Optimizing Color Saturation in Colloidal Photonic Crystals by Control of Absorber Amount and Distribution

Nanostructured materials that mimic structural coloration in nature can be synthetically created by colloidal self‐assembly. To maximize optical effects, the natural world integrates melanin as a broadband absorber to remove incoherently scattered light. Polydopamine (PDA) is used as a synthetic analog of natural melanin to systematically investigate the influence of absorber quantity and distribution on color saturation in colloidal photonic crystals. The absorbing PDA is integrated into two distinct ways: homogenous colloidal crystals are produced from core–shell particles with incrementally increasing polydopamine shells, and heterogeneous colloidal crystals are formed by co‐assembly of varying ratios of polystyrene (PS) and absorbing PS@PDA particles. The chromaticity is quantified by converting the measured spectra to reconstructed colors in the L*a*b* color sphere to identify structures with optimal color saturation. Simulations based on the discrete dipole approximation (DDA) indicate that a homogeneous absorber distribution is most efficient in creating saturated structural coloration. Experiments, however, demonstrate that the heterogeneous absorber incorporation outperforms the homogeneous strategies, as it allows for a more precise adjustment of the absorber content in the required concentration range. These results underline the importance of incorporating absorbers and indicate efficient ways in which colloidal photonic crystals with saturated structural colors can be prepared.


Introduction
The animal kingdom is replete with striking and vibrant colors.These colors arise from different physical principles.Absorption pigments exploit the photophysical properties of dye molecules, in which electronic transitions cause absorption at DOI: 10.1002/admi.202300986certain wavelengths of the visible spectrum so that the complementary color is observed. [1,2]Common pigments in the animal kingdom are melanin, [3] which also colors the skin and hair of humans, [4] or carotenoids, [5] which are responsible for the pink plumage of flamingos [6] or the yellow or red coloration of canaries. [7,8]Color can also be created by the interaction of visible light with defined nanostructured materials with dimensions in the range of the wavelength of visible light.[30][31] If visible light interacts with such a structure, it is either transmitted or reflected at each of these elements.][34][35] Most often in the natural world, such nanostructures are combined with absorbing pigments, in particular the broadband absorbing melanin. [3]The peacock with its unique colorful tail feathers, caused by rodshaped melanosomes, [24,25,[36][37][38][39][40] the morpho butterfly with its distinctive Christmas tree structure [17,41,42] or the iridescent feathers Copyright Glenn Bartley.b) Amelanotic white Stellar's Jay, which lacks melanin granules.Reproduced with permission. [47]Copyright 2006, Company of Biologists (Photograph by Bill Schmoker).c) Vertical deposition method to assemble PS@PDA core-shell particles into a 3D PC. d) Two strategies to incorporate the absorbing PS@PDA particles.I: Homogeneous 3D PCs consisting of a single type of particle, wherein the layer thickness of PDA progressively increases, resulting in enhanced absorbing properties; II: Heterogeneous 3D PCs formed by mixing bare PS with coated PS@PDA particles.The amount of absorber is increased by introducing more PS@PDA particles into the system. of a hummingbird [43][44][45] are prominent examples of this efficient combination of structural coloration with absorbing melanin.
Structural imperfections or defects, which inevitably arise in naturally grown systems, cause incoherent light scattering that results in a spatial redistribution of white light.The maxima in reflection arising from interference effects are thus superimposed with a background of stray light, causing less intense, whitish observable colors.The addition of broadband absorbers removes part of this stray light and thus enhances the contrast between the photonic stop band and background.As a result, saturated reflection colors with high chromaticity are observed. [46]he significance of melanin in creating saturated structural colors is beautifully illustrated by the plumage of a normal and an amelanotic Stellar's Jay [47] (Figure 1a,b).The former incorporates a basal melanin layer, which is not present in the amelanotic individual and likely carries a mutation affecting melanin synthesis or deposition.The blue coloration of the Stellar's Jay is completely lost in the amelanotic, which exhibits a white, colorless plumage, despite the similar nanostructural architecture of the feathers. [10,47]he importance of absorbing materials to enhance visible structural coloration has been identified in artificial structurally colored materials, [10,48] in particular in self-assembling systems formed from colloidal building blocks, where defects and irregularities cannot be completely prevented.Typically, such absorbers are incorporated in the form of carbon black, [49][50][51] which is added in small quantities to increase color saturation.[51][52][53] To mitigate this limitation, absorbing building blocks that can directly form colloidal crystals are required.Poly-dopamine (PDA) and polystyrene-polydopamine core-shell particles (PS@PDA) have been identified as an ideal system for this purpose because they can be synthesized with high uniformity and controlled absorption properties. [46,54,55]PDA has structural similarities with natural melanin and is thus known as synthetic melanin. [46]The combination of polystyrene (PS) as a core with PDA as a shell is especially advantageous.Polystyrene colloidal particles can be synthesized with high accuracy and low polydispersity, providing control of the particle size.The PDA shells can be grown on these particles with high precision, [46,54,55] allowing for the design of well-controlled building blocks to form colloidal photonic crystals.Such PS@PDA core-shell particles have been used for the development of structurally-colored colloidal crystals with much improved optical properties, [52,[56][57][58][59][60][61] in analogy to the use of melanin in the natural world.
The ability to directly encode absorbing properties at the primary particle levels also opens the possibility to systematically explore structure-property relations and address fundamental questions on the design of structurally colored materials.In particular, while synthetic and natural examples clearly underline the need to incorporate absorbing material, it remains unclear how much absorbing material creates the most saturated color and how the absorber is most efficiently integrated within the structurally colored material.In this article, we address this question by fabricating colloidal photonic crystals from building blocks with defined absorption properties and then systematically analyzing, quantifying, and comparing the resultant optical properties as a function of absorber content and the method of absorber incorporation.
To this end, we synthesize such PS@PDA core-shell particles and use the vertical convective assembly method [62][63][64] to create colloidal 3D photonic crystals (3D PCs), as illustrated in Figure 1c.The precision with which these particles can be synthesized allows us to isolate the effect of absorber concentration by generating building blocks with different absorbing properties (encoded by the PDA shell thickness) while maintaining comparable particle sizes and thus spectral positions of the photonic stop band.
We integrate the absorbing materials in two distinct ways with the goal of exploring the most efficient way of incorporating the absorber (Figure 1d).In the first method (Figure 1dI), we produce homogeneous 3D PCs comprising a single batch of PS@PDA particles, as commonly used in literature. [46,54,55]In this case, the absorbing PDA is uniformly distributed throughout the photonic structure, similar to the natural examples of rodshaped melanosomes [24,25,[36][37][38][39][40] in peacocks, or melanin platelets in hummingbirds. [43]The PDA content within these crystals is incrementally increased by adjusting the PDA-shell thickness while maintaining a constant total diameter.In the second approach (Figure 1dII), plain PS and PS@PDA particles with nearidentical sizes are mixed to create heterogeneous 3D PCs.In this case, the increase of absorber is achieved by varying the mixing ratio of both particle types.[52][53]65] While both, homogeneous and heterogeneous incorporation of absorber have been shown to increase color saturation, [46,[49][50][51][52][53][54][55] the most efficient way of absorber addition, and the ideal absorber concentration to yield the most saturated structurally colored materials have yet to be established.

Synthesis
We first synthesized the required colloidal building blocks to assemble the 3D PCs with controlled absorber content.Using surfactant-free emulsion polymerization, [66] we produced a set of PS particles with different sizes and low sizedispersity (Figures 2a, S1, Supporting Information).These particles were subsequently coated with a PDA shell of controlled thickness. [54,56,67]The formed shells were uniform and provided core-shell particles with a narrow size distribution (Figure 2b).Using dopamine (DA) concentrations between 0.25 g L −1 to 1 g L −1 enabled us to vary the PDA-shell thickness from 2 nm to 15 nm (Figure 2c), and thus to control both, the diameter, and the absorber content.Notably, the synthesis produced uniform and reproducible shell thicknesses with variations in the range of a few nanometers, providing the required control of particle diameters to form 3D PCs.Photographs of representative dispersions show the pronounced change in the optical properties (Figure 2c).The dispersion of plain PS particles is white due to scattering, while dispersions of PS@PDA particles appear brownish and rapidly become darker as the shell thickness increases, as a consequence of the increased broadband light absorption of the PDA shell.The zeta potential of these dispersions remained approximately constant independent of the PDA-layer thickness (Figure S2, Supporting Information).
We adjusted the size of the PS core particles in a way that all PS@PDA particles had similar overall sizes of d ≈290 nm so that the spectral properties of 3D PCs formed from the different building blocks were comparable.We chose this particle size because 3D PCs formed from such particles display a photonic stop band in the red part of the visible spectrum, at around 600 nm.Although PDA is a broadband absorber, the absorbing properties decrease with increasing wavelength, [59] so introducing PDA into red structurally colored 3D PCs should exhibit a weaker suppression of the main reflection peak while more efficiently absorbing Rayleigh scattering in the blue region, resulting in a saturated red color.

Microscopic Investigations
We first prepared homogeneous 3D PCs with controlled absorber content (Figure 1) using core-shell particles with a comparable total diameter (d≈284 nm) and PDA-shells with thicknesses Figure 3. Microscopic optical properties of homogeneous 3D PCs made from PS@PDA particles with different PDA-shell thicknesses.a-e) From left to right: Macrophotographs, light microscopy and SEM images of 3D PCs prepared from PS@PDA with increasing shell thickness as specified in the cartoon.f-h) Reflectance spectra taken by correlative micro-spectroscopy of all homogeneous 3D PCs with identical thickness, shown for 5,7, and 9 layers.between 2.5 nm and 13 nm (Figure 2). Figure 3a-e (left) shows photographs of the 3D PCs prepared by convective self-assembly of these colloidal building blocks. [46,54,55]The 3D PC formed from bare PS colloidal particles exhibited the most uniform appearance and evenly covered the entire substrate.In contrast, 3D PCs made from PS@PDA colloidal particles covered the substrates less evenly and left void spaces.This difference indicates a lower propensity of the PDA-coated particles to self-assemble, which increases with increasing PDA layer thickness in agreement with reports in the literature, [57] possibly due to reduced colloidal stability.However, it is noteworthy that all particles formed ordered 3D PCs in substrate regions covered with particles, contrasting reports in the literature where PS@PDA particles with larger PDA layer thicknesses tended to form colloidal glasses without long-range order. [57]As expected, the coloration of the 3D PCs became progressively darker with increasing PDA layer thickness due to the increased light absorption by the thicker PDA-shells.Taking advantage of the ability to create ordered 3D PCs for all particle types, we systematically compared the optical properties as a function of absorber content in otherwise similar structures.To quantify the changes in optical appearance, we used correlative microscopy to measure spectral properties in 3D PCs with identical numbers of layers.Correlative microscopy enables the acquisition of microscope images, spectra, and scanning electron microscopy (SEM) images at identical locations, and thus links the optical effects with the structural properties of the 3D PCs.We exemplarily identified areas of 3D PCs formed from seven stacked layers by visually following the steps formed by increasing layers at the initial crystal growth front using a method described in a previous publication. [68]In brief, changes in the crystal thickness related to a step-wise increase in the number of layers can be accurately traced in an optical microscope via a visible change in reflectance color (see microscopy images in Figure 3), and confirmed by the correlative SEM image.By simply counting the number of steps starting at the boundary of the formed colloidal crystal, we can precisely determine the number of layers at the desired measurement region.When measuring all samples at the same number of layers, the microscopic images and spectra can be compared even if the 3D PCs vary in coverage at a macroscopic level.The correlative light microscopy images of an area with 7 layers in the 3D PC revealed characteristic microscopic cracks and defects that were not visible in the macroscopic photographs but caused incoherent scattering that reduces color saturation.Similar to the macroscopic color appearance, the microscopic color transition from orange to increasingly darker red can be observed as the layer thickness of PDA increases.However, this shift was much less pronounced compared to the macroscopic appearance.In particular, the whitish macroscopic color of the 3D PC formed from bare PS particles is not as apparent in microscopy, as scattered stray light is not effectively collected by the objective lens with a limited numerical aperture (NA = 0.5).These differences underscore the importance of utilizing multiple characterization methods to accurately analyze the structural and spectral properties of self-assembled structurally colored materials.
Comprehensive SEM investigations of regions with 7 layers (Figure 3), 5 layers, and 9 layers (Figures S3-S5, Supporting Information) showed well-ordered particle arrangements with large cracks between individual crystal domains for all PS@PDA particles.The correlative microscopy approach thus enables us to isolate the effect of absorbers on the optical properties because all spectra are recorded at 3D PCs with identical thicknesses and comparable structural uniformity.Figure 3f-h exemplarily shows reflectance spectra for samples consisting of 5, 7, and 9 layers of the colloidal crystals (30 × 30 μm spot size).These spectra show a prominent photonic stop band in the red part of the spectrum ( max ≈ 610 nm) and additional Fabry-Perot resonances. [69]The spectral position of the photonic stop band slightly varies between the individual samples as a result of small differences in the average diameters of the particles with different PDA content in the crystal packing (Figure S6, Supporting Information).With increasing layer thickness, the intensity of the photonic stop band increases. [68]When comparing samples with different absorber content (i.e., increasing PDA-shell thickness), two general trends are observed.First, the intensity of the photonic stop band decreases with increasing absorber content.Concomitantly, the intensity of the reflected light in other spectral regions decreases as well.This decrease is most pronounced at small wavelengths ( < 500 nm), where the reflected light intensity decreases from around 6 % for the sample with bare PS particles to approx. 2 % for the samples with high PDA content.This microscopic investigation shows that the optical properties of the samples can be accurately compared, i.e., that the structural quality of the different 3D PCs is comparable.As the reduction of the photonic stop band with increasing absorber content is pronounced while the reduction of reflected light at other spectral regions seems subtle, one would intuitively assume that the optical properties of the bare PS crystal would provide the most colorful sample.However, this interpretation does not give justice to the whitish appearance, underlining the need for a macroscopic characteri-zation, where scattering effects are more thoroughly considered, which we will discuss below.
We next prepared heterogeneous 3D PCs where the absorbing PS@PDA particles were mixed with bare PS particles to control the total amount of absorber (Figure 1).We used the PS@PDA particles with the largest shell thickness (t≈14 nm, total diameter d≈291 nm) and produced heterogeneous 3D PCs by mixing this particle population in different ratios with plain PS particles of comparable diameter (d≈292 nm) (Figure 4a-e).Compared to the homogeneous case, the heterogeneous 3D PCs showed a more uniform coverage over a much larger absorber concentration range, likely caused by the presence of colloidally stable, bare PS particles in the dispersion.Only for very large amounts of PS@PDA particles (above 50 %), an uneven substrate coverage within the colloidal crystal was observed (Figure 4d,e).The macroscopic color of the formed 3D PCs transitioned from a slightly whitish appearance of the bare PS 3D PC via an intense red color to a darker appearance at high absorber concentration.Similar to the homogeneous 3D PCs, we used correlative microspectroscopy to quantify the optical properties as a function of the absorber content.Light microscope images in areas identified to have 7 layers of particles (Figure 4a-e), showed a color transition toward darker red with increasing absorber content.As above, the whitish appearance of the bare particles was not visible in the microscopic images due to the limited angular collection of light.The SEM images (Figure 4, Figures S7-S9, Supporting Information) of the different colloidal crystals showed a well-ordered selfassembled structure for all compositions, underlining the high precision with which the diameter of both particle populations was matched.Notably, with increasing PDA content, the density of cracks seemingly increased, indicating the decreased propensity of PS@PDA particles to self-assemble.The different particle populations can be distinguished by their contrast in SEM images, where PS@PDA particles appear lighter.The two particles in these binary crystals appear to form small clusters, suggesting a partial phase separation.However, such patterns, including the clusters can be reproduced by random mixing of two populations of individual components (Figure S10, Supporting Information).In addition, isolated particles of each species can be clearly identified in the SEM images.These observations suggest that the particles formed mixed crystals with randomized positions.
The comparable structural properties again enabled us to accurately compare the spectral properties of the 3D PCs formed with different PS@PDA content at identical layer thicknesses using correlative micro-spectroscopy, and thus, to isolate the effect of absorber content.Figure 4f-h shows microscopic reflectance spectra of samples with 5, 7, and 9 layers of the colloidal crystals (30×30 μm spot size).Similar to the homogeneous crystals, the intensity of the photonic stop band increased with increasing layer thickness.In the heterogeneous mixtures, however, the photonic stop bands remained nearly at the same spectral position, since all samples were composed of identical particle populations.The two general trends observed for the homogeneous 3D PCs were reproduced in the heterogeneous case.A pronounced decrease in the intensity of the photonic stop band was observed with an increasing amount of absorbing PS@PDA particles.In addition, the reflection at other spectral positions, most notably at lower wavelengths, was also reduced with increasing PS@PDA particles.PCs prepared from mixed colloidal dispersions of PS (d≈292 nm total particle size) and PS@PDA (t≈14 nm shell thickness; d≈291 nm total particle size) particles, with increasing amount of the PS@PDA as specified in the cartoon.f-h) Reflectance spectra taken by correlative micro-spectroscopy of all heterogeneous 3D PCs with increasing ratios of PS@PDA core-shell particles measured at PC areas with identical thickness, shown for 5, 7, and 9 layers.

Macroscopic Investigations
Next, we investigated the macroscopic optical properties of both types of colloidal crystals with controlled absorber composition.We prepared thick sample specimens with large (yet undefined) numbers of colloidal crystal layers to maximize the reflected light intensity.To this end, we cast the respective colloidal dispersions into preformed PDMS molds bonded to a glass substrate and dried the dispersions within these molds.Figure 5a,d shows photographs of the resultant 3D PC structures in the PDMS mold for homogeneous and heterogeneous samples, respectively (more details in Figures S11 and S12, Supporting Information).In both cases, the whitish appearance of the bare PS 3D PC is clearly visible.increasing absorber content, the observable structural coloration first becomes brighter and then subsequently darkens.Notably, for the homogeneous 3D PCs (Figure 5a), this darkening is already pronounced for the sample with a shell thickness of 4.5 nm and only the sample with the smallest shell thickness (2.5 nm) shows a bright, red color.In contrast, the heterogeneous 3D PCs show a much more gradual transition to more intense, and subsequently darkening color, as the absorber can be added in smaller steps and thus the overall content more accurately controlled.Noteworthily, even in samples with comparable absorber contents, e.g., in the homogeneous samples made with 7 nm shell thickness (PDA vol.ratio ≈14 %) and the heterogeneous sample made at a 50 % ratio (PDA vol.ratio ≈13 %), the color of the latter sample appears more saturated.
To quantify the suppression of incoherently scattered light by the absorbing PDA content, we used a UV-VIS spectrometer with an integrating sphere to collect the reflected light as well as the scattered light over all angles.Figure 5e,f shows the recorded and heterogeneous (bottom row) 3D PCs formed with increasing absorber content, as specified in the cartoons.b,c) Abstracted colors perceived by the human eye of the different samples, created with the RGB values that were determined from the reflectance spectra using the CIE standard illuminant D65, shown for the homogeneous.b) and heterogeneous.c) 3D PCs.The heterogeneous 3D PCs were made from 292 nm PS particles mixed with 291 nm PS@PDA nanoparticles with a shell thickness of t≈14 nm.e,f) Macroscopic reflectance spectra using an integrating sphere, for homogeneous (increasing PDA shell thickness).e) and heterogeneous (decreasing PS ratio).f) 3D PCs with varying absorber content.g) Quantification of spectroscopic data showing the intensity of the photonic stop band, the intensity of scattered light at 400 nm, and the difference between the two for the heterogeneous 3D PCs.
spectra for homogeneous and heterogeneous 3D PCs with different PDA content, respectively.
For the homogeneous 3D PCs, a rapid decrease of the photonic stop band intensity with increasing absorber content was observed, especially evident between the sample with 2.5 nm layer thickness and the one with 4.5 nm, which was much more pronounced compared to the microscopic analysis using identical layers.This may be caused by the increased layer thickness compared to the previous samples and a reduction in the quality of the colloidal crystals, such as increasing crack density for the PS@PDA core-shell particles, corroborating the impression of the samples in Figure 3.In addition, the effect of incoherently scattered light, only slightly observable in the microspectroscopic analysis, was pronounced in the macroscopic measurements.While the 3D PC made from bare PS particles showed large intensities of scattered light, especially at wavelengths below the photonic stop band, this scattering was already effectively suppressed for the sample with the smallest PDAshell.
The heterogeneous 3D PCs showed a more uniform reduction of the intensity of the photonic stop band with increasing amount of absorbing PS@PDA particles, indicating that it is more efficient to incorporate smaller amounts of PDA using this method.Notably, the intensity of incoherently scattered light was already much suppressed for samples with small PDA content, evidenced by the reduction of reflected light intensities below the photonic stop band.In Figure 5g, we quantify this different effect of the absorber on the intensity of the photonic stop band (I phot , intensity of peak) and the intensity of incoherently scattered light (I scat ), measured at a wavelength of 400 nm.Overall, samples with absorber particle contents below approx.15 % (85 % PS and more) showed an efficient reduction in scattering while the photonic stop bands retained a high intensity.This is clearly seen by the difference I phot -I scat , which reached its maximum value at these absorber concentrations (blue triangles in Figure 5g) (results for the homogeneous system can be seen in Figure S13, Supporting Information).
While the consideration of I phot -I scat is a useful measure to track the evolution of scattering and photonic stop bands, it does not necessarily represent how these spectral properties translate into human color perception.To investigate this perceived color impression, the measured reflectance spectra can be translated to L*a*b* values, which define a color sphere that characterizes the color based on its lightness (L*), ranging from 0 to 100, represented along the vertical axis of the sphere.Additionally, the contributions of green versus red (a*) and blue versus yellow (b*) to the color are described by the a* and b* values.While both axes have no defined end, for simplicity we limit them to ±100 as the resulting sphere is sufficient to display and track the color within this work (Figure 6a).Based on this formalism, we converted the measured spectra of the macroscopic samples into reconstructed colors (Figure 5b,c). [70]These reconstructed colors closely match the qualitative impression of the photographs, underlining the accuracy of the approach.This, in turn, allows us to quantitatively analyze and compare the color of the different samples.
Figure 6b,c shows the position of homogeneous (Figure 6b) and heterogeneous (Figure 6c) samples on the L*a*b* color sphere.The progression of the observed color in both distributions was also apparent here.The samples showed an increased darkening as they traversed the vertical axis of the sphere from higher L*-values to lower L*-values.Both types of samples also showed an increase in the content of the color components, evidenced by an increasing horizontal distance from the origin in the graphs.The heterogeneous samples showed a more gradual transition, reflecting the more uniform spectral changes.In the following, we quantified these changing positions on the color sphere via the chromaticity C*.The chromaticity describes the horizontal distance of the color to the origin of the color sphere and can therefore be used as a means to describe the perceived saturation of the color.It is computed from the L*a*b* values using Formula 1.With a further increase in PDA-shell thickness, the chromaticity rapidly decreased as the samples became progressively darker.Figure 6e shows the chromaticity of heterogeneous 3D PCs.Again, the PS sample showed a small chromaticity.Upon inclusion of absorbing PS@PDA particles, the chromaticity gradually increased until it reached a peak at approx.20% PS@PDA particles, after which it began to decline.This strategy thus provides a quantitative measure that accurately captures the macroscopic optical impression of the photographs in Figure 5.In Figure 6f, we compared both samples by calculating the volumetric PDA fraction of each particle.In both types of samples, the maximum chromaticity occurred at a volumetric proportion of PDA equivalent to approx.5 %.
Our optical analyses underline that it is challenging to accurately evaluate color properties based solely on measured spectra without the additional translation to perceived color.This is because the entire off-peak spectrum, in addition to the peak intensity, plays a critical role in determining the macroscopically observed color.This difference in evaluation can be seen by the discrepancy between the optimal absorber component determined from the intensity ratio of the photonic stop band and scattering at 400 nm (Figure 5g), and from the L*a*b* formalism (Figure 6).Color optimization must therefore be performed in a more encompassing way, using, e.g., parameters derived from the L*a*b* color space.Moreover, multiple analytical techniques are required for a coherent analysis.Microscopic examinations allow accurate comparison between samples but provide only a limited estimate of the color, as the incoherently scattered light is not well collected by the objective lens.Macroscopic examinations, on the other hand, allow for clear color observation and comparison but fall short on the exact comparison between samples, because the measurement is not performed at identical thicknesses.

Simulations of Photonic Properties
We numerically investigated the influence of absorbing moieties on the resultant structural coloration to corroborate the experimental results, address the uncertainties created by experimental limitations, and thus clarify the effect of absorber distribution within the 3D PCs.To this end, we used the discrete dipole approximation (DDA) [71,72] and set up a close-packed colloidal crystal slab with an ideal face-centered cubic (fcc) structure.The colloidal crystal consisted of seven layers of particles, each layer comprising ca. 10 × 10 particles, resulting in a total of 627 particles within the crystal (Figure 7a).Each particle was formed by 1762 active dipoles, resulting in a total number of 1.7622 × 10 6 dipoles (1 104 890 active and 657 310 inactive dipoles) in the simulation cell.The crystal consisted of two types of particles.Pure scattering particles with a real part of the refractive index n = 1.59 and an imaginary part k = 0, mimicking the PS particles in the experimental system.Absorbing properties, mimicking the PS@PDA particles in the experiment, were introduced by increasing the (wavelength-dependent) k-value.For the homogeneous system, where each particle had a small, but uniform contribution to absorption, k was set as the k-value of PDA [59] and was controlled by a preceding constant c, which can vary from 0 % (representing only PS particles) to 100% (indicating particles with the thickest PDA-layer).For the heterogeneous system, a fraction of the individual particles were assigned absorbing properties matching those of the experimental particles with maximal absorption properties (Figure 7a, FigureS14, Supporting Information).Note that the core-shell nature of the particles in the experimental system was not resolved in the numerical scheme to reduce computational effort.Instead, the absorber content was assumed to be homogeneous within the particle and the dielectric properties were calculated as a homogeneous medium based on the volume ratio of PS and PDA in the experimental particles.This ensured that each particle could be modeled with a minimal number of dipoles and thus allowed maximizing the system size that could be addressed in simulation.In the simulation cell, the incident light propagates in a negative z direction (Figure 7b) and impinges perpendicular to the 3D PC.The scattered light is collected by a monitor plane on the imaginary observation sphere with adjustable dimensions.In our case, the opening angle of the monitor was set to 180°, mimicking the total reflectance collected by the integrating sphere (Figures 5 and 6). Figure 7c exemplarily shows the simulated reflectance spectra for the 3D PCs with heterogeneous absorber distribution.Similar to the experimental case, the main peak of the bare PS colloidal crystal exhibited the highest intensity, which gradually decreased as the number of absorbing particles increased.Concomitantly, incoherently scattered light that is captured in spectral regions of the photonic stop band, was most pronounced for the 3D PC composed solely of PS and increasingly suppressed with the addition of absorbing particles.These results qualitatively reproduced the experimental behavior of the heterogeneous 3D PCs.Correspondingly, the comparison between the peak intensities of the photonic stop band (I phot ; taken at  = 635-660 nm) and the incoherently scattered light (I scat ; taken at  = 400 nm) showed a comparable behavior, with a maximum value I phot -I scat at 5-10 % absorbing particles.Similar results were found for the homogeneous system (Figure S15, Supporting Information).
We next compared the simulated spectra with the experimental data (Figure 8).The differences in system size between experiment and simulation prevent a direct comparison.We therefore matched the intensity of the photonic stop band for the sample of bare PS as an internal reference.The ob-tained spectra from simulations and experiments with different absorber concentrations qualitatively agreed.Notable differences in peak width (larger in simulation) and Fabry-Perot resonances (only occurring in simulation) can be assigned to the different setups of the experimental and numerical system.The simulated 3D PCs consisted of only 7 layers of particles, whereas a significantly higher number of layers were present in the UV-VIS experiments, accounting for the decreased peak width.In addition, while the simulated 3D PCs had a uniform thickness, and thus supported Fabry-Perot resonances, the sample in the experiment was heterogeneous in thickness.Note that the sample measured with micro-spectroscopy on uniform sample thicknesses (Figure 4) exhibited similar Fabry-Perot resonances.
The chromaticity of the simulated heterogeneous 3D PCs, extracted from the simulated spectra via the L*a*b* values, exhibited a similar trend as the experimental case, but their dependence on the absorber content was less pronounced compared to that of the experimental 3D PCs (Figure 8b).This disparity can be attributed to the limited number of particles and the regular nature without defects used in the simulations.Since incoherent scattering predominantly arises from the interaction of light with defects and irregularities, the impact of absorbing properties was more pronounced in the experimental system, where cracks and imperfect stacking cannot be avoided.However, both chromaticity curves showed an optimal absorber concentration in the range of 20-25 % absorbing particles.
We used the regular and uniform simulated 3D PCs to quantitatively compare the optical properties of 3D PCs with homogeneous and heterogeneous absorber distribution (Figure 8c).Both types of simulated samples had qualitatively similar optical properties, with matching peak shapes, widths, and Fabry-Perot resonances.However, there were differences in peak intensity.The intensity of the photonic stop band of the heterogeneous 3D PCs was systematically smaller than that of the homogeneous distribution, with a simultaneous stronger suppression of the incoherently scattered light.This difference in absorption properties translates to the color properties of the samples, assessed via their chromaticity values.Figure 8d shows that the homogeneous 3D PCs in simulations exhibited a higher chromaticity in the maximum region of the curve.This behavior indicates that a homogeneous absorber distribution is more efficient compared to the heterogeneous case.

Discussion
Our simulations indicate that a homogeneous absorber distribution via particles coated with a thin absorbing shell results in more intense colors when compared to a heterogeneous distribution of the absorbing material.We attribute the higher intensity of the photonic stop band in the homogeneous case to the higher regularity and thus higher quality of the photonic crystal.In the heterogeneous case, the random mixing of particles with slightly different optical properties (changing k values) alters the periodicity and thus broadens the photonic stop band.These predictions from simulations, however, are not reflected in the experimental realization, where the heterogeneous case provided samples with generally higher chromaticity values.In this case, the quality of the 3D PCs formed from homogeneous PS@PDA particles seemed lower, likely due to the lower colloidal stability of the PDA-coated samples.This, in turn, weakens the photonic properties.In the heterogeneous 3D PCs, the high proportion of bare PS particles seemingly improved the order during the self-assembly process and thus produced higher-quality photonic crystals.
Moreover, both experiments and simulations demonstrate that the optimal color intensity is already achieved with small absorber contents.Achieving these targeted compositions in a homogeneous system is experimentally challenging, as they require very thin PDA-shells in the range of single nanometers.The heterogeneous addition of absorbing particles in a mixed crystal, however, offers a straightforward solution to this challenge.The ratio of absorbing particles can be precisely varied, and thus even small amounts can be reliably incorporated with high levels of control.In addition, the mixed colloidal crystals showed higher regularity compared to the colloidal crystals assembled from PS@PDA particles due to the large amount of colloidally stable PS particles, affording more control in tailoring the photonic properties.Together, these experimental considerations make it easier to adjust and optimize the resultant structural color via heterogeneous absorber addition.

Conclusion
The most vivid color effects observed in the natural world are based on structural coloration.As the function of these colors is often to warn and scare predators, or to attract mates, it can be assumed that evolution has optimized their intensity and chromaticity by the incorporation of melanin as a broadband absorber to remove incoherently scattered light.In a bioinspired approach, we fundamentally investigated the role of absorber concentration and distribution within colloidal photonic crystals to form structural coloration with optimal saturation.
We implement absorbing properties via homogeneous 3D PCs formed entirely from PS@PDA particles with defined absorbing shell thicknesses, and by heterogeneous 3D PCs formed by mixing highly absorbing PS@PDA particles with bare PS particles.In simulations, the homogeneous case forms more regular photonic crystals and thus produces higher-quality structural coloration.In experiments, however, heterogeneous 3D PCs outperform the homogeneous strategies as it is easier to self-assemble ordered systems and adjust the absorber content in the required, small concentration range.
PS Particles: Polystyrene colloidal particles were synthesized by surfactant-free emulsion polymerization. [73]In brief, the synthesis was conducted in a 500 mL round-bottom flask containing 240 g water, stirred with a magnetic bar, and heated in an oil bath at 80 °C.After adding purified styrene (6-7 g), acrylic acid (0.1 g) and ammonium persulfate (0.1 g), both dissolved in 5 mL water, were added.The reaction was stopped after 22 h.The particles were cleaned by centrifugation (11 000 rpm for ≈17 min) and redispersion with a vortex shaker and an ultrasonic bath (80 W).
PS@PDA Particles: To a solution of 0.1 wt.-%PS colloidal particles, tris(hydroxymethyl) aminomethane (0.302 g), and dopamine HCl (0.25-1.0 g L −1 ) were added, and the dispersion was stirred for 22 h at room temperature.The particle dispersions were cleaned by centrifugation (11 000 rpm for ≈15-20 min) and redispersion with a vortex shaker and an ultrasonic bath (80 W) until the supernatant was clear and free of polydopamine residues (evidenced by a transparent color).The thickness of the PDA layer was calculated by subtracting the diameter of the PS-core particle from the diameter of the PS@PDA particle.
Zeta Potential: The zeta potential ( ) of PS and PS@PDA colloidal particles were determined in 1 mM KCl aqueous solution using a Nano-ZS Zetasizer (MALVERN).
5.0.0.1.PCs: All 3D photonic crystals were produced via convective assembly. [62,74,75]Glass slides were cut vertically and cleaned with a Piranha solution (30 mL sulfuric acid and 10 mL 10% hydrogen peroxide) without additional heating.The glass slides were put into glass vials) containing the dispersion of colloidal particles (0.1 -0.2 wt.-%), placed in an oven and the continuous water phase evaporated (60 °C for 6 days).Care was taken to avoid vibration during the deposition process.
Macroscopic Optical Characterizations: To prepare thick samples for the UV-VIS measurements, PDMS molds were created.The 3.5 cm × 2 cm × 0.5 cm molds were cut out of PDMS, and a hole (2 cm × 0.8 cm × 0.5 cm) was cut into the molds.Glass slides were cleaned with ethanol.Both, glass slides and PDMS molds were activated in a plasma oven (48 s) and bonded together.600 μL of particle dispersion (1.0 wt.-%) was filled into the mold and dried in an oven (80 °C for 8 h).UV-VIS spectra were recorded with a UV-VIS LAMBDA 950 spectrometer (Perkin Elmer) equipped with a 150 mm integrating sphere.A diffuse white standard (Labsphere) was utilized as the standard reference.
Photos: The macroscopic photographs of the samples were taken with a Sony Alpha 6400 combined with a Godox V1s flash.Both were installed in a setup with an angle of the flash of 13°.
Microscopic Optical Characterization: Light microscope images and spectra were taken with a correlative microscope setup, a Zeiss Axio Z2 optical microscope (Zeiss, Germany) that is combined with an MCS CCD UV-NIR spectrometer (Zeiss, Germany).A silver mirror (Thorlabs) was used as a 100 % reference.The measured spot size was consistently 30×30 μm.SEM images were taken with a GeminiSEM 500 (Zeiss, Germany) and correlated with the optical microscope using a specialized sample holder and software (Zeiss, Germany).
Transfer of Spectra to CIE L*a*b* Color Values: The calculations were done according to established procedures using the CIE 2006 convention with standard daylight illumination (D65), a 10°observer angle, and a step size of 5 nm. [66]imulations: To perform simulations, numerical simulation techniques were utilized to describe the electromagnetic scattering of particulate products via the time-harmonic Maxwell's equation.To obtain a solution of the latter, the discrete dipole approximation (DDA) method was employed. [71]The simulated object was a PC with particles arranged in a fcc structure with a closed packing.Seven layers with roughly 10 by 10 particles per layer were considered, resulting in a total of 627 particles in the system.
The incident light propagated in the negative z direction.Two types of particles were considered: pure scatterers made of PS and absorbers made of PS@PDA.The wavelength-dependent refractive index of PS was determined from the literature, [76] while the refractive index of PS@PDA particles was given by n PS + ick PDA .Here, n PS corresponds to the real part of the refractive index of PS, k PDA represented the imaginary part of the refractive index of PDA, [59] and c∈[0,1] was a tunable constant.To control the total amount of absorbers in the system, the parameter c was varied defining the imaginary part of the combined refractive index in the homogeneous setting.In the heterogeneous setting, the ratio of PS and PS@PDA particles in the PC were changed.To test the influence of the positions of the absorbers, five random scenarios for each configuration were simulated and the average result was calculated.
All DDA simulations were performed using the open-source code ADDA. [77]The dipole distance was set to 18.2 nm, resulting in ≈1.76 × 10 6 total dipoles.Based on the polarization vector obtained from each DDA simulation, the wavelength-dependent reflectance was computed by integrating the scattering energy in a predefined monitor.The monitor was placed in the back direction of the incident field and had an opening angle of 180°.

Figure 1 .
Figure 1.Absorber incorporation to enhance structural coloration.a) Normal Stellar's Jay which displays a blue plumage.Reproduced with permission.Copyright Glenn Bartley.b) Amelanotic white Stellar's Jay, which lacks melanin granules.Reproduced with permission.[47]Copyright 2006, Company of Biologists (Photograph by Bill Schmoker).c) Vertical deposition method to assemble PS@PDA core-shell particles into a 3D PC. d) Two strategies to incorporate the absorbing PS@PDA particles.I: Homogeneous 3D PCs consisting of a single type of particle, wherein the layer thickness of PDA progressively increases, resulting in enhanced absorbing properties; II: Heterogeneous 3D PCs formed by mixing bare PS with coated PS@PDA particles.The amount of absorber is increased by introducing more PS@PDA particles into the system.

Figure 2 .
Figure 2. Synthesis of the PDA@PS core-shell particles with defined absorber content.a) SEM image of PS particles with a size of 267 nm.b) SEM image of PS@PDA particles with a size of 291 nm.c) Dispersions containing 0.2 wt.-% particles with different absorber content (from left to right, bare, 2, 7, 10, and 13 nm PDA-shell thickness).d) Control of PDA-shell thicknesses via the concentration of DA in the synthesis.

Figure 4 .
Figure 4. Microscopic optical properties of heterogeneous 3D PCs.a-e) From left to right: Macrophotographs, light microscopy, and SEM images of 3D PCs prepared from mixed colloidal dispersions of PS (d≈292 nm total particle size) and PS@PDA (t≈14 nm shell thickness; d≈291 nm total particle size) particles, with increasing amount of the PS@PDA as specified in the cartoon.f-h) Reflectance spectra taken by correlative micro-spectroscopy of all heterogeneous 3D PCs with increasing ratios of PS@PDA core-shell particles measured at PC areas with identical thickness, shown for 5, 7, and 9 layers.

Figure 5 .
Figure 5. Macroscopic characterization of structural color as a function of the absorber position.a,d) Macrophotographs of homogeneous (top row)and heterogeneous (bottom row) 3D PCs formed with increasing absorber content, as specified in the cartoons.b,c) Abstracted colors perceived by the human eye of the different samples, created with the RGB values that were determined from the reflectance spectra using the CIE standard illuminant D65, shown for the homogeneous.b) and heterogeneous.c) 3D PCs.The heterogeneous 3D PCs were made from 292 nm PS particles mixed with 291 nm PS@PDA nanoparticles with a shell thickness of t≈14 nm.e,f) Macroscopic reflectance spectra using an integrating sphere, for homogeneous (increasing PDA shell thickness).e) and heterogeneous (decreasing PS ratio).f) 3D PCs with varying absorber content.g) Quantification of spectroscopic data showing the intensity of the photonic stop band, the intensity of scattered light at 400 nm, and the difference between the two for the heterogeneous 3D PCs.

Figure 6 .
Figure 6.Characterization of the perceived color via L*a*b* color space and chromaticity.a) L*a*b* color sphere used to determine the perceived color properties (1st quadrant is missing to better visualize the colors of the samples).b,c) Perceived color of the different 3D PCs plotted into the room of the L*a*b* color sphere, for homogeneous.b) and heterogeneous 3D PCs.c), respectively.d,e) Chromaticity of the homogeneous.d) and heterogeneous.e) 3D PCs extracted from the L*a*b* values and the reflectance spectra.f) Comparison of the chromaticity of both types of samples (homogeneous: light red; heterogeneous: dark red) based on the overall amount of PDA in the colloidal crystals.

Figure 7 .
Figure 7. Numerical investigation of the optical properties of 3D PCs as a function of absorber concentration using the DDA method.a) Model of the 3D PCs for the simulation scheme.The magnitude of the k-value of the refractive index was varied to include absorbing properties.For homogeneous 3D PCs, each particle was assigned a (variable) k-value (Figure S14, Supporting Information).For heterogeneous 3D PCs, the ratio of two particle populations was varied.b) Simulation cells used to capture and analyze the reflected light intensity.c) Simulated spectra for heterogeneous 3D PCs, with reflectance values given in arbitrary units.d) Comparison between the intensities of the photonic stop band (I phot ) and incoherently scattered light (I scat ) as a function of absorber content.

Figure
Figure6dshows the calculated chromaticity values of the homogeneous 3D PCs.The sample with bare PS particles exhibited a minimal chromaticity (C* = 8), corroborating the whitish appearance in the photographs.With the addition of the smallest PDA-shell (2.5 nm), the chromaticity rose sharply to C* = 81.With a further increase in PDA-shell thickness, the chromaticity rapidly decreased as the samples became progressively darker.Figure6eshows the chromaticity of heterogeneous 3D PCs.Again, the PS sample showed a small chromaticity.Upon inclusion of absorbing PS@PDA particles, the chromaticity gradually increased until it reached a peak at approx.20% PS@PDA particles, after which it began to decline.This strategy thus provides a quantitative measure that accurately captures the macroscopic optical impression of the photographs in Figure5.In Figure6f, we compared both samples by calculating the volumetric PDA fraction of each particle.In both types of samples, the maximum chromaticity occurred at a volumetric proportion of PDA equivalent to approx.5 %.Our optical analyses underline that it is challenging to accurately evaluate color properties based solely on measured spectra without the additional translation to perceived color.This is because the entire off-peak spectrum, in addition to the peak intensity, plays a critical role in determining the macroscopically observed color.This difference in evaluation can be seen by the discrepancy between the optimal absorber component determined from the intensity ratio of the photonic stop band and scattering at 400 nm (Figure5g), and from the L*a*b* formalism (Figure6).Color optimization must therefore be performed in a more encompassing way, using, e.g., parameters derived from the L*a*b* color space.Moreover, multiple analytical techniques are required for a coherent analysis.Microscopic examinations allow accurate comparison between samples but provide only a limited estimate of the color, as the incoherently scattered light is not well collected by the objective lens.Macroscopic examinations, on the other hand, allow for clear color observation and comparison but fall short on the exact comparison between samples, because the measurement is not performed at identical thicknesses.

Figure 8 .
Figure 8.Comparison of spectra and chromaticity of simulated 3D PCs.a) Spectra of the simulated heterogeneous 3D PCs (solid curves) and the heterogeneous experimental crystals measured by the UV-VIS (dashed curves), and b) comparison of their chromaticity.c) Comparison of spectra for the simulated homogeneous (solid curves) and heterogeneous (dashed curves) 3D PCs.The homogeneous 3D PCs with a percentage of the k-value of PDA are compared with heterogeneous 3D PCs with a percentage of absorbent particles with a refractive index of PDA.d) Chromaticity of homogeneous and heterogeneous simulated 3D PCs over the average percentage of k-value of PDA in all particles.