating of the Au particles may also have a positive effect – eliminating the direct interface of the metal with the active layer thus decreasing the quenching of excitons in such metal-polymer interface, an effect not accounted for in our simulations.
Finally, we are ready to interpret the last detail in the experimental result, namely the split at the main plasmon resonance (650-730nm, Fig. 2b), which is absent in the simulation (Fig. 3b). Similar split, with one major peak and a smaller peak at longer wavelength, is regained in simulation for slightly different periodicity of 350nm (Fig. 7a). This split can be retraced to two interacting resonances: the highly localized quadrupole resonance exhibiting the larger peak at 665nm, Fig. 7b, and a second, with smaller enhancement, peak at 695nm, Fig. 7c. The second resonance is less localized and is similar to the patch cavity mode, but at this wavelength it is generated between the plasmonic particle and the top surface. Such resonance is thus sensitive to periodicity, as discussed previously in relation to the nano-patch mode which explains the dependence of this resonance shape as a function of the array density in Fig. 4. The slightly different periodicity between the simulation and experiment required to observe this split may be attributed to a difference in the simulated and actual permittivity of the active layer influencing the effective period. The third large peak at 745nm is again the patch antenna resonance discussed in connection with Fig. 3.
In conclusion, we have shown both experimentally and theoretically an increase in the EQE of OPV cells after embedding ordered arrays of Au NPs extending into the active layer. We identified two enhancement mechanisms, enhanced absorption driven by local field enhancement of the plasmon resonance and enhanced absorption driven by a cavity mode of a circular nano patch antenna. Based on an analysis of these mechanisms we derived design guidelines for optimal MNP properties. Particles should protrude into the active layer in order to maximize the near field enhancement. Optimal particle density is roughly inversely proportional to particle cross section at the plasmon resonance leading to minimum absorption and reflection losses at wavelength outside the resonance. A thin coating layer around the Au particles causes some reduction in near field enhancement, but may also reduce exciton quenching at the metal polymer interface. It is important to remember that the polymerization of the photoactive layer itself can be modified by changing the interfaces – thus adding MNP may result in structural related spectral changes. However, due to the small fill factor (5%), due to the fact that the metal disks are covered by thin buffer layer, even when they are residing in the active layer, and due to the fact that the enhancement is obtained exactly at the calculated plasmon resonances (independent of the original spectrum as we obtained by measurements of a different photoactive blend-not reported here), makes our interpretations highly plausible.
References and links
1. X. Yang, J. Loos, S. C. Veenstra, W. J. H. Verhees, M. M. Wienk, J. M. Kroon, M. A. J. Michels, and R. A. J. Janssen, “Nanoscale morphology of high-performance polymer solar cells,” Nano Lett. 5(4), 579–583 (2005). [CrossRef] [PubMed]
2. P. Peumans, A. Yakimov, and S. R. Forrest, “Small molecular weight organic thin-film photodetectors and solar cells,” J. Appl. Phys. 93(7), 3693 (2003). [CrossRef]
3. G. Li, V. Shrotriya, J. Huang, Y. Yao, T. Moriarty, K. Emery, and Y. Yang, “High-efficiency solution processable polymer photovoltaic cells by self-organization of polymer blends,” Nat. Mater. 4(11), 864–868 (2005). [CrossRef]
4. K. Kim, J. Liu, M. A. G. Namboothiry, and D. L. Carroll, “Roles of donor and acceptor nanodomains in 6% efficient thermally annealed polymer photovoltaics,” Appl. Phys. Lett. 90(16), 163511 (2007). [CrossRef]
5. T. Ameri, G. Dennler, C. Lungenschmied, and C. J. Brabec, “Organic tandem solar cells: A review,” Energy Environ. Scie. 2(4), 347 (2009). [CrossRef]
6. V. Andersson, K. Tvingstedt, and O. Inganas, “Optical modeling of a folded organic solar cell,” J. Appl. Phys. 103(9), 094520 (2008). [CrossRef]
7. S. H. Park, A. Roy, S. Beaupre, S. Cho, N. Coates, J. S. Moon, D. Moses, M. Leclerc, K. Lee, and A. J. Heeger, “Bulk heterojunction solar cells with internal quantum efficiency approaching 100%,” Nat. Photonics 3(5), 297–302 (2009). [CrossRef]
8. W. Y. Wong, X. Z. Wang, Z. He, A. B. Djurisić, C. T. Yip, K. Y. Cheung, H. Wang, C. S. K. Mak, and W. K. Chan, “Metallated conjugated polymers as a new avenue towards high-efficiency polymer solar cells,” Nat. Mater. 6(7), 521–527 (2007). [CrossRef] [PubMed]
10. V. E. Ferry, M. A. Verschuuren, H. B. T. Li, R. E. I. Schropp, H. A. Atwater, and A. Polman, “Improved red-response in thin film a-Si:H solar cells with soft-imprinted plasmonic back reflectors,” Appl. Phys. Lett. 95(18), 183503 (2009). [CrossRef]
11. K. Nakayama, K. Tanabe, and H. A. Atwater, “Plasmonic nanoparticle enhanced light absorption in GaAs solar cells,” Appl. Phys. Lett. 93(12), 121904 (2008). [CrossRef]
12. R. A. Pala, J. White, E. Barnard, J. Liu, and M. L. Brongersma, “Design of Plasmonic Thin-Film Solar Cells with Broadband Absorption Enhancements,” Adv. Mater. 21(34), 3504–3509 (2009). [CrossRef]
13. C. Hägglund, M. Zach, G. Petersson, and B. Kasemo, “Electromagnetic coupling of light into a silicon solar cell by nanodisk plasmons,” Appl. Phys. Lett. 92(5), 053110 (2008). [CrossRef]
14. F. J. Beck, A. Polman, and K. R. Catchpole, “Tunable light trapping for solar cells using localized surface plasmons,” J. Appl. Phys. 105(11), 4310 (2009). [CrossRef]
15. S. Kim, S. Na, J. Jo, D. Kim, and Y. Nah, “Plasmon enhanced performance of organic solar cells using electrodeposited Ag nanoparticles,” Appl. Phys. Lett. 93(7), 073307 (2008). [CrossRef]
16. F. Chen, J. Wu, C. Lee, Y. Hong, C. Kuo, and M. H. Huang, “Plasmonic-enhanced polymer photovoltaic devices incorporating solution-processable metal nanoparticles,” Appl. Phys. Lett. 95(1), 013305 (2009). [CrossRef]
17. A. J. Morfa, K. L. Rowlen, T. H. Reilly III, M. J. Romero, and J. van de Lagemaat, “Plasmon-enhanced solar energy conversion in organic bulk heterojunction photovoltaics,” Appl. Phys. Lett. 92(1), 013504 (2008). [CrossRef]
18. T. H. Reilly, J. van de Lagemaat, R. C. Tenent, A. J. Morfa, and K. L. Rowlen, “Surface-plasmon enhanced transparent electrodes in organic photovoltaics,” Appl. Phys. Lett. 92(24), 243304 (2008). [CrossRef]
19. D. Duche, P. Torchio, L. Escoubas, F. Monestier, J. Simon, F. Flory, and G. Mathian, “Improving light absorption in organic solar cellsby plasmonic contribution,” Sol. Energy Mater. Sol. Cells 93(8), 1377–1382 (2009). [CrossRef]
20. A. P. Kulkarni, K. M. Noone, K. Munechika, S. R. Guyer, and D. S. Ginger, “Plasmon-enhanced charge carrier generation in organic photovoltaic films using silver nanoprisms,” Nano Lett. 10(4), 1501–1505 (2010). [CrossRef] [PubMed]
22. Y. A. Akimov, W. S. Koh, and K. Ostrikov, “Enhancement of optical absorption in thin-film solar cells through the excitation of higher-order nanoparticle plasmon modes,” Opt. Express 17(12), 10195–10205 (2009). [CrossRef] [PubMed]
23. K. R. Catchpole and A. Polman, “Design principles for particle plasmon enhanced solar cells,” Appl. Phys. Lett. 93(19), 191113 (2008). [CrossRef]
24. Lumerical Solutions, Inc. Optical and photonic design and engineering software products.
25. F. Monestier, J. Simon, P. Torchio, L. Escoubas, F. Flory, S. Bailly, R. de Bettignies, S. Guillerez, and C. Defranoux, “Modeling the short circuit current density of polymer solar cells based on P3HT:PCBM blend,” Sol. Energy Mater. Sol. Cells 91(5), 405–410 (2007). [CrossRef]
26. E. D. Palik Handbook of Optical Constants of Solids (Academic, New York, 1985).
27. C. F. Bohren, and D. R. Huffman, Absorption and scattering of light by small particles, (Wiley, New York, 1983).
28. C. Manolatou, and F. Rana, “Subwavelength Nanopatch Cavities for Semiconductor Plasmon Lasers, ” arXiv: 0712.1489v1 (2007).
29. N. C. Lindquist, W. A. Luhman, S. Oh, and R. J. Holmes, “Plasmonic nanocavity arrays for enhanced efficiency in organic photovoltaic cells,” Appl. Phys. Lett. 93(12), 123308 (2008). [CrossRef]
30. A. Alù and N. Engheta, “Effect of small random disorders and imperfections on the performance of arrays of plasmonic nanoparticles,” N. J. Phys. 12(1), 013015 (2010). [CrossRef]