In densely populated areas, ground mounted photovoltaic power plants compete with agriculture for cultivable land. Agrivoltaic systems allow the combination of these two forms of land use by deliberately designed light sharing. In this contribution, we present a spectrally selective solar cell, for use in agrivoltaic systems, greenhouses, and photo-bioreactors. Our concept benefits from a solar cell with a transmission spectrum which can be easily tuned for the specific absorption requirements of algae and plants. This is achieved by a Fabry-Perot-type multilayer resonator as a back reflector, which determines the transmission and absorption spectrum of the solar cell. We demonstrate the extent of how this transmission spectrum can be engineered by varying the layer thicknesses of the reflector and we show how the reflecting metal layers in the back reflector influence the transmission and photocurrent generation of the spectrally selective solar cell. Finally, we analyze the optical loss mechanisms of the solar cell layer stack to address further optimization potential. Our work offers a spectrally selective solar cell which can be easily adjusted for the requirements of combining photovoltaic and photosynthesis.
© 2021 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
Renewable energy plays a central role in the effort to meet the increasing energy demand without the use of fossil resources. Photovoltaic (PV) power generation in particular offers a huge potential to meet the demand. Although this technology is well known for roof top installations and large PV power plants, alternative applications have also been realized to further utilize solar energy. PV integration in building facades and windows [1–5], in floating installations [6–8] as well as in cars [9–12] are just a few concepts. Another application is the combination of photovoltaic and photosynthesis, as introduced by Goetzberger et al. in 1982 . Their idea was to install spaced rows of PV modules on agricultural land. This ensured only temporary shadowing during the day and the possibility to cultivate crops amidst the PV modules. Recently, this approach has been studied in several field tests [14–19]. Instead of installing opaque modules in spaced rows, spectrally-selective PV modules can be used to cover entire areas of farm land, greenhouses or photo-bioreactors [20–23]. The semi-transparent approach provides the opportunity to utilize the spectrally selective absorption of light in algae and green plants. They show a peaked rate of photosynthesis for blue and red light, where the pigment chlorophyll is absorbing . This leaves a “green gap” between the absorption peaks, which can be utilized for photovoltaic applications. Spectrally selective solar cells (SSSC) use the light from the green gap and infrared to generate electricity, while blue and red light is transmitted. This enables a high utilization of the solar spectrum since power can be generated from the light not required by the plants for photosynthesis. The spectrally selective approach was first realized by using luminescent solar concentrators . By embedding them in panes of a greenhouse, a significant growth of tomatoes and micro algae for spectrally selective illumination was demonstrated [21,25]. However, this approach still needs opaque solar cells for electricity generation and the light-guiding towards the solar cells leads to optical losses.
Recently, we proposed an alternative SSSC concept , which is based on a cavity enhanced amorphous germanium (a-Ge:H) solar cell [26–29]. This type of solar cell confines light in an ultra-thin absorber due to non-trivial phase shifts at the interfaces between the semiconductor and the electrical contacts consisting of a reflective metal layer and a transparent conductive oxide, respectively. Due to the strong optical confinement and the high absorption coefficient of a-Ge:H the absorber thickness can be reduced to ∼5-10 nm while still achieving an efficiency of 5% . We chose a-Ge:H instead of a-Si:H as absorber material because of its higher absorption coefficient for wavelengths above 500 nm. This allows high transmission below 500 nm for absorber thicknesses of 5 to 10 nm, while ensuring high photocurrent generation of the non-transmitted wavelengths.
Semi transparency can be achieved by replacing the silver layer with a transparent conductive oxide or an oxide/metal/oxide (OMO) multilayer. These multilayers are transparent and conductive, so that charge carrier extraction is ensured [30–32]. Furthermore, OMOs are stable for annealing temperatures below 300°C . However, they do not show strong interferences which are necessary for the deliberately designed light sharing. Therefore, we introduce an additional metal layer in the oxide/metal/oxide multilayer stack to achieve a spectrally selective reflector as shown in Fig. 1(b). The metal-oxide-metal-oxide (MOMO) multilayer stack is based on a Fabry-Perot-resonator, which is well known for its color filtering properties [34–36]. The oxide layer in between the two highly reflecting metal layers is an optical cavity in which constructive interference causes transmission and destructive interference results in reflection. For integration into the solar cell the MOMO reflector can be designed in such a way, that selected wavelengths are reflected into the absorber layer and used for charge carrier generation, while the remaining wavelengths are transmitted through the devices. In addition, the metal layers ensure good conductivity (<10 Ω/□) of the electrode so that charge carriers can be extracted. Here, silver is used as a metal due to its high conductivity  and aluminum doped zinc oxide (AZO) as transparent conductive oxide due to its conductivity and favorable low absorption . The absorbing part consist of an ultrathin nip solar cell with an intrinsic a-Ge:H layer of ∼10 nm thickness, which is sandwiched by n- and p-doped silicon layers [see Fig. 1(a)]. In addition, intrinsic silicon buffer layers are placed between the doped layers and the absorber. The chosen materials have the advantage, that only well-established, industry proven thin film deposition processes like plasma enhanced chemical vapor deposition and magnetron sputtering are used for the deposition. In Fig. 1(c) the optical appearance of the SSSC is demonstrated.
In this publication we demonstrate the spectral selective properties of a MOMO reflector, which can be easily adjusted by only changing layer thicknesses. This flexibility is used to achieve a transmission spectrum of the MOMO reflector that corresponds to the absorption spectrum of chlorophyll a (chl a) and chlorophyll b (chl b), respectively. We focus on the influence of increasing the finesse of the Fabry-Perot-cavity by adjusting and optimizing the metal thickness on the optical and electrical properties of the solar cell. Furthermore, the optical losses of the solar cell are identified which need to be overcome in future for improved transmission and power conversion efficiency.
2.1. Sample preparation and characterization
The silicon and germanium layers were deposited on 10 × 10 cm2 float glass substrates (Saint Gobain) with a 400 nm AZO front contact via plasma enhanced vapor deposition. The MOMO back contacts were added via dc magnetron sputtering at room temperature through a mask for 1 × 1 cm2 cells. After the depositions, the solar cells were annealed in 1h steps at 130°C in ambient atmosphere for 3h in total. The detailed fabrication process is described in Ref. . The deposition time for the silver layers was 8, 15 and 22 s with a deposition rate of ∼1 nm/s, determined by profilometry measurements. The AZO layers were deposited with a rate of 1.4 nm/s aiming for layer thicknesses of tAZO1 = 271 nm and tAZO2 = 85 nm. Optical characterization was performed using a Cary 5000 spectrophotometer from Agilent with an integrating sphere. The current-voltage-curves were recorded with a WACOM dual lamp solar simulator according to standard test conditions (AM1.5G spectrum, 1000 W/m2, 25 °C). The external quantum efficiency was measured with an RR-2100 system from LOT Oriel. The sheet resistance of the MOMO reflector samples was measured with a Jandel RM3-AR four point probe system.
Optical simulations were performed by 1D transfer matrix method using the software package CODE/Scout. An optical model of the solar cell layer stack and MOMO reflector, depicted in Fig. 1, were created. In the solar cell layer stack, the doping and buffer layers on the n- as well as on the p-side are treated as one combined layer respectively, since the difference in refractive indices is negligible. Optical constants n (refractive index) and k (absorption index) for all materials except for silver were generated from optical models, which were fitted to measured reflection and transmission spectra of single layers on glass substrates (see Fig. S1 in Supplement 1). However, due to de-wetting effects various morphologies other than a flat layer silver are possible, including a connected network of Ag grains, which can result in plasmonic effects . Furthermore, it was shown, that the optical constants of silver thin films, in contrast to bulk material, depends on the layer thickness . By splitting the silver layer virtually in two parts with different optical constants, different silver layer thicknesses can be approximated reasonably well without individual optical constants for every film thickness. To reduce the complexity of the simulation for the stand-alone MOMO layer, the n and k values for silver provided by Palik et al.  were used (Figs. 2–5).
It should be noted that for fitting of the solar cell stack, only the layer thicknesses of silver and AZO in the MOMO where changed individually, whereas remaining layer thicknesses were the same for all SSSC. The fitted layer thicknesses were used to calculate the absorption of each individual layer of the SSSCs.
3. Results and discussion
3.1. Spectrally selective reflector
The spectrally selective reflector, as shown in Fig. 1(b), is the basis of the SSSC concept. As mentioned above, a major advantage of the MOMO reflector is the wide range of transmission spectra accessible with its design. According to the Fabry-Perot-theory, the transmission peak position and the free spectral range (FSR, distance between two consecutive transmission maxima for a given thickness) are a result of a phase shift, accumulated during a round trip of light in the cavity. This phase shift depends on the reflection phase shifts at the interfaces, on the refractive index of the cavity material and on the thickness of the layer sandwiched by the mirrors. The finesse (ratio between FSR and full-width-at-half-maximum) of the transmission peaks is defined by the reflectivity of the mirrors in the resonator. The aim of this analysis is to use these properties to design a MOMO with high transmission between 400 and 470 nm and between 640 and 680 nm, where chlorophyll absorbs light . Therefore, we studied the influence of the AZO layer thicknesses tAZO1 and tAZO2 and the influence of the silver layer thicknesses tAg1 and tAg2 by simulating transmission and reflection spectra of the reflector.
Figures 2(a) and 2(b) shows calculated transmission (T) and reflection (R) spectra of a MOMO reflector as a function of tAZO1 and wavelength (λ), respectively. As expected, the graphs show that with increasing tAZO1 thickness transmission maxima and reflection minima appear in the UV and successively shift to longer wavelengths. It is important to note, that the FSR for maxima at a given wavelength changes with the layer. For example, a transmission maximum at 650 nm can be achieved for tAZO1 thicknesses of 110 nm, 275 nm, and 440 nm with consecutive transmission maxima at 400 nm, 455 nm, and 500 nm, respectively. This can be used to fit the transmission spectrum to the absorption spectrum of chlorophyll. The highest achievable transmission is 80% between 550 and 750 nm, depending on the thickness of the layer tAZO1. Figure 2(b) shows that the reflection maxima lie between the transmission maxima in Fig. 2(a). For visible light a reflection of 80% is achievable while the highest reflection (88%) can be observed in the infrared range.
The influence of tAZO2 on the optical properties is shown in Figs. 2(c) and 2(d). Similar to tAZO1, an increasing film thickness of tAZO2 leads to transmission peaks which arise in the UV and shift to longer wavelengths. However, they are less pronounced, since they are part of an asymmetric cavity between silver (Ag2) and air. These peaks are superimposed on to the peaks originating from AZO1. Therefore, this layer allows a fine adjustment of the transmission peak which is beneficial for an optimal matching of the chlorophyll absorption spectrum. For example, between tAZO2 = 80 and 100 nm an additional shoulder occurs at λ = 400 nm enhancing the peak width.
Figure 3 shows the influence of the thickness of the two silver layers, tAg1 and tAg2, on the reflection and transmission spectra. AZO thicknesses of tAZO1 = 274 nm and tAZO2 = 85 nm were chosen due to transmission peaks at 400 nm, 450 nm, and 645 nm. As shown in Fig. 3(a), a simultaneous thickness increase of both silver layers (tAg1 = tAg2) leads to an increasing finesse of the transmission peaks since the reflectivity of the layers increases. Simultaneously, the transmission peak height is reduced and the reflection in between is enhanced. Therefore, the silver layers can be used to adjust the ratio of reflected light for photocurrent generation and transmitted light for plant growth. We found an overall optimal thickness to be at approximately tAg1 = tAg2 = 15 nm as it leads to high reflection for λ = 550 nm and λ > 800 nm and high transmission for λ = 450 nm and λ = 650 nm. A discussion of the effects of unequal silver thicknesses (tAg1 ≠ tAg2) can be found in Supplement 1. However, only smaller transmission values can be achieved for the MOMO with tAg1 ≠ tAg2 (Fig. S2 in Supplement 1) and, therefore, we focus on tAg1 = tAg2.
The tuneability of the MOMO can be used to achieve transmission spectra matching the absorption of either chl a or chl b, since significant growth of algae and tomatoes can be achieved for spectral selective illumination at the absorption maxima [21,25]. Figure 4 shows the absorption spectra (norm. α) of chl a and chl b and transmission spectra calculated from our optical model for two configurations of the MOMO reflector. As the absorption spectra of chl a and chl b differ in the spectral positions of the absorption peaks, we designed for each chlorophyll pigment a reflector with variations in the AZO layer thicknesses. The first one, matching the absorption of chl a, is achieved with tAZO1 = 106 nm and tAZO2 = 121 nm. As can be seen in Fig. 4(a), the first and second order resonance of the transmission at 425 nm and 660 nm coincide with the absorption of chl a. Even the shoulder at ∼400 nm is included. The transmission in the “green gap” is reduced to 20%. The second reflector configuration with tAZO1 = 271 nm and tAZO2 = 85 nm in Fig. 4(b) uses the second and third order resonance, since it has to match the smaller FSR of the chl b absorption peaks. It exhibits four transmission peaks at 400 nm, 450 nm, 645 nm and 1350 nm, while the latter peak has no effect on the chl b absorption and therefore is not shown in the figure. A silver thickness of 15 nm was chosen for both reflectors due to the high reflection between high transmission peaks. Depending on requirements of different plants or algae, other silver thicknesses can also be advantageous, since these determine the ratio of reflected and transmitted light.
For validation of our simualtion we fabricated MOMO reflectors of the second configuration for chl b, since higher transmission values were achieved in the simulation. Three MOMO reflectors are deposited using tAg1 = tAg2 = 8 nm, 15 nm and 22 nm. Figure 5 shows the corresponding calculated and measured spectra of the three reflectors. As can be seen, the peak positions of the spectrum were predicted successfully using the optical model. However, additional absorption by interface and plasmonic effects lead to a slight mismatch in intensity and are not accurately represented by the simple Ag model (employing optical constants data from Palik et al.). The two-silver layer model used to simulate the complete SSSC stack results in a more accurate prediction of peak height (shown in Fig S3 and Table S1 in Supplement 1). As expected from the simulations, the Ag thickness affects the height of the two transmission peaks, as well as the amount of reflection between them, showing the increasing finesse of the resonator with thicker silver layers. This demonstrates the possibility to adjust the ratio between the light for the plants and the light for the solar cell and verifies our modeling.
3.2. Spectrally selective solar cell
The three fabricated MOMO reflectors shown in Fig. 5 in the previous section were integrated into ultrathin germanium solar cells in order to achieve SSSCs. The sheet resistance of the MOMO layers is between 2 and 10 Ω/□ depending on the silver thickness. A front contact of 400 nm AZO on glass was used in combination with a p-i-n solar cell stack which has an approximately 10 nm thick a-Ge:H absorber as depicted in Fig. 1.
Figure 6 shows the external quantum efficiency (EQE), the total absorption (A, presented as 1-R-T), the reflection (R, presented as 1-R), and transmission (T) spectra of the three SSSCs and an opaque reference solar cell with a silver reflector. All four solar cells show broadband EQE between 350 nm and 1050 nm. The EQE of the reference cell reaches the highest values, followed by the SSSCs with tAg1,2 = 22 nm and tAg1,2 = 15 nm which have similar EQEs. The solar cell with tAg1,2 = 8 nm has the lowest EQE of the set. In the wavelength range of 400 to 700 nm the EQE of the solar cells is highest, when the total absorption is maximal. When the total absorption and the transmission of the SSSCs are compared side by side, it can be seen that high transmission leads to reduced absorption in the SSSCs. The reduced total absorption leads to a reduced EQE. This demonstrates that the photocurrent generation mirrors the spectral selectivity of the solar cell. Therefore, the MOMO reflector can be used in the ultrathin solar cell for deliberately designed light sharing. Compared to the MOMO reflectors in Fig. 5 the transmission peaks of the SSSC are shifted slightly to smaller wavelengths due to the silicon-silver interface in the solar cell. Furthermore, interference effects of the front contact and the absorbing layers lead to a small shift. However, the SSSC still transmits light in the wavelength range, where chlorophyll absorbs light. A small adjustment of the MOMO reflector in the SSSC would be necessary to match the transmission spectra of the MOMO on glass.
The influence of the silver layers in the MOMO reflector is clearly visible in the transmission spectra of the SSSC. With decreasing silver thickness, the transmission peak increases, as it was observed in Fig. 5 for the MOMO reflector without solar cell. The blue transmission peak at 450 nm achieves a transmission between 4% and 16%, while the red transmission peak centered at approximately 620 nm shows a transmission between 34% and 48%. The latter peak shows additionally a shift to longer wavelengths with thicker silver layers. This is caused by the increasing reflectivity of the MOMO reflector leading to a higher absorption within the resonant cavity in the wavelength range between 580 nm and 620 nm. Additionally, the reflection losses are increased in this wavelength range. These reflection losses are also observed for the opaque solar cell and hence they are not an issue of the MOMO reflector integration. The decrease in EQE between 600 nm and 800 nm in the SSSCs compared to the opaque cell, is mainly dominated by the transmission of light in this region. It is also influenced by the increased reflection in that region and by parasitic absorption, which will be discussed in the next section.
Table 1 compares the photocurrent calculated from the EQE with the photocurrent from the IV curves (shown in Fig. S4 and Table S2 in Supplement 1). Both methods show for increasing silver thicknesses an increasing photocurrent generation due to the increasing reflectivity of the MOMO reflector. Furthermore, the Voc extracted from the IV curves is shown in Table 1. It should be emphasized that the reflector type has a negligible influence on the open-circuit voltage (Voc), as it is for all cells between 458 mV and 467 mV. It can be concluded that the charge carrier extraction efficiency is comparable for all four back contacts. Therefore, the MOMO reflector is suitable as back contact in solar cells. A silver thickness of tAg1,2 = 15 nm is most suitable for application in the SSSC as it shows the best balance of high transmission, high photocurrent generation and a high charge carrier extraction.
3.3. Analysis of current losses in the SSSC
Figure 7 shows the internal quantum efficiency (IQE = EQE/(1-R-T)) of the solar cells. It can be seen, that the IQE of the SSSCs is reduced at the transmission peaks and in the IR part of the spectrum if compared to the IQE of the opaque cell. At 560 nm the difference in the IQE of the opaque and the SSSCs is only 6%. This remaining difference can be explained by small additional parasitic absorption losses due to the integration of the MOMO reflector. This shows that no significant electronic losses occur due to the integration of the MOMO reflector. In wavelength ranges where the SSSC transmits light, the difference of the IQE of the opaque cell and the SSSC is much higher. Here, additional parasitic absorption losses in the MOMO reflector occur.
In order to quantify the parasitic absorption losses, the absorption of each layer in the solar cell was calculated as described in the methods section and in Supplement 1. Figure 8 shows a stacked area plot of the layer absorptions together with the reflection and transmission of the entire solar cell stack based on the simulation of the solar cell stack. Here the solar cell stack with tAg1,2 = 15 nm, tAZO1 = 271 nm, and tAZO2 = 85 nm was used. As expected, the germanium layer in the center of the solar cell shows high absorption in the wavelength range 300 - 1000 nm. Below 600 nm absorption occurs predominantly in the silicon layers, whereas above 600 nm the absorption in the silver layers is more prominent. The high absorption in the silicon layers is a result of the high absorption coefficient of the silicon at wavelengths smaller than 500 nm, and the reason for the reduced transmission below 20% at 450 nm, like shown in Fig. 6(b). In contrast, the transmission around 620 nm is almost not affected by the silicon layers as their absorption coefficient is small in this wavelength range. The absorption in the silver and the reflection losses explain the small EQE between 700 and 800 nm in comparison to the opaque cell.
The absorption in the silver layer can be explained by two absorption mechanisms. For thin silver layers with a thickness below or near the percolation limit, plasmon resonances on the surface can occur due the formation of clusters during growth. This is the cause of light absorption and scattering [43,44]. Furthermore, the absorption of light in the silver layers increases with increasing silver thickness. This leads to a higher chance of photon absorption due to its interaction with the free elections of the metal . Therefore, the silver layer thickness has to be chosen just above the percolation limit to minimize absorption losses. This limit can be reduced by adding seed layers like germanium or silver oxide [39,45]. Besides the parasitic absorption losses in the silver, the reflection losses must be reduced. This can be achieved by minimizing the reflection at the air-glass and glass AZO interface. Additionally, the reflection losses can be minimized by applying AZO nanorods  or optimized oxide/metal/oxide front contacts.
A closer look on Fig. 8 shows, that the layer absorption of the a-Ge:H at 400 nm is approximately 10%, while the EQE in Fig. 6 indicates a photon absorption efficiency of approximately 35% for the solar cell. This indicates photocurrent generation in the intrinsic silicon buffer layers as it was shown in other studies . In Fig. 9 the calculated layer absorption is compared to the experimental EQE. A significant contribution to the EQE by the silicon layers is evident, especially for the small wavelength below 500 nm. This is due to the high absorption coefficient of silicon for small wavelengths. It can be reasonably assumed that mainly the intrinsic silicon layers, i.e. only half of the total thickness of silicon, are active in current generation, since the doped layers are very defect rich and lead to parasitic absorption.
This opens up the possibility of further optimization of the approach by replacing the doped silicon layers with transparent, charge carrier selective contacts. Recently published reviews discuss appropriate materials, which were already applied for different solar cell concepts [48,49]. For example, molybdenum oxide as hole-selective contact or titanium oxide as electron selective contact are promising materials.
In this paper we demonstrated the concept of a spectrally selective solar cell based on a semi-transparent solar cell and a Fabry-Perot-type metal/oxide/metal/oxide (MOMO) reflector. The design of the MOMO reflector offers the flexibility to adjust transmission and reflection easily by changing layer thicknesses. We presented the possibility to match the absorption spectrum of chlorophyll a and b for the simultaneous use of photovoltaic power generation and photosynthesis. Three different silver thicknesses were studied in the MOMO reflector by adapting the silver layer thickness, the ratio between light for illumination of algae or plants and for photocurrent generation can be changed. Finally, the optical losses in the solar cells were discussed and paths for improvement were suggested. The solar cell shows huge potential for combining photovoltaics with photosynthesis to reach new applications of solar cells on bio-reactors, greenhouses, or agricultural land.
Bundesministerium des Innern, für Bau und Heimat (SWD-10.08.18.7-17.02).
The authors thank D. Berends for sputter deposition of the AZO front contacts and for helpful discussion. The authors thank also C. Lattyak, M. Götz, U. Banik, and H. Meddeb for helpful discussion.
The authors declare no conflicts of interest.
See Supplement 1 for supporting content.
1. M. Debbarma, K. Sudhakar, and P. Baredar, “Comparison of BIPV and BIPVT: A review,” Resource-Efficient Technol. 3(3), 263–271 (2017). [CrossRef]
2. G. Y. Yun, M. McEvoy, and K. Steemers, “Design and overall energy performance of a ventilated photovoltaic façade,” Sol. Energy 81(3), 383–394 (2007). [CrossRef]
3. J.-H. Yoon, J. Song, and S.-J. Lee, “Practical application of building integrated photovoltaic (BIPV) system using transparent amorphous silicon thin-film PV module,” Sol. Energy 85(5), 723–733 (2011). [CrossRef]
4. E. Biyik, M. Araz, A. Hepbasli, M. Shahrestani, R. Yao, L. Shao, E. Essah, A. C. Oliveira, T. del Caño, E. Rico, J. L. Lechón, L. Andrade, A. Mendes, and Y. B. Atlı, “A key review of building integrated photovoltaic (BIPV) systems,” Eng. Sci. Technol. Int J. 20(3), 833–858 (2017). [CrossRef]
5. C. J. Traverse, R. Pandey, M. C. Barr, and R. R. Lunt, “Emergence of highly transparent photovoltaics for distributed applications,” Nat. Energy 2(11), 849–860 (2017). [CrossRef]
6. A. Sahu, N. Yadav, and K. Sudhakar, “Floating photovoltaic power plant: A review,” Renewable Sustainable Energy Rev. 66, 815–824 (2016). [CrossRef]
7. C. Ferrer-Gisbert, J. J. Ferrán-Gozálvez, M. Redón-Santafé, P. Ferrer-Gisbert, F. J. Sánchez-Romero, and J. B. Torregrosa-Soler, “A new photovoltaic floating cover system for water reservoirs,” Renewable Energy 60, 63–70 (2013). [CrossRef]
8. K. Trapani and M. Redón Santafé, “A review of floating photovoltaic installations: 2007-2013,” Prog. Photovolt. Res. Appl. 23(4), 524–532 (2015). [CrossRef]
9. C. Pisanti, “Design and energetic evaluation of a mobile photovoltaic roof for cars,” Energy Procedia 81, 182–192 (2015). [CrossRef]
10. T. Masuda, K. Araki, K. Okumura, S. Urabe, Y. Kudo, K. Kimura, T. Nakado, A. Sato, and M. Yamaguchi, “Static concentrator photovoltaics for automotive applications,” Sol. Energy 146, 523–531 (2017). [CrossRef]
11. D. Sato, K. H. Lee, K. Araki, T. Masuda, M. Yamaguchi, and N. Yamada, “Design of low-concentration static III-V/Si partial CPV module with 27.3% annual efficiency for car-roof application,” Prog. Photovolt. Res. Appl. 27(6), 501–510 (2019). [CrossRef]
12. M. Kühnel, B. Hanke, S. Geißendörfer, K. von Maydell, and C. Agert, “Energy forecast for mobile photovoltaic systems with focus on trucks for cooling applications,” Prog. Photovolt. Res. Appl. 25(7), 525–532 (2017). [CrossRef]
13. A. Goetzberger and A. Zastrow, “On the coexistence of solar-energy conversion and plant cultivation,” Int. J. Sol. Energy 1(1), 55–69 (1982). [CrossRef]
14. A. Weselek, A. Ehmann, S. Zikeli, I. Lewandowski, S. Schindele, and P. Högy, “Agrophotovoltaic systems: applications, challenges, and opportunities. A review,” Agron. Sustainable Dev. 39(4), 35 (2019). [CrossRef]
15. C. Dupraz, H. Marrou, G. Talbot, L. Dufour, A. Nogier, and Y. Ferard, “Combining solar photovoltaic panels and food crops for optimising land use: Towards new agrivoltaic schemes,” Renewable Energy 36(10), 2725–2732 (2011). [CrossRef]
16. B. Valle, T. Simonneau, F. Sourd, P. Pechier, P. Hamard, T. Frisson, M. Ryckewaert, and A. Christophe, “Increasing the total productivity of a land by combining mobile photovoltaic panels and food crops,” Appl. Energy 206, 1495–1507 (2017). [CrossRef]
17. S. Ravi, J. Macknick, D. Lobell, C. Field, K. Ganesan, R. Jain, M. Elchinger, and B. Stoltenberg, “Colocation opportunities for large solar infrastructures and agriculture in drylands,” Appl. Energy 165, 383–392 (2016). [CrossRef]
18. D. Majumdar and M. J. Pasqualetti, “Dual use of agricultural land: Introducing ‘agrivoltaics’ in Phoenix Metropolitan Statistical Area, USA,” Landsc. Urban Plan. 170, 150–168 (2018). [CrossRef]
19. P. R. Malu, U. S. Sharma, and J. M. Pearce, “Agrivoltaic potential on grape farms in India,” Sustain. Energy Technol. Assess. 23, 104–110 (2017). [CrossRef]
20. C. S. Allardyce, C. Fankhauser, S. M. Zakeeruddin, M. Grätzel, and P. J. Dyson, “The influence of greenhouse-integrated photovoltaics on crop production,” Sol. Energy 155, 517–522 (2017). [CrossRef]
21. M. E. Loik, S. A. Carter, G. Alers, C. E. Wade, D. Shugar, C. Corrado, D. Jokerst, and C. Kitayama, “Wavelength-Selective Solar Photovoltaic Systems: Powering Greenhouses for Plant Growth at the Food-Energy-Water Nexus,” Earth's Future 5(10), 1044–1053 (2017). [CrossRef]
22. H. Shi, R. Xia, G. Zhang, H.-L. Yip, and Y. Cao, “Spectral Engineering of Semitransparent Polymer Solar Cells for Greenhouse Applications,” Adv. Energy Mater. 9(5), 1803438 (2019). [CrossRef]
23. M. Götz, N. Osterthun, K. Gehrke, M. Vehse, and C. Agert, “Ultrathin Nano-Absorbers in Photovoltaics: Prospects and Innovative Applications,” Coatings 10(3), 218 (2020). [CrossRef]
24. H.-W. Heldt and B. Piechulla, Plant biochemistry (Academic, 2010).
25. A. M. Detweiler, C. E. Mioni, K. L. Hellier, J. J. Allen, S. A. Carter, B. M. Bebout, E. E. Fleming, C. Corrado, and L. E. Prufert-Bebout, “Evaluation of wavelength selective photovoltaic panels on microalgae growth and photosynthetic efficiency,” Algal Res. 9, 170–177 (2015). [CrossRef]
26. V. Steenhoff, M. Theuring, M. Vehse, K. von Maydell, and C. Agert, “Ultrathin Resonant-Cavity-Enhanced Solar Cells with Amorphous Germanium Absorbers,” Adv. Opt. Mater. 3(2), 182–186 (2015). [CrossRef]
27. V. Steenhoff, A. Neumüller, O. Sergeev, M. Vehse, and C. Agert, “Integration of a-Ge: H nanocavity solar cells in tandem devices,” Sol. Energy Mater. Sol. Cells 145, 148–153 (2016). [CrossRef]
28. C. Lattyak, R.-E. Ravekes, V. Steenhoff, M. Vehse, and C. Agert, “Ultrathin Resonant-Cavity-Enhanced Amorphous Germanium Solar Cells on ZnO Honeycomb Electrodes,” IEEE J. Photovoltaics 8(1), 3–7 (2018). [CrossRef]
29. H. Meddeb, N. Osterthun, M. Götz, O. Sergeev, K. Gehrke, M. Vehse, and C. Agert, “Quantum confinement-tunable solar cell based on ultrathin amorphous germanium,” Nano Energy 76, 105048 (2020). [CrossRef]
30. M. Theuring, M. Vehse, K. von Maydell, and C. Agert, “AZO-Ag-AZO transparent electrode for amorphous silicon solar cells,” Thin Solid Films 558, 294–297 (2014). [CrossRef]
31. M. Theuring, S. Geissendörfer, M. Vehse, K. von Maydell, and C. Agert, “Thin metal layer as transparent electrode in n-i-p amorphous silicon solar cells,” EPJ Photovoltaics 5, 55205 (2014). [CrossRef]
32. N. Neugebohrn, K. Gehrke, K. Brucke, M. Götz, and M. Vehse, “Multifunctional metal oxide electrodes: Colour for thin film solar cells,” Thin Solid Films 685, 131–135 (2019). [CrossRef]
33. Y. Sugimoto, K. Igarashi, S. Shirasaki, and A. Kikuchi, “Thermal durability of AZO/Ag(Al)/AZO transparent conductive films,” Jpn. J. Appl. Phys. 55(4S), 04EJ15 (2016). [CrossRef]
34. Y.-T. Yoon and S.-S. Lee, “Transmission type color filter incorporating a silver film based etalon,” Optics Express 18(5), 5344–5349 (2010). [CrossRef]
35. C.-S. Park, V. R. Shrestha, S.-S. Lee, and D.-Y. Choi, “Trans-reflective color filters based on a phase compensated etalon enabling adjustable color saturation,” Sci. Rep. 6(1), 25496 (2016). [CrossRef]
36. Y. Kim, J. Son, S. Shafian, K. Kim, and J. K. Hyun, “Semitransparent Blue, Green, and Red Organic Solar Cells Using Color Filtering Electrodes,” Adv. Opt. Mater. 6(13), 1800051 (2018). [CrossRef]
37. D. R. Lide, CRC handbook of chemistry and physics: a ready-reference book of chemical and physical data (CRC, 1995).
38. C. Guillen and J. Herrero, “Structure, optical and electrical properties of Al: ZnO thin films deposited by DC sputtering at room temperature on glass and plastic substrates,” Phys. Status Solidi A 206(7), 1531–1536 (2009). [CrossRef]
39. J. Yun, “Ultrathin Metal films for Transparent Electrodes of Flexible Optoelectronic Devices,” Adv. Funct. Mater. 27(18), 1606641 (2017). [CrossRef]
40. P. Zhao, W. Su, R. Wang, X. Xu, and F. Zhang, “Properties of thin silver films with different thickness,” Phys. E 41(3), 387–390 (2009). [CrossRef]
41. E. D. Palik, Handbook of optical constants of solids (Academic, 1998), Vol. 3.
42. S. Helfrich, “Absorption spectrum of chlorophyll a and chlophyll b” (2018), retrieved 23.07., 2020, https://commons.wikimedia.org/wiki/File:Chlorophyll_Absorption_Spectrum.svg.
43. F. Le, D. W. Brandl, Y. A. Urzhumov, H. Wang, J. Kundu, N. J. Halas, J. Aizpurua, and P. Nordlander, “Metallic nanoparticle arrays: a common substrate for both surface-enhanced Raman scattering and surface-enhanced infrared absorption,” ACS Nano 2(4), 707–718 (2008). [CrossRef]
44. W. J. Cho, Y. Kim, and J. K. Kim, “Ultrahigh-density array of silver nanoclusters for SERS substrate with high sensitivity and excellent reproducibility,” ACS Nano 6(1), 249–255 (2012). [CrossRef]
45. G. Zhao, W. Shen, E. Jeong, S. G. Lee, S. M. Yu, T. S. Bae, G. H. Lee, S. Z. Han, J. Tang, E. A. Choi, and J. Yun, “Ultrathin Silver Film Electrodes with Ultralow Optical and Electrical Losses for Flexible Organic Photovoltaics,” ACS Appl. Mater. Interfaces 10(32), 27510–27520 (2018). [CrossRef]
46. R.-E. Nowak, M. Vehse, O. Sergeev, T. Voss, M. Seyfried, K. von Maydell, and C. Agert, “ZnO Nanorods with Broadband Antireflective Properties for Improved Light Management in Silicon Thin-Film Solar Cells,” Adv. Opt. Mater. 2(1), 94–99 (2014). [CrossRef]
47. M. Götz, M. Lengert, N. Osterthun, K. Gehrke, M. Vehse, and C. Agert, “Switchable Photocurrent Generation in an Ultrathin Resonant Cavity Solar Cell,” ACS Photonics 7(4), 1022–1029 (2020). [CrossRef]
48. P. Gao, Z. Yang, J. He, J. Yu, P. Liu, J. Zhu, Z. Ge, and J. Ye, “Dopant-Free and Carrier-Selective Heterocontacts for Silicon Solar Cells: Recent Advances and Perspectives,” Adv. Sci. 5(3), 1700547 (2018). [CrossRef]
49. T. G. Allen, J. Bullock, X. Yang, A. Javey, and S. De Wolf, “Passivating contacts for crystalline silicon solar cells,” Nat. Energy 4(11), 914–928 (2019). [CrossRef]