Many insect species rely on the polarization properties of object-reflected light for vital tasks like water or host detection. Unfortunately, typical glass-encapsulated photovoltaic modules, which are expected to cover increasingly large surfaces in the coming years, inadvertently attract various species of water-seeking aquatic insects by the horizontally polarized light they reflect. Such polarized light pollution can be extremely harmful to the entomofauna if polarotactic aquatic insects are trapped by this attractive light signal and perish before reproduction, or if they lay their eggs in unsuitable locations. Textured photovoltaic cover layers are usually engineered to maximize sunlight-harvesting, without taking into consideration their impact on polarized light pollution. The goal of the present study is therefore to experimentally and computationally assess the influence of the cover layer topography on polarized light pollution. By conducting field experiments with polarotactic horseflies (Diptera: Tabanidae) and a mayfly species (Ephemeroptera: Ephemera danica), we demonstrate that bioreplicated cover layers (here obtained by directly copying the surface microtexture of rose petals) were almost unattractive to these species, which is indicative of reduced polarized light pollution. Relative to a planar cover layer, we find that, for the examined aquatic species, the bioreplicated texture can greatly reduce the numbers of landings. This observation is further analyzed and explained by means of imaging polarimetry and ray-tracing simulations. The results pave the way to novel photovoltaic cover layers, the interface of which can be designed to improve sunlight conversion efficiency while minimizing their detrimental influence on the ecology and conservation of polarotactic aquatic insects.
Funding: This work was supported by the grant NKFIH K-123930 received by Gábor Horváth from the Hungarian National Research, Development and Innovation Office. Ádám Egri was supported by the Hungarian Economic Development and Innovation Operational Programme (GINOP-2.3.2-15-2016-00057), the Hungarian National Research, Development and Innovation Office (grant PD_19-131738) and the János Bolyai Research Scholarship of the Hungarian Academy of Sciences. Benjamin Fritz acknowledges the support of the Karlsruhe School of Optics and Photonics (KSOP, www.ksop.idschools.kit.edu ). Furthermore, this study was supported by the German Federal Ministry for Economic Affairs and Energy (BMWi) under contract number 0324179 (CISHiTec).
Copyright: © 2020 Fritz et al. This is an open access article distributed under the terms of the Creative Commons Attribution License , which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
In this work we demonstrate that microtextured photovoltaic cover layers can strongly reduce their attractiveness to the mayfly species Ephemera danica (Müller, 1764) and horseflies (Diptera: Tabanidae), two typical polarotactic aquatic insect taxa [ 29 ] functioning as indicators of polarized light pollution. We performed field experiments with microtextured polymeric coatings incorporating the surface texture of rose petals. These bioreplicated layers improve both light and water management in photovoltaic devices [ 9 , 11 , 12 ]. We show here that they also strongly reduce the attractiveness to mayflies and horseflies. This observation is explained on the basis of their reflection-polarization properties measured with imaging polarimetry, and of raytracing simulations used to analyze the influence of the surface texture topography on the generation of horizontally polarized reflected light.
We note, however, that although a commercialized antireflective, nanoporous solar glass reduced polarized light pollution for polarotactic horseflies, its attractiveness to certain mayflies was drastically increased relative to that of a smooth reference layer [ 31 ]. The reason for this may be that the reflection-polarization characteristics of the black absorber covered with the nanoporous antireflective glass mimicked a ripple-free, smooth water surface for mayflies which prefer calm water bodies to lay their eggs, where their larvae can develop safely [ 27 , 31 ]. Considering insect conservation, such an approach might therefore cause even more harm to certain insect species than a smooth glass cover.
If, for viewing directions near the Brewster’s angle the degree of polarization of reflected light is lower than the threshold of polarization sensitivity and/or if the direction of polarization of reflected light deviates from the horizontal by more than a given threshold value (both thresholds depend on the species considered), a polarotactic insect does not get attracted by the surface-reflected polarized light [ 29 ]. Thus, the use of appropriately fine-textured photovoltaic cover layers can reduce the maladaptive attractiveness, and thus polarized light pollution, by decreasing the degree of polarization and changing the angle of polarization of reflected light.
So far, the reduction of polarized light pollution of photovoltaic panels has been realized in two ways: i) By painting a grid pattern of narrow (1–2 mm width) white lines on the panel surface [ 34 , 35 ]. This is based on the unattractiveness of striped or spotted animal coats to polarotactic insects [ 24 – 26 ]. ii) with a nanoporous antireflective layer on the glass cover to decrease the degree of linear polarization of reflected light [ 31 ]. Method i) can strongly reduce polarized light pollution, however, the light-collecting area then decreases by 1–5%, depending on the density and width of the white lines [ 34 ]. Method ii) does not have this disadvantage.
A) Dark photovoltaic modules coated by a reflecting planar cover layer act as polarization traps for polarotactic insects (left) if the photovoltaic-reflected light is partially or completely horizontally polarized. An appropriate texturing of the cover layer strongly reduces polarized light pollution and improves sunlight-harvesting (right). B-C) Scanning electron microscope images of the rose petal replicated cover layer analyzed herein, and incorporating the microtexture of rose petals into a polymeric material (PMMA). Its measured surface reflectance spectrum is shown for both a blackened rear side (D) and for a Cu(In,Ga)Se 2 (CIGS) thin-film solar cell coupled to it (E). An untextured (”planar”) cover layer is used as a reference and both normal and highly oblique angle of incidences are considered. The coloured areas surrounding the (solid or dotted) curves indicate the standard deviation over N = 4 individual measurements.
Many insect species use polarization of light reflected from natural or artificial terrestrial surfaces for object identification or water detection [ 22 – 28 ]. From the Brewster’s angle, water surfaces reflect light with typical degrees of linear polarization 15% ≤ d ≤ 90% and angles of polarization -10° ≤ α ≤ +10° relative to the horizontal direction, depending on their brightness/darkness, colour and surface roughness [ 20 – 23 , 29 – 33 ]. Water surfaces usually reflect nearly horizontally polarized light [ 33 ]. Aquatic insects in general (belonging to the following orders with aquatic or semiaquatic species: Coleoptera, Collembola, Diptera, Ephemeroptera, Hemiptera, Hymenoptera, Lepidoptera, Mecoptera, Megaloptera, Neuroptera, Odonata, Plecoptera, Trichoptera) have therefore evolved to identify water bodies by the horizontal polarization of water-reflected light [ 22 , 23 ]. This strategy can result in a maladaptive attraction of polarotactic aquatic insects to smooth artificial surfaces like the glass/plactic covers of solar panels, because these surfaces can reflect a similar polarization pattern as water surfaces [ 23 , 30 , 31 ]. This can result in that aquatic insects are unable to escape from the horizontally polarized signal reflected from photovoltaic panels acting as polarization traps [ 34 ] ( Fig 1A ). If the egg-batches of these insects (e.g. mayflies, dragonflies, stone flies, caddis flies) are laid onto photovoltaic modules, they irremediably perish because of dehydration [ 23 ]. Since the larvae of these insects develop in water/mud for a few months/years, hydration by dew or rain drops on the solar panels cannot ensure the survival of eggs. This effect is harmful for the aquatic insect populations concerned, and therefore is called polarized light pollution [ 30 ].
A properly textured front surface of photovoltaic solar panels should allow the following characteristics: (i) A low sunlight reflectance irrespective of the illumination conditions and a high absorption of the collected light in the photovoltaic active layer, both leading to a high energy yield [ 1 – 3 ]. (ii) Radiative cooling that improves the power conversion efficiency and the reliability of the solar panels [ 4 , 5 ]. (iii) Self-cleaning, which decreases the maintenance costs associated to the removal of soiling particles [ 6 – 8 ]. Although many different multifunctional cover layers have been developed [ 9 – 12 ], their impact on insect ecology and conservation is largely unexplored. The study of this impact is important due to the global insect crisis [ 13 – 17 ] and to the worldwide deployment of photovoltaic installations [ 18 , 19 ]. These facts urge to understand how the reflection-polarization characteristics of photovoltaic-covered habitats can affect the behaviour of polarization-sensitive insects, especially water-seeking polarotactic aquatic insects being maladaptively attracted to asphalt roads [ 20 ] or dark car paints [ 21 ], for example.
Polymeric cover layers replicating the surface texture of rose petals, previously developed for improving sunlight-harvesting in solar cells [11, 12], were fabricated over large areas to experimentally assess their impact on polarized light pollution in the field. The scanning electron microscope images in Fig 1B and 1C show the topography of these replicas. These hot-embossed layers exhibit high structural fidelity with respect to their original bio-template, and replicate both the densely packed array of (epidermal cell) microcones with their mean aspect ratio = 0.6 [9] as well as the (cuticular) nano wrinkles atop, which are kept intact over the double replication process. Their outstanding light-harvesting capabilities are highlighted in Fig 1D and 1E. The rose petal texturing leads to a broadband and angle-tolerant decrease of surface reflectance compared with planar cover layers. For increasing angle of incidence, its optical benefit becomes even more pronounced. Thus, at an angle of incidence = 70°, the overall reflectance of CIGS solar cells covered with rose petal textured layers can be kept under 10%, a value comparable to the reflectance of planar configurations under normal incidence (see Fig 1E). Overall, the measured reflectance spectra highlight the potential of hot-embossed rose petal replicas as photovoltaic light-harvesting layers.
In a pilot experiment we used only the rose-petal-replica matte black test surface, which however, did not attract any mayflies due to its low polarizing ability (see Fig 6). Therefore, we continued our field experiment with the use of two additional control surfaces that polarize the reflected light much stronger. From the pilot experiment we conclude that the rose petal replica reduces significantly the polarized light pollution for mayflies, even if it stands alone.
On each experimental day between 18:30 and 19:15, mainly male Ephemera danica mayflies were observed around the test surfaces. Then, in the second phase of swarming, the majority of landing specimens were females often with visible egg-batches. Sometimes they laid their egg-batches on the test surfaces. S1 Table contains the time-resolved number of landings of Ephemera danica on the three different test surfaces.
Fig 5B shows the numbers of mayfly landings, the statistical analysis of which was performed for 14 min intervals (N = 49). The smooth black plastic (SBP) was the most attractive to mayflies, the attractiveness of the glass-covered rose petal (GRP) was weaker, and the rose petal (RP) was practically unattractive to Ephemera danica. The Friedman test was highly significant (p < 0.0001, Friedman chi2 = 116.97, df = 3, Kendall’s W = 0.6045). According to the Wilcoxon signed-rank test, the mayfly attractiveness of these three test surfaces differed highly significantly (RP versus SBP: p < 0.0001, RP versus GRP: p < 0.0001, SBP versus GRP: p < 0.0001, N = 49). The main reason for these findings can be explained by the facts that (i) Ephemera danica detects water with the horizontal polarization of water-reflected light [20, 32], (ii) the higher the degree of horizontal polarization of surface-reflected light, the more attractive the respective surface is to polarotactic mayflies [20, 29], and (iii) the degree of horizontal polarization of light reflected from our test surfaces decreased in the order: SBP > GRP > RP (see Fig 6). Egg laying by mayflies was observed only on smooth black plastic and glass-covered rose petal, but not on rose petal.
Similarly to the experiment with mayflies, in the horsefly experiment we first presented only the rose-petal-replica matte black test surface, which however, did not attract any horsefly due to its low polarizing ability and because in sunshine it reflected not always horizontally polarized light (Figs 2, 4, S1 Fig). Thus, the alone-standing rose petal replica reduces significantly the polarized light pollution for horseflies, too. Therefore, we added two more polarizing control surfaces in the continuation of the horsefly experiment.
Fig 5D shows the numbers of horsefly landings (detailed in S2 Table), the statistical analysis of which was performed for 29 min intervals (N = 28). As was the case for mayflies, the rose petal was again found to be practically unattractive. Consequently, one can conclude that the rose petal replicated photovoltaic light-harvesting layer greatly reduced attractiveness of the two representative indicator species (mayflies and horseflies) tested, and therefore, exerts no significant polarized light pollution. This unattractiveness was caused by the reflection-polarization characteristics of the rose petal replicated test surface, which are distinctly different from those of a planar top layer. The differences in degree of polarization are rather small, therefore differences in the attractiveness may be driven mainly by the angle of polarization. We emphasize, however, that the reactions of polarization-sensitive insects are affected by both degree and angle of polarization. Whereas a planar interface produces exclusively (partial) horizontal polarization in reflected sky/sunlight, the rose petal replica only exhibits horizontal polarization if viewed with the sun shining from in front of the observer (see Figs 2 and 4). Since an approaching insect will always observe a targeted surface from multiple viewing angles and can track the change in polarization direction during its flight, a surface that only produces horizontal polarization for a narrow range of viewing directions is unlikely to be confused as a body of water [22, 23].
The Friedman test was again highly significant (p < 0.0001, Friedman chi2 = 18.489, df = 2, Kendall’s W = 0.3796). According to the Wilcoxon signed-rank test, the horsefly attractiveness of the rose petal (RP) was significantly smaller than that of the glass-covered rose petal (GRP) and smooth black plastic (SBP), but the attractiveness of GRP and SBP did not differ significantly (RP versus SBP: p = 0.0035, RP versus GRP: p = 0.00028, SBP versus GRP: p = 1, N = 28). In our field experiments we used the glass-covered rose petal as an important control test surface for the bare rose petal. The former has the same substrate as the latter but with a smooth and thus strongly polarizing glass covering. The reflective (glass-covered) treatments tend to have a higher intensity, yet still attracted more insects (given that horseflies in particular tend to be attracted by darker objects), implying that polarization and not intensity was responsible for the different reactions of the studied polarotactic insects.
To the best of our knowledge, the polarized light pollution caused by a photovoltaic light-harvesting layer has so far only been investigated for the case of antireflective solar glass incorporating nanopores [31]. Although the attractiveness to horseflies could thereby be reduced, mayflies were actually significantly more attracted to the solar glass compared to a planar reference cover layer. We note that the respective experimental site was identical to our field studies with mayflies. The microtextured layer investigated herein therefore exhibits the novel property of minimizing polarized light pollution at least for the studied mayflies and horseflies. Comparing the degree and angle of polarization patterns, we suspect that the reason for this broadband applicability compared to the antireflective layer studied in reference [31] can be attributed to the fact that the rose petal microtexture reduces the degree of polarization of reflected light with much greater extents than the nanoporous antireflective solar glass, depending on the angle of reflection. Thus, the rose petal treatment was much less attractive to water-seeking mayflies and water- or host-seeking horseflies under the studied illumination conditions.
The horsefly attractiveness of glass-covered rose petal did not differ significantly from that of smooth black plastic. On the other hand, the mayfly attractiveness of glass-covered rose petal was significantly smaller than that of smooth black plastic. The reasons for this attractiveness difference may be the different illumination conditions in the field experiments and the species-dependent reactions of horseflies and mayflies to these two different black planar test surfaces.
It was experimentally shown that the disordered microtexture replicated from rose petals greatly reduced polarized light pollution. By comparison of its reflection-polarization characteristics with those of a microlens array foil (see Fig 2), we also concluded that such dense arrays of micron-sized textural elements lead to a distinct pattern in the angle of polarization for different observer viewing directions, irrespective of the exact curvature of the individual building blocks. In what follows, we analyze more systematically the influence of the texture topography on polarized light pollution to derive general guidelines for the design of photovoltaic cover layers. To this end, a ray-tracing based numerical assessment of the polarized light pollution (depending on the angle of incidence) caused by microcone arrays of varying aspect ratio as well as varying degrees of both height and positional disorder (see Fig 7C–7E) were carried out according to the methodology described in section 2.7.
PPT
PowerPoint slide
PNG
larger image
TIFF
original image
Download:
Fig 7.
Illustration of a ray-tracing-based simulation approach for studying the properties of light reflected from densely packed and disordered microtextures.A) Definition of the spherical coordinate system for characterizing the observer’s viewing directions in the farfield. The planar rectangular and transparent light source is marked by an orange rectangle. It emitted 108 parallel and unpolarized rays of identical initial power with random starting positions into the central part of the microtexture models. B) At every observer viewing direction (φ, θ), the incoming rays’ (exemplary ray paths are drawn as faint orange lines) propagation direction k defines the z-axis of a local coordinate system that is used for measuring the angle of polarization (noted AOP in the figure). The y-axis is chosen parallel to the local φ = constant line. The angle of polarization of the electric field vector of light represented by an orange double-headed arrow is then always measured relative to the local x-axis (mathematically positive). The array of parallel orange lines under the orange rectangle represents parallel sun rays arriving from an intermediate elevation near an azimuth of 180°. C)-E) depict exemplary microtexture models (cone’s aspect ratio = 0.6, full tiling of the base) illustrating three extreme cases, namely the disorder-free configuration (C), the maximum degree of height disorder (D) and the maximum disorder in the cones’ horizontal position (E).
https://doi.org/10.1371/journal.pone.0243296.g007
As reported in [29], surfaces reflecting light with a degree of polarization d > 15% and a polarization direction deviating by maximum ±10° from the horizontal cause a maladaptive attractiveness to mayflies and horseflies because of misidentification as a water surface. These thresholds can be exploited to quantify the polarized light pollution (and its dependence on several structural parameters) that is caused by the reflection of direct sunlight from microtextured surfaces. We assume that the textured surface is placed horizontally on the ground so that it can be observed from all positions on the 2π hemisphere (see Fig 7A and 7B). Furthermore, on the basis of numerous field experiments [e.g. 20, 21, 27–29, 31, 34], we assume that none of the possible observer viewing directions is favoured by the insects for spotting water surfaces. From the farfield distributions of reflected light intensity, degree and angle of polarization, we then calculated three numbers for characterizing the polarized light pollution:
With a focus on aspect ratio = 0.3, 0.6 (close to the average aspect ratio of rose petal epidermal cells [9]) and 1, the results of these computations are displayed in Fig 8. We found that for the whole ranges of the aspect ratio and the angle of incidence considered, it can first be concluded that both disorder in cone positioning and disorder in cone height lead to a larger portion of the observer hemisphere being hit by reflected light. This is especially pronounced for disorder in cone height, since the slight variations in local angle of incidence for neighbouring cones leads to a peak broadening effect in the reflected light farfield intensity (see S2–S4 Figs and orange curves in Fig 8A–8C). Except for the cases of (at least close to) perfectly ordered and cones with low aspect ratio, the microtextured surfaces are able to spread the incident parallel light over most of farfield viewing directions (φ, θ). Considering the portion of the hemisphere from which the reflecting surface could be wrongfully detected as water, an analogous statement for the influence of disorder can not be made. However, we note that (with the exception of aspect ratio = 0.3 at angle of incidence = 0°) this portion is smaller than 10%, and even decreases below 1% for high angle of incidences. The calculated fractions of possible farfield viewing directions (φ, θ) that lead to wrongful detection as a water surface relative to the whole hemisphere are summarized in Fig 8D. The shaded area in Fig 8D indicates the full range of numerical results that were found when ramping up both height and position disorder. For the entire ranges of the aspect ratio, angle of incidence and disorder type and amounts that were investigated herein, the conditions for a misdetection as water are only met for a very limited range on the observer hemisphere: according to Fig 8D, across all angle of incidences, less than 10% of the 2π hemisphere (a solid angle withe area smaller than 0.2π) would fool the insects. This remarkable robustness leads to the conclusion that the ability of microtextured cover layers, including polymeric petal replicas, to drastically diminish polarized light pollution is not strongly susceptible to changes in illumination conditions as well as to variations of geometrical parameters like the aspect ratio or the type and degree of structural disorder.
PPT
PowerPoint slide
PNG
larger image
TIFF
original image
Download:
Fig 8.
Numerical assessment of the influence of structural disorder on the polarized light pollution caused by microtextured surfaces.For aspect ratio AR = 0.3 (A), 0.6 (B), and 1 (C), the two characteristic solid angle ratios Sillum/2π and Sattract/Sillum are displayed as a function of the angle of incidence AoI. D) Depicts the solid angle fraction Sattract/2π versus AoI for which horseflies and mayflies would detect the reflecting surface as a water surface with respect to the entire hemisphere of possible observer viewing directions for aspect ratio AR = 0.6. The shaded area in panel D indicates the full range of numerical results that were found when ramping up both height and position disorder.
https://doi.org/10.1371/journal.pone.0243296.g008
For the design of multifunctional microtextured photovoltaic cover layers, the target of causing minimal polarized light pollution therefore only constitutes a weak constraint on the design parameters. With a clear advantage over nanoporous antireflective light-harvesting solar glass layers [31] in terms of reducing polarized light pollution, our microtextured rose petal imitating surfaces therefore seem to be a promising pathway towards, at the same time, achieving outstanding light-harvesting properties [9–12] as well as introducing a self-cleaning scheme [11] and diminishing maladaptive attractiveness to polarotactic aquatic insects.
If you have any questions on skylight solar panel. We will give the professional answers to your questions.