Structural comparison of Mel-Syn and Mel-BSFFourier-transform infrared (FTIR) spectroscopy was performed on the as-obtained powder materials to analyze the structural differences between Mel-Syn and Mel-BSF (Fig. 1c). The bands for both samples are as follows: \(\widetilde{\nu }\)max / cm−1 3610 m (free OH in H2O), 3400br (NH stretching, indole), 3200br (OH), 2900 m (CH stretching, alkane, only Mel-BSF), 1715 m (C=O), 1580 m (indole ring vibration), 1390 m (CN in-plane vibration), 1240 m (CN in-ring stretching). The spectrum of Mel-Syn matches with previous literature findings43,44,45. The spectrum of Mel-BSF mostly resembles that of Mel-Syn, except for additional bands around 2900 cm−1. These correspond to aliphatic CH-stretching and can stem from aliphatic lipids and protein coat in natural melanin sources as has been documented previously43. Additionally, lipids were observed and utilized in the last instar larvae and prepupae state of Hermetia illucens46. To further analyze the nature of these lipids, filtered Mel-BSF samples were prepared, and the isolated lipids were separately investigated using FTIR. The details of the filtering procedure are discussed in the methods section and the resulting spectrum can be seen in Supplementary Fig. 1. The bands \(\widetilde{\nu }\)max / cm−1 3270br (OH) and 2920s/2850 s (CH stretching, alkane) are identical to those found in surface lipids of insect bodies found in the literature; the signals in the fingerprint area match the filter paper substrate and the signals corresponding to redox-active moieties are not strongly present, confirming the presence of simple polyethylene-segments prominent for cuticular lipids in the Mel-BSF samples47. The presence of the lipids is further confirmed by X-ray photoelectron spectroscopy (XPS) studies as discussed further below.Next, UV-Vis absorption spectroscopy was performed on aqueous solutions of the melanin samples. (Fig. 1d). The absorption spectra of both samples demonstrate the behaviour to be expected from eumelanin48, which confirms the sample quality and melanin nature of the materials. As eumelanin absorption results in an exponential decay as a function of the wavelength, no classic listing of absorption peaks and coefficients can be given easily. This unusual melanin absorption spectrum shape gives the material a black-brown colour, and has been led back to both molecular absorption and scattering49. Although the spectrum is reminiscent to the spectra of scattering inorganic compounds, it is caused by the heterogenous species ensemble and redox processes of melanin. The resulting broadening of the chromophores around 500 nm has been attributed to a mix of real absorption and Mie scattering50. It is to note that Mel-BSF shows lesser overall absorption than Mel-Syn, possibly due to the presence of diluting species like the lipids embedding the absorbing melanin. This trend is consistent across a range of concentrations of samples evaluated (Supplementary Fig. 2).For further confirmation of the powder surface composition, Raman spectroscopy was performed on Mel-BSF powder and a filtered Mel-BSF film casted onto a Si wafer. The normalized spectra are shown in Supplementary Fig. 3. The unfiltered Mel-BSF sample shows a peak at 1100 cm−1 that can be correlated to symmetric aliphatic C-C stretching, while the rest of the signals are combined into a very broad band starting at around 1250 cm−1 51. After filtration, peaks at 1350 cm−1 and 1550 cm−1 can be observed that match the aromatic C-N stretching mode as well as the stretching vibrations of the aromatic C = C bonds observable in the eumelanin structure. Finally, a shoulder peak at 1690 cm−1 can be observed corresponding to a weak contribution of quinone C = O stretching from the melanin semi- and indolequinone moieties52. The melanin-typical peaks in the filtered Mel-BSF spectrum match literature findings, thus the Raman data further confirms aliphatic species that are present in the unfiltered Mel-BSF, resulting in peak intensities of the redox active moieties being decreased and not being clearly distinguishable compared to the filtered Mel-BSF data.Besides the overall composition, techniques like X-ray powder diffraction (XRD) analysis can be used to obtain information about the structural ordering of melanin, which is described to have an amorphous, disordered structure in literature53. The recorded XRD data can be seen in Supplementary Fig. 4. Both Mel-Syn and Mel-BSF show a broad peak at slightly shifted 2 θ values (centered at about 26.8° for Mel-Syn and 30.0° for Mel-BSF) that match previous findings for melanin and amorphous/disordered materials54. Furthermore, Mel-BSF shows an additional broad peak presumably caused by the additionally aliphatic species centered around 16.7°. Based on this peak position, a spacing of 3.3 Å can be calculated for Mel-Syn, and for Mel-BSF, a spacing of 3.0 Å is obtained. The differences between the spacings can be explained through residual molecules like lipids mixed with melanin oligomers modifying interlayer distances, as well as potential differences in the DHI/DHICA ratio affecting the three-dimensional structure of the melanin samples due to different numbers of cross-linking positions being available54. An apparent ratio can be obtained from XPS analysis which will be discussed further below. To summarize, the XRD findings confirm the amorphous, disordered structure of Mel-BSF, as well as further support the presence of aliphatic structures in Mel-BSF.Using field emission scanning electron microscopy (FE-SEM), the surface structures of Mel-Syn and Mel-BSF powders can be compared (Fig. 2) further visually. The materials show a different morphology. For Mel-Syn (Fig. 2a, b), inconsistently shaped flakes between 100 nm and 300 nm in size can be observed that match previous literature findings55. For Mel-BSF (Fig. 2c, d), differently shaped flakes on the micron scale with an average size of (6 ± 0.5) µm are apparent, although some are as large as about 30 µm. Small protrusions can be observed with sizes between 5 nm and 10 nm on the Mel-BSF flakes. Two different batches of Mel-BSF were investigated with similar results (Supplementary Fig. 5). The close-up of the Mel-BSF flakes of the second batch also shows the presence of the surface-protrusions in higher magnification.Fig. 2: SEM images of melanin powders.a SEM image of Mel-Syn powder at low magnification. b SEM image of Mel-Syn at high magnification. Nanoscale particles can be observed. c SEM image of Mel-BSF powder at low magnification with an inset with a comparable scale bar to the Mel-Syn data. Micron-sized flakes can be seen. d SEM image of Mel-BSF at high magnification with nanostructured surface protrusions being visible on the flakes. All images obtained using secondary electron imaging and an acceleration voltage of 5 kV and using a few nm of Pt coating on the powders on a double-sided conductive carbon tape substrate to prevent charging.This type of surface morphology was not observed for Mel-Syn. In combination with the differences observed in the FTIR and Raman spectra of the two materials indicating aliphatic surface species for Mel-BSF, the surface protrusions could potentially be explained through the presence of lipid assemblies on the sample surface. Similar interactions between polymers and lipids influencing surface structures have been observed before for polymeric lipid hybrid nanoparticles56.To further characterize the surface nature of the two different materials, N2 gas adsorption isotherms were recorded of Mel-Syn and Mel-BSF powders, and the results were analyzed using the method of Brunauer, Emmett, and Teller (BET) (Supplementary Fig. 6 and Supplementary Note 1), a method used in previous literature works for type II isotherm melanin surface area determination57. The specific surface areas (Σsyn = 10.0 m2 g−1 and ΣBSF = 0.5 m2 g−1) and the positive BET-constant C and monolayer capacity νm listed in Supplementary Fig. 6 fit into typical literature values for BET surface areas of melanin from different sources57. The work of Crippa et al.57. Also analyzed synthetic melanin powder synthesized from tyrosine by oxidation with hydrogen peroxide from Sigma-Aldrich without further sample modification or composites added, and the obtained specific surface area of 18.04 m2 g−1 matches our findings for Mel-Syn. The lower surface area of the micron-scale BSF particle flakes aligns with the SEM findings of larger flakes and indicates the possibility of Mel-BSF powder being microscale particles with nanoscale attributes. The combination of both nanoparticles and microparticles may offer enhanced electrochemical activities58. Good reliability, scalability, and early-cycle Coulombic efficiency of the macroscopic particles coupled with the improved overall cycle stability and faster charging capabilities of the surface nanostructures through augmented surface redox centre accessibility seem to be promising attributes for energy storage applications58. As described in Supplementary Note 1, additional analysis of the gas adsorption isotherms using the Barrett-Joyner-Halenda (BJH) method for pore structure analysis59 yields the total micropore volumes (V0,syn = 0.03 cm3 g−1 and V0,BSF = 0.4 × 10−3 cm3 g−1) and average pore diameters (dsyn = 26 nm and dBSF = 16 nm) for Mel-Syn and Mel-BSF. Interestingly, in combination with the flat adsorption isotherm in Supplementary Fig. 6a, the low pore volume and size indicate Mel-BSF being almost a nonporous material. In contrast to that, Mel-Syn shows porousness comparable to amorphous carbon, matching previous literature findings57.To complete the comparison of surface structures between Mel-Syn and Mel-BSF, XPS was utilized for obtaining the elemental composition and binding states of both materials. High-resolution spectra were recorded for Mel-Syn (Fig. 3a–c) and Mel-BSF (Fig. 3d–f). The complete survey spectra are found in supplementary Supplementary Fig. 7.Fig. 3: Spectroscopic investigation of melanin films.a–c High resolution XPS data of films of synthetic melanin on an Au/glass substrate. d–f High resolution XPS data of films of natural melanin from the black soldier fly (Hermetia illucens) on an Au/glass substrate. For both samples, the C 1 s/O 1 s/N 1 s spectra are shown next to each other. All displayed data includes a smart background fit. g, h UP spectra of melanin films (Mel-Syn, black, Mel-BSF, red) on Au/glass. Shown in blue are the tangential functions used to calculate the HOMO energies of the samples. i Custom 3D-printed resin cell set-up for XPS at different potentials.They exhibit the presence of C, O, N, and Au for both melanin samples, confirming the presence of the films on the Au/glass substrates. The high-resolution spectral components are listed in Table 1. The Mel-Syn data corresponds to previous literature findings, but fewer peaks were used during fitting than previously reported, to prevent over-fitting60. The fitted peaks all have a full width at half maximum (FWHM) value of at least roughly 1 eV, confirming the peaks being real and not stemming from instrumental noise. The C 1s-data of Mel-Syn shows the presence of sp2 (EB = 284.3 eV) and sp3 carbon species (EB = 284.9 eV), as well as C-O / C-N units (EB = 285.8 eV) and C = O units (EB = 287.7 eV) that are found in either the quinone or hydroquinone form of the eumelanin polymer structure (Fig. 1). This is confirmed by corresponding peaks in the O 1s-data (EB = 531.0 eV for O = C and EB = 532.2 eV for O-C) and the indole nitrogen peak visible in the N 1s-data (EB = 399.6 eV). The shake-up π-π* satellite (EB = 290.2 eV) was not fitted due to insufficient FWHM values. For Mel-Syn, the sp2 carbon to sp3 carbon ratio can be determined as roughly 10:3. In contrast, for the Mel-BSF, the ratio is about 10:9. Similar to the FTIR-results, this can be attributed to an additional aliphatic lipid species being strongly present in the Mel-BSF samples. Correlated to the high carbon content, the O 1 s and N 1 s signal intensities are decreased indicated by increased noise in comparison to the corresponding Mel-Syn data. Still, the binding energies stay the same within the experimental error when comparing both samples. This indicates that the additional lipid species have no strong polar groups which could potentially influence the binding environments of other carbon species. Next, Mel-BSF was checked for sulfur species in case of trace pheomelanin remnants being present as previously reported for other bio-melanin samples61,62,63,64. No sulfur content was found, indicating the absence of pheomelanin moieties (Supplementary Fig. 8). Finally, the XPS data were checked for metal ion species that could potentially contaminate and influence the properties of prepared melanin electrodes. In the survey spectrum data from Supplementary Fig. 7, no signals of Al or Fe contaminants could be observed. Ca and Mn signals were absent as well, which could potentially stem from the food structure or physiological composition of the black soldier fly. Only a very small Na Auger peak (EB = 497.0 eV) is present for Mel-BSF (but absent for Mel-Syn), possibly stemming from the black soldier fly food structure, but more likely to be a remnant of the agents used in the black soldier fly melanin extraction protocol12. As the Na peak is negligibly small compared to the relevant melanin signals, it can be assumed there was no significant influence on the Mel-BSF electrode properties caused by metal ion contamination.Table 1 Assignment of the XPS components of the high-resolution spectra of melanin-films on Au/glass substrates displayed in Fig. 3Using ultraviolet photoelectron spectroscopy (UPS), films of the two different melanin samples on Au/glass were then investigated to determine the energies of the respective highest occupied molecular orbital (HOMO) (Fig. 3g, h). The signal intensity I is plotted against the binding energy EB with reference to the Fermi energy EF. The samples were referenced against a clean Au surface (Supplementary Fig. 9). The intersection of the first signal shoulder tangent with the baseline gives a HOMO energy of (1.3 ± 0.2) eV for Mel-Syn and (4.2 ± 0.2) eV for Mel-BSF. The Mel-Syn HOMO energy corresponds to previous literature values in the typical range for aromatic compounds65. The Mel-BSF HOMO energy is shifted to higher binding energies, confirming the assumption of the presence of lipids in the samples, as the direction of the energy shift is typical for long carbon-chain aliphatic structures66. The data presented here were obtained through the accumulation of 10 spectra for each sample, and specifically the Mel-BSF spectrum can be discussed further looking at the missing features at the observed energy values compared to Mel-Syn. Similar missing features in this energy region have been observed for aliphatic compounds like hexane in the work of Seki and Inokuchi66, further confirming the presence of surface lipid species potentially covering or overlapping with melanin-specific spectral features.Ex-situ electrochemical comparison of the materialsTo understand the electrochemistry of the BSF biopolymer material, XPS was conducted ex-situ at a potential of −0.8 V and 1.2 V in a three-electrode system with a polished Zn sheet CE versus an Ag/AgCl reference electrode (RE) (Fig. 3i). These potentials were found to correspond to melanin reduction and oxidation potentials of the samples on the Au/glass substrate using the RTIL electrolyte and were selected to not lead to any decomposition reactions (Supplementary Fig. 10). The potentials are shifted from the ones observed in the coin cell device tests, as those tests were performed in a two-electrode set-up versus Zn/Zn2+ as the counter/reference electrode. To ensure probing of the electrode surface was possible with minimal contamination from sample transport between set-ups, or from the de-crimping processes of a coin cell set-up leaving behind residue of the separator, the ex-situ XPS experiments were performed in this three-electrode system with the adjusted potentials. The experiments were performed on both Mel-Syn (Supplementary Fig. 11 and Supplementary Note 2) and Mel-BSF (Supplementary Fig. 12) WEs respectively, using Zn(TFSI)2 in EMIM TFSI as the electrolyte. The mechanism of quinone reduction (oxidization) in ionic liquids is well-studied and further details can be found in the literature67. The reduction is a mechanism involving the formation of quinone radical ions followed by ion pairing with either protons or the electrolyte ions. During oxidization, the protons are paired up with or intercalated with the electrolyte solvent by shifting the hydrogen bonding68. The ex-situ XPS analysis of Mel-Syn was able to confirm the assumed main chemical structure at high (low) potentials having an increased content in quinone (hydroquinone) moieties21.The exact assignment of the XPS peaks at different potentials including leftover ionic liquid peaks69 is displayed in Table 2. Again, the shake-up π-π* satellite (EB = 290.2 eV) was not fitted due to insufficient FWHM values. The C 1s-data show that when oxidized, the C-O peak shifts by 1.7 eV and the C = O peak by 0.7 V (\({E}_{{\rm{B}}}^{{\rm{ox}}}\) = 287.5 eV and \({E}_{{\rm{B}}}^{{\rm{ox}}}\) = 288.4 eV) to higher binding energies corresponding to a more polar environment. This shift is large enough to exclude surface charging from being the only reason for this shift. Thus, it is indeed a true shift confirming oxidation/reduction of the sample.Table 2 XPS component-assignment of the high-resolution spectra displayed in Supplementary Fig 11For the Mel-BSF results (Supplementary Fig. 12), no strong correlating C 1 s and O 1 s shifts of the peaks could be observed for the biopolymer at different potentials. This could be the result of the lipids covering the direct surface of the redox active material.Finally, the binding energy area of EB = 1200 eV to EB = 850 eV of the survey spectra at different potentials can be analyzed in more detail, in order to compare the Zn intercalation for Mel-Syn (Supplementary Fig. 13a) and Mel-BSF (Supplementary Fig. 13b) (Supplementary Note 3). The close-up spectra confirm the expected intercalation of Zn after the application of a negative potential, as two new peaks related to Zn 2p appear, matching prior literature findings for Zn and organic materials in ionic liquid electrolytes70. For Mel-Syn, more noticeable than for Mel-BSF, some Zn remains irreversibly trapped also during charging at positive potentials, as the potential binding sites are not only the oxidized/reduced quinone/hydroquinone moieties, but also the porphyrin-nitrogen atoms and carboxylic acid groups present in the melanin structure which are mostly unaffected by the application of the selected potentials. A chemical equation summarizing the zinc charge storage mechanism analog to quinone polymer systems found in literature71 can be seen in Supplementary Fig. 14. The mechanism is expected to be fundamentally identical for Mel-BSF and Mel-Syn due to the same general melanin monomer structure being present in both materials. However, differences in pore structure and film consistency between Mel-Syn and Mel-BSF can lead to differences in redox centre accessibility and Zn intercalation, thus influencing charge storage.ZIHCs using ionic liquid electrolyteIonic liquid electrolyte ZIHC coin cells were built with melanin mixed with carbon black and PVDF on carbon paper as WE, and polished Zn as the CE. As a separator, glass fiber filter paper was used where the electrolyte (0.31 g Zn(TFSI)2 in 1 mL EMIM TFSI) was drop-casted before assembly (Fig. 4).Fig. 4: Electrochemical investigation of devices.a–c CV, EIS, and GCD data of 2-electrode ZIHC coin cells prepared using Mel-Syn. d–f CV, EIS, and GCD data of identical devices prepared using Mel-BSF as the working electrode. The cells use ionic liquid electrolyte and Zn counter electrodes. g, h Calculated specific charge/discharge capacity (CC/CD) values of the coin cells. i Illustration of a ZIHC coin cell assembly using Mel-BSF/-Syn as working electrode material.For the CV of Mel-Syn (Fig. 4a), a potential window from 0.1 V to 2.5 V was chosen to prevent electrolyte decomposition reactions that were observed outside the selected potential window. For Mel-BSF (Fig. 4d), the window had to be reduced by 0.5 V due to the earlier occurrence of decomposition reactions employing low currents, possibly caused by a decreased apparent DHICA to DHI ratio compared to Mel-Syn and a modified adhesion of the composite materials caused by the additional lipids. The broad oxidation peak found between 1.05 V and 2 V and the broad reduction peak around 1.5 V both include all monomer variations (DHI/DHICA, IQ, QI, SQ, HQ) oxidizing/being reduced at similar potentials, resulting in the combined broad peaks5.The additional pair of peaks for potentials higher than 2.0 V for Mel-Syn can be attributed to the intercalation of TFSI ions into the electrode material that are not present for Mel-BSF due to the reduced potential window. This intercalation has been observed in other systems using RTIL electrolytes72. For both samples, the oxidation peaks move to the right at increased scan rates, as the slow charge transport mechanism (Fig. 1) reacts sluggishly to the change in system potential. Comparing the Mel-BSF and Mel-Syn CV data at slower scan rates (~5 mV s−1) using the Dunn method (Supplementary Fig. 15 and Supplementary Note 4)73, the difference in the contributions of the capacities becomes clear. Mel-BSF demonstrates a higher capacitive contribution at high scan rates compared to Mel-Syn. Due to the presence of aliphatic compounds increasing film stability and thus redox centre availability, the Mel-BSF film has a faster redox reaction process than Mel-Syn.Comparing the electrochemical impedance spectroscopy (EIS) data of Mel-Syn (Fig. 4b) with that of Mel-BSF (Fig. 4e) qualitatively gives insight into the charge transport of the systems. As the composite material includes components promoting electronic conductivity like carbon black and the ionic conductivity associated with melanin, the resulting conductivity will be referred to as mixed conductivity in the following. Only the Mel-BSF data shows an electrochemical semi-circle at mid to high-frequency values corresponding to the charge transfer properties of the organic working electrode. This can be attributed to the decreasing charge transfer resistance, plausibly explained with enhanced charge transfer through the lipids that improve film consistency. They possibly increase the spread of the other composite materials like carbon black to augment redox centre availability, electron mobility, and thus mixed conductivity, an effect that has been made use of in the literature in comparable systems74. As carbon black is naturally hydrophobic and organophilic, its incorporation in membranes or films including hydrophobic lipids reduces carbon particle agglomeration and causes an augment in electronically conductive conduits through the composite material. The Mel-Syn data only shows a small electrochemical semi-circle at very high-frequency values, potentially corresponding to anode activation losses and resistance caused by the separator. A micrograph of the high-frequency region for Mel-Syn can be seen in Supplementary Fig. 16.Even with the reduced potential window, Mel-BSF showed much higher current values than Mel-Syn during CV measurements, making it promising to investigate the capacity of the complete ZIHC devices using galvanostatic charge/discharge (GCD) measurements. The collated results are presented for Mel-Syn (Fig. 4c) and Mel-BSF (Fig. 4f) with the corresponding calculated capacity values (Fig. 4g, h) and the experimental set-up (Fig. 4i). The energy density ε and power density PD values together with the GCD cycling results over 4900 cycles, including the Coulombic efficiency η (Supplementary Fig. 17), were calculated and recorded. The maximum values for both sample types at low mass loadings (<0.5 mg) are collated in Table 3. The calculated values were obtained by integration of the GCD data. Average and standard error values for higher mass loading samples (up to 1 mg) are shown in Supplementary Table 1 with further discussion in Supplementary Note 5.Table 3 Electrochemical data of Mel-Syn and Mel-BSF Zn coin cells using ionic liquid electrolyteThe measurements between samples were normalized by the redox active melanin mass of the samples. For all current density values, Mel-BSF shows a higher charge-discharge time resulting in both higher specific capacity values at specific current densities for the charge and discharge states of the material (\({C}_{{\rm{D}},\max }^{{\rm{Syn}}}\) = 17.3 mAh g−1; \({C}_{{\rm{D}},\max }^{{\rm{BSF}}}\) = 91.8 mAh g−1), and thus higher energy density values (\({\varepsilon }_{\max }^{{\rm{Syn}}}\) = 20.7 Wh kg−1 at a power density of \({P}_{{\rm{D}}}^{{\rm{Syn}}}\) = 65.8 W kg−1; \({\varepsilon }_{\max }^{{\rm{BSF}}}\) = 87.2 Wh kg−1 at a power density of \({P}_{{\rm{D}},\max }^{{\rm{BSF}}}\) = 72.8 W kg−1). This is also a result of the increase in accessibility of redox centres. The lipids increase the dispersion of melanin molecules, and thus film consistency. The lipids interact with the other composite materials, allowing a higher amount of total charge transfer.Even after 4900 GCD cycles at 5 A g−1, Mel-BSF shows a high Coulombic efficiency of 99 %, aligning with the expected good reversibility for Mel-BSF from the XPS analysis. A good efficiency around 100 % and high cycling stability have also been reported in similar literature systems employing porphyrin-based polymers and biopolymer-based carbon materials, showing the competitiveness of Mel-BSF in ZIHCs18,75. For Mel-Syn long cycling at high currents, a low Coulombic efficiency of 80 % is observed.The potentials vs. specific discharge capacity curves obtained from the GCD data for the two materials at 0.1 A g−1 (Fig. 5a, b) clearly show that the discrepancy between charge and discharge capacity is much higher for Mel-Syn, matching the reduced Coulombic efficiency observed for high currents during the long cycling over 4900 cycles when compared to Mel-BSF. This also matches the XPS Zn 2p observations. Furthermore, the maximum capacity values for Mel-BSF and Mel-Syn match the reported data in Table 3.Fig. 5: Performance analysis and comparison of melanin ZIHCs.a Galvanostatic charge/discharge potentials of Mel-Syn plotted versus specific capacity. b Galvanostatic charge/discharge potentials of Mel-BSF plotted versus specific capacity. The maximum capacity values match the Table 3 data. c Charge transfer mechanisms of Mel-Syn and d Mel-BSF films using the model from Bard76. The left depicts a redox species S travelling through pinholes in Mel-Syn films. The right showcases S reacting to P at a Mel-BSF film, giving an electron that can hop from one melanin redox moiety to the next, reducing species O to R, until the carbon paper electrode is reached. e Ragone plots of presented devices compared to literature SCs and ZIHCs. The Mel-Syn and Mel-BSF devices are compared to melanin, PANI, and PDA literature systems. Literature information compiled from refs. 7,18,19,77.The differences between the performances of the devices prepared with Mel-Syn and Mel-BSF films due to different surface structures and film consistency behaviours can be described accurately using the simple model of processes at modified electrode interfaces proposed by Bard76 (Fig. 5c, d) and compared with other literature SCs and ZIHCs (Fig. 5e)7,18,19,77.The charge transfer in Mel-Syn films can be described as partial movement of a redox active species S (e.g., Zn ions) through pinholes or cracks to the substrate where it can be reduced. On the other hand, the transfer in Mel-BSF films is best described with electron diffusion caused by the redox active species Zn2+ reacting to a product P (Zn) at the film interface and subsequent electron hopping from one indolequinone moiety to the next through the uniform film with high redox centre accessibility, and finally to the substrate.Light microscopy images of Mel-Syn and Mel-BSF composite films (with PVDF and carbon black) were recorded at different spots, and commonly occurring representative parts of the samples that underline the explanation using the model of Bard can be seen (Fig. 5c, d). The left of the schematic depicts the described charge transfer of a redox active species to the electrode through pinholes and cracks in the surface of a Mel-Syn film, while the right side shows a model for the charge transfer for Mel-BSF. The bulk and surface melanin-intercalated species with no access to pinholes proceed to react with the melanin in the film on the conductive carbon paper electrode as described previously (Fig. 1), giving an electron that continues a cycle of further electron donation, hopping from one quinone/semiquinone/hydroquinone unit to the next, until the substrate is reached, resulting in higher electron storage capabilities for the Mel-BSF films compared to Mel-Syn films.It has been shown in the literature that an increased interlayer spacing results in increased accessible surface area for the electrolyte, leading to improved energy density78. The XRD results do show an increased spacing for Mel-Syn (3.3 Å) compared to Mel-BSF (3.0 Å), but as the difference is relatively small compared to the size of Zn-ionic liquid electrolytes (solvated radius over 10 Å79), this does not affect the accessibility of redox centres. The BJH-method pore analysis showed increased total pore volume and a slightly larger average pore diameter for Mel-Syn (26 nm) compared to Mel-BSF (16 nm). While the increased porousness may increase redox centre accessibility, the pore structures or wettability may not be favorable for enhanced charged storage. Works like that of Siwicka et al.80. Have shown the porousness of melanin materials can reach values as high as 215 m2 g−1 depending on the fabrication technique. Future work on improving both the film stability of melanin materials, and optimizing properties like the porousness, could be carried out.To further prove the importance of the lipids increasing the melanin dispersion in the films, filtered Mel-BSF samples were prepared to remove the leftover lipids and they were measured with an identical set-up as for the previously discussed synthetic and unfiltered BSF-melanin (Supplementary Fig. 18). While the accessible potential window and qualitative shape of the CV curve stayed relatively similar after filtration, the charge and discharge times were shortened at the same current density values despite having a slightly larger accessible potential window. This results in a reduced specific discharge capacity of 4.7 mAh g−1 at 0.1 A g−1 observed for the filtered material. The prior FTIR investigations of the filter remnant have clearly shown that the material removed through filtration is of redox-inactive nature, as only features typical for lipids could be observed. Thus, the reason for the drop in performance in the filtered sample cannot be the removal of the main redox active compound but is indeed likely the decreased film stability and thus redox centre accessibility caused by the absence of said lipids.Lastly, to confirm the melanin redox centres being the main contributors for charge storage, the CV and GCD tests for Mel-Syn in the Zn coin cell set-up were further recorded from 0.1 V to 2.0 V to match the potential window of Mel-BSF (Supplementary Fig. 19). The CV curves now match those of Mel-BSF, and there are no additional electrochemical processes dominating the response as observed at both low and high scan speeds. From the GCD data, a maximum specific discharge capacity of 16.0 mAh g−1 at 0.1 A g−1 is obtained for Mel-Syn employing a reduced potential window. This value is only slightly lower than the 17.3 mAh g−1 observed when measured up to a potential window of 2.5 V, further indicating that the melanin redox centres are the main contributors to the electrode capacity.Previously, for aqueous electrolyte (NH4CH3COO in H2O at pH 5.5) systems using eumelanin films on carbon paper, cyclic voltammetry data measurements yielded a specific capacitance of 167 F/g (or 24 mAh g−1)7. This value corresponds to the similar order of magnitude of the extracted GCD data of our Mel-Syn ZIHCs with a specific capacity of 17.3 mAh g−1. The naturally extracted, unfiltered Mel-BSF surpasses this with a specific capacity of 91.8 mAh g−1. For unmodified and soluble melanin derivatives in solid-state energy storage systems using a polyvinyl alcohol-phosphoric acid gel electrolyte in a symmetric cell set-up, power densities of around 1 W kg−1 with an energy density of 2 Wh kg−1 were determined previously81. The Mel-BSF data surpassed both this performance and the performance of the devices using melanin thin-films under solar light reported by Xu et al.77. Supercap-like devices made from self-synthesized melanin on carbon paper, but with catalytic NH3, and using NaCH3COO in H2O at pH 5 as electrolyte, showed power density values of 3800 W kg−1 77 which is higher than both Mel-Syn and Mel-BSF in ZIHC coin cell devices, but showed a lower energy density of 12 mWh kg−1. The Mel-Syn and Mel-BSF data and comparative melanin literature was collated in a Ragone plot (Fig. 5e). Additionally, similar organic electrode systems are presented to underline the competitiveness of the Mel-BSF devices. PANI ZIHCs in 2 M ZnSO4 from Cui et al.18. And PDA ZIHC complete devices from Cong et al.19. Show a comparable operating range at reduced maximal energy density (48 Wh kg−1 for PANI and 119 Wh kg−1 for PDA) and maximal power density (2000 W kg−1 for PANI and 941 W kg-1 for PDA) when compared to Mel-BSF. This evaluation against recent works shows that the devices presented in this work are comparable to commonly used energy storage materials and, more specifically, other ZIHCs using typical organic working electrodes.This work evidently reveals that the naturally extracted Mel-BSF biopolymer shows favorable electrochemical properties potentially stemming from additive lipids and from a beneficial surface morphology consisting of nanoscale features on micro-structured particles. The lipid-melanin interaction has endowed superior electrochemical properties and film processability, which are highly promising for energy storage applications.
