- Email: [email protected]
High-efficiency white-light solar window using waveguide glass plate
Energy & Buildings 202 (2019) 109341
Contents lists available at ScienceDirect
Energy & Buildings journal homepage: www.elsevier.com/locate/enbuild
High-efficiency white-light solar window using waveguide glass plate Ganghoo Lee a, Myunghun Shin a,∗, Gi yong Lee b, Hyungduk Ko b,∗ a b
School of Electronics and Information Engineering, Korea Aerospace University, Goyang-city, Gyeonggi-do 412-791, Republic of Korea Korea Institute of Science and Technology (KIST), Hwarangno 14-gil, Seongbuk-gu, Seoul 136-791, Republic of Korea
a r t i c l e
i n f o
Article history: Received 26 June 2018 Revised 12 June 2019 Accepted 24 July 2019 Available online 25 July 2019 Keywords: Building-integrated photovoltaics (BIPVs) Solar window CRI Waveguide plate
a b s t r a c t A high-efficiency white-light solar window is proposed for building-integrated photovoltaic (BIPV) applications. In the solar window, incident light is scattered at a waveguide plate and guided into GaAs cell arrays at the edges of the window frame. The optical characteristics of the waveguide plate are designed and evaluated using a ray-tracing simulation, and the solar window is fabricated and assembled with three-dimensional printing. The solar window exhibits an almost constant transmittance, with an average of about 21.6% in the visible range of 40 0–80 0 nm and a color rendering index of about 97.8 for sunlight, which is a neutral color that does not distort the colors of indoor objects under sunlight through the window. The solar window achieves a geometric concentration gain of 1.14 and edge collection ratio (ECR) of about 27.0%, which is constantly obtainable regardless of the incident angle of sunlight, and it exhibits an efficiency of 6.368%, a very high value for conventional transparent solar cells and modules. The method for optimizing the window efficiency and effective indoor illuminance by increasing the size of the solar window is also presented. These results can contribute to the further development of highperformance and large-area BIPV windows. © 2019 Elsevier B.V. All rights reserved.
1. Introduction Global energy consumption is increasing every year [1]. More than 40% of the world’s electricity consumption is used in buildings and electricity demand has been steadily increasing for commercial or residential buildings [2]. In addition, greenhouse gases, which induce climate change, are emitted by various human activities, a substantial amount of which are generated in high-density cities [3–5]. Therefore, it is critical that renewable energy technologies are utilized in urban areas [6–8]. Building-integrated photovoltaic (BIPV) technology can directly provide electrical power to buildings without the emission of greenhouse gases. It does not require an additional site for installation, so BIPV panels can be used as a direct part of buildings, such as roofs, facades, walls, and windows [9,10]. Among them, the BIPV window is not only economically valuable in that it has the widest installable area for buildings, but it is also technically challenging
Abbreviations: BIPV, building-integrated photovoltaic; ECR, edge collection ratio; EQE, external quantum efficiency; VOC , open-circuit voltage; FF, fill factor; η, power conversion efficiency; PCB, printed circuit board; ADF, angularly distributed functions; FWHM, full-width at half-maximum; CRI, color rendering index; FoM, figure of merit. ∗ Corresponding authors. E-mail addresses: [email protected] (M. Shin), [email protected] (H. Ko). https://doi.org/10.1016/j.enbuild.2019.109341 0378-7788/© 2019 Elsevier B.V. All rights reserved.
in that it should exhibit both a high transparency and high efficiency [11–13]. BIPV windows should not only produce electricity but also provide the necessary indoor light through the windows, with appropriate brightness and natural color. For BIPV windows, thinfilm-based transparent PV modules have been researched, including amorphous silicon, cadmium telluride, copper indium gallium selenide, and dye-sensitized and organic solar cells [14–17]. However, these transparent PV modules still exhibit a low conversion efficiency less than 6% and their colors are limited to the spectral characteristics of their own absorber materials [14]. In general, white light is a neutral color to the human eye because it can illuminate indoor objects without distorting their color, but realizing a neutral white color is more difficult than realizing other colors with conventional thin-film-based solar windows [18–20]. Waveguide-type solar cells can also be used as BIPV windows. Luminescent solar concentrators absorb high-energy photons of shorter wavelengths in luminescent materials in a waveguide structure and radiate low-energy photons in all directions; among the re-emitted light, only the light emitted within a total internal reflection angle is guided into the cells at the edge of the waveguide [21,22]. Diffusive flat panels use nanoparticles to diffuse and scatter the incident light instead of luminescent materials [23,24]. However, these types of solar modules still exhibit a lower efficiency less than 5% and their spectral characteristics are distorted by the
2
G. Lee, M. Shin and G.y. Lee et al. / Energy & Buildings 202 (2019) 109341
Fig. 1. Schematics of the white solar window: (a) structure of the white solar window, (b) laminated structure of diffuser sheet and patterned glass, and (c) scanning electron microscopy image of the rear surface of the patterned glass.
absorption of luminescent materials or metallic nanoparticles. Further, a sophisticated and lens-based solar window could improve the spectral properties, but it still requires improvements in the configuration and performance optimization of the system [25]. In this paper, we propose a waveguide-type solar module transmitting white light for BIPV solar windows, which consists of a waveguide glass plate and high-efficiency GaAs solar cell arrays. We designed and evaluated the solar module using a ray-tracing simulation method and fabricated and assembled the frame of the solar window using a three-dimensional (3D) printer. We investigated the performance of the module at collecting incident light to the cell arrays at the edge of the module frame and characterized the performance of the solar module for the spectral properties. 2. Design and characterization of white solar window using ray-tracing simulation
The waveguide glass plate is a structure in which four diffuser sheets and patterned glass plates are laminated, as shown in Fig. 1b. It is a weakly guided plate structure where a part of the incident light is trapped in the glass plate and absorbed to contribute to power generation and a part of the incident light is transmitted and used to illuminate the interior of the building. The diffuser sheet diffuses the incident light at the front surface regardless of the polarity and direction of the light. The glass patterns scatter the light at angles larger than the critical angle for total internal reflection to guide the light in the glass plate, but they can also leak the guided light. We note that in our preliminary experiments, the waveguide plates with only diffuse sheets or with only patterned glasses exhibited serious scattering loss or severe leak loss, respectively. The device only worked when the diffuse sheet and patterned glasses were properly laminated, as shown in Fig. 1b. 2.2. Fabrication of white solar window
2.1. Structure of white solar window A schematic of the proposed white-light solar window is illustrated in Fig. 1a. It consists of a waveguide glass plate as the window and solar cell arrays at the edges of the window frame. Incident light on the front face of the window is diffused and scattered at the front or rear surfaces of the window and guided to the solar cell arrays through the waveguide glass plates. Then, the solar cell arrays absorb the guided light to generate electric power.
The structure of the waveguide plate is illustrated in Fig. 1b. We prepared four diffuser sheets (LG-d90 commercial diffuser) and glasses (BK7) with a size of approximately 22 × 22 mm and thickness of about 0.3 and 1 mm, respectively. On the rear surface of the glass, a periodic groove is patterned with a period of about 400 μm and depth of about 80 μm using high-power pulse lasers, as shown in Fig. 1c. The four diffuser and patterned-glass plates were laminated and assembled as the waveguide plate, as shown
G. Lee, M. Shin and G.y. Lee et al. / Energy & Buildings 202 (2019) 109341
in Fig. 1b, where the pattern directions are perpendicular to each other. In the solar cell arrays, we used high-efficiency single-junction GaAs cells. The GaAs solar cells were fabricated by growing the layered structure sequentially on an n-type (n-) GaAs substrate in the following order: 300-nm n-GaAs contact and buffer, 50-nm n-GaInP back surface field, 350 0-nm n-GaAs base, 50 0-nm p-type (p-) GaAs emitter, 30-nm p-GaInP window, and 60-nm Si3 N4 anti-reflection coating. The aperture area and total size of the cell were 5.5 × 5.5 and 7.0 × 5.5 mm, respectively [26]. Among the fabricated cells, we selected the cells with similar performance parameters: short-circuit current density of 28.7 mA/cm2 , opencircuit voltage (VOC ) of about 1.01 V, fill factor (FF) of about 86%, and power conversion efficiency (η) of about 24.9%. The selected cells were mounted on a printed circuit board (PCB) designed to connect the cells in parallel. We also constructed additional PCBs, PCB-i, where each GaAs cell can be addressed individually. Although we used in-house GaAs cells in this work, high-efficiency back contact cells can be more beneficial in reducing the cost and assembly effort [27,28]. The base and cover of the window frame of Fig. 1a were manufactured using a 3D printer (Form 2, Formlabs). We designed the frame where GaAs-arrayed PCBs can be self-aligned at the four sides of the waveguide plate. As the edge of the waveguide plate should be inserted into the frame, the effective light receiving area (aperture area) was designed to be approximately 20 × 20 mm smaller than the total area of the waveguide plate (∼22 × 22 mm). In addition, the necessary jigs and tools for the measurements were fabricated with the 3D printer.
3
Fig. 2. Measurement setup for spectral effective collection ratio (ECR) using integrating sphere: (a) reference (PIN ), (b) transmittance (PT ), and (c) edge collection (PE ) for the spectral incident light.
K3100, McScience Inc.). The beam spot size from the spectral light source was maintained in a square of about 1 × 3 mm. The solar window was irradiated by the spectral light, and we measured the input reference (PIN ), transmittance (PT ), and edge collection (PE ) powers at one edge of the solar window, as shown in Fig. 2. The transmittance of the window is defined as PT /PIN , and the ECR is defined as four times PE /PIN . The setup of Fig. 2 was also used to analyze the spatial response of the solar window using a 532-nm laser source, instead.
3. Results and Discussion 3.1. Transmittance and edge collection ratio
2.3. Ray optics simulation for white solar window For the design and evaluation of the optical properties of the white solar window, we performed Monte–Carlo ray tracing using LightTools 8.5 (Synopsis Inc.). Basically, depending on the n-k parameters of the material and geometric structure of the device, the ray-tracing simulation can track the transmission, reflection, and absorption of light incident on the solar window in the form of rays. We modeled the periodic grooves at the rear surface of the glass plate (refractive index of 1.5168) as the shape of observation, as shown in Fig. 1c (400-μm period and 80-μm depth). The optical properties of the diffuser sheets, the scattering characteristics, are very important because they are used to reveal the random behavior of the light at the diffuser sheet for the raytracing simulation. First, we measured the total and diffused transmittance and reflectance using an integrating sphere and the angularly distributed functions (ADF) of transmittance and reflectance using a goniometer system. The scattering characteristics of the diffuser sheet are 40° at full-width at half-maximum (FWHM) for both the transmittance and reflectance with green laser light of 532 nm in wavelength, the measured ADF of which was input into the ray-tracing tool for the simulation. 2.4. Characterization of white solar window The solar window generates electrical power by forwarding the light incident on the front of the window to the frame edges. Thus, the amount of light collected at the edges is a key performance parameter of the solar window. In this work, we defined the effective collection ratio (ECR), which is the ratio of the collected light power at all the frame edges to the total incident power on the front of the solar window. Fig. 2 presents the measurement setup for the spectral ECR using an integrating sphere. Using a monochromator, halogen lamp, and internal photon counter, we configured a spectral light source to control the incident photon to converted electron ratio (ICPE
We performed Monte–Carlo ray-tracing simulations to estimate the transmittance and ECR for the number of waveguide plates and compared the simulation results with the measurements. The simulation and measurement results presented in Fig. 3 are in good agreement for the transmittance (Fig. 3a) and ECR (Fig. 3b) at 532 nm. As the number of waveguide plates increases, the transmittance decreases, but the ECR increases. The ray-tracing simulation also indicates that the reflectance increases with the number of the waveguide plates because of the high probability to bounce back from the air-gap interface of the glass plate. Using the setup of Fig. 2, we measured the transmittance and ECR for the wavelength range of 40 0–110 0 nm; in Fig. 3c and d, the average transmittance and ECR are 23.8% and 27%, respectively. As shown in the transmittance of Fig. 3c, the solar window has an almost constant transmittance over the wide wavelength range. When sunlight passes through the solar window, the color rendering index (CRI) of the transmitted light is calculated to be about 97.8 (color temperature ∼ 4860 K), indicating white light of a neutral color. This implies that the colors of any indoor object will not be distorted by the transmitted light, which can hardly be achieved with conventional transparent thin-film solar cells [29–31]. The ECR constant in a wide spectral range up to 1100 nm (Fig. 3d) also implies that various types of solar cells can be used for the solar cell arrays; even multi-junction cells, known to be very sensitive to the spectral characteristics of PV systems, can be used. The solar cells applied to the solar windows should be optimized in the future. Fig. 3c and d show the measured transmittance and ECR of the solar window when four waveguide plates are used. As the wavelength increases, the transmittance and ECR tend to increase slightly; it is because the scattering of light generally decreases with increasing wavelength [32]. The diffuser sheet and the patterned glass constituting the waveguide plate have typical characteristics of scattering; the abrupt drops in transmittance and ECR at ∼ 400 nm are also attributed to the severe scattering loss of the diffuse sheet.
4
G. Lee, M. Shin and G.y. Lee et al. / Energy & Buildings 202 (2019) 109341
Fig. 3. Simulation and measurement results: (a) transmittance and (b) ECR at wavelength of 532 nm as a function of the number (N) of waveguide plates, and (c) measured transmittance and (d) ECR of four waveguide plates for the wavelength range of 40 0–110 0 nm.
Fig. 4. Two-dimensional (2D) images by using a linear scanning system with a photodiode: (a) direction of scanning and (b) captured 2D images.
3.2. Spatial characteristics As the waveguide plate of this work is designed to weakly guide incident light to the frame edges, we must understand how the beam profile changes in the waveguide plates. We prepared a two-dimensional (2D) image scanning system. In the 2D image scanning system, the photodiode receiving area is approximately 1 mm2 and the scanning interval of the linear scanner was set to 1 mm. Using the photodiode on the linear scanner, we captured the 2D images of the incident, transmitted, and edge-collected light of a 532-nm laser, as shown in Fig. 4a. The incident laser beam, collimated with a beam diameter of approximately 3 mm (FWHM), was diffused in the waveguide plate, and the beam waist increased to
approximately 6 mm (FWHM) at the rear side (rear in Fig. 4b) and about 16.4 × 5.7 mm (edge 1–4 in Fig. 4b, respectively) at the edge of the window plate. The 2D profiles of Fig. 4b indicate that the outgoing light from the edge is nonuniformly distributed. Thus, the array of solar cells at the edge should be carefully arranged to reduce the power loss due to the mismatch of the output power of each solar cell. It also implies that the incident light undergoes a guiding loss until it is transferred to the solar cell arrays, and the loss can vary depending on the location of the incident light and position of the solar cell. In order to investigate the output power dependency on the position of the solar cell and the location of the incident light, we used the cell array on PCB-i, at which the individual cell can be
G. Lee, M. Shin and G.y. Lee et al. / Energy & Buildings 202 (2019) 109341
5
Fig. 5. 2D charts and 2D distribution of the ECR: (a) measured and (b) simulated ECR observed at each GaAs cell for the location of the 532-nm incident light, and (c) measured and (d) simulated overall ECR for the location of the 532-nm incident light.
independently accessed to read the output power of each cell for different locations of incident light. At first, we divided the illuminated areas on the solar window into 4 × 4 blocks, and then focused the light from a 532-nm laser to the center of each block. We measured the ECR at each cell (named as cell A, B, C, and D in this work) at one edge of the solar window. Fig. 5a shows the variation in the measured ECR at each cell at the location of incident light; as expected, the cells generate more power when the locations of incident light are closer, which is well matched with the ray tracing simulation, as shown in Fig. 5b. Fig. 5c and d show the simulated and measured 2D distributions of ECR at the location of the incident light of 532 nm, both of which are in good agreement. These results can provide the reason for the ECR (∼27% in Fig. 3d) of the 532-nm light entering at the center being slightly smaller than the average ECR (∼31%) of all 16 points in Fig. 5d. 3.3. System performance We designed and assembled the white solar window using the 3D printer. Fig. 6a presents the setup of the assembly of the so-
lar window with the top cover open. Under the standard recommended condition of AM1.5G, we measured the module efficiency. Fig. 6b presents the measured current–voltage characteristics of the solar window confirmed at the Korea Institute of Energy Research (KIER); a short-circuit current of 32.27 mA, VOC of about 0.95 V, FF of about 86 %, and η of about 6.368% are obtained at the module with an aperture area of approximately 398 mm2 . The geometric concentration gain of the solar window is approximately 1.14, which is induced from the areas of all four waveguide plate edges (four sets of 22 × 4 mm) and the aperture area (20 × 20 mm) of the module. Using the solar spectrum of AM1.5G and the measured ECR of Fig. 3d, the collected power density at the edges is estimated as about 228 mW/cm2 and the concentration ratio (up to 1100 nm) is calculated as about 26.5%. Considering the concentration ratio and the η (24.9%) of the GaAs cell, the module η should be about 6.59%, but the air gap between the waveguide edges and GaAs cells and the spacing between the GaAs cells can induce additional coupling loss. Thus, the coupling loss of the waveguide plate to the cell arrays is expected to be about 3.3% at present, but it can be further reduced if the optical
6
G. Lee, M. Shin and G.y. Lee et al. / Energy & Buildings 202 (2019) 109341
Fig. 6. Fabricated solar window: (a) assembly of the solar window with a GaAs cell array on a printed circuit board and white-light waveguide glass plate, (b) current– voltage characteristics of the solar window measured by the Korea Institute of Energy Research, and (c) schematic of the measurement of spectral ECR as a function of the incident angle of light, and the measured spectral ECR results for the incident angle.
system focusing the outgoing light to the solar cells and the single longitudinal solar cells are developed in the future. The most important limitation of the solar windows is that once they are installed in a building, it is usually not easy to change the angle of incidence or the placement of the solar cell. In particular, solar cells that generate power by changing the path of sunlight, such as light-collecting solar cells, should have very different performance characteristics depending on the position of the sun. Thus, we measure the spectral characteristics of the ECR for the angle of incident light. As shown in Fig. 6c, we measured the edge collected power at each direction of the edges for the incident angle by using the integrating sphere setup shown in Fig. 2c. While rotating the integrator in 10° increments, we measured the ECR value up to 70°, which is the maximum angle measurable with the measurement setup without disturbing the incident light. The total ECR obtained by summing up the powers collected at all the edges is shown at Fig. 6d. It implies that the waveguide glass plate structure shown in Fig. 1b can constantly collect power from the sun regardless of the incident angle (or the position of the sun), and it can be also expected that the proposed device has very stable performance for BIPV window applications. 3.4. Size scalability of solar window The proposed solar window utilizes the weakly guided plate structure so that a loss is expected depending on the window size. In order to analyze the loss according to the window area, we
performed ray tracing simulation to estimate the transmittance (T) and ECR at 532 nm for the size of the square solar window, each side of which is W (mm). As shown in Fig. 7b, T increases gradually, but the ECR decreases rapidly with the increase in the side length W because of the leak loss in the waveguide glass plates. Therefore, the size of the proposed solar window cannot be simply scaled up. In large windows, the proposed structure could be applied to the limited area of the window located less than ∼10 cm away from the window frame, taking into account the results in Fig. 7b. As an alternative, large-scale windows comprising the proposed small windows arranged like tiles can be considered, where the solar cell arrays are installed inside the grids, as shown in Fig. 7a. The window efficiency (ηW ) of such grid solar windows can be determined as a function of the side length of the window (W, transparent area contributes to power generation) and the grid frame length (F, opaque area does not contribute to power generation). A smaller F value yields better results, but the F value depends on the maturity of the solar cell and the module manufacturing technology. By considering the ECR tendency for the size of window (Fig. 7b) and the performance characteristics of the GaAs cells, we calculated ηW for the variation in W and F. For the calculation, the measured spectral ECR (Fig. 3d), the external quantum efficiency (ref [26]) of the GaAs cell, and the standard recommended spectral data of AM1.5 G were multiplied and integrated, and the used reference window area was defined as the square of (W + F). As shown in Fig. 7c, at certain side length (optimized side length, Wopt ), the maximum ηW can be
G. Lee, M. Shin and G.y. Lee et al. / Energy & Buildings 202 (2019) 109341
7
Fig. 7. Grid solar window and simulated system performance for the window size, where the size of a scaled-up window is defined by the lengths of the side (W) and frame (F) of the square window: (a) schematic of the grid solar window, and (b) transmittance and ECR at 532 nm (at F = 0), (c) window efficiency (ηW ), (d) effective illuminance, and (e) figure of merits (FoM = ηW × effective illuminance) for the grid solar windows.
obtained by compensating the loss area of the grids; when F of the grid in our device is approximately 4–8 mm, Wopt can be ∼20 mm. The solar windows should not only produce electricity, but should also provide sufficient brightness indoors. The illuminance from a window can be calculated by multiplying and integrating the measured spectral transmittance (Fig. 3d), the standard recommended spectral data of AM1.5G, and luminosity function [29], and the effective illuminance is calculated as a function of the transparent (W) and opaque (F) areas of the grid windows, as shown in Fig. 7d. The overall performance of the solar window should be evaluated in terms of both power generation and indoor illuminance. Therefore, we define the figure of merit (FoM) as the product of the window efficiency (Fig. 7c) and the effective illuminance (Fig. 7d); the FoM in Fig. 7e shows that Wopt can be differently
determined with F, and the Wopt for maximum FoM is longer than that of maximum ηW because of the property of effective illuminance. We want to note that although Wopt can be differently optimized depending on the light-guiding ability, spectral loss, transparency, and collection characteristics of the solar windows, Fig. 7e presents an alternative optimization procedure for the grid solar windows. For waveguide type of solar windows, the imperfection in the light guiding property inevitably shows a certain amount of guiding loss. Therefore, optimization is required between the aperture area contributing to transparency and power generation and the peripheral area not contributing to transparency and power generation. The proposed FoM can be used to optimize the efficiency and illuminance, which are the most important properties of a solar window.
8
G. Lee, M. Shin and G.y. Lee et al. / Energy & Buildings 202 (2019) 109341
4. Conclusions In this study, a high-efficiency white-light solar window was developed. The waveguide plates consist of a diffuser and patterned glass, which scatters and guides the incident light to GaAs cell arrays at the window frame edges. These were designed and evaluated using the ray-tracing simulation method. We constructed and assembled the solar window using 3D printing. The spectral and spatial characteristics of the solar window were investigated and measured with an integrating sphere and 2D scanning image system. The solar window exhibited a module η of 6.368%, a very high value for transparent solar cells and modules. It also exhibited an almost constant average transmittance of about 21.6% in the visible range of 40 0–80 0 nm and a CRI value of about 97.8 for sunlight, allowing the human eye to see indoor objects without significant color distortion. It still exhibited a small geometric concentration gain of 1.14, an average ECR of about 27.0%, and a small coupling loss of approximately 3.3% in the 40 0–110 0 nm range. In addition, the proposed solar window could perform stably regardless of the incident angle of sunlight. We also presented the method to increase the size of the solar window for optimizing the window efficiency and effective indoor illuminance. These results can contribute to the further development of improved solar windows with a wider spectral capability and better optical system, with solar cell arrays optimized for waveguide-type solar windows. Declaration of Competing Interest None. Acknowledgments The authors appreciate the Korea Institute of Energy Research (KIER) for officially evaluating the efficiency of the fabricated device. Funding This work was supported by the Energy Technology Development Program of the Korea Institute of Energy Technology Evaluation and Planning (KETEP) [grant number 20153030012870]; Pioneer Research Center Program through the National Research Foundation of Korea funded by the Ministry of Science, ICT, and Future Planning [grant number NRF-2013M3C1A3065040]; KIST institutional research program (2E29300). References [1] N. Antonakakis, I. Chatziantoniou, G. Filis, Energy consumption, CO2 emissions, and economic growth: an ethical dilemma, Renew. Sustain. Energy Rev. 68 (2017) 808–824. [2] A. Allouhi, Y.E. Fouih, T. Kouskou, A. Jamil, Y. Zeraouli, Y. Mourad, Energy consumption and efficiency in buildings: current status and future trends, J. Clean Prod. 109 (2015) 118–130. [3] A. Dimoudi, C. Tompa, Energy and environmental indicators related to construction of office buildings, Resour. Conserv. Recycl. 53 (2008) 86–95. [4] M. Suzuki, T. Oka, Estimation of life cycle energy consumption and CO2 emission of office buildings in Japan, Energy Build. 28 (1998) 33–41.
[5] A. Hunt, P. Watkiss, Climate change impacts and adaptation in cities: a review of the literature, Clim. Change 104 (2011) 13–49. [6] I. Dincer, Renewable energy and sustainable development: a crucial review, Renew. Sustain. Energy Rev. 4 (20 0 0) 157–175. [7] N.L. Panwar, S.C. Kaushik, S. Kothari, Role of renewable energy sources in environmental protection: a review, Renew. Sustain. Energy Rev. 15 (2011) 1513–1524. [8] A.K. Akella, R.P. Saini, M.P. Sharma, Social, economical and environmental impacts of renewable energy systems, Renew. Energy 34 (2009) 389–396. [9] B.P. Jelle, C. Breivik, The path to the building integrated photovoltaics of tomorrow, Energy Proc. 20 (2012) 78–87. [10] I. Cerón, E. Camaño-Martin, F.J. Neila, ‘State-of-the-art’ of building integrated photovoltaic products, Renew. Energy 58 (2013) 127–133. [11] E. Biyik, M. Araz, A. Hepbasli, M. Shahrestani, R. Yao, L. Shao, et al., A key review of building integrated photovoltaic (BIPV) systems, Eng. Sci. Technol. Int. J. 20 (2017) 833–858. [12] Y.T. Chae, J. Kim, H. Park, B. Shin, Building energy performance evaluation of building integrated photovoltaic (BIPV) window with semi-transparent solar cells, Appl. Energy 129 (2014) 217–227. [13] B.P. Jelle, C. Breivik, H.D. Røkenes, Building integrated photovoltaic products: a state-of-the-art review and future research opportunities, Sol. Energy Mater. Sol. Cells 100 (2012) 69–96. [14] M. Pagliaro, R. Ciriminna, G. Palmisano, BIPV: merging the photovoltaic with the construction industry, Prog. Photovolt. 18 (2010) 61–72. [15] J.H. Yoon, J. Song, S.J. Lee, Practical application of building integrated photovoltaic (BIPV) system using transparent amorphous silicon thin-film PV module, Sol. Energy 85 (2011) 723–733. [16] C.A. Wolden, J. Kurtin, J.B. Baxter, I. Repins, S.E. Shaheen, J.T. Torvik, et al., Photovoltaic manufacturing: present status, future prospects, and research needs, J. Vac. Sci. Technol. A 29 (2011) 030801. [17] M.A. Green, K. Emery, Y. Hishikawa, W. Warta, E.D. Dunlop, Solar cell efficiency tables (version 48), Prog. Photovolt. 24 (2016) 905–913. [18] A.K. Shukla, K. Sudhakar, P. Baredar, Recent advancement in BIPV product technologies: a review, Energy Build. 140 (2017) 188–195. [19] K.S. Chen, J.F. Salinas, H.L. Yip, L. Huo, J. Hou, A.K.Y. Jen, Semi-transparent polymer solar cells with 6% PCE, 25% average visible transmittance and a color rendering index close to 100 for power generating window applications, Energy Environ. Sci. 5 (2012) 9551–9557. ˇ [20] M. Cekon , R. Slávik, Total solar transmittance quantifying of transparent insulation building materials based on real climate outdoor measurements, Energy Procedia 132 (2017) 243–248. [21] L.H. Slooff, E.E. Bende, A.R. Burgers, T. Budel, M. Pravettoni, R.P. Kenny, et al., A luminescent solar concentrator with 7.1% power conversion efficiency, Phys. Stat. Sol. RRL 2 (2008) 257–259. [22] M.G. Debije, P.P.C. Verbunt, Thirty years of luminescent solar concentrator research: solar energy for the built environment, Adv. Energy Mater. 2 (2012) 12–35. [23] R.-T. Chen, J.L.H. Chau, G.-L. Hwang, Design and fabrication of diffusive solar cell window, Renew. Energy 40 (2012) 24–28. [24] J.M. Kim, P.S. Dutta, Optical efficiency–concentration ratio trade-off for a flat panel photovoltaic system with diffuser type concentrator, Sol. Energy Mater. Sol. Cells 103 (2012) 35–40. [25] J.H. Karp, E.J. Tremblay, J.E. Ford, Planar micro-optic solar concentrator, Opt. Express 18 (2010) 1122–1133. [26] Y.H. Ghymn, K. Jung, M. Shin, H. Ko, A luminescent down-shifting and moth-eyed anti-reflective film for highly efficient photovoltaic devices, Nanoscale 7 (2015) 18642. [27] K. Yoshikawa, H. Kawasaki, W. Yoshida, T. Irie, K. Konishi, K. Nakano, et al., Silicon heterojunction solar cell with interdigitated back contacts for a photoconversion efficiency over 26%, Nat. Energy 2 (2017) 17032. [28] M.K.M. Desa, S. Sapeai, A.W. Azhari, K. Sopian, M.Y. Sulaiman, A. Amin, et al., Silicon back contact solar cell configuration: a pathway towards higher efficiency, Renew. Sustain. Energy Rev. 60 (2016) 1516–1532. [29] X. Guo, K.W. Houser, A review of colour rendering indices and their application to commercial light sources, Light. Res. Technol. 36 (2004) 183–197. [30] N. Lynn, L. Mohanty, S. Wittkopf, Color rendering properties of semi-transparent thin-film PV modules, Build. Environ. 54 (2012) 148–158. [31] G. Wyszecki, W.S. Stiles, in: Color Science: Concepts and Methods, Quantitative Data and Formulae, 3, Wiley, Chap., 1982, p. 3. [32] A.J. Cox, A.J. DeWeerd, J. Linden, An experiment to measure Mie and Rayleigh total scattering cross sections, Am. J. Phys. 70 (2002) 620.
Contents lists available at ScienceDirect
Energy & Buildings journal homepage: www.elsevier.com/locate/enbuild
High-efficiency white-light solar window using waveguide glass plate Ganghoo Lee a, Myunghun Shin a,∗, Gi yong Lee b, Hyungduk Ko b,∗ a b
School of Electronics and Information Engineering, Korea Aerospace University, Goyang-city, Gyeonggi-do 412-791, Republic of Korea Korea Institute of Science and Technology (KIST), Hwarangno 14-gil, Seongbuk-gu, Seoul 136-791, Republic of Korea
a r t i c l e
i n f o
Article history: Received 26 June 2018 Revised 12 June 2019 Accepted 24 July 2019 Available online 25 July 2019 Keywords: Building-integrated photovoltaics (BIPVs) Solar window CRI Waveguide plate
a b s t r a c t A high-efficiency white-light solar window is proposed for building-integrated photovoltaic (BIPV) applications. In the solar window, incident light is scattered at a waveguide plate and guided into GaAs cell arrays at the edges of the window frame. The optical characteristics of the waveguide plate are designed and evaluated using a ray-tracing simulation, and the solar window is fabricated and assembled with three-dimensional printing. The solar window exhibits an almost constant transmittance, with an average of about 21.6% in the visible range of 40 0–80 0 nm and a color rendering index of about 97.8 for sunlight, which is a neutral color that does not distort the colors of indoor objects under sunlight through the window. The solar window achieves a geometric concentration gain of 1.14 and edge collection ratio (ECR) of about 27.0%, which is constantly obtainable regardless of the incident angle of sunlight, and it exhibits an efficiency of 6.368%, a very high value for conventional transparent solar cells and modules. The method for optimizing the window efficiency and effective indoor illuminance by increasing the size of the solar window is also presented. These results can contribute to the further development of highperformance and large-area BIPV windows. © 2019 Elsevier B.V. All rights reserved.
1. Introduction Global energy consumption is increasing every year [1]. More than 40% of the world’s electricity consumption is used in buildings and electricity demand has been steadily increasing for commercial or residential buildings [2]. In addition, greenhouse gases, which induce climate change, are emitted by various human activities, a substantial amount of which are generated in high-density cities [3–5]. Therefore, it is critical that renewable energy technologies are utilized in urban areas [6–8]. Building-integrated photovoltaic (BIPV) technology can directly provide electrical power to buildings without the emission of greenhouse gases. It does not require an additional site for installation, so BIPV panels can be used as a direct part of buildings, such as roofs, facades, walls, and windows [9,10]. Among them, the BIPV window is not only economically valuable in that it has the widest installable area for buildings, but it is also technically challenging
Abbreviations: BIPV, building-integrated photovoltaic; ECR, edge collection ratio; EQE, external quantum efficiency; VOC , open-circuit voltage; FF, fill factor; η, power conversion efficiency; PCB, printed circuit board; ADF, angularly distributed functions; FWHM, full-width at half-maximum; CRI, color rendering index; FoM, figure of merit. ∗ Corresponding authors. E-mail addresses: [email protected] (M. Shin), [email protected] (H. Ko). https://doi.org/10.1016/j.enbuild.2019.109341 0378-7788/© 2019 Elsevier B.V. All rights reserved.
in that it should exhibit both a high transparency and high efficiency [11–13]. BIPV windows should not only produce electricity but also provide the necessary indoor light through the windows, with appropriate brightness and natural color. For BIPV windows, thinfilm-based transparent PV modules have been researched, including amorphous silicon, cadmium telluride, copper indium gallium selenide, and dye-sensitized and organic solar cells [14–17]. However, these transparent PV modules still exhibit a low conversion efficiency less than 6% and their colors are limited to the spectral characteristics of their own absorber materials [14]. In general, white light is a neutral color to the human eye because it can illuminate indoor objects without distorting their color, but realizing a neutral white color is more difficult than realizing other colors with conventional thin-film-based solar windows [18–20]. Waveguide-type solar cells can also be used as BIPV windows. Luminescent solar concentrators absorb high-energy photons of shorter wavelengths in luminescent materials in a waveguide structure and radiate low-energy photons in all directions; among the re-emitted light, only the light emitted within a total internal reflection angle is guided into the cells at the edge of the waveguide [21,22]. Diffusive flat panels use nanoparticles to diffuse and scatter the incident light instead of luminescent materials [23,24]. However, these types of solar modules still exhibit a lower efficiency less than 5% and their spectral characteristics are distorted by the
2
G. Lee, M. Shin and G.y. Lee et al. / Energy & Buildings 202 (2019) 109341
Fig. 1. Schematics of the white solar window: (a) structure of the white solar window, (b) laminated structure of diffuser sheet and patterned glass, and (c) scanning electron microscopy image of the rear surface of the patterned glass.
absorption of luminescent materials or metallic nanoparticles. Further, a sophisticated and lens-based solar window could improve the spectral properties, but it still requires improvements in the configuration and performance optimization of the system [25]. In this paper, we propose a waveguide-type solar module transmitting white light for BIPV solar windows, which consists of a waveguide glass plate and high-efficiency GaAs solar cell arrays. We designed and evaluated the solar module using a ray-tracing simulation method and fabricated and assembled the frame of the solar window using a three-dimensional (3D) printer. We investigated the performance of the module at collecting incident light to the cell arrays at the edge of the module frame and characterized the performance of the solar module for the spectral properties. 2. Design and characterization of white solar window using ray-tracing simulation
The waveguide glass plate is a structure in which four diffuser sheets and patterned glass plates are laminated, as shown in Fig. 1b. It is a weakly guided plate structure where a part of the incident light is trapped in the glass plate and absorbed to contribute to power generation and a part of the incident light is transmitted and used to illuminate the interior of the building. The diffuser sheet diffuses the incident light at the front surface regardless of the polarity and direction of the light. The glass patterns scatter the light at angles larger than the critical angle for total internal reflection to guide the light in the glass plate, but they can also leak the guided light. We note that in our preliminary experiments, the waveguide plates with only diffuse sheets or with only patterned glasses exhibited serious scattering loss or severe leak loss, respectively. The device only worked when the diffuse sheet and patterned glasses were properly laminated, as shown in Fig. 1b. 2.2. Fabrication of white solar window
2.1. Structure of white solar window A schematic of the proposed white-light solar window is illustrated in Fig. 1a. It consists of a waveguide glass plate as the window and solar cell arrays at the edges of the window frame. Incident light on the front face of the window is diffused and scattered at the front or rear surfaces of the window and guided to the solar cell arrays through the waveguide glass plates. Then, the solar cell arrays absorb the guided light to generate electric power.
The structure of the waveguide plate is illustrated in Fig. 1b. We prepared four diffuser sheets (LG-d90 commercial diffuser) and glasses (BK7) with a size of approximately 22 × 22 mm and thickness of about 0.3 and 1 mm, respectively. On the rear surface of the glass, a periodic groove is patterned with a period of about 400 μm and depth of about 80 μm using high-power pulse lasers, as shown in Fig. 1c. The four diffuser and patterned-glass plates were laminated and assembled as the waveguide plate, as shown
G. Lee, M. Shin and G.y. Lee et al. / Energy & Buildings 202 (2019) 109341
in Fig. 1b, where the pattern directions are perpendicular to each other. In the solar cell arrays, we used high-efficiency single-junction GaAs cells. The GaAs solar cells were fabricated by growing the layered structure sequentially on an n-type (n-) GaAs substrate in the following order: 300-nm n-GaAs contact and buffer, 50-nm n-GaInP back surface field, 350 0-nm n-GaAs base, 50 0-nm p-type (p-) GaAs emitter, 30-nm p-GaInP window, and 60-nm Si3 N4 anti-reflection coating. The aperture area and total size of the cell were 5.5 × 5.5 and 7.0 × 5.5 mm, respectively [26]. Among the fabricated cells, we selected the cells with similar performance parameters: short-circuit current density of 28.7 mA/cm2 , opencircuit voltage (VOC ) of about 1.01 V, fill factor (FF) of about 86%, and power conversion efficiency (η) of about 24.9%. The selected cells were mounted on a printed circuit board (PCB) designed to connect the cells in parallel. We also constructed additional PCBs, PCB-i, where each GaAs cell can be addressed individually. Although we used in-house GaAs cells in this work, high-efficiency back contact cells can be more beneficial in reducing the cost and assembly effort [27,28]. The base and cover of the window frame of Fig. 1a were manufactured using a 3D printer (Form 2, Formlabs). We designed the frame where GaAs-arrayed PCBs can be self-aligned at the four sides of the waveguide plate. As the edge of the waveguide plate should be inserted into the frame, the effective light receiving area (aperture area) was designed to be approximately 20 × 20 mm smaller than the total area of the waveguide plate (∼22 × 22 mm). In addition, the necessary jigs and tools for the measurements were fabricated with the 3D printer.
3
Fig. 2. Measurement setup for spectral effective collection ratio (ECR) using integrating sphere: (a) reference (PIN ), (b) transmittance (PT ), and (c) edge collection (PE ) for the spectral incident light.
K3100, McScience Inc.). The beam spot size from the spectral light source was maintained in a square of about 1 × 3 mm. The solar window was irradiated by the spectral light, and we measured the input reference (PIN ), transmittance (PT ), and edge collection (PE ) powers at one edge of the solar window, as shown in Fig. 2. The transmittance of the window is defined as PT /PIN , and the ECR is defined as four times PE /PIN . The setup of Fig. 2 was also used to analyze the spatial response of the solar window using a 532-nm laser source, instead.
3. Results and Discussion 3.1. Transmittance and edge collection ratio
2.3. Ray optics simulation for white solar window For the design and evaluation of the optical properties of the white solar window, we performed Monte–Carlo ray tracing using LightTools 8.5 (Synopsis Inc.). Basically, depending on the n-k parameters of the material and geometric structure of the device, the ray-tracing simulation can track the transmission, reflection, and absorption of light incident on the solar window in the form of rays. We modeled the periodic grooves at the rear surface of the glass plate (refractive index of 1.5168) as the shape of observation, as shown in Fig. 1c (400-μm period and 80-μm depth). The optical properties of the diffuser sheets, the scattering characteristics, are very important because they are used to reveal the random behavior of the light at the diffuser sheet for the raytracing simulation. First, we measured the total and diffused transmittance and reflectance using an integrating sphere and the angularly distributed functions (ADF) of transmittance and reflectance using a goniometer system. The scattering characteristics of the diffuser sheet are 40° at full-width at half-maximum (FWHM) for both the transmittance and reflectance with green laser light of 532 nm in wavelength, the measured ADF of which was input into the ray-tracing tool for the simulation. 2.4. Characterization of white solar window The solar window generates electrical power by forwarding the light incident on the front of the window to the frame edges. Thus, the amount of light collected at the edges is a key performance parameter of the solar window. In this work, we defined the effective collection ratio (ECR), which is the ratio of the collected light power at all the frame edges to the total incident power on the front of the solar window. Fig. 2 presents the measurement setup for the spectral ECR using an integrating sphere. Using a monochromator, halogen lamp, and internal photon counter, we configured a spectral light source to control the incident photon to converted electron ratio (ICPE
We performed Monte–Carlo ray-tracing simulations to estimate the transmittance and ECR for the number of waveguide plates and compared the simulation results with the measurements. The simulation and measurement results presented in Fig. 3 are in good agreement for the transmittance (Fig. 3a) and ECR (Fig. 3b) at 532 nm. As the number of waveguide plates increases, the transmittance decreases, but the ECR increases. The ray-tracing simulation also indicates that the reflectance increases with the number of the waveguide plates because of the high probability to bounce back from the air-gap interface of the glass plate. Using the setup of Fig. 2, we measured the transmittance and ECR for the wavelength range of 40 0–110 0 nm; in Fig. 3c and d, the average transmittance and ECR are 23.8% and 27%, respectively. As shown in the transmittance of Fig. 3c, the solar window has an almost constant transmittance over the wide wavelength range. When sunlight passes through the solar window, the color rendering index (CRI) of the transmitted light is calculated to be about 97.8 (color temperature ∼ 4860 K), indicating white light of a neutral color. This implies that the colors of any indoor object will not be distorted by the transmitted light, which can hardly be achieved with conventional transparent thin-film solar cells [29–31]. The ECR constant in a wide spectral range up to 1100 nm (Fig. 3d) also implies that various types of solar cells can be used for the solar cell arrays; even multi-junction cells, known to be very sensitive to the spectral characteristics of PV systems, can be used. The solar cells applied to the solar windows should be optimized in the future. Fig. 3c and d show the measured transmittance and ECR of the solar window when four waveguide plates are used. As the wavelength increases, the transmittance and ECR tend to increase slightly; it is because the scattering of light generally decreases with increasing wavelength [32]. The diffuser sheet and the patterned glass constituting the waveguide plate have typical characteristics of scattering; the abrupt drops in transmittance and ECR at ∼ 400 nm are also attributed to the severe scattering loss of the diffuse sheet.
4
G. Lee, M. Shin and G.y. Lee et al. / Energy & Buildings 202 (2019) 109341
Fig. 3. Simulation and measurement results: (a) transmittance and (b) ECR at wavelength of 532 nm as a function of the number (N) of waveguide plates, and (c) measured transmittance and (d) ECR of four waveguide plates for the wavelength range of 40 0–110 0 nm.
Fig. 4. Two-dimensional (2D) images by using a linear scanning system with a photodiode: (a) direction of scanning and (b) captured 2D images.
3.2. Spatial characteristics As the waveguide plate of this work is designed to weakly guide incident light to the frame edges, we must understand how the beam profile changes in the waveguide plates. We prepared a two-dimensional (2D) image scanning system. In the 2D image scanning system, the photodiode receiving area is approximately 1 mm2 and the scanning interval of the linear scanner was set to 1 mm. Using the photodiode on the linear scanner, we captured the 2D images of the incident, transmitted, and edge-collected light of a 532-nm laser, as shown in Fig. 4a. The incident laser beam, collimated with a beam diameter of approximately 3 mm (FWHM), was diffused in the waveguide plate, and the beam waist increased to
approximately 6 mm (FWHM) at the rear side (rear in Fig. 4b) and about 16.4 × 5.7 mm (edge 1–4 in Fig. 4b, respectively) at the edge of the window plate. The 2D profiles of Fig. 4b indicate that the outgoing light from the edge is nonuniformly distributed. Thus, the array of solar cells at the edge should be carefully arranged to reduce the power loss due to the mismatch of the output power of each solar cell. It also implies that the incident light undergoes a guiding loss until it is transferred to the solar cell arrays, and the loss can vary depending on the location of the incident light and position of the solar cell. In order to investigate the output power dependency on the position of the solar cell and the location of the incident light, we used the cell array on PCB-i, at which the individual cell can be
G. Lee, M. Shin and G.y. Lee et al. / Energy & Buildings 202 (2019) 109341
5
Fig. 5. 2D charts and 2D distribution of the ECR: (a) measured and (b) simulated ECR observed at each GaAs cell for the location of the 532-nm incident light, and (c) measured and (d) simulated overall ECR for the location of the 532-nm incident light.
independently accessed to read the output power of each cell for different locations of incident light. At first, we divided the illuminated areas on the solar window into 4 × 4 blocks, and then focused the light from a 532-nm laser to the center of each block. We measured the ECR at each cell (named as cell A, B, C, and D in this work) at one edge of the solar window. Fig. 5a shows the variation in the measured ECR at each cell at the location of incident light; as expected, the cells generate more power when the locations of incident light are closer, which is well matched with the ray tracing simulation, as shown in Fig. 5b. Fig. 5c and d show the simulated and measured 2D distributions of ECR at the location of the incident light of 532 nm, both of which are in good agreement. These results can provide the reason for the ECR (∼27% in Fig. 3d) of the 532-nm light entering at the center being slightly smaller than the average ECR (∼31%) of all 16 points in Fig. 5d. 3.3. System performance We designed and assembled the white solar window using the 3D printer. Fig. 6a presents the setup of the assembly of the so-
lar window with the top cover open. Under the standard recommended condition of AM1.5G, we measured the module efficiency. Fig. 6b presents the measured current–voltage characteristics of the solar window confirmed at the Korea Institute of Energy Research (KIER); a short-circuit current of 32.27 mA, VOC of about 0.95 V, FF of about 86 %, and η of about 6.368% are obtained at the module with an aperture area of approximately 398 mm2 . The geometric concentration gain of the solar window is approximately 1.14, which is induced from the areas of all four waveguide plate edges (four sets of 22 × 4 mm) and the aperture area (20 × 20 mm) of the module. Using the solar spectrum of AM1.5G and the measured ECR of Fig. 3d, the collected power density at the edges is estimated as about 228 mW/cm2 and the concentration ratio (up to 1100 nm) is calculated as about 26.5%. Considering the concentration ratio and the η (24.9%) of the GaAs cell, the module η should be about 6.59%, but the air gap between the waveguide edges and GaAs cells and the spacing between the GaAs cells can induce additional coupling loss. Thus, the coupling loss of the waveguide plate to the cell arrays is expected to be about 3.3% at present, but it can be further reduced if the optical
6
G. Lee, M. Shin and G.y. Lee et al. / Energy & Buildings 202 (2019) 109341
Fig. 6. Fabricated solar window: (a) assembly of the solar window with a GaAs cell array on a printed circuit board and white-light waveguide glass plate, (b) current– voltage characteristics of the solar window measured by the Korea Institute of Energy Research, and (c) schematic of the measurement of spectral ECR as a function of the incident angle of light, and the measured spectral ECR results for the incident angle.
system focusing the outgoing light to the solar cells and the single longitudinal solar cells are developed in the future. The most important limitation of the solar windows is that once they are installed in a building, it is usually not easy to change the angle of incidence or the placement of the solar cell. In particular, solar cells that generate power by changing the path of sunlight, such as light-collecting solar cells, should have very different performance characteristics depending on the position of the sun. Thus, we measure the spectral characteristics of the ECR for the angle of incident light. As shown in Fig. 6c, we measured the edge collected power at each direction of the edges for the incident angle by using the integrating sphere setup shown in Fig. 2c. While rotating the integrator in 10° increments, we measured the ECR value up to 70°, which is the maximum angle measurable with the measurement setup without disturbing the incident light. The total ECR obtained by summing up the powers collected at all the edges is shown at Fig. 6d. It implies that the waveguide glass plate structure shown in Fig. 1b can constantly collect power from the sun regardless of the incident angle (or the position of the sun), and it can be also expected that the proposed device has very stable performance for BIPV window applications. 3.4. Size scalability of solar window The proposed solar window utilizes the weakly guided plate structure so that a loss is expected depending on the window size. In order to analyze the loss according to the window area, we
performed ray tracing simulation to estimate the transmittance (T) and ECR at 532 nm for the size of the square solar window, each side of which is W (mm). As shown in Fig. 7b, T increases gradually, but the ECR decreases rapidly with the increase in the side length W because of the leak loss in the waveguide glass plates. Therefore, the size of the proposed solar window cannot be simply scaled up. In large windows, the proposed structure could be applied to the limited area of the window located less than ∼10 cm away from the window frame, taking into account the results in Fig. 7b. As an alternative, large-scale windows comprising the proposed small windows arranged like tiles can be considered, where the solar cell arrays are installed inside the grids, as shown in Fig. 7a. The window efficiency (ηW ) of such grid solar windows can be determined as a function of the side length of the window (W, transparent area contributes to power generation) and the grid frame length (F, opaque area does not contribute to power generation). A smaller F value yields better results, but the F value depends on the maturity of the solar cell and the module manufacturing technology. By considering the ECR tendency for the size of window (Fig. 7b) and the performance characteristics of the GaAs cells, we calculated ηW for the variation in W and F. For the calculation, the measured spectral ECR (Fig. 3d), the external quantum efficiency (ref [26]) of the GaAs cell, and the standard recommended spectral data of AM1.5 G were multiplied and integrated, and the used reference window area was defined as the square of (W + F). As shown in Fig. 7c, at certain side length (optimized side length, Wopt ), the maximum ηW can be
G. Lee, M. Shin and G.y. Lee et al. / Energy & Buildings 202 (2019) 109341
7
Fig. 7. Grid solar window and simulated system performance for the window size, where the size of a scaled-up window is defined by the lengths of the side (W) and frame (F) of the square window: (a) schematic of the grid solar window, and (b) transmittance and ECR at 532 nm (at F = 0), (c) window efficiency (ηW ), (d) effective illuminance, and (e) figure of merits (FoM = ηW × effective illuminance) for the grid solar windows.
obtained by compensating the loss area of the grids; when F of the grid in our device is approximately 4–8 mm, Wopt can be ∼20 mm. The solar windows should not only produce electricity, but should also provide sufficient brightness indoors. The illuminance from a window can be calculated by multiplying and integrating the measured spectral transmittance (Fig. 3d), the standard recommended spectral data of AM1.5G, and luminosity function [29], and the effective illuminance is calculated as a function of the transparent (W) and opaque (F) areas of the grid windows, as shown in Fig. 7d. The overall performance of the solar window should be evaluated in terms of both power generation and indoor illuminance. Therefore, we define the figure of merit (FoM) as the product of the window efficiency (Fig. 7c) and the effective illuminance (Fig. 7d); the FoM in Fig. 7e shows that Wopt can be differently
determined with F, and the Wopt for maximum FoM is longer than that of maximum ηW because of the property of effective illuminance. We want to note that although Wopt can be differently optimized depending on the light-guiding ability, spectral loss, transparency, and collection characteristics of the solar windows, Fig. 7e presents an alternative optimization procedure for the grid solar windows. For waveguide type of solar windows, the imperfection in the light guiding property inevitably shows a certain amount of guiding loss. Therefore, optimization is required between the aperture area contributing to transparency and power generation and the peripheral area not contributing to transparency and power generation. The proposed FoM can be used to optimize the efficiency and illuminance, which are the most important properties of a solar window.
8
G. Lee, M. Shin and G.y. Lee et al. / Energy & Buildings 202 (2019) 109341
4. Conclusions In this study, a high-efficiency white-light solar window was developed. The waveguide plates consist of a diffuser and patterned glass, which scatters and guides the incident light to GaAs cell arrays at the window frame edges. These were designed and evaluated using the ray-tracing simulation method. We constructed and assembled the solar window using 3D printing. The spectral and spatial characteristics of the solar window were investigated and measured with an integrating sphere and 2D scanning image system. The solar window exhibited a module η of 6.368%, a very high value for transparent solar cells and modules. It also exhibited an almost constant average transmittance of about 21.6% in the visible range of 40 0–80 0 nm and a CRI value of about 97.8 for sunlight, allowing the human eye to see indoor objects without significant color distortion. It still exhibited a small geometric concentration gain of 1.14, an average ECR of about 27.0%, and a small coupling loss of approximately 3.3% in the 40 0–110 0 nm range. In addition, the proposed solar window could perform stably regardless of the incident angle of sunlight. We also presented the method to increase the size of the solar window for optimizing the window efficiency and effective indoor illuminance. These results can contribute to the further development of improved solar windows with a wider spectral capability and better optical system, with solar cell arrays optimized for waveguide-type solar windows. Declaration of Competing Interest None. Acknowledgments The authors appreciate the Korea Institute of Energy Research (KIER) for officially evaluating the efficiency of the fabricated device. Funding This work was supported by the Energy Technology Development Program of the Korea Institute of Energy Technology Evaluation and Planning (KETEP) [grant number 20153030012870]; Pioneer Research Center Program through the National Research Foundation of Korea funded by the Ministry of Science, ICT, and Future Planning [grant number NRF-2013M3C1A3065040]; KIST institutional research program (2E29300). References [1] N. Antonakakis, I. Chatziantoniou, G. Filis, Energy consumption, CO2 emissions, and economic growth: an ethical dilemma, Renew. Sustain. Energy Rev. 68 (2017) 808–824. [2] A. Allouhi, Y.E. Fouih, T. Kouskou, A. Jamil, Y. Zeraouli, Y. Mourad, Energy consumption and efficiency in buildings: current status and future trends, J. Clean Prod. 109 (2015) 118–130. [3] A. Dimoudi, C. Tompa, Energy and environmental indicators related to construction of office buildings, Resour. Conserv. Recycl. 53 (2008) 86–95. [4] M. Suzuki, T. Oka, Estimation of life cycle energy consumption and CO2 emission of office buildings in Japan, Energy Build. 28 (1998) 33–41.
[5] A. Hunt, P. Watkiss, Climate change impacts and adaptation in cities: a review of the literature, Clim. Change 104 (2011) 13–49. [6] I. Dincer, Renewable energy and sustainable development: a crucial review, Renew. Sustain. Energy Rev. 4 (20 0 0) 157–175. [7] N.L. Panwar, S.C. Kaushik, S. Kothari, Role of renewable energy sources in environmental protection: a review, Renew. Sustain. Energy Rev. 15 (2011) 1513–1524. [8] A.K. Akella, R.P. Saini, M.P. Sharma, Social, economical and environmental impacts of renewable energy systems, Renew. Energy 34 (2009) 389–396. [9] B.P. Jelle, C. Breivik, The path to the building integrated photovoltaics of tomorrow, Energy Proc. 20 (2012) 78–87. [10] I. Cerón, E. Camaño-Martin, F.J. Neila, ‘State-of-the-art’ of building integrated photovoltaic products, Renew. Energy 58 (2013) 127–133. [11] E. Biyik, M. Araz, A. Hepbasli, M. Shahrestani, R. Yao, L. Shao, et al., A key review of building integrated photovoltaic (BIPV) systems, Eng. Sci. Technol. Int. J. 20 (2017) 833–858. [12] Y.T. Chae, J. Kim, H. Park, B. Shin, Building energy performance evaluation of building integrated photovoltaic (BIPV) window with semi-transparent solar cells, Appl. Energy 129 (2014) 217–227. [13] B.P. Jelle, C. Breivik, H.D. Røkenes, Building integrated photovoltaic products: a state-of-the-art review and future research opportunities, Sol. Energy Mater. Sol. Cells 100 (2012) 69–96. [14] M. Pagliaro, R. Ciriminna, G. Palmisano, BIPV: merging the photovoltaic with the construction industry, Prog. Photovolt. 18 (2010) 61–72. [15] J.H. Yoon, J. Song, S.J. Lee, Practical application of building integrated photovoltaic (BIPV) system using transparent amorphous silicon thin-film PV module, Sol. Energy 85 (2011) 723–733. [16] C.A. Wolden, J. Kurtin, J.B. Baxter, I. Repins, S.E. Shaheen, J.T. Torvik, et al., Photovoltaic manufacturing: present status, future prospects, and research needs, J. Vac. Sci. Technol. A 29 (2011) 030801. [17] M.A. Green, K. Emery, Y. Hishikawa, W. Warta, E.D. Dunlop, Solar cell efficiency tables (version 48), Prog. Photovolt. 24 (2016) 905–913. [18] A.K. Shukla, K. Sudhakar, P. Baredar, Recent advancement in BIPV product technologies: a review, Energy Build. 140 (2017) 188–195. [19] K.S. Chen, J.F. Salinas, H.L. Yip, L. Huo, J. Hou, A.K.Y. Jen, Semi-transparent polymer solar cells with 6% PCE, 25% average visible transmittance and a color rendering index close to 100 for power generating window applications, Energy Environ. Sci. 5 (2012) 9551–9557. ˇ [20] M. Cekon , R. Slávik, Total solar transmittance quantifying of transparent insulation building materials based on real climate outdoor measurements, Energy Procedia 132 (2017) 243–248. [21] L.H. Slooff, E.E. Bende, A.R. Burgers, T. Budel, M. Pravettoni, R.P. Kenny, et al., A luminescent solar concentrator with 7.1% power conversion efficiency, Phys. Stat. Sol. RRL 2 (2008) 257–259. [22] M.G. Debije, P.P.C. Verbunt, Thirty years of luminescent solar concentrator research: solar energy for the built environment, Adv. Energy Mater. 2 (2012) 12–35. [23] R.-T. Chen, J.L.H. Chau, G.-L. Hwang, Design and fabrication of diffusive solar cell window, Renew. Energy 40 (2012) 24–28. [24] J.M. Kim, P.S. Dutta, Optical efficiency–concentration ratio trade-off for a flat panel photovoltaic system with diffuser type concentrator, Sol. Energy Mater. Sol. Cells 103 (2012) 35–40. [25] J.H. Karp, E.J. Tremblay, J.E. Ford, Planar micro-optic solar concentrator, Opt. Express 18 (2010) 1122–1133. [26] Y.H. Ghymn, K. Jung, M. Shin, H. Ko, A luminescent down-shifting and moth-eyed anti-reflective film for highly efficient photovoltaic devices, Nanoscale 7 (2015) 18642. [27] K. Yoshikawa, H. Kawasaki, W. Yoshida, T. Irie, K. Konishi, K. Nakano, et al., Silicon heterojunction solar cell with interdigitated back contacts for a photoconversion efficiency over 26%, Nat. Energy 2 (2017) 17032. [28] M.K.M. Desa, S. Sapeai, A.W. Azhari, K. Sopian, M.Y. Sulaiman, A. Amin, et al., Silicon back contact solar cell configuration: a pathway towards higher efficiency, Renew. Sustain. Energy Rev. 60 (2016) 1516–1532. [29] X. Guo, K.W. Houser, A review of colour rendering indices and their application to commercial light sources, Light. Res. Technol. 36 (2004) 183–197. [30] N. Lynn, L. Mohanty, S. Wittkopf, Color rendering properties of semi-transparent thin-film PV modules, Build. Environ. 54 (2012) 148–158. [31] G. Wyszecki, W.S. Stiles, in: Color Science: Concepts and Methods, Quantitative Data and Formulae, 3, Wiley, Chap., 1982, p. 3. [32] A.J. Cox, A.J. DeWeerd, J. Linden, An experiment to measure Mie and Rayleigh total scattering cross sections, Am. J. Phys. 70 (2002) 620.