- Email: [email protected]
Broadband photon management of subwavelength structures surface for full-spectrum utilization of solar energy
Energy Conversion and Management 152 (2017) 22–30
Contents lists available at ScienceDirect
Energy Conversion and Management journal homepage: www.elsevier.com/locate/enconman
Broadband photon management of subwavelength structures surface for full-spectrum utilization of solar energy
MARK
⁎
Yuanpei Xu, Yimin Xuan , Xianglei Liu School of Energy and Power Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, China
A R T I C L E I N F O
A B S T R A C T
Keywords: Composite subwavelength structures Full spectrum Photon management Photovoltaic-thermoelectric hybrid systems
In this work, advanced photon-management composite subwavelength structures are fabricated to manage solar energy in the full-spectrum (300–2500 nm) wavelength range for the application of the photovoltaic-thermoelectric hybrid systems to fully utilize the solar energy. This proposed photon-management method will simultaneously realize efficient light trapping for the photons with above-bandgap energy in solar cells and the utilization of the below-bandgap photons and the waste heat resulting from the thermalization effect in solar cells with bottom thermoelectric devices. The typical structure is designed: Top ordered hexagonal nanohole arrays can trap above-bandgap photons to enhance the absorption in the silicon wafer, and TiO2/SiO2 bilayer films are deposited on the bottom side of the wafer to improve the transmission of the below-bandgap photons. ∼97% total absorptance for wavelengths of 300–1100 nm is achieved with optimized diameters of central nanoholes. The total transmittance, on the other hand, is improved to ∼60% from 1200 nm to 2500 nm. The results indicate that the structures realize the appropriate allocation of photons within different wavelength ranges to different devices for sufficient utilization of full-spectrum solar energy. This novel full-spectrum photon management benefits from the strong scattering effect among the nanoholes and the gradient refractive index of bilayer films. Moreover, the photon-management performance shows angle-independent and polarization-insensitive characteristics. This method can be applied to various kinds of solar cells for photovoltaic-thermoelectric hybrid systems and may provide thoughts for other solar harvesting applications.
1. Introduction Solar energy, which is a kind of renewable and abundant energy, has gained wide attention due to the intense needs for clean energy. Various utilization methods of solar energy have been proposed. At present, the photovoltaic (PV) device is one of the most popular approaches for solar energy application. However, the utilization efficiency of full-spectrum (300–2500 nm) solar energy reaching the ground is limited by the bandgap of solar cells. The PV devices can only take use of incident solar energy above bandgap. Furthermore, part of the absorbed light in solar cells will be converted to heat due to the recombination of electron-hole pairs in the active layer or at the surface, which is called thermalization loss [1]. This recombination process will cause the increase of the temperature of solar cells, and further brings down the usage of solar energy. To break these limitations in PV devices, a novel PV-thermoelectric (ThE) hybrid system has been proposed to improve the utilization efficiency of full-spectrum solar energy, which combines PV devices and ThE devices into a hybrid PV-ThE system. A concentrated PV-ThE hybrid system was systematically
⁎
analyzed by Lamba and Kaushik [2] with the influences of thermocouple numbers, the irradiance, PV and ThE current. Hsush et al. [3] proposed a CuInGaSe2 (CIGS)/ThE device with the efficient of 22% by employing passive light-trapping ZnO nanowires. Zhang and Yin [4,5] systematically investigated the influences of the thermal resistances on the hybrid system under the circumstance of high concentration ratio. In this system, the first-line goal is to allocate the full-spectrum solar energy to different devices. The solar energy above the bandgap must be absorbed by PV devices efficiently, and the other part should transmit to ThE devices. Photon management is a reasonable approach to realize the performance in hybrid systems. Recently, all sorts of photon-management subwavelength structures have been widely investigated to trap light in PV devices, including anti-reflection films, biomimetic structures, nanopillars, nanoholes, nanocones, nanopyramids, and plasmonic structures. The sizes of these structures are comparable to the wavelengths of solar spectrum and can lead to various excellent light-trapping effects due to light-matter interactions. Anti-reflection coatings were proposed the earliest to suppress the reflection by smoothing the change of refractive index from the air to
Corresponding author. E-mail address: [email protected] (Y. Xuan).
http://dx.doi.org/10.1016/j.enconman.2017.09.036 Received 5 July 2017; Received in revised form 13 September 2017; Accepted 13 September 2017 0196-8904/ © 2017 Elsevier Ltd. All rights reserved.
Energy Conversion and Management 152 (2017) 22–30
Y. Xu et al.
materials. Makableh et al. [6] fabricated an optimized ZnO film on the GaAs solar cell through the sol–gel method. The enhancement could be seen in spectral properties, conversion efficiency and the external quantum efficiency (EQE). Similar indium oxide films front contact was used to enhance the opto-electrical properties of heterojunction solar cells [7]. Kanda et al. [8] introduced an Al2O3/TiO2 double layer coating, and the reflection in the visible wavelength range was suppressed efficiently. Elshorbagy et al. [9] designed a customized Si3N4 coating with a 15.2% enhancement of the short circuit current. Various structured solar cells have developed rapidly with the discovery of moth-eye structures in solar cells, which could be attributed to the effective gradient refractive index, including nanocones and nanopyramids. The influences of the sizes of the moth-eye structure on the optical properties of thin-film silicon have been studied [10,11]. The moth-eye pattern could also be prepared on the surface of the Polydimethyl siloxane (PDMS) or polymethyl methacrylate (PMMA) film through the nanoimprint technology for the application of organic solar cells [12]. The same technology is employed to fabricate nanocone structures for high-efficiency CdS/CdTe solar cells [13]. Li et al. [14] applied the nanopyramid structure to a tandem silicon solar cell, which led to a remarkable improvement of the short circuit current density in both the top and bottom cells. Moreover, Zada et al. [15] fabricated nanostructures on the surface of a TiO2 film inspired by cicada wings with angle-independent anti-reflection properties. Nanopillars, nanoholes and gratings are typical structures for trapping light in solar cells as well. The light-trapping effect in these varieties of structures is attributed to the strong scattering between structures to increase light path in solar cells and generate effective optical coupling. He et al. [16] found that both the absorption and open circuit voltage were enhanced for organic-silicon heterojunction solar cell. Chen et al. [17] studied the light-trapping effects of nanohole arrays with different depths on the performance of silicon PV devices and found that the depth of 700 nm was sufficient for the optimal light-trapping effect. Besides these structures based on dielectric and semiconductor materials, metallic structures, such as metallic nanoparticles, gratings and metamaterials, have been revealed to provide advanced light trapping with specific plasmon effect. Clavero [18] gave the fundamentals of the hot-electron generation and regeneration process in plasmonic structures, which was the key point to get high-efficiency photovoltaics. Hsu et al. [19] studied the performance of different shapes of silver nanoplates embedded in organic and perovskite solar cells. With the aid of the surface plasmon resonance of the nanoplates, the power conversion efficiency was enhanced in both kinds of solar cells. Massiot et al. [20] used fishnet plasmonic absorber to excite surface plasmons for the perfect absorption in the GaAs solar cell. The application of these subwavelength structures is a quite efficient trick to enhance the PV efficiency and reduce the cost. Although these plasmonic structures can improve the absorption in materials, they will cause an unavoidable heating problem, which is an undesired loss for PV devices. This kind of loss was discussed by Vora et al. [21] to obtain the maximum useful absorption. Nevertheless, most researches only focus on the wavelength range above the bandgap for PV applications, and there are few reports to study the utilization of below-bandgap photons in the hybrid systems. It is a big challenge to get high transmission in the wavelength below the bandgap simultaneously with omnidirectional and polarization-insensitive properties [22,23]. Therefore, subwavelength structures should be developed to provide proper full-spectrum photon management for PV-ThE applications. High absorption for the abovebandgap wavelengths is to ensure the high efficiency of solar cells. The efficient transmission should be obtained to be the heat source of the ThE module, as well as the heat in the solar cells. The improved conversion efficiency of solar energy can be achieved for this hybrid system with the assistance of the proposed photon management. Based on this prior and crucial research work, in our present study, we propose a new method of photon management for full-spectrum light harvesting in PV-ThE hybrid applications. This photon
management can be realized through fabricated subwavelength structures with high absorption in the above-bandgap wavelength range and high transmission in the rest wavelengths, which also presents angleindependent and polarization-insensitive characteristics. In this way, both the full-spectrum solar energy and heat in solar cells will attribute to the power conversion efficiency of the hybrid system. The photonmanagement subwavelength structures consist of top novel hexagonal nanohole arrays and bottom bilayer films on a 400–μm-thick crystalline silicon (c-Si) wafer. On the top surface, novel hexagonal nanohole arrays are fabricated through the large-area polystyrene (PS) self-assembly method and Bosch deep silicon etching process. On the bottom side of the wafer, TiO2/SiO2 bilayer films are used to increase the transmittance for wavelengths of 1100–2500 nm. With fabricated composite subwavelength structures, near 97% total absorptance in the wavelength range of 300–1100 nm and about 60% total transmittance from 1200 nm to 2500 nm are obtained simultaneously. The lighttrapping property in the above-bandgap wavelength range is analyzed through the finite difference time domain (FDTD) simulations. The impacts of incident angles and polarization states on the spectral characteristics are also investigated. 2. Experimental and simulation methods 2.1. Experimental process At first, a c-Si wafer (p-type, bandgap: 1.12 eV (∼1100 nm), resistivity: 0.05–0.1 Ω cm, doping level: 10e16 cm−3) is dealt with the hydrophilic treatment for the preparation of the etching mask, which is soaked in a mixed solution consisted of concentrated sulfuric acid (H2SO4) and the hydrogen peroxide (H2O2) (the volume fraction of H2SO4 (98 wt%): H2O2 (30 wt%) is 3:1) at 90 °C for 15 min, and then ultrasonically cleaned by the acetone, ethanol and deionized (DI) water, successively. The area of the c-Si wafer is 30 mm × 30 mm. Over recent years, various types of technologies have been proposed for the manufacture of micro/nanostructures, such as metal-assisted catalyst etching (MACE), nanoimprinting, photolithography, electron beam lithography. Han et al. [24] gave a review of recent advances in MACE of silicon and the applications of these fabricated structures for sensors, energy storage and conversion. Nanoimprint technology is a prospective method for the fabrication of large-area and periodic structures [25] with enhanced light absorption to improve the power conversion efficiency. Sivasubramaniam and Alkaisi [26] obtained inverted nanopyramid structures through the combination of the interference lithography and relevant pattern transfer technologies. The electron beam lithography is a quite flexible method to get various complex and composite structures [27], whereas the cost of this technology is too expensive to be employed to the commercial production. Al-Douri et al. [28] also created CdS random nanostructures through simple sol–gel spin coating technique with remarkable optical properties. In our work, the Bosch deep silicon etching process [29] is employed to fabricate the novel top hexagonal nanohole arrays with periodic polystyrene (PS) masks made through the large-area PS sphere self-assembly method [30]. The schematic of the fabrication process is provided in Fig. 1. The PS spheres have unique characteristics as they can float on the water surface. On this basis, the PS monolayer can be obtained through the self-assembly process on the water surface. The original PS spheres are stored in ethanol, and the concentration and diameter are 2.5 w/v% and 600 nm, respectively, and this ethanol suspension is driven to the water surface with an injector. The spheres then can spread out slowly on the water surface according to the Marangoni effect. In order to obtain a high-quality PS monolayer, the PS dispersed suspension is 10 times diluted by 50 wt% ethanol solution and the diluted suspension is ultrasonic dispersed to avoid the agglomeration of PS spheres. The injection rate is 0.15 mm per minute, which is controlled by a pump. At the end of the injection process, 20 drops of Sodium Dodecyl Sulfonate (SDS) solution is dropped to the water surface to accomplish the self23
Energy Conversion and Management 152 (2017) 22–30
Y. Xu et al.
Fig. 1. Schematic of the fabrication process of the hexagonal nanohole arrays.
arrays with the same Bosch etching process. Magnetron sputtering is used to fabricate the TiO2/SiO2 bilayer films. Differently, the deposition of the TiO2 film uses Ti target and direct current sputtering at 400 W power and 0.8 Pa pressure. The flow rate Ar:O2 is 8:1. For the SiO2 film, the glass target is used with radiofrequency sputtering. The power is 200 W, and the pressure is 0.8 Pa with the same flow rate.
assembly process. After 5 min standing, the dispersive PS spheres become a close-packed periodic PS monolayer. This water-based floating PS monolayer is transferred to the surface of the c-Si substrate to create a mask for etching. The acreage of monolayer relies on the area of the water surface. The next is the etching process to fabricate the hexagonal nanohole arrays. In view of the verticality of the structures and the security of the etching gases, Bosch dry etching method is employed, and the reactant gases are chosen as SiF6 and C4F8. The etching process with the assistance of SiF6 gas is isotropous. It will have both horizontal and vertical etching at the same time with single SiF6 gas. Hence the C4F8 gas is added to passivate silicon to create a protective layer on the sidewalls. By using the two gases alternately during the inductively coupled plasma (ICP) etching process, the vertical nanopillar array will be gained with high depth-to-width ratio. The flow rate and etching power of the SiF6 and C4F8 gases are both 100 sccm and 600/15 W, respectively. Differently, the etching times for SiF6 and C4F8 are 9 s and 7 s, respectively. Repeat the above two etching process 20 times, and ordered nanopillar arrays are obtained. The diameter of the PS spheres we used is 600 nm uniformly, and this can be adjusted by oxygen plasma etching (PE). The etching power and flow rate we used are 200 W and 100 sccm, respectively. After reducing the diameter of the PS spheres, a silver film is sputtered on the present surface by the magnetron sputtering at 60 W power and 0.7 Pa pressure for 100 s. The PS monolayer photomask is then transformed to the silver ring mask after a lift-off step of the PS spheres. And finally, it comes to the hexagonal nanohole
2.2. Structural characteristics The fabricated structures are represented through field emission scanning electron microscope (SEM). An electron beam is shot to the surface of the sample, and these electrons interact with the extranuclear electrons of the nuclei of the sample. The extranuclear electrons will be divorced from the atoms and become the secondary electrons. These secondary electrons are amplified through a multiplier and captured by a fluorescent screen. With the aid of very sophisticated software, the crystal structures can be characterized. 2.3. Spectral measurements The absorptance A and transmittance T are measured based on a spectrophotometer (Agilent Cary 5000) in the full-spectrum wavelength range of 300–2500 nm. The measurement schematic of the spectrophotometer system is given in Fig. 2 [31]. The reflectance R at normal incidence is obtained by setting sample at the bottom of the integrating Fig. 2. The measurement schematic of the integrated system of Agilent Cary 5000 UV–VIS–NIR spectrophotometer.
24
Energy Conversion and Management 152 (2017) 22–30
Y. Xu et al.
A = 1−R
sphere (sample holder 2). When setting the sample at the center (sample holder 1), the sum of R and T with the change of incident angles can be obtained by rotating the sample holder 1. The light source can be transferred to TM or TE wave by adding a polarizer behind the reflector 1 with different angles. The spectral properties of a blank sample made of the polytetrafluoroethylene is measured as the base line for practical samples to ensure the accuracy of the measured data. The Photometric accuracy of this system is less than or equal to ± 0.00025 A, and the measured accuracy of R and T from this spectrophotometer can be less than 0.3% [32]. In order to measure R at oblique incidence, a black-body box is placed on the bottom of the sample to get rid of the influences of transmission. The spectral properties (RTM, RTE, TTM and TTE) at different polarization states (transverse magnetic (TM) and transverse electric (TE) wave) are measured by tuning polarization angles of the polarizer. The oblique-incident absorptance Aobl and transmittance Tobl can be obtained through the following formula
Aobl =
Tobl =
(1−(RTM + TTM )) + (1−(RTE + TTE )) 2
((RTM + TTM )−RTM ) + ((RTE + TTE )−RTE ) 2
The refractive index (n) and extinction coefficient (k) of crystalline silicon are from Palik [36] and given in Fig. 3(c). The top monitor can get the reflectance from the surface. 3. Results and discussion 3.1. Effects of the top hexagonal nanohole arrays As we know, the photon-management property of subwavelength structures is closely linked to the sizes of structures. Single subwavelength structures cannot meet the full-spectrum photon-management demands. Therefore, it is quite essential to fabricate composite structures to manage incident photons in different wavelength ranges. For the fabrication of subwavelength structures, nanopillar arrays are the basic structures for the manufacture of many other kinds of subwavelength structures. However, for single periodic nanopillar array, the absorption enhancement is in a narrow wavelength range, and the enhanced wavelengths are heavily dependent on the periods and diameters of nanopillars [37]. Many efforts have been done to get broadband absorption for nanopillar arrays, like combining nanopillars with different diameters, heights and arrangements. In this work, based on the structures of periodic nanopillars, novel hexagonal nanohole arrays are fabricated as shown in Fig. 4(a). From SEM images in Fig. 4(b), it can be seen that the surrounding nanoholes result from the first Bosch etching process with monolayer PS photomask. The central nanoholes are the results of the second Bosch etching process with the silver ring mask, and the diameter of central nanoholes can be adjusted by the oxygen PE process. The unit array consists of a central hole and six smaller holes around the central hole. This structure is also built on the single nanopillar array. The hexagonal distribution profits from the hexagonal arrangement of the PS spheres. In our work, four different cases of hexagonal nanohole arrays are fabricated with different diameters of central nanoholes as shown in the SEM images of Fig. 5(a). As mentioned above, the sizes of central holes are dependent on the oxygen PE process of PS spheres, and times for the second oxygen PE etching process are 35 s, 100 s, 150 s and 200 s, respectively. The height of the central nanohole array is about 800 nm and the period of the central nanohole array is about 600 nm. The absorptance and transmittance are measured from 300 nm to 2500 nm (see Fig. 5(b)), which is consistent with the AM1.5G solar spectrum. For nanopillars, the light-trapping property is mainly due to the scattering between the nanopillars to increase the light path [38]. The light can scatter between these nanopillars to let the reflected light enter the silicon again. In this way, the light path in the silicon is enhanced, which can improve the chances of the light to be absorbed by materials. The enhanced wavelengths are closely related to the periods and diameters of the nanopillars, as well as those of the single nanohole arrays
(1)
(2)
2.4. Simulation methods The simulations are carried out via the FDTD method, which is used to solve Maxwell’s curl equations to calculate the spectral characteristics in the relevant wavelength range. In this method, the frequency domain equation can be expressed as [33]
D (ω) = ε (ω)·E (ω)
(3)
where the electric field is E. The magnetic flux density is D. Frequency is ω and the complex permittivity is ε. The time step is determined by the following formula [34]
Δt = δ /(2·c )
(5)
(4)
where the smallest spatial step is δ and the speed of light in vacuum is c = 3 × 108 m/s. The condition of numerical convergence is that the residual error is less than 1e − 5 (0.001%). The model for simulation is shown in Fig. 3(a). In the simulation, the light source is set as a plane wave with wavelengths from 300 nm to 1100 nm. Periodic boundary conditions (PBC) are used with a unit structure in the x-axis and y-axis directions. In the z-axis direction, two perfect matched layers (PML) are employed as shown in Fig. 3(b). The photons arriving at the monitor will be absorbed entirely. The top monitor is used to get the reflectance R and the bottom PML can absorb all the photons in silicon, as the 400 μm-thick-silicon can also absorb the internal photons in the materials. The absorptance A can be obtained from [35]
Fig. 3. (a) Structural model and (b) boundary conditions of the FDTD simulation. (c) The optical constants of crystalline silicon.
25
Energy Conversion and Management 152 (2017) 22–30
Y. Xu et al.
Fig. 4. (a) SEM image of the ordered hexagonal nanohole arrays at 20°. (b) Top view of the SEM image of the ordered hexagonal nanohole arrays.
between the sizes of structures and incident wavelengths. The thickness of the sidewalls is much less than the wavelengths, which can further weaken the links between absorptance and the parameters of hexagonal nanoholes and obtain a much more broadband absorption from 300 nm to 1100 nm. To evaluate the photon-management effect more clearly, the total absorptance Atotal and transmittance Ttotal can be calculated by [10]
[38,39]. The absorption enhancement of nanopillars is mainly in the visible wavelength range with these parameters. It also can be seen that the absorption in the ultraviolet (UV) and near-infrared (NIR) range is relatively weak with smaller diameters of the central nanoholes. With increased diameters of central nanoholes, the absorptance is enhanced in the whole above-bandgap wavelength range. It is worth mentioning that, the larger central holes represent the thinner silicon sidewalls. As mentioned above, the absorption is closely related to the relationships
Fig. 5. (a) SEM images of hexagonal nanohole arrays with different diameters of central nanohole (∼460 nm, 380 nm, 290 nm, and 230 nm). (b) Full-spectrum absorption and transmission spectra of the cases in (a) with different diameters of central nanoholes.
26
Energy Conversion and Management 152 (2017) 22–30
Y. Xu et al.
Atotal =
Ttotal =
1100 nm hc A (λ ) I (λ ) dλ λ 1100 nm hc int300 nm λ I (λ ) dλ
chosen as 380 nm uniformly, which is corresponding to those in the second case of Fig. 5(a). The simulated results are given in Fig. 6(b) and are compared with the experimental results, and two kinds of results match quite well. Simulated results of bare silicon are very consistent with the experimental results. The absorption-enhanced wavelength range is also mainly in the visible range. The only difference is that the simulated results have several absorption peaks. This is due to the relevant resonances at specific wavelengths in the nanoholes caused by the scattering. For this kind of single nanopillar or nanohole arrays, the links between absorptance and parameters of structures are unavoidable. However, for the fabricated samples, both the heights of sidewalls and diameters of central nanoholes have slight randomness, especially for the nanoholes with thinner sidewalls. The irregular distribution will alleviate the size effects, which exhibits a smoothed variation. Furthermore, to see the scattering effects in the nanoholes more clearly, the cross-sectional electric field in the nanoholes is also displayed at the wavelengths of those absorption peaks (see Fig. 6(c), wavelength λ ∼ 576 nm, 700 nm, 758 nm, 830 nm, 880 nm). The electric distribution demonstrates the absorption mechanisms in the nanohole array. Like deep gratings, there exist several cavities in the central nanoholes, which display as enhanced electric field intensity. These cavities also affect the electric distribution in the surrounding nanoholes. That is to say that the light can be scattered between the two kinds of nanoholes. Originally, for the close-packed single nanopillars, only a little amount of incident light can enter the smaller nanoholes to generate the scattering effect. The added central nanoholes introduce more light into the holes to generate the scattering effect compared with single smaller nanoholes, which become a transmission media of the scattering light between those smaller nanoholes. The belowbandgap photons can also enter the structures, which enhances transmission in the NIR wavelength range, and full-spectrum anti-reflection
∫300 nm
1100 nm hc T (λ ) I (λ ) dλ λ 1100 nm hc int300 nm λ I (λ ) dλ
(6)
∫300 nm
(7)
where the incident wavelengths is λ. The absorptance and transmittance changing with wavelengths are A(λ) and T(λ), respectively. The Plank constant is h = 6.625 × 10−34 J s. The intensity of the incident AM1.5G (global air mass = 1.5) light source is I(λ). About 97% total absorptance can be obtained with the sample of the first case (named case a) in Fig. 5(a). For single nanopillar array, the scattering effect only occurs between the nanopillars. When coming to the hollow nanopillars, the incident light will be scattered in the central holes as well. Also, the added central nanoholes let more incident light enter solar cells, which indicates a better anti-reflection property. As the c-Si wafer takes no use of solar energy from 1100 nm to 2500 nm, this part of energy in the c-Si will transmit or be reflected to the air again. According to the absorptance of different cases, the first case in Fig. 5(a) has the best absorption and transmission properties. The increased diameters of central nanoholes let more incident light enter solar cells, and give rise to a stronger scattering between the two different kinds of nanoholes, which exhibits remarkable full-spectrum photon management in case a. With the top novel hexagonal nanohole arrays, remarkable full-spectrum photon management can be got via appropriate diameters of the central nanoholes. To reveal the absorption mechanisms of the hexagonal nanohole arrays, simulations are carried out by the FDTD method. The period and height are set at 600 nm and 800 nm, respectively, which can be seen in the SEM image in Fig. 6(a). The diameters of central nanoholes are
Fig. 6. (a) Cross-section SEM image of the hexagonal arrays. (b) The comparison of absorptance from 300 nm to 1000 nm between simulated results of periodic nanopillars, hexagonal nanoholes and experimental results. (c) The cross-sectional electric field intensity around the hexagonal nanohole arrays at different wavelengths (∼576 nm, 700 nm, 758 nm, 830 nm and 880 nm).
27
Energy Conversion and Management 152 (2017) 22–30
Y. Xu et al.
3.2. Effects of the bottom TiO2/SiO2 films In order to fully use the NIR solar energy for ThE devices, the transmittance from 1100 nm to 2500 nm should be further improved. Due to the suppressed reflection at the top silicon-air interface, more photons will enter into the wafer. However, only part of the NIR energy can transmit to the bottom ThE device, and it also suffers from a transmission loss at the bottom silicon-air interface, which is caused by reflected photons at this interface. Efforts should be taken to eliminate the reflection loss at the material interface to enhance transmission. The polished top surface was used to fabricate the hexagonal nanohole arrays. The bottom side of the c-Si wafer is unpolished, and it is inappropriate to fabricate the PS monolayer mask for etching. In this way, In consideration of the gradient refractive index method [41], film structures are employed at the rough bottom side with the refractive index between air and silicon. TiO2 and SiO2 are chosen as film materials, which have the refractive index of ∼2.5 and ∼1.5, respectively. The thicknesses of fabricated TiO2 and SiO2 films are both about 50 nm. The absorptance and transmittance are measured by the spectrophotometer as shown in Fig. 7. Due to the gradient refractive index at the silicon-air interface, the transmittance is enhanced to about 60% from 1200 nm to 2500 nm compared with the transmittance of the sample without bottom films, and the films have no influences on the absorptance. In this way, high transmission (1200 nm–2500 nm) and high absorption (300 nm–1100 nm) can be obtained simultaneously
Fig. 7. The comparison of absorption and transmission spectra between case a and case a with bottom TiO2/SiO2 anti-reflection bilayer films.
properties are achieved significantly. Similarly, there can be seen “channelling model” in the study [40], which leads to the electric field intensity enhancement in the nanoholes. This model will finally result in broadband absorption in the in-between sidewalls and the substrate due to the high extinction coefficient of silicon. Passivation should be carried out to handle the high surface recombination at the sidewalls for the future fabrication of solar cells.
Fig. 8. Full-spectrum absorption and transmission spectra of the hexagonal nanohole arrays of case a & films with the change of (a) incident angles (0°–60°) and (b) polarization states (TE and TM) at different incident angles (20°, 40° and 60°). (c) Total absorption and transmission spectra with the change of incident angles and polarizations.
28
Energy Conversion and Management 152 (2017) 22–30
Y. Xu et al.
[6] Makableh YF, Vasan R, Sarker JC, Nusir AI, Seal S, Manasreh MO. Enhancement of GaAs solar cell performance by using a ZnO sol-gel anti-reflection coating. Sol Energy Mater Sol Cells 2014;123:178–82. [7] Anh Huy Tuan L, Ahn S, Kim S, Han S, Kim S, Park H, et al. A statistical approach for the optimization of indium tin oxide films used as a front contact in amorphous/ crystalline silicon heterojunction solar cells. Energy Convers Manage 2014;87:191–8. [8] Kanda H, Uzum A, Harano N, Yoshinaga S, Ishikawa Y, Uraoka Y, et al. Al2O3/TiO2 double layer anti-reflection coating film for crystalline silicon solar cells formed by spray pyrolysis. Energy Sci Eng 2016;4. [9] Elshorbagy MH, Abdel-Hady K, Kamal H, Alda J. Broadband anti-reflection coating using dielectric Si3N4 nanostructures. Application to amorphous-Si-H solar cells. Opt Commun 2017;390:130–6. [10] Zhang Y, Xuan Y. Biomimetic omnidirectional broadband structured surface for photon management in photovoltaic-thermoelectric hybrid systems. Sol Energy Mater Sol Cells 2016;144:68–77. [11] Xu Y, Xuan Y, Yang L. Full-spectrum photon management of solar cell structures for photovoltaic-thermoelectric hybrid systems. Energy Convers Manage 2015;103:533–41. [12] Leem JW, Kim S, Lee SH, Rogers JA, Kim E, Yu JS. Efficiency enhancement of organic solar cells using hydrophobic antireflective inverted moth-eye nanopatterned PDMS films. Adv Energy Mater 2014;4. [13] Tsui KH, Lin Q, Chou H, Zhang Q, Fu H, Qi P, et al. Low-cost, flexible, and selfcleaning 3D nanocone anti-reflection films for high-efficiency photovoltaics. Adv Mater 2014;26:2805–11. [14] Li G, Li H, Ho JYL, Wong M, Kwok HS. Nanopyramid structure for ultrathin c-Si tandem solar cells. Nano Lett 2014;14:2563–8. [15] Zada I, Zhang W, Li Y, Sun P, Cai N, Gu J, et al. Angle dependent antireflection property of TiO2 inspired by cicada wings. Appl Phys Lett 2016;109. [16] He J, Yang Z, Liu P, Wu S, Gao P, Wang M, et al. Enhanced electro-optical properties of nanocone/nanopillar dual-structured arrays for ultrathin silicon/organic hybrid solar cell applications. Adv Energy Mater 2016;6. [17] Chen T-G, Yu P, Chen S-W, Chang F-Y, Huang B-Y, Cheng Y-C, et al. Characteristics of large-scale nanohole arrays for thin-silicon photovoltaics. Prog Photovolt 2014;22:452–61. [18] Clavero C. Plasmon-induced hot-electron generation at nanoparticle/metal-oxide interfaces for photovoltaic and photocatalytic devices. Nat Photon 2014;8:95–103. [19] Hsu H-L, Juang T-Y, Chen C-P, Hsieh C-M, Yang C-C, Huang C-L, et al. Enhanced efficiency of organic and perovskite photovoltaics from shape-dependent broadband plasmonic effects of silver nanoplates. Sol Energy Mater Sol Cells 2015;140:224–31. [20] Massiot I, Vandamme N, Bardou N, Dupuis C, Lemaitre A, Guillemoles J-F, et al. Metal nanogrid for broadband multiresonant light-harvesting in ultrathin GaAs layers. Acs Photon 2014;1:878–84. [21] Vora A, Gwamuri J, Pala N, Kulkarni A, Pearce JM, Gueney DO. Exchanging ohmic losses in metamaterial absorbers with useful optical absorption for photovoltaics. Sci Rep 2014;4. [22] Dincer F, Akgol O, Karaaslan M, Unal E, Sabah C. Polarization angle independent perfect metamaterial absorbers for solar cell applications in the microwave, infrared, and visible regime. Prog Electromagn Res - Pier 2014;144:93–101. [23] Rufangura P, Sabah C. Dual-band perfect metamaterial absorber for solar cell applications. Vacuum 2015;120:68–74. [24] Han H, Huang ZP, Lee W. Metal-assisted chemical etching of silicon and nanotechnology applications. Nano Today 2014;9:271–304. [25] Chen JD, Cui CH, Li YQ, Zhou L, Ou QD, Li C, et al. Single-junction polymer solar cells exceeding 10% power conversion efficiency. Adv Mater 2015;27:1035–41. [26] Sivasubramaniam S, Alkaisi MM. Inverted nanopyramid texturing for silicon solar cells using interference lithography. Microelectron Eng 2014;119:146–50. [27] Nicaise SM, Cheng JJ, Kiani A, Gradecak S, Berggren KK. Control of zinc oxide nanowire array properties with electron-beam lithography templating for photovoltaic applications. Nanotechnology 2015;26. [28] Al-Douri Y, Khasawneh Q, Kiwan S, Hashim U, Abd Hamid SB, Reshak AH, et al. Structural and optical insights to enhance solar cell performance of CdS nanostructures. Energy Convers Manage 2014;82:238–43. [29] Chang CL, Wang YF, Kanamori Y, Shih JJ, Kawai Y, Lee CK, et al. Etching submicrometer trenches by using the Bosch process and its application to the fabrication of antireflection structures. J Micromech Microeng 2005;15:580–5. [30] Gao P, He J, Zhou S, Yang X, Li S, Sheng J, et al. Large-area nanosphere self-assembly by a micro-propulsive injection method for high throughput periodic surface nanotexturing. Nano Lett 2015;15:4591–8. [31] Photometric accuracy of the Agilent DRA in the NIR region using the double aperture technique. < http://www.agilent.com/cs/library/datasheets/public/auvvis06.pdf > . [32] User guide for Cary 5000 absorption spectrometer with external DRA 1800 attachment. < http://mmrc.caltech.edu/Cary%20UV-Vis%20Int.Sphere/manuals/ Cary5000_User_Guide.pdf > . [33] Taflove A, Umashankar KR. The finite-difference time-domain method for numerical modeling of electromagnetic wave interactions. Electromagnetics 1990;10:105–26. [34] Yang L, Xuan Y, Han Y, Tan J. Investigation on the performance enhancement of silicon solar cells with an assembly grating structure. Energy Convers Manage 2012;54:30–7. [35] Soum-Glaude A, Bousquet I, Thomas L, Flamant G. Optical modeling of multilayered coatings based on SiC(N)H materials for their potential use as high-temperature solar selective absorbers. Sol Energy Mater Sol Cells 2013;117:315–23. [36] Palik ED. Handbook of optical constants of solids; 1985.
with the combination of top and bottom subwavelength structures. 3.3. Effects of the incident angles and polarization states Due to the different incident angles and polarization states of sunlight, it is essential to ensure the photon-management performance of composite subwavelength structures have omnidirectional and polarization-insensitive properties. Therefore, according to the characteristics of natural light, the full-spectrum photon management under the circumstances of oblique incidence and different polarization states are further investigated. It can be seen from Fig. 8(a) and (b) that both the absorptance and transmittance change little with the increase of incident angles. Apparently, the absorption and transmission spectra in the cases of TM and TE polarizations almost coincide at same incident angles. The total absorptance for the wavelengths of 300–1100 nm and the total transmittance from 1100 nm and 2500 nm are also calculated and are shown in Fig. 8(c). Both total absorptance and transmittance are almost unaffected by the TM or TE polarization. The total absorption and transmittance can maintain above 95% and 52%, respectively, when the incident angles are below 50°. It is worth noting that, the total transmittance has an apparent drop when it comes to the oblique incidence. Also, the polarization-insensitive property of transmittance is inferior to that of absorptance. Therefore, the top hexagonal nanohole arrays perform better than the film structures for the angle-independent and polarization-insensitive requirements. It suggests that this kind of hexagonal nanohole array is quite usable for photon management in different conditions of sunlight. In this way, the fabricated subwavelength structures can manage the full-spectrum incident light efficiently for different types of work conditions. 4. Conclusions In summary, a new approach of full-spectrum photon management is proposed for PV-ThE hybrid systems via fabricated novel composite subwavelength structures, which consist of top novel hexagonal nanohole arrays and the bottom sputtered TiO2/SiO2 bilayer anti-reflection films. Remarkable light trapping in the above-bandgap wavelengths (∼97% total absorption) is achieved due to the scattering effects between the two kinds of nanohole arrays, thereby enhancing the light propagation path in silicon. The transmittance in the NIR wavelength range can reach about 60% from 1200 nm to 2500 nm by employing gradient refractive index. Besides, the performance is omnidirectional and polarization-insensitive. The proposed design can help to allocate the photons to different devices smartly for the PV-ThE hybrid systems and has a great potential in enhancing the utilization efficiency of solar energy by using this hybrid system with various kinds of solar cells. This full-spectrum photon-management approach may pave the way for other solar energy harvesting applications. Acknowledgements This work was financially supported by the National Natural Science Foundation of China (Grant No. 51590901). References [1] Park K-T, Shin S-M, Tazebay AS, Um H-D, Jung J-Y, Jee S-W, et al. Lossless hybridization between photovoltaic and thermoelectric devices. Sci Rep 2013;3. [2] Lamba R, Kaushik SC. Modeling and performance analysis of a concentrated photovoltaic-thermoelectric hybrid power generation system. Energy Convers Manage 2016;115:288–98. [3] Hsueh T-J, Shieh J-M, Yeh Y-M. Hybrid Cd-free CIGS solar cell/TEG device with ZnO nanowires. Prog Photovolt 2015;23:507–12. [4] Zhang J, Xuan Y. Investigation on the effect of thermal resistances on a highly concentrated photovoltaic-thermoelectric hybrid system. Energy Convers Manage 2016;129:1–10. [5] Yin E, Li Q, Xuan Y. Thermal resistance analysis and optimization of photovoltaicthermoelectric hybrid system. Energy Convers Manage 2017;143:188–202.
29
Energy Conversion and Management 152 (2017) 22–30
Y. Xu et al.
solar photovoltaics. Nano Lett 2010;10:1012–5. [40] Gomard G, Peretti R, Callard S, Meng X, Artinyan R, Deschamps T, et al. Blue light absorption enhancement based on vertically channelling modes in nano-holes arrays. Appl Phys Lett 2014;104. [41] Southwell WH. Gradient-index antireflection coatings. Opt Lett 1983;8:584–6.
[37] Zhong S, Zeng Y, Huang Z, Shen W. Superior broadband antireflection from buried Mie resonator arrays for high-efficiency photovoltaics. Sci Rep 2015;5. [38] Garnett E, Yang P. Light trapping in silicon nanowire solar cells. Nano Lett 2010;10:1082–7. [39] Han SE, Chen G. Optical absorption enhancement in silicon nanohole arrays for
30
Contents lists available at ScienceDirect
Energy Conversion and Management journal homepage: www.elsevier.com/locate/enconman
Broadband photon management of subwavelength structures surface for full-spectrum utilization of solar energy
MARK
⁎
Yuanpei Xu, Yimin Xuan , Xianglei Liu School of Energy and Power Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, China
A R T I C L E I N F O
A B S T R A C T
Keywords: Composite subwavelength structures Full spectrum Photon management Photovoltaic-thermoelectric hybrid systems
In this work, advanced photon-management composite subwavelength structures are fabricated to manage solar energy in the full-spectrum (300–2500 nm) wavelength range for the application of the photovoltaic-thermoelectric hybrid systems to fully utilize the solar energy. This proposed photon-management method will simultaneously realize efficient light trapping for the photons with above-bandgap energy in solar cells and the utilization of the below-bandgap photons and the waste heat resulting from the thermalization effect in solar cells with bottom thermoelectric devices. The typical structure is designed: Top ordered hexagonal nanohole arrays can trap above-bandgap photons to enhance the absorption in the silicon wafer, and TiO2/SiO2 bilayer films are deposited on the bottom side of the wafer to improve the transmission of the below-bandgap photons. ∼97% total absorptance for wavelengths of 300–1100 nm is achieved with optimized diameters of central nanoholes. The total transmittance, on the other hand, is improved to ∼60% from 1200 nm to 2500 nm. The results indicate that the structures realize the appropriate allocation of photons within different wavelength ranges to different devices for sufficient utilization of full-spectrum solar energy. This novel full-spectrum photon management benefits from the strong scattering effect among the nanoholes and the gradient refractive index of bilayer films. Moreover, the photon-management performance shows angle-independent and polarization-insensitive characteristics. This method can be applied to various kinds of solar cells for photovoltaic-thermoelectric hybrid systems and may provide thoughts for other solar harvesting applications.
1. Introduction Solar energy, which is a kind of renewable and abundant energy, has gained wide attention due to the intense needs for clean energy. Various utilization methods of solar energy have been proposed. At present, the photovoltaic (PV) device is one of the most popular approaches for solar energy application. However, the utilization efficiency of full-spectrum (300–2500 nm) solar energy reaching the ground is limited by the bandgap of solar cells. The PV devices can only take use of incident solar energy above bandgap. Furthermore, part of the absorbed light in solar cells will be converted to heat due to the recombination of electron-hole pairs in the active layer or at the surface, which is called thermalization loss [1]. This recombination process will cause the increase of the temperature of solar cells, and further brings down the usage of solar energy. To break these limitations in PV devices, a novel PV-thermoelectric (ThE) hybrid system has been proposed to improve the utilization efficiency of full-spectrum solar energy, which combines PV devices and ThE devices into a hybrid PV-ThE system. A concentrated PV-ThE hybrid system was systematically
⁎
analyzed by Lamba and Kaushik [2] with the influences of thermocouple numbers, the irradiance, PV and ThE current. Hsush et al. [3] proposed a CuInGaSe2 (CIGS)/ThE device with the efficient of 22% by employing passive light-trapping ZnO nanowires. Zhang and Yin [4,5] systematically investigated the influences of the thermal resistances on the hybrid system under the circumstance of high concentration ratio. In this system, the first-line goal is to allocate the full-spectrum solar energy to different devices. The solar energy above the bandgap must be absorbed by PV devices efficiently, and the other part should transmit to ThE devices. Photon management is a reasonable approach to realize the performance in hybrid systems. Recently, all sorts of photon-management subwavelength structures have been widely investigated to trap light in PV devices, including anti-reflection films, biomimetic structures, nanopillars, nanoholes, nanocones, nanopyramids, and plasmonic structures. The sizes of these structures are comparable to the wavelengths of solar spectrum and can lead to various excellent light-trapping effects due to light-matter interactions. Anti-reflection coatings were proposed the earliest to suppress the reflection by smoothing the change of refractive index from the air to
Corresponding author. E-mail address: [email protected] (Y. Xuan).
http://dx.doi.org/10.1016/j.enconman.2017.09.036 Received 5 July 2017; Received in revised form 13 September 2017; Accepted 13 September 2017 0196-8904/ © 2017 Elsevier Ltd. All rights reserved.
Energy Conversion and Management 152 (2017) 22–30
Y. Xu et al.
materials. Makableh et al. [6] fabricated an optimized ZnO film on the GaAs solar cell through the sol–gel method. The enhancement could be seen in spectral properties, conversion efficiency and the external quantum efficiency (EQE). Similar indium oxide films front contact was used to enhance the opto-electrical properties of heterojunction solar cells [7]. Kanda et al. [8] introduced an Al2O3/TiO2 double layer coating, and the reflection in the visible wavelength range was suppressed efficiently. Elshorbagy et al. [9] designed a customized Si3N4 coating with a 15.2% enhancement of the short circuit current. Various structured solar cells have developed rapidly with the discovery of moth-eye structures in solar cells, which could be attributed to the effective gradient refractive index, including nanocones and nanopyramids. The influences of the sizes of the moth-eye structure on the optical properties of thin-film silicon have been studied [10,11]. The moth-eye pattern could also be prepared on the surface of the Polydimethyl siloxane (PDMS) or polymethyl methacrylate (PMMA) film through the nanoimprint technology for the application of organic solar cells [12]. The same technology is employed to fabricate nanocone structures for high-efficiency CdS/CdTe solar cells [13]. Li et al. [14] applied the nanopyramid structure to a tandem silicon solar cell, which led to a remarkable improvement of the short circuit current density in both the top and bottom cells. Moreover, Zada et al. [15] fabricated nanostructures on the surface of a TiO2 film inspired by cicada wings with angle-independent anti-reflection properties. Nanopillars, nanoholes and gratings are typical structures for trapping light in solar cells as well. The light-trapping effect in these varieties of structures is attributed to the strong scattering between structures to increase light path in solar cells and generate effective optical coupling. He et al. [16] found that both the absorption and open circuit voltage were enhanced for organic-silicon heterojunction solar cell. Chen et al. [17] studied the light-trapping effects of nanohole arrays with different depths on the performance of silicon PV devices and found that the depth of 700 nm was sufficient for the optimal light-trapping effect. Besides these structures based on dielectric and semiconductor materials, metallic structures, such as metallic nanoparticles, gratings and metamaterials, have been revealed to provide advanced light trapping with specific plasmon effect. Clavero [18] gave the fundamentals of the hot-electron generation and regeneration process in plasmonic structures, which was the key point to get high-efficiency photovoltaics. Hsu et al. [19] studied the performance of different shapes of silver nanoplates embedded in organic and perovskite solar cells. With the aid of the surface plasmon resonance of the nanoplates, the power conversion efficiency was enhanced in both kinds of solar cells. Massiot et al. [20] used fishnet plasmonic absorber to excite surface plasmons for the perfect absorption in the GaAs solar cell. The application of these subwavelength structures is a quite efficient trick to enhance the PV efficiency and reduce the cost. Although these plasmonic structures can improve the absorption in materials, they will cause an unavoidable heating problem, which is an undesired loss for PV devices. This kind of loss was discussed by Vora et al. [21] to obtain the maximum useful absorption. Nevertheless, most researches only focus on the wavelength range above the bandgap for PV applications, and there are few reports to study the utilization of below-bandgap photons in the hybrid systems. It is a big challenge to get high transmission in the wavelength below the bandgap simultaneously with omnidirectional and polarization-insensitive properties [22,23]. Therefore, subwavelength structures should be developed to provide proper full-spectrum photon management for PV-ThE applications. High absorption for the abovebandgap wavelengths is to ensure the high efficiency of solar cells. The efficient transmission should be obtained to be the heat source of the ThE module, as well as the heat in the solar cells. The improved conversion efficiency of solar energy can be achieved for this hybrid system with the assistance of the proposed photon management. Based on this prior and crucial research work, in our present study, we propose a new method of photon management for full-spectrum light harvesting in PV-ThE hybrid applications. This photon
management can be realized through fabricated subwavelength structures with high absorption in the above-bandgap wavelength range and high transmission in the rest wavelengths, which also presents angleindependent and polarization-insensitive characteristics. In this way, both the full-spectrum solar energy and heat in solar cells will attribute to the power conversion efficiency of the hybrid system. The photonmanagement subwavelength structures consist of top novel hexagonal nanohole arrays and bottom bilayer films on a 400–μm-thick crystalline silicon (c-Si) wafer. On the top surface, novel hexagonal nanohole arrays are fabricated through the large-area polystyrene (PS) self-assembly method and Bosch deep silicon etching process. On the bottom side of the wafer, TiO2/SiO2 bilayer films are used to increase the transmittance for wavelengths of 1100–2500 nm. With fabricated composite subwavelength structures, near 97% total absorptance in the wavelength range of 300–1100 nm and about 60% total transmittance from 1200 nm to 2500 nm are obtained simultaneously. The lighttrapping property in the above-bandgap wavelength range is analyzed through the finite difference time domain (FDTD) simulations. The impacts of incident angles and polarization states on the spectral characteristics are also investigated. 2. Experimental and simulation methods 2.1. Experimental process At first, a c-Si wafer (p-type, bandgap: 1.12 eV (∼1100 nm), resistivity: 0.05–0.1 Ω cm, doping level: 10e16 cm−3) is dealt with the hydrophilic treatment for the preparation of the etching mask, which is soaked in a mixed solution consisted of concentrated sulfuric acid (H2SO4) and the hydrogen peroxide (H2O2) (the volume fraction of H2SO4 (98 wt%): H2O2 (30 wt%) is 3:1) at 90 °C for 15 min, and then ultrasonically cleaned by the acetone, ethanol and deionized (DI) water, successively. The area of the c-Si wafer is 30 mm × 30 mm. Over recent years, various types of technologies have been proposed for the manufacture of micro/nanostructures, such as metal-assisted catalyst etching (MACE), nanoimprinting, photolithography, electron beam lithography. Han et al. [24] gave a review of recent advances in MACE of silicon and the applications of these fabricated structures for sensors, energy storage and conversion. Nanoimprint technology is a prospective method for the fabrication of large-area and periodic structures [25] with enhanced light absorption to improve the power conversion efficiency. Sivasubramaniam and Alkaisi [26] obtained inverted nanopyramid structures through the combination of the interference lithography and relevant pattern transfer technologies. The electron beam lithography is a quite flexible method to get various complex and composite structures [27], whereas the cost of this technology is too expensive to be employed to the commercial production. Al-Douri et al. [28] also created CdS random nanostructures through simple sol–gel spin coating technique with remarkable optical properties. In our work, the Bosch deep silicon etching process [29] is employed to fabricate the novel top hexagonal nanohole arrays with periodic polystyrene (PS) masks made through the large-area PS sphere self-assembly method [30]. The schematic of the fabrication process is provided in Fig. 1. The PS spheres have unique characteristics as they can float on the water surface. On this basis, the PS monolayer can be obtained through the self-assembly process on the water surface. The original PS spheres are stored in ethanol, and the concentration and diameter are 2.5 w/v% and 600 nm, respectively, and this ethanol suspension is driven to the water surface with an injector. The spheres then can spread out slowly on the water surface according to the Marangoni effect. In order to obtain a high-quality PS monolayer, the PS dispersed suspension is 10 times diluted by 50 wt% ethanol solution and the diluted suspension is ultrasonic dispersed to avoid the agglomeration of PS spheres. The injection rate is 0.15 mm per minute, which is controlled by a pump. At the end of the injection process, 20 drops of Sodium Dodecyl Sulfonate (SDS) solution is dropped to the water surface to accomplish the self23
Energy Conversion and Management 152 (2017) 22–30
Y. Xu et al.
Fig. 1. Schematic of the fabrication process of the hexagonal nanohole arrays.
arrays with the same Bosch etching process. Magnetron sputtering is used to fabricate the TiO2/SiO2 bilayer films. Differently, the deposition of the TiO2 film uses Ti target and direct current sputtering at 400 W power and 0.8 Pa pressure. The flow rate Ar:O2 is 8:1. For the SiO2 film, the glass target is used with radiofrequency sputtering. The power is 200 W, and the pressure is 0.8 Pa with the same flow rate.
assembly process. After 5 min standing, the dispersive PS spheres become a close-packed periodic PS monolayer. This water-based floating PS monolayer is transferred to the surface of the c-Si substrate to create a mask for etching. The acreage of monolayer relies on the area of the water surface. The next is the etching process to fabricate the hexagonal nanohole arrays. In view of the verticality of the structures and the security of the etching gases, Bosch dry etching method is employed, and the reactant gases are chosen as SiF6 and C4F8. The etching process with the assistance of SiF6 gas is isotropous. It will have both horizontal and vertical etching at the same time with single SiF6 gas. Hence the C4F8 gas is added to passivate silicon to create a protective layer on the sidewalls. By using the two gases alternately during the inductively coupled plasma (ICP) etching process, the vertical nanopillar array will be gained with high depth-to-width ratio. The flow rate and etching power of the SiF6 and C4F8 gases are both 100 sccm and 600/15 W, respectively. Differently, the etching times for SiF6 and C4F8 are 9 s and 7 s, respectively. Repeat the above two etching process 20 times, and ordered nanopillar arrays are obtained. The diameter of the PS spheres we used is 600 nm uniformly, and this can be adjusted by oxygen plasma etching (PE). The etching power and flow rate we used are 200 W and 100 sccm, respectively. After reducing the diameter of the PS spheres, a silver film is sputtered on the present surface by the magnetron sputtering at 60 W power and 0.7 Pa pressure for 100 s. The PS monolayer photomask is then transformed to the silver ring mask after a lift-off step of the PS spheres. And finally, it comes to the hexagonal nanohole
2.2. Structural characteristics The fabricated structures are represented through field emission scanning electron microscope (SEM). An electron beam is shot to the surface of the sample, and these electrons interact with the extranuclear electrons of the nuclei of the sample. The extranuclear electrons will be divorced from the atoms and become the secondary electrons. These secondary electrons are amplified through a multiplier and captured by a fluorescent screen. With the aid of very sophisticated software, the crystal structures can be characterized. 2.3. Spectral measurements The absorptance A and transmittance T are measured based on a spectrophotometer (Agilent Cary 5000) in the full-spectrum wavelength range of 300–2500 nm. The measurement schematic of the spectrophotometer system is given in Fig. 2 [31]. The reflectance R at normal incidence is obtained by setting sample at the bottom of the integrating Fig. 2. The measurement schematic of the integrated system of Agilent Cary 5000 UV–VIS–NIR spectrophotometer.
24
Energy Conversion and Management 152 (2017) 22–30
Y. Xu et al.
A = 1−R
sphere (sample holder 2). When setting the sample at the center (sample holder 1), the sum of R and T with the change of incident angles can be obtained by rotating the sample holder 1. The light source can be transferred to TM or TE wave by adding a polarizer behind the reflector 1 with different angles. The spectral properties of a blank sample made of the polytetrafluoroethylene is measured as the base line for practical samples to ensure the accuracy of the measured data. The Photometric accuracy of this system is less than or equal to ± 0.00025 A, and the measured accuracy of R and T from this spectrophotometer can be less than 0.3% [32]. In order to measure R at oblique incidence, a black-body box is placed on the bottom of the sample to get rid of the influences of transmission. The spectral properties (RTM, RTE, TTM and TTE) at different polarization states (transverse magnetic (TM) and transverse electric (TE) wave) are measured by tuning polarization angles of the polarizer. The oblique-incident absorptance Aobl and transmittance Tobl can be obtained through the following formula
Aobl =
Tobl =
(1−(RTM + TTM )) + (1−(RTE + TTE )) 2
((RTM + TTM )−RTM ) + ((RTE + TTE )−RTE ) 2
The refractive index (n) and extinction coefficient (k) of crystalline silicon are from Palik [36] and given in Fig. 3(c). The top monitor can get the reflectance from the surface. 3. Results and discussion 3.1. Effects of the top hexagonal nanohole arrays As we know, the photon-management property of subwavelength structures is closely linked to the sizes of structures. Single subwavelength structures cannot meet the full-spectrum photon-management demands. Therefore, it is quite essential to fabricate composite structures to manage incident photons in different wavelength ranges. For the fabrication of subwavelength structures, nanopillar arrays are the basic structures for the manufacture of many other kinds of subwavelength structures. However, for single periodic nanopillar array, the absorption enhancement is in a narrow wavelength range, and the enhanced wavelengths are heavily dependent on the periods and diameters of nanopillars [37]. Many efforts have been done to get broadband absorption for nanopillar arrays, like combining nanopillars with different diameters, heights and arrangements. In this work, based on the structures of periodic nanopillars, novel hexagonal nanohole arrays are fabricated as shown in Fig. 4(a). From SEM images in Fig. 4(b), it can be seen that the surrounding nanoholes result from the first Bosch etching process with monolayer PS photomask. The central nanoholes are the results of the second Bosch etching process with the silver ring mask, and the diameter of central nanoholes can be adjusted by the oxygen PE process. The unit array consists of a central hole and six smaller holes around the central hole. This structure is also built on the single nanopillar array. The hexagonal distribution profits from the hexagonal arrangement of the PS spheres. In our work, four different cases of hexagonal nanohole arrays are fabricated with different diameters of central nanoholes as shown in the SEM images of Fig. 5(a). As mentioned above, the sizes of central holes are dependent on the oxygen PE process of PS spheres, and times for the second oxygen PE etching process are 35 s, 100 s, 150 s and 200 s, respectively. The height of the central nanohole array is about 800 nm and the period of the central nanohole array is about 600 nm. The absorptance and transmittance are measured from 300 nm to 2500 nm (see Fig. 5(b)), which is consistent with the AM1.5G solar spectrum. For nanopillars, the light-trapping property is mainly due to the scattering between the nanopillars to increase the light path [38]. The light can scatter between these nanopillars to let the reflected light enter the silicon again. In this way, the light path in the silicon is enhanced, which can improve the chances of the light to be absorbed by materials. The enhanced wavelengths are closely related to the periods and diameters of the nanopillars, as well as those of the single nanohole arrays
(1)
(2)
2.4. Simulation methods The simulations are carried out via the FDTD method, which is used to solve Maxwell’s curl equations to calculate the spectral characteristics in the relevant wavelength range. In this method, the frequency domain equation can be expressed as [33]
D (ω) = ε (ω)·E (ω)
(3)
where the electric field is E. The magnetic flux density is D. Frequency is ω and the complex permittivity is ε. The time step is determined by the following formula [34]
Δt = δ /(2·c )
(5)
(4)
where the smallest spatial step is δ and the speed of light in vacuum is c = 3 × 108 m/s. The condition of numerical convergence is that the residual error is less than 1e − 5 (0.001%). The model for simulation is shown in Fig. 3(a). In the simulation, the light source is set as a plane wave with wavelengths from 300 nm to 1100 nm. Periodic boundary conditions (PBC) are used with a unit structure in the x-axis and y-axis directions. In the z-axis direction, two perfect matched layers (PML) are employed as shown in Fig. 3(b). The photons arriving at the monitor will be absorbed entirely. The top monitor is used to get the reflectance R and the bottom PML can absorb all the photons in silicon, as the 400 μm-thick-silicon can also absorb the internal photons in the materials. The absorptance A can be obtained from [35]
Fig. 3. (a) Structural model and (b) boundary conditions of the FDTD simulation. (c) The optical constants of crystalline silicon.
25
Energy Conversion and Management 152 (2017) 22–30
Y. Xu et al.
Fig. 4. (a) SEM image of the ordered hexagonal nanohole arrays at 20°. (b) Top view of the SEM image of the ordered hexagonal nanohole arrays.
between the sizes of structures and incident wavelengths. The thickness of the sidewalls is much less than the wavelengths, which can further weaken the links between absorptance and the parameters of hexagonal nanoholes and obtain a much more broadband absorption from 300 nm to 1100 nm. To evaluate the photon-management effect more clearly, the total absorptance Atotal and transmittance Ttotal can be calculated by [10]
[38,39]. The absorption enhancement of nanopillars is mainly in the visible wavelength range with these parameters. It also can be seen that the absorption in the ultraviolet (UV) and near-infrared (NIR) range is relatively weak with smaller diameters of the central nanoholes. With increased diameters of central nanoholes, the absorptance is enhanced in the whole above-bandgap wavelength range. It is worth mentioning that, the larger central holes represent the thinner silicon sidewalls. As mentioned above, the absorption is closely related to the relationships
Fig. 5. (a) SEM images of hexagonal nanohole arrays with different diameters of central nanohole (∼460 nm, 380 nm, 290 nm, and 230 nm). (b) Full-spectrum absorption and transmission spectra of the cases in (a) with different diameters of central nanoholes.
26
Energy Conversion and Management 152 (2017) 22–30
Y. Xu et al.
Atotal =
Ttotal =
1100 nm hc A (λ ) I (λ ) dλ λ 1100 nm hc int300 nm λ I (λ ) dλ
chosen as 380 nm uniformly, which is corresponding to those in the second case of Fig. 5(a). The simulated results are given in Fig. 6(b) and are compared with the experimental results, and two kinds of results match quite well. Simulated results of bare silicon are very consistent with the experimental results. The absorption-enhanced wavelength range is also mainly in the visible range. The only difference is that the simulated results have several absorption peaks. This is due to the relevant resonances at specific wavelengths in the nanoholes caused by the scattering. For this kind of single nanopillar or nanohole arrays, the links between absorptance and parameters of structures are unavoidable. However, for the fabricated samples, both the heights of sidewalls and diameters of central nanoholes have slight randomness, especially for the nanoholes with thinner sidewalls. The irregular distribution will alleviate the size effects, which exhibits a smoothed variation. Furthermore, to see the scattering effects in the nanoholes more clearly, the cross-sectional electric field in the nanoholes is also displayed at the wavelengths of those absorption peaks (see Fig. 6(c), wavelength λ ∼ 576 nm, 700 nm, 758 nm, 830 nm, 880 nm). The electric distribution demonstrates the absorption mechanisms in the nanohole array. Like deep gratings, there exist several cavities in the central nanoholes, which display as enhanced electric field intensity. These cavities also affect the electric distribution in the surrounding nanoholes. That is to say that the light can be scattered between the two kinds of nanoholes. Originally, for the close-packed single nanopillars, only a little amount of incident light can enter the smaller nanoholes to generate the scattering effect. The added central nanoholes introduce more light into the holes to generate the scattering effect compared with single smaller nanoholes, which become a transmission media of the scattering light between those smaller nanoholes. The belowbandgap photons can also enter the structures, which enhances transmission in the NIR wavelength range, and full-spectrum anti-reflection
∫300 nm
1100 nm hc T (λ ) I (λ ) dλ λ 1100 nm hc int300 nm λ I (λ ) dλ
(6)
∫300 nm
(7)
where the incident wavelengths is λ. The absorptance and transmittance changing with wavelengths are A(λ) and T(λ), respectively. The Plank constant is h = 6.625 × 10−34 J s. The intensity of the incident AM1.5G (global air mass = 1.5) light source is I(λ). About 97% total absorptance can be obtained with the sample of the first case (named case a) in Fig. 5(a). For single nanopillar array, the scattering effect only occurs between the nanopillars. When coming to the hollow nanopillars, the incident light will be scattered in the central holes as well. Also, the added central nanoholes let more incident light enter solar cells, which indicates a better anti-reflection property. As the c-Si wafer takes no use of solar energy from 1100 nm to 2500 nm, this part of energy in the c-Si will transmit or be reflected to the air again. According to the absorptance of different cases, the first case in Fig. 5(a) has the best absorption and transmission properties. The increased diameters of central nanoholes let more incident light enter solar cells, and give rise to a stronger scattering between the two different kinds of nanoholes, which exhibits remarkable full-spectrum photon management in case a. With the top novel hexagonal nanohole arrays, remarkable full-spectrum photon management can be got via appropriate diameters of the central nanoholes. To reveal the absorption mechanisms of the hexagonal nanohole arrays, simulations are carried out by the FDTD method. The period and height are set at 600 nm and 800 nm, respectively, which can be seen in the SEM image in Fig. 6(a). The diameters of central nanoholes are
Fig. 6. (a) Cross-section SEM image of the hexagonal arrays. (b) The comparison of absorptance from 300 nm to 1000 nm between simulated results of periodic nanopillars, hexagonal nanoholes and experimental results. (c) The cross-sectional electric field intensity around the hexagonal nanohole arrays at different wavelengths (∼576 nm, 700 nm, 758 nm, 830 nm and 880 nm).
27
Energy Conversion and Management 152 (2017) 22–30
Y. Xu et al.
3.2. Effects of the bottom TiO2/SiO2 films In order to fully use the NIR solar energy for ThE devices, the transmittance from 1100 nm to 2500 nm should be further improved. Due to the suppressed reflection at the top silicon-air interface, more photons will enter into the wafer. However, only part of the NIR energy can transmit to the bottom ThE device, and it also suffers from a transmission loss at the bottom silicon-air interface, which is caused by reflected photons at this interface. Efforts should be taken to eliminate the reflection loss at the material interface to enhance transmission. The polished top surface was used to fabricate the hexagonal nanohole arrays. The bottom side of the c-Si wafer is unpolished, and it is inappropriate to fabricate the PS monolayer mask for etching. In this way, In consideration of the gradient refractive index method [41], film structures are employed at the rough bottom side with the refractive index between air and silicon. TiO2 and SiO2 are chosen as film materials, which have the refractive index of ∼2.5 and ∼1.5, respectively. The thicknesses of fabricated TiO2 and SiO2 films are both about 50 nm. The absorptance and transmittance are measured by the spectrophotometer as shown in Fig. 7. Due to the gradient refractive index at the silicon-air interface, the transmittance is enhanced to about 60% from 1200 nm to 2500 nm compared with the transmittance of the sample without bottom films, and the films have no influences on the absorptance. In this way, high transmission (1200 nm–2500 nm) and high absorption (300 nm–1100 nm) can be obtained simultaneously
Fig. 7. The comparison of absorption and transmission spectra between case a and case a with bottom TiO2/SiO2 anti-reflection bilayer films.
properties are achieved significantly. Similarly, there can be seen “channelling model” in the study [40], which leads to the electric field intensity enhancement in the nanoholes. This model will finally result in broadband absorption in the in-between sidewalls and the substrate due to the high extinction coefficient of silicon. Passivation should be carried out to handle the high surface recombination at the sidewalls for the future fabrication of solar cells.
Fig. 8. Full-spectrum absorption and transmission spectra of the hexagonal nanohole arrays of case a & films with the change of (a) incident angles (0°–60°) and (b) polarization states (TE and TM) at different incident angles (20°, 40° and 60°). (c) Total absorption and transmission spectra with the change of incident angles and polarizations.
28
Energy Conversion and Management 152 (2017) 22–30
Y. Xu et al.
[6] Makableh YF, Vasan R, Sarker JC, Nusir AI, Seal S, Manasreh MO. Enhancement of GaAs solar cell performance by using a ZnO sol-gel anti-reflection coating. Sol Energy Mater Sol Cells 2014;123:178–82. [7] Anh Huy Tuan L, Ahn S, Kim S, Han S, Kim S, Park H, et al. A statistical approach for the optimization of indium tin oxide films used as a front contact in amorphous/ crystalline silicon heterojunction solar cells. Energy Convers Manage 2014;87:191–8. [8] Kanda H, Uzum A, Harano N, Yoshinaga S, Ishikawa Y, Uraoka Y, et al. Al2O3/TiO2 double layer anti-reflection coating film for crystalline silicon solar cells formed by spray pyrolysis. Energy Sci Eng 2016;4. [9] Elshorbagy MH, Abdel-Hady K, Kamal H, Alda J. Broadband anti-reflection coating using dielectric Si3N4 nanostructures. Application to amorphous-Si-H solar cells. Opt Commun 2017;390:130–6. [10] Zhang Y, Xuan Y. Biomimetic omnidirectional broadband structured surface for photon management in photovoltaic-thermoelectric hybrid systems. Sol Energy Mater Sol Cells 2016;144:68–77. [11] Xu Y, Xuan Y, Yang L. Full-spectrum photon management of solar cell structures for photovoltaic-thermoelectric hybrid systems. Energy Convers Manage 2015;103:533–41. [12] Leem JW, Kim S, Lee SH, Rogers JA, Kim E, Yu JS. Efficiency enhancement of organic solar cells using hydrophobic antireflective inverted moth-eye nanopatterned PDMS films. Adv Energy Mater 2014;4. [13] Tsui KH, Lin Q, Chou H, Zhang Q, Fu H, Qi P, et al. Low-cost, flexible, and selfcleaning 3D nanocone anti-reflection films for high-efficiency photovoltaics. Adv Mater 2014;26:2805–11. [14] Li G, Li H, Ho JYL, Wong M, Kwok HS. Nanopyramid structure for ultrathin c-Si tandem solar cells. Nano Lett 2014;14:2563–8. [15] Zada I, Zhang W, Li Y, Sun P, Cai N, Gu J, et al. Angle dependent antireflection property of TiO2 inspired by cicada wings. Appl Phys Lett 2016;109. [16] He J, Yang Z, Liu P, Wu S, Gao P, Wang M, et al. Enhanced electro-optical properties of nanocone/nanopillar dual-structured arrays for ultrathin silicon/organic hybrid solar cell applications. Adv Energy Mater 2016;6. [17] Chen T-G, Yu P, Chen S-W, Chang F-Y, Huang B-Y, Cheng Y-C, et al. Characteristics of large-scale nanohole arrays for thin-silicon photovoltaics. Prog Photovolt 2014;22:452–61. [18] Clavero C. Plasmon-induced hot-electron generation at nanoparticle/metal-oxide interfaces for photovoltaic and photocatalytic devices. Nat Photon 2014;8:95–103. [19] Hsu H-L, Juang T-Y, Chen C-P, Hsieh C-M, Yang C-C, Huang C-L, et al. Enhanced efficiency of organic and perovskite photovoltaics from shape-dependent broadband plasmonic effects of silver nanoplates. Sol Energy Mater Sol Cells 2015;140:224–31. [20] Massiot I, Vandamme N, Bardou N, Dupuis C, Lemaitre A, Guillemoles J-F, et al. Metal nanogrid for broadband multiresonant light-harvesting in ultrathin GaAs layers. Acs Photon 2014;1:878–84. [21] Vora A, Gwamuri J, Pala N, Kulkarni A, Pearce JM, Gueney DO. Exchanging ohmic losses in metamaterial absorbers with useful optical absorption for photovoltaics. Sci Rep 2014;4. [22] Dincer F, Akgol O, Karaaslan M, Unal E, Sabah C. Polarization angle independent perfect metamaterial absorbers for solar cell applications in the microwave, infrared, and visible regime. Prog Electromagn Res - Pier 2014;144:93–101. [23] Rufangura P, Sabah C. Dual-band perfect metamaterial absorber for solar cell applications. Vacuum 2015;120:68–74. [24] Han H, Huang ZP, Lee W. Metal-assisted chemical etching of silicon and nanotechnology applications. Nano Today 2014;9:271–304. [25] Chen JD, Cui CH, Li YQ, Zhou L, Ou QD, Li C, et al. Single-junction polymer solar cells exceeding 10% power conversion efficiency. Adv Mater 2015;27:1035–41. [26] Sivasubramaniam S, Alkaisi MM. Inverted nanopyramid texturing for silicon solar cells using interference lithography. Microelectron Eng 2014;119:146–50. [27] Nicaise SM, Cheng JJ, Kiani A, Gradecak S, Berggren KK. Control of zinc oxide nanowire array properties with electron-beam lithography templating for photovoltaic applications. Nanotechnology 2015;26. [28] Al-Douri Y, Khasawneh Q, Kiwan S, Hashim U, Abd Hamid SB, Reshak AH, et al. Structural and optical insights to enhance solar cell performance of CdS nanostructures. Energy Convers Manage 2014;82:238–43. [29] Chang CL, Wang YF, Kanamori Y, Shih JJ, Kawai Y, Lee CK, et al. Etching submicrometer trenches by using the Bosch process and its application to the fabrication of antireflection structures. J Micromech Microeng 2005;15:580–5. [30] Gao P, He J, Zhou S, Yang X, Li S, Sheng J, et al. Large-area nanosphere self-assembly by a micro-propulsive injection method for high throughput periodic surface nanotexturing. Nano Lett 2015;15:4591–8. [31] Photometric accuracy of the Agilent DRA in the NIR region using the double aperture technique. < http://www.agilent.com/cs/library/datasheets/public/auvvis06.pdf > . [32] User guide for Cary 5000 absorption spectrometer with external DRA 1800 attachment. < http://mmrc.caltech.edu/Cary%20UV-Vis%20Int.Sphere/manuals/ Cary5000_User_Guide.pdf > . [33] Taflove A, Umashankar KR. The finite-difference time-domain method for numerical modeling of electromagnetic wave interactions. Electromagnetics 1990;10:105–26. [34] Yang L, Xuan Y, Han Y, Tan J. Investigation on the performance enhancement of silicon solar cells with an assembly grating structure. Energy Convers Manage 2012;54:30–7. [35] Soum-Glaude A, Bousquet I, Thomas L, Flamant G. Optical modeling of multilayered coatings based on SiC(N)H materials for their potential use as high-temperature solar selective absorbers. Sol Energy Mater Sol Cells 2013;117:315–23. [36] Palik ED. Handbook of optical constants of solids; 1985.
with the combination of top and bottom subwavelength structures. 3.3. Effects of the incident angles and polarization states Due to the different incident angles and polarization states of sunlight, it is essential to ensure the photon-management performance of composite subwavelength structures have omnidirectional and polarization-insensitive properties. Therefore, according to the characteristics of natural light, the full-spectrum photon management under the circumstances of oblique incidence and different polarization states are further investigated. It can be seen from Fig. 8(a) and (b) that both the absorptance and transmittance change little with the increase of incident angles. Apparently, the absorption and transmission spectra in the cases of TM and TE polarizations almost coincide at same incident angles. The total absorptance for the wavelengths of 300–1100 nm and the total transmittance from 1100 nm and 2500 nm are also calculated and are shown in Fig. 8(c). Both total absorptance and transmittance are almost unaffected by the TM or TE polarization. The total absorption and transmittance can maintain above 95% and 52%, respectively, when the incident angles are below 50°. It is worth noting that, the total transmittance has an apparent drop when it comes to the oblique incidence. Also, the polarization-insensitive property of transmittance is inferior to that of absorptance. Therefore, the top hexagonal nanohole arrays perform better than the film structures for the angle-independent and polarization-insensitive requirements. It suggests that this kind of hexagonal nanohole array is quite usable for photon management in different conditions of sunlight. In this way, the fabricated subwavelength structures can manage the full-spectrum incident light efficiently for different types of work conditions. 4. Conclusions In summary, a new approach of full-spectrum photon management is proposed for PV-ThE hybrid systems via fabricated novel composite subwavelength structures, which consist of top novel hexagonal nanohole arrays and the bottom sputtered TiO2/SiO2 bilayer anti-reflection films. Remarkable light trapping in the above-bandgap wavelengths (∼97% total absorption) is achieved due to the scattering effects between the two kinds of nanohole arrays, thereby enhancing the light propagation path in silicon. The transmittance in the NIR wavelength range can reach about 60% from 1200 nm to 2500 nm by employing gradient refractive index. Besides, the performance is omnidirectional and polarization-insensitive. The proposed design can help to allocate the photons to different devices smartly for the PV-ThE hybrid systems and has a great potential in enhancing the utilization efficiency of solar energy by using this hybrid system with various kinds of solar cells. This full-spectrum photon-management approach may pave the way for other solar energy harvesting applications. Acknowledgements This work was financially supported by the National Natural Science Foundation of China (Grant No. 51590901). References [1] Park K-T, Shin S-M, Tazebay AS, Um H-D, Jung J-Y, Jee S-W, et al. Lossless hybridization between photovoltaic and thermoelectric devices. Sci Rep 2013;3. [2] Lamba R, Kaushik SC. Modeling and performance analysis of a concentrated photovoltaic-thermoelectric hybrid power generation system. Energy Convers Manage 2016;115:288–98. [3] Hsueh T-J, Shieh J-M, Yeh Y-M. Hybrid Cd-free CIGS solar cell/TEG device with ZnO nanowires. Prog Photovolt 2015;23:507–12. [4] Zhang J, Xuan Y. Investigation on the effect of thermal resistances on a highly concentrated photovoltaic-thermoelectric hybrid system. Energy Convers Manage 2016;129:1–10. [5] Yin E, Li Q, Xuan Y. Thermal resistance analysis and optimization of photovoltaicthermoelectric hybrid system. Energy Convers Manage 2017;143:188–202.
29
Energy Conversion and Management 152 (2017) 22–30
Y. Xu et al.
solar photovoltaics. Nano Lett 2010;10:1012–5. [40] Gomard G, Peretti R, Callard S, Meng X, Artinyan R, Deschamps T, et al. Blue light absorption enhancement based on vertically channelling modes in nano-holes arrays. Appl Phys Lett 2014;104. [41] Southwell WH. Gradient-index antireflection coatings. Opt Lett 1983;8:584–6.
[37] Zhong S, Zeng Y, Huang Z, Shen W. Superior broadband antireflection from buried Mie resonator arrays for high-efficiency photovoltaics. Sci Rep 2015;5. [38] Garnett E, Yang P. Light trapping in silicon nanowire solar cells. Nano Lett 2010;10:1082–7. [39] Han SE, Chen G. Optical absorption enhancement in silicon nanohole arrays for
30