Vacuum 152 (2018) 285e290

Contents lists available at ScienceDirect

Vacuum

journal homepage: www.elsevier.com/locate/vacuum

Enhancing efficiency with plasmonic behavior of double metal system

* Nipon Deka a, Maidul Islam a, Prashant K. Sarswat b, , Gagan Kumar a a Department of Physics, Indian Institute of Technology Guwahati, Guwahati, 781039, India b Department of Metallurgical Engineering, University of Utah, Salt Lake City, UT, 84112, USA article info abstract

Article history: The use of special arrangement of nanoparticle (Ag Np) arrays is presented in order to enhance Received 20 November 2017 light trapping capability of the thin film solar cells due to their ability to couple incident sunlight into Received in revised form localized modes. The spatial distribution of the in the matrix and their relative position are 10 March 2018 the important factors governing the performance of the device. Double nanoparticle system (DNS) over Accepted 19 March 2018 the substrate, which is not a commonly studied system, is explored and the performance (ab- Available online 20 March 2018 sorption enhancement) was compared with periodically arranged single Ag Nps system. Finite difference time domain (FDTD) simulation has been performed to investigate the effect of array patterning on the absorption of solar radiation. The presence of DNS was found responsible for increased coupling of photons into plasmonic modes, and consequently increased absorption of photons into the substrate over a broad range of the solar spectrum. © 2018 Elsevier Ltd. All rights reserved.

1. Introduction plasmonic nanostructures have the capability to enhance absorp- tion near the localized resonance, which in turn resulted In the last and recent decade, researchers have paid keen overall power conversion enhancement. Hence, plasmonic solar attention to solar cell technology as important and strategic cell based research is one of the hot and demanding field of interest research area [1e4]. The salient features of the photovoltaic science in recent decades. In 2005, Tian et al. reported their seminal work and technology and associated motivation of the research in this on the conversion of the photocurrent from absorbed visible light segment are due to low-maintenance, cost-effective, clean, and by metal nanoparticles [11]. It was observed that plasmonically environment-friendly nature of this technology. Furthermore, solar induced particles incorporated in a TiO2 matrix [12] can cause cells are long lasting sources of energy which are renewable and the charge separation and efficient conversion of photocurrent can provide sustainable power. To increase the conversion effi- resulting from the absorbance spectra of AueTiO2 films was ciency of the solar cell to the photocurrent, one needs structure claimed. Later on, in year 2008, enhancements of photocurrent in which has high absorption. In this regard, metal nanoparticles have GaAs solar cells using Ag nanoparticles was reported [13]. In their very high absorption and scattering cross-section features which research, plasmonic effect for the absorption and resulting photo- are very useful for solar cells, depending upon their size and shape current change for the different size/shape of Ag nanoparticles [5,6]. Optical properties of plasmonic [7] nanoparticles strongly (from hemispheres to cones while increasing the height) was depend on their size and shapes therefore appropriate size and investigated. In later years, Munday et al. proposed integrated ab- shape are crucial parameters to achieve a highly efficient solar cell. sorption enhancement from the combination of the metallic grat- There are several ways to insert metal nanoparticles (MNP) in ings and an antireflection (AR) coating [14]. It was observed that a solar cell absorber layers [2,8e10]. The relative position and combination of gratings and traditional AR coatings can exceed the orientation of the nanoparticles in solar cell are important pa- absorption enhancement rather than either metallic gratings or AR rameters determining the solar cell performance. Metal based coatings. Scattering efficiency and absorption losses are the crucial factors in the case of metal nanoparticles based solar cells applications. In the starting of current decade, it was shown that among the metal * Corresponding author. 135 S, 1460 E, Room 412, Salt Lake city, UT, 84112, USA. nanoparticles, silver is highly efficient for the scattering and E-mail address: [email protected] (P.K. Sarswat). https://doi.org/10.1016/j.vacuum.2018.03.026 0042-207X/© 2018 Elsevier Ltd. All rights reserved. 286 N. Deka et al. / Vacuum 152 (2018) 285e290 absorption, mainly due to its large scattering efficiency and low DNS is denoted by 0p0 and the diameter of each metallic nano- absorption losses. These MNPs have offered greatest efficiency particle is denoted by 0d0. The internal gap between the two metallic 0 improvements in the variety of solar cells [15,16]. In 2012, Reineck nanoparticles in the DNS is 0i : Two major analysis are performed et al. reported the generation of photocurrents in a solid state de- where the absorption enhancement is determined by varying the vice, comprised of spherical silver and gold nanoparticles deposited periodicity of the DNS and by varying the internal gap within the on TiO2ecoated substrate [17]. It was observed that incident photon DNS at different values of the diameter ‘d’ of the metallic nano- to electron conversion efficiency (IPCE), absorptivity, and current particles. The diameters that we considered were 0 density for Au nanoparticles were higher than that of Ag 0d ¼ 75 nm; 100 nm and 150 nm: Field monitors were placed on nanoparticles. the surface of the substrate and 0:5 mm below the surface. Their The efficiency enhancement of PbS quantum dot/ZnO nanowire purpose is to calculate transmission of the incident radiation at solar cells, using plasmonic behavior of silver nanocubes, was also those regions to generate the absorption profile. The periodicity of 0 proposed in 2015 [18]. This research was based on efficiency the DNS was varied by considering 0p ¼ 350 nm; 400 nm; 450 nm; enhancement of solar cell by controlling the position and amount of and 500 nm ðcentre to centreÞ and the internal gap was varied by 0 Ag nanocubes in a large scale while PbS quantum dots exhibited considering 0i ¼ 5 nm; 10 nm; 15 nm; and 20 nm: The wavelength weak absorption without Ag nanocubes. To find out a perfect range for the simulation is chosen from 400 nm to 900 nm: absorber of visible light, in 2016, Ullah et al. work on plasmonic effect based perfect absorber for solar cell applications relies on a concept that perfect absorber should be free from the reflection as 2.1. Numerical methods well as transmission through the structure [19]. They proposed a fi theoretical polarization independent plasmonic absorber design, A3D nite difference time domain (FDTD) numerical method consist of gold nanorings on the top of the spacer and a [21] was employed to model the optical effects of metallic nano- fi gold reflector that explored nearly unity absorption in the visible particles suspended on top of the thin lm solar cells. The FDTD fi and near infrared spectrum. More recently, Duan et al. reported the results were used as input to the nite-element time-domain solver light trapping and the light absorption for two different shapes of of the commercially available CST Microwave Studio software. The silver nanoparticles (spheroidal and hemispheroidal) in order to method solves Maxwell's equations over the entire cell. In the FDTD fi improve the light absorption in thin film silicon solar cell. The simulation, the material is de ned in every simulation grid point. fi fi difference of light absorption distribution between the antireflec- All effects related to electromagnetic elds including rear eld, fi fi fi tion coating and the body was studied [20]. radiating near eld, and scattered eld (or far eld) effects were Though a lot of research works have been done in this area, still taken into account. it is a great challenge to improve the integrated solar cell efficiency Metallic nanoparticles having sizes comparable to the wave- over the whole visible spectrum. In this letter, we propose a plas- length of illuminating light scatter the light preferentially along the monic structure made of silver double nanoparticle system (DNS) incident direction of wave propagation. In this work, we analysed a separated by a small gap suspended onto a silicon substrate. Here, structure consisting of periodic repetition of a system of metallic we varied the size of nanoparticles for optimization and studied double nanoparticle system (DNS), considering the main parame- integrated quantum efficiency by varying the gap between the ters taken as variables for the optimization studies. Fig. 1 represents double nanoparticles, in order to enhance utility of visible spec- a cubical unit cell consisting of two metallic spherical nanoparticles trum. In the first section, we propose our geometry and numerical having an inter-particle separation. This unit cell simulates one simulations to extract the absorption of the structure. In subse- period of a square array of closed packed particles suspended on quent sections, we first discuss the absorption enhancement and top of a silicon solar cell. The FDTD program places the unit cell in a fi resulting integrated absorption enhancement in single and double square box with speci c boundary conditions (BCs) on each face. nano particles system. Then we discuss the power loss in both the On the z-direction boundary conditions are perfectly matched layer ð ; Þ cases and finally we find out an optimized structure for our in- (PML). On the side boundaries x y periodic BCs are used to model fi vestigations in double nano particles system. the in nite periodicity of the structures and the normally incident plane wave source. Here, symmetric and anti-symmetric BCs are used to reduce the required memory size and computation time. 2. Geometric model To simulate solar illumination, a broadband plane wave source is placed in the air medium above the structure. The source band- Fig. 1 shows the double nanoparticle system (DNS) suspended width is 400 1100 nm, since the AM1:5 solar photon flux outside periodically on a silicon substrate for solar cells. The period of the this wavelength range is small, and it corresponds to the most

Fig. 1. (a) Depicts schematic of a 3D geometry of solar cell structure having double nano particle system (DNS). The p and i in the schematic stand for periodicity and the internal gap between double nanoparticles, respectively. Sunlight is coming from above the structure. (b) It depicts side-view of the substrate with DNS on its top. N. Deka et al. / Vacuum 152 (2018) 285e290 287 significant portion of the photocurrent spectrum of silicon-based photovoltaic devices. QE ðlÞ The power absorbed per unit volume (P ) in each element of ðlÞ¼ particle abs g ðlÞ the structure is calculated using the resulting electric field distri- QEbare bution established in their material [22]: ðlÞ ðlÞ fi where QEparticle and QEbare represents the quantum ef ciency 1 of the solar cell in presence and in absence of nanoparticles over the P ¼ ujEj2imagðεÞ abs 2 silicon substrate. We compare the performance of double nanoparticles system where jEj2 is the electric field intensity, u is the angular frequency (DNS) with that of a single nanoparticle system (SNS). Fig. 2 shows of the light and imagðεÞ is the imaginary part of dielectric permit- the comparison of simulated absorption enhancement measured tivity. Pabs is normalized by the source power to obtain the for SNS and DNS with each particle of 100 nm diameter and 20 nm absorption. internal gap between the two particles with a periodicity of The number of absorbed photons per unit volume can then be 400 nm. We can see that employing DNS has an advantage as calculated by: compared to SNS. At 480 nm the absorption enhancement is well above unity for DNS as compared to SNS and remains so over the P 1jEj2imagðεÞ entire spectrum. This means that the DNS is more effective in g ¼ abs ¼ 2 ħu ħ trapping sunlight as compared to SNS. The inset in Fig. 2 shows the electric field around the SNS and DNS at 480 nm which clearly where ħu is the energy of one photon. The generation rate is the supports the absorption enhancement in case of DNS. The overall integration of g over the simulation spectrum. The quantum effi- performance of the structure over the entire solar spectrum can be ciency of a solar cell is defined by: evaluated by considering the integrated absorption enhancement (G) which is defined as: ðlÞ ðlÞ¼Pabs QE ðlÞ Pin IQE ¼ particle ðlÞ ðlÞ G where Pin and Pabs are the powers of the incident light and IQEbare absorbed light within the solar cell, respectively, at a wavelength l. If the value of integrated absorption enhancement (G) is greater Using the quantum efficiency, integrated quantum efficiency, IQE; is than unity, it means that IQE > IQE and this further dem- defined as: particle bare onstrates that the absorption of sunlight has increased due to Z l presence of metallic nanoparticles over the silicon substrate. ðlÞ ðlÞ l QE IAM1:5 d Further, the percentage absorption enhancement of the struc- ¼ hcZ IQE l ture can be calculated as: I : ðlÞdl hc AM1 5 IQEparticle IQEbare where h is the Planck's constant, c is the speed of light in free space Percentage Enhancement ¼ *100% IQEbare and IAM1:5 is the intensity of AM1:5 solar spectrum. To see how the efficiency of solar cell with metal nanoparticles is improved These values are listed in Table 1 for different combinations. comparing with a bare solar cell, following quantities are calculated:

ðlÞ ðlÞ ðlÞ¼QEparticle ðlÞ¼QEparticle g ðlÞ Absorption Enhancement [23]: g ðlÞ QEbare QEbare

And, Integrated absorption Enhancement: G ¼ IQEparticle IQEbare

If we assume that all electron-hole pair contributes to photo- current, the short circuit current density is given by Ref. [22]: Z l J ¼ e QEðlÞI : ðlÞdl sc hc AM1 5 where e is the charge on an electron. We evaluate the performance of the metallic nanoparticles by comparing the performance of a bare silicon surface as the absorber layer.

3. Results and discussions

3.1. Absorption enhancement

In order to describe the enhancing efficiency of the metallic nanoparticles, the absorption enhancement (g) is plotted against Fig. 2. (a) Represents comparison of absorption enhancement (g) between SNS and l wavelength ( ) from 400 1100 nm. The absorption enhancement DNS. Inset shows the electric field profiles for SNS and DNS. In case of DNS, electric is defined as: field profile is showing coupling between the double nano particles system. 288 N. Deka et al. / Vacuum 152 (2018) 285e290

Table 1 Enhancement factors for single and double nanoparticle based Plasmonic solar cells.

System Overall Enhancement (G) Percentage Enhancement

Single nanoparticle 1.124 12.4% Diameter ¼ 100 nm Periodicity ¼ 400 nm Double Nanoparticle System 1.829 82.9% Diameter ¼ 100 nm Periodicity ¼ 400 nm Internal Gap ¼ 20 nm

3.2. Power loss profile 3.3. Theory

To better understand the LSPR of the SNS and DNS we determine When the plasmon resonance [25] of one particle overlaps the power loss of incident light in the absorber medium (silicon) we spectrally with the resonance of another particle, the two plasmon calculated the power loss profiles in the substrate. The results for resonances will couple to one another. This coupling can lead to the the two different nanoparticle system is shown in a Fig. 3.To emergence of new plasmon modes and can further enhance the calculate these power loss profiles we used the finite element time local fields. The near field coupling results from the Coulomb domain solver of the commercially available CST Microwave Studio interaction between the surface charges on the two particles. When software. The simulation geometry consists of a silicon substrate near field coupling is strong, large charge dipoles can develop and suspended silver nanoparticles on top of the substrate. In the across the gap between the particles, so that the local fields in the simulations, we considered the silicon substrate thickness of gap can be much greater than the sum of the local fields that would 400 nm while varying the diameter of the silver nanoparticle as be produced by the isolated particles. This replacement of particles 100 nm and 150 nm. A waveguide port at distance of 1 mm from by point dipoles is correct since the electric field near the nano- the top of the substrate was used as a plane wave source with an particles is almost homogeneous for long enough distances be- amplitude of 1 V=m. The time-averaged power loss Qðx; yÞ was tween them. When an external field is incident upon the two point calculated as [24], dipoles it may lead to either symmetric or asymmetric plasmon oscillation. When the nanospheres approach each other the modes with 1 2 resonance electric permittivity tending to ∞ are conditioned by Qðx; yÞ¼ cε0na : jEðx; yÞj 2 asymmetric plasmon oscillations and are referred to as L-Modes [26]. The modes with resonance values of electrical permittivity where c is the speed of light in free space, ε is the permittivity of 0 tending to a finite limit are called T-modes and those modes having free space, a is the absorption coefficient, n is the real part of the zero value of resonance electrical permittivity are called M-modes. complex refractive index, and E is the electric field. The spatial distribution of the wave function reveals that the L The average power loss profiles within the silicon substrate in modes and the T modes corresponds generally to the structure of the case of SNS and DNS for an incident wavelength of 495 nm are plasmon wave function of isolated spheres whereas in case of M shown as in Fig. 3 (a), (b), (c) and (d) respectively. For comparison, modes the positive and negative charges are localized near the gap we also calculated the average power loss density in the silicon between the nanospheres. The M modes arise only when R12 1:2 layer in the two cases. In the case of SNS, this was 1:223 2R , where R is the centre to centre distance of the DNS and R is the 1021 W=m3, whereas in the case of the DNS it was found to be 12 radius of the particle. The M-modes are called “bound states of 2:355 1021 W=m3, i.e. an increase of 92:5%. To provide a plasmon atoms” or “plasmon molecules”. comprehensive picture of the power loss profiles, we also present The wave function of unbounded plasmon atoms then power loss profiles in a different perspective (in the x z ðL and T modesÞ remain more or less distributed over the plane) in Fig. 3 (c) and (d) for the two cases discussed above.

Fig. 3. (a) and (b) Simulated power loss profile for SNS and DNS respectively at incident wavelength 495 nm. (c) and (d) represent the same in the x-z view. N. Deka et al. / Vacuum 152 (2018) 285e290 289 volume of both nanospheres while the potential of a plasmon molecule ðM modeÞ is localized strongly near the region between the nanospheres. The expression for the enhancement factor G of the resonant local field is [27]:

8 R 1 Gz 3 i ðε"ðuÞÞ where, “R” is the radius of the nanospheres ‘i’ is the width of the gap between the nanoparticles and ε"ðuÞ is the imaginary part of the permittivity of silver nanoparticles.

3.4. Optimized structure

Fig. 5. Contour plot representing the integrated absorption enhancement (G) as a The optimization procedure is performed by varying the function of particle diameter (D) and periodicity within the DNS. The color bar rep- diameter of the metal nanoparticles and the internal gap within resents the values of integrated absorption enhancement (G). (For interpretation of the the DNS to determine the maximum possible absorption of the references to color in this figure legend, the reader is referred to the Web version of incident light within the silicon substrate. Fig. 4 shows the result this article.) of the geometry optimization procedure for a 400 nm thick silicon substrate. The contour plot for the integrated absorption 4. Conclusions enhancement (G) shows that for a DNS consisting of particles of 100 nm diameter the optimum internal gap between the particles The performance of double nanoparticle system (DNS) sus- is between 18 nm to 22 nm for efficient absorption of sunlight. For pended over the silicon substrate has been compared with the a DNS consisting of particles having larger diameter ð150 nmÞ the single nanoparticle system (SNS) and was found much better as integrated absorption enhancement (G) is maximum in the range compared to the latter. The absorption enhancement ðgÞ in a case of of 20 nm to 30 nm internal gap but the value would be less than DNS as compared to SNS at wavelength greater than 480 nm causes that of particles with 100 nm diameter. Thus the most efficient efficient trapping of solar radiation by the DNS. The integrated structure would be a silicon substrate of thickness 400 nm having absorption enhancement ðGÞ in a case of SNS and DNS are found to double nanoparticles system (DNS) suspended over it with a be 1:124 and 1:829; respectively, which corresponds to 12:4% periodicity of 400 nm, the particle diameter equal to 100 nm and and 82:9% efficiency enhancement in trapping solar radiation as the internal gap between the two particles in the DNS equal to compared to the bare structure. This result is supported by the 18 nm to 22 nm. power loss profiles which shows the average power loss density in The second phase of the optimization procedure was performed DNS and SNS are of 2:355 1021 W=m3 and 1:223 by varying the periodicity of the DNS over the silicon substrate. The 1021 W=m3, respectively which results in 92:5% increment in contour plot in Fig. 5 represents the variation of integrated ab- absorption of solar radiation due to DNS. Finally, we found an sorption enhancement ðGÞ at different values of periodicity by optimized geometry by varying the internal gap within the DNS varying the diameter of the metallic nanoparticles. At a periodicity and the most efficient geometry turned out to be one with silver of around 500 nm and for a diameter of 100 nm the absorption nanoparticles of diameter 100 nm and having an internal gap enhancement ðGÞ is maximum. ranged from 18 nm to 20 nm:

Acknowledgements

We gratefully acknowledge the financial support from the SERB, Department of Science and Technology, India (SB/FTP/PS-051/ 2014).

References

[1] J.D. Winans, C. Hungerford, K. Shome, L.J. Rothberg, P.M. Fauchet, Plasmonic effects in ultrathin solar cells: performance improvements with Ag nanoparticles on the front, the back, and both, Optic Express 23 (3) (2015) A92eA105. [2] H.A. Atwater, A. Polman, Plasmonics for improved photovoltaic devices, Nat. Mater. 9 (3) (2010) 205e213. [3] C. Uhrenfeldt, T.F. Villesen, A. Tetu, B. Johansen, A.N. Larsen, Broadband photocurrent enhancement and light-trapping in thin film Si solar cells with periodic Al nanoparticle arrays on the front, Optic Express 23 (11) (2015) A525eA538. [4] C. Sun, X. Wang, Efficient light trapping structures of thin film silicon solar cells based on silver nanoparticle arrays, Plasmonics 10 (6) (2015) 1307e1314. Fig. 4. Contour plot representing the integrated absorption enhancement (G) as a [5] M. Pelton, J. Aizpurua, G. Bryant, Metal-nanoparticle plasmonics, Laser Photon. function of particle diameter (D) and internal gap (i) within the DNS. The color bar Rev. 2 (3) (2008) 136e159. represents the values of integrated absorption enhancement (G). (For interpretation of [6] K.L. Kelly, E. Coronado, L.L. Zhao, G.C. Schatz, The optical properties of metal fi the references to color in this gure legend, the reader is referred to the Web version of nanoparticles: the influence of size, shape, and dielectric environment, Journal this article.) of Physical Chemistry B-Condensed Phase 107 (3) (2003) 668e677. 290 N. Deka et al. / Vacuum 152 (2018) 285e290

[7] S. Mohapatra, Y.K. Mishra, A.M. Warrier, R. Philip, S. Sahoo, A.K. Arora, 217e226. D.K. Avasthi, Plasmonic, low-frequency Raman, and nonlinear optical-limiting [17] P. Reineck, G.P. Lee, D. Brick, M. Karg, P. Mulvaney, U. Bach, A solid-state studies in copperesilica nanocomposites, Plasmonics 7 (1) (2012) 25e31. plasmonic solar cell via metal nanoparticle self-assembly, Adv. Mater. 24 (35) [8] N.C. Jeong, C. Prasittichai, J.T. Hupp, Photocurrent enhancement by surface (2012) 4750e4755. plasmon resonance of silver nanoparticles in highly porous dye-sensitized [18] T. Kawawaki, H. Wang, T. Kubo, K. Saito, J. Nakazaki, H. Segawa, T. Tatsuma, solar cells, Langmuir 27 (23) (2011) 14609e14614. Efficiency enhancement of PbS quantum Dot/ZnO nanowire bulk- [9] M.D. Brown, T. Suteewong, R.S.S. Kumar, V. D'Innocenzo, A. Petrozza, heterojunction solar cells by plasmonic silver nanocubes, ACS Nano 9 (4) M.M. Lee, U. Wiesner, H.J. Snaith, Plasmonic dye-sensitized solar cells using (2015) 4165e4172. core shell metal insulator nanoparticles, Nano Letters 11 (2) (2010) [19] H. Ullah, A.D. Khan, A. Ullah, I. Ullah, M. Noman, in: Plasmonic Perfect 438e445. Absorber for Solar Cell Applications, Emerging Technologies (ICET), 2016 In- [10] D.H. Wang, D.Y. Kim, K.W. Choi, J.H. Seo, S.H. Im, J.H. Park, O.O. Park, ternational Conference on, IEEE, 2016, pp. 1e5. A.J. Heeger, Enhancement of donoreacceptor polymer bulk heterojunction [20] Z. Duan, M. Li, T. Mwenya, Y. Li, D. Song, Morphology optimization of silver solar cell power conversion efficiencies by addition of Au nanoparticles, nanoparticles used to improve the light absorption in thin-film silicon solar Angew. Chem. 123 (24) (2011) 5633e5637. cells, Plasmonics (2017) 1e7. [11] Y. Tian, T. Tatsuma, Mechanisms and applications of plasmon-induced charge [21] P.K. Sarswat, N. Deka, S. Jagan Mohan Rao, M.L. Free, G. Kumar, Surface separation at TiO2 films loaded with gold nanoparticles, J. Am. Chem. Soc. 127 texture-induced enhancement of optical and photoelectrochemical activity of (20) (2005) 7632e7637. Cu2ZnSnS4 photocathodes, J. Electron. Mater. 46 (8) (2017) 5308e5318. [12] D. Bhattacharyya, P.K. Sarswat, M. Islam, G. Kumar, M. Misra, M.L. Free, [22] W. Ren, G. Zhang, Y. Wu, H. Ding, Q. Shen, K. Zhang, J. Li, N. Pan, X. Wang, Geometrical modifications and tuning of optical and surface plasmon reso- Broadband absorption enhancement achieved by optical layer mediated nance behaviour of Au and Ag coated TiO 2 nanotubular arrays, RSC Adv. 5 plasmonic solar cell, Optic Express 19 (27) (2011) 26536e26550. (86) (2015) 70361e70370. [23] G.D. Shilpa, T.K. Subramanyam, K. Sreelakshmi, Uttarakumari, Study and [13] K. Nakayama, K. Tanabe, H.A. Atwater, Plasmonic nanoparticle enhanced light Optimization of Metal Nanoparticles for the Enhanced Efficiency absorption in GaAs solar cells, Appl. Phys. Lett. 93 (12) (2008) 121904. Solar Cells, IOP Conf Ser Mater Sci Eng 149 (1) (2016) 012074. [14] J.N. Munday, H.A. Atwater, Large integrated absorption enhancement in [24] N. Ida, Engineering Electromagnetics, Springer International Publishing, 2015. plasmonic solar cells by combining metallic gratings and antireflection coat- [25] G. Kumar, P.K. Sarswat, Interaction of surface plasmon polaritons with ings, Nano Letters 11 (6) (2010) 2195e2201. nanomaterials, in: C.D. Geddes (Ed.), Reviews in Plasmonics 2015, Springer [15] Z. Ouyang, X. Zhao, S. Varlamov, Y. Tao, J. Wong, S. Pillai, Nanoparticle- International Publishing, 2016, pp. 103e129. Cham. enhanced light trapping in thin-film silicon solar cells, Prog. Res. [26] V. Klimov, Nanoplasmonics, Pan Stanford Publishing, 2014. Appl. 19 (8) (2011) 917e926. [27] S. Franzen, Intrinsic limitations on the jEj4 dependence of the enhancement [16] Y. Yang, S. Pillai, H. Mehrvarz, H. Kampwerth, A. Ho-Baillie, M. Green, factor for surface-enhanced Raman scattering, J. Phys. Chem. C 113 (15) Enhanced light trapping for high efficiency crystalline solar cells by the (2009) 5912e5919. application of rear surface , Sol. Energy Mater. Sol. Cell. 101 (2012)