Blueberries as a Source of Energy: Physical Chemistry Characterization of Their Anthocyanins as Dye-Sensitized Solar Cells’ Sensitizers

: This work aimed to show the possibility of applying anthocyanins extracted from blueberries following a straightforward path as potential impregnation dyes in dye-sensitized solar cells (DSSCs), particularly in the presence of co-adsorbents, such as silver nanoparticles, as an alternative in order to proﬁt from large amounts of discarded fruits. Following a simple procedure, anthocyanins (mainly delphinidin-3-glucoside) were obtained from blueberries (Southern Highbush type). Com-plete characterization was carried out in order to prove the utility of delphinidin-3-glucoside as a sensitizer in DSSCs. The analyzed anthocyanin is suitable for sensitizing because of its high molar absorptivity values within the visible region of the light spectra, the adsorption ability to a FTO/TiO 2 electrode (FTO, ﬂuorine-doped tin oxide) as conﬁrmed by Fourier transform infrared (FTIR) as well as thermogravimetry coupled to differential scanning calorimetry (TG-DSC), a potential oxidation value near 1 V, and adequate thermal as well as light stabilities. Moreover, the cell’s conversion efﬁciency is improved in the presence of silver nanoparticles, reaching 0.24% (nearly a 25% increase). The sum of all these characteristics points to the application of delphinidin-3-glucoside as a sensitizer in DSSCs, offering a technological use with potential interest for countries where agricultural production offers an abundant origin of extraction.


Introduction
Dye-sensitized solar cells (DSSCs) are based on transparent and colorful materials. They provide an alternative to silicon-based devices for indoor applications and BIPVs (building-integrated photovoltaics) due to their transparency and beauty or for long-term wireless power for Internet of things (IoT) devices and wireless sensors (where the need for current or voltage is low) [1][2][3][4][5][6][7]. Dyes can be extracted from natural sources or obtained from synthetic paths, and reach conversion efficiency depending on their origin.
The total efficiency of DSSCs depends strongly on the photoanode composition, where the dye plays a vital role in the transfer processes [12]. Strong dye adsorption onto the TiO 2 is required for efficient electron injection from the pigment to the semiconductor. The selection of photosensitizers in dye-sensitized cells is based on the high molar absorptivity of visible light at different wavelengths, the stability towards the temperature or light, and the costs associated with them being obtained. Among the evaluated sensitizers, several dyes extracted from natural resources (anthocyanins, carotenoids, flavonoids, and chlorophyll) have been reported and represent an attractive alternative regarding their Solar 2023, 3 284 abundance, costs, and eco-friendly characteristics [13][14][15][16][17][18]. Conversion efficiencies for natural pigments are much lower than those obtained for synthesized dyes, but they still constitute a research focus [19][20][21][22][23][24][25][26]. Although DSSCs based on synthetic dyes are more efficient, their use of expensive metals, such as N3 and N719, has several drawbacks [27]. DSSCs based on natural dyes could offer an exciting alternative to profit from tons of biomass originating from fruits that are discarded due to different reasons (i.e., them being overripe, crushed, or too small), from invasive seaweeds or produced as secondary products associated with tree felling, among others [28][29][30][31]. Moreover, pigment extraction from these sources could be as easy as using a suitable solvent [32][33][34].
DSSCs can operate indoors and under light-diffused conditions or cloudy skies. This being the case, they can be used as photovoltaic windows or integrated into greenhouses [35][36][37][38]. Additionally, applying this technology could offer an energy generation alternative, particularly for cold regions. When sensitizers are natural dyes, low temperatures constitute an ally regarding thermal stability.
Silver nanoparticles have been applied in DSSCs, increasing cell efficiency [39][40][41]. They enhance the electron transference towards the photoanode surface by covering bare spots on the electrode or as a part of a net with the dyes. Nanoparticles could be added mixed with the pigment as an impregnation cocktail, or by following a sequential step.
This work aimed to explore the use of anthocyanins extracted from blueberries (Southern Highbush type), primarily delphinidin-3-glucoside (Scheme 1), as sensitizers in DSSCs. Solar 2023, 3, FOR PEER REVIEW 2 sensitizers, several dyes extracted from natural resources (anthocyanins, carotenoids, flavonoids, and chlorophyll) have been reported and represent an attractive alternative regarding their abundance, costs, and eco-friendly characteristics [13][14][15][16][17][18]. Conversion efficiencies for natural pigments are much lower than those obtained for synthesized dyes, but they still constitute a research focus [19][20][21][22][23][24][25][26]. Although DSSCs based on synthetic dyes are more efficient, their use of expensive metals, such as N3 and N719, has several drawbacks [27]. DSSCs based on natural dyes could offer an exciting alternative to profit from tons of biomass originating from fruits that are discarded due to different reasons (i.e., them being overripe, crushed, or too small), from invasive seaweeds or produced as secondary products associated with tree felling, among others [28][29][30][31]. Moreover, pigment extraction from these sources could be as easy as using a suitable solvent [32][33][34]. DSSCs can operate indoors and under light-diffused conditions or cloudy skies. This being the case, they can be used as photovoltaic windows or integrated into greenhouses [35][36][37][38]. Additionally, applying this technology could offer an energy generation alternative, particularly for cold regions. When sensitizers are natural dyes, low temperatures constitute an ally regarding thermal stability.
Silver nanoparticles have been applied in DSSCs, increasing cell efficiency [39][40][41]. They enhance the electron transference towards the photoanode surface by covering bare spots on the electrode or as a part of a net with the dyes. Nanoparticles could be added mixed with the pigment as an impregnation cocktail, or by following a sequential step.
This work aimed to explore the use of anthocyanins extracted from blueberries (Southern Highbush type), primarily delphinidin-3-glucoside (Scheme 1), as sensitizers in DSSCs. The use of anthocyanins as sensitizers has been widely reported and constitutes one of the best options for this use among natural resources [13,[42][43][44][45][46][47][48]. Nevertheless, exploring new sources from which to obtain these compounds and examine technological benefits could interest countries where agricultural production offers an abundant extraction origin.
For co-sensitized cells assembled with delphinidin-3-glucoside and silver round nanoparticles, we achieved a conversion efficiency of 0.24% under a power of 1 sun, AM 1.5.

Extraction and Purification of Dyes
Fresh samples from blueberries of the Southern Highbush type were obtained from the market. Blueberries were washed with Milli-Q water and dried at 40 °C. Two extraction options were tested: from the entire fruit or from the peel. The fruit or the peel were mechanically crushed into a mortar and mixed with liquid nitrogen. The remaining solid was dried once more at 40 °C. Then, 1 g of solid was mixed with 10 mL of solvent (Milli-Q Scheme 1. Structure of delphinidin-3-glucoside (Glu = glucose).
The use of anthocyanins as sensitizers has been widely reported and constitutes one of the best options for this use among natural resources [13,[42][43][44][45][46][47][48]. Nevertheless, exploring new sources from which to obtain these compounds and examine technological benefits could interest countries where agricultural production offers an abundant extraction origin.
For co-sensitized cells assembled with delphinidin-3-glucoside and silver round nanoparticles, we achieved a conversion efficiency of 0.24% under a power of 1 sun, AM 1.5.

Extraction and Purification of Dyes
Fresh samples from blueberries of the Southern Highbush type were obtained from the market. Blueberries were washed with Milli-Q water and dried at 40 • C. Two extraction options were tested: from the entire fruit or from the peel. The fruit or the peel were mechanically crushed into a mortar and mixed with liquid nitrogen. The remaining solid was dried once more at 40 • C. Then, 1 g of solid was mixed with 10 mL of solvent (Milli-Q water, acidified methanol pH = 3.5 or acidified ethanol pH = 3.5, or absolute ethanol). The mixtures were vortexed and refrigerated at 4 • C in the dark overnight. Finally, the solutions were centrifuged and the solid was discarded.
After the extractions, the solid-free centrifuged solutions were purified via the use of C18 columns, employing an acetonitrile-methanol (ratio of 3:2) mixture as a mobile phase. The eluted solution was partially evaporated via the use of a Savant SPD 2010 Speed Vac Concentrator to increase the concentration. The purified concentrated solutions were then characterized through different techniques and applied to sensitize the DSSCs.

Characterization
FTIR spectra in the range of 400-4000 cm −1 were obtained at room temperature through employing a Shimadzu infrared spectrometer model IR-Prestige 21, averaging ten scans at a nominal resolution of 4 cm −1 . In the case of anthocyanins, a few drops of the concentrated solutions (obtained as described in Section 2.1) were deposited in a Petri dish and dried overnight. After this, the remaining solids were scratched with a spatula and thoroughly mixed with KBr in an agate mortar to obtain an approximately 1% mixture (w/w). In this way, 13 mm discs were prepared in a Pike Crush IR at a pressure of 10 tons.
Different FTIR spectra were recorded: 1-for TiO 2 (anatase form), 2-for the purified anthocyanin-containing solution (from the dried drops, as explained before), and 3-for the dye after incubation for 24 h onto TiO 2 electrodes. To obtain the dye incubate on TiO 2 samples, 100 µL of the dye-containing solutions was deposited on the FTO/TiO 2 electrodes and left at room temperature to dryness. After 24 h in darkness, a disc was prepared through the use of KBr and some milligrams of the scraped electrode surface via the described procedure.
Cyclic voltammetry was employed to evaluate the redox behavior through the use of screen-printed disposable Au-pc (polycrystalline) electrodes (DROPSENS) at a potential scan rate of 0.05 Vs −1 , with a pseudo-reference potential of E = 0.238 V vs. SHE (standard hydrogen electrode). The supporting electrolyte was ethanol/0.1 M NaClO 4 in milli-Q water (50/50).
For thermogravimetry (TG) and differential scanning calorimetry (DSC), 449F5 Netzsch simultaneous thermal analysis equipment was employed, using aluminum crucibles. The equipment was previously calibrated using metals as standards under heat flow and temperature. Synthetic air was used as an atmosphere with a 50 mL/min flow rate. We used a 10 • C/min heating rate. Three samples were analyzed: DEL3GLU, DEL3GLU adsorbed to TiO 2 , and TiO 2 . These were obtained in the same way as for FTIR, except that the DEL3GLU was obtained from a lyophilized aqueous crude extract from blueberry peels.
In addition, the stability of the DEL3GLU extracted with ethanol and water was evaluated by heating the solution at temperatures within the range of 40 to 80 • C in the presence and absence of spherical Ag nanoparticles of 100 nm (NP100). In this case, the thermal stability of the compound was followed by recording the absorbance values at 550 nm, i.e., the maximum of the visible spectra.

DSSC Assembly and Characterization
DSSCs were assembled following a sandwich configuration, with FTO/TiO 2 (SOLA-RONIX test kit, 20 mm × 20 mm sized, active area of mesoporous TiO 2 0.36 cm 2 ) and FTO/Pt (20 mm × 20 mm sized, screen printed with SOLARONIX's Pt Platinum Catalyst, as offered at the test cell kits) as the working and counter-electrodes, respectively. Both electrodes were purchased from SOLARONIX and used without further treatment. The titania opaque electrodes were printed on TCO22-7/LI glass (7 Ω/sq fluorine-doped tin oxide coating on one side), with active titania and reflective titania. The active titania was about 10 microns thick and comprised anatase nanoparticles of about 15-20 nm (porositỹ 10-20 nm). The reflective titania was about 2-3 microns and composed of about 80-120 nm nanoparticles. Before use, the FTO/TiO 2 electrode was heated at 500 • C for 30 min. The electrolyte was a 50 mM iodide/tri-iodide (I − /I 3 − ) in acetonitrile (SOLARONIX Iodolyte AN-50). Three different sensitizers were applied to assemble the DSSCs: pure ethanolic extract, pure ethanolic extract with silver nanoparticles, and aqueous crude extract.
FTO/TiO 2 electrodes were immersed overnight into the DEL3GLU-containing solution to adsorb the dye and then rinsed thoroughly with ethanol. A sequential approach was followed when silver nanoparticles were used. The dye solution was applied first, followed by the immersion of the photoanode in the nanoparticle solution [40]. Spherical Ag nanoparticles of 100 nm (NP100) (Nanocomposix ® , 0.02 mg mL −1 of silver with 2 mmol L −1 of sodium citrate) were utilized.
The assembled DSSCs were characterized by measuring the current density vs. voltage (J-V) profiles and recording the electrochemical impedance spectroscopy (EIS) data with a CHI 604E potentiostat. J-V analyses were accomplished at a potential scan rate (v) of 0.05 Vs −1 , at room temperature, in darkness and using a solar simulator from ABET Technologies (1 sun or 100 mW cm −2 , AM 1.5). EIS experiments were performed between 0 and 0.5 V and inside the frequency range of 0.1 Hz to 3 MHz (in darkness).
Some additional absorbance evaluations were registered by using a reflection probe connected to a SPELEC (DROPSENS) instrument to explore the co-adsorption of anthocyanins and NP100. For this, the FTO/TiO 2 electrodes were dipped overnight inside the cocktail solutions (i.e., a mixture of purified anthocyanins and NP100) and rinsed with ethanol before being measured within the 350 to 650 nm range.

Extraction and Purification
As detailed in Section 2.1, different extraction conditions were applied and followed via visible spectroscopy. Concentrations were calculated accordingly to Beer's law.
The highest delphinidin-3-glucoside (DEL3GLU) contents were obtained from extractions from the whole fruit using water as an extraction solvent and from the peel using acidic methanol. For these cases, the obtained concentrations were 12 × 10 −5 M. On the opposite end, the lowest delphinidin-3-glucoside amounts were obtained using ethanol as the extraction solvent starting from the peel or the whole fruit, and almost the same concentration of 2 × 10 −5 M was calculated.
The absorbance spectra for the anthocyanin-containing solutions confirmed the presence of chlorophylls (peak at 666 nm) and anthocyanins (peaks at 520 and 565 nm), and therefore a C18 column was used to purify them [17,51]. Additionally, as observed in Figure 1a, the absorbance peak is very wide in ethanol, and even two superimposed peaks are detected. In neutral to slightly acidified media, anthocyanins are involved in equilibrium with their quinonoid form, and therefore two species could be seen [52][53][54].
Comparing the spectra recorded from the solution with those obtained from the adsorbed compounds onto the FTO/TiO 2 electrodes is interesting. Anthocyanins and chlorophylls are detected in both types of spectra (Figure 1a,b). The spectra also showed the presence of NP100 originating from the sensitizing impregnation cocktail. The ratio between the absorbance of the peak at ca. 550 nm related to the one at 666 nm changes and decreases after adsorption. This shows that chlorophylls, when present, compete with DEL3GLU for the semiconductor surface, but that their adsorption is more favorable, even in the presence of silver NP100. The presence of adsorbed NP100 could be assessed from the absorbance peaks at 380 and 420 nm due to the surface plasmon absorption band, also pointing to nanoparticle aggregation [55]. At this point, it is necessary to clarify some details. When ethanol is used to extract, chlorophylls in the raw solution are detected for the peak absorbance at 666 nm. When adsorbed to the TiO 2 , they compete with the anthocyanins for the electrode surface. For this reason, chlorophylls are removed from the raw solution via the use of a C18 column to increase the cell's efficiency. Comparing the spectra recorded from the solution with those obtained from the adsorbed compounds onto the FTO/TiO2 electrodes is interesting. Anthocyanins and chlorophylls are detected in both types of spectra (Figure 1a,b). The spectra also showed the presence of NP100 originating from the sensitizing impregnation cocktail. The ratio between the absorbance of the peak at ca. 550 nm related to the one at 666 nm changes and decreases after adsorption. This shows that chlorophylls, when present, compete with DEL3GLU for the semiconductor surface, but that their adsorption is more favorable, even in the presence of silver NP100. The presence of adsorbed NP100 could be assessed from the absorbance peaks at 380 and 420 nm due to the surface plasmon absorption band, also pointing to nanoparticle aggregation [55]. At this point, it is necessary to clarify some details. When ethanol is used to extract, chlorophylls in the raw solution are detected for the peak absorbance at 666 nm. When adsorbed to the TiO2, they compete with the anthocyanins for the electrode surface. For this reason, chlorophylls are removed from the raw solution via the use of a C18 column to increase the cell's efficiency.

FTIR
After evaporation to dryness, KBr discs were prepared to contain the purified anthocyanins and the red-colored compound after incubation for 24 h onto FTO/TiO2 electrodes ( Figure 2, Table 1). Then, those samples prepared through scratching the TiO2 contained less DEL3GLU due to the presence of an amount of the semiconductor. Remarkably, the signals' relative intensities are irrelevant since the KBr discs were prepared through weighing 1 mg of the sample to be analyzed.

FTIR
After evaporation to dryness, KBr discs were prepared to contain the purified anthocyanins and the red-colored compound after incubation for 24 h onto FTO/TiO 2 electrodes ( Figure 2, Table 1). Then, those samples prepared through scratching the TiO 2 contained less DEL3GLU due to the presence of an amount of the semiconductor. Remarkably, the signals' relative intensities are irrelevant since the KBr discs were prepared through weighing 1 mg of the sample to be analyzed.
Despite this, for discs prepared from TiO 2 after incubation overnight in the anthocyanin solution, the presence of the signals assigned to DEL3GLU demonstrated the adsorption. Moreover, if the position of some of the signals moved to different frequencies, the functional group responsible for the signal was involved in the bonding with the semiconductor. Despite this, for discs prepared from TiO2 after incubation overnight in the anthocyanin solution, the presence of the signals assigned to DEL3GLU demonstrated the adsorption. Moreover, if the position of some of the signals moved to different frequencies,  The double signal peak at 1715 cm −1 and 1647 cm −1 is due to C=O groups. In the TiO 2 samples, the signal at 1647 cm −1 detected in the dye moved to 1637 cm −1 after the adsorption of DEL3GLU. Establishing bonds between the dye and the semiconductor caused this displacement. To understand some details, it is important to consider the structures involved in the equilibrium. In anthocyanins, flavylium cations (red colored) and the quinonoid form (red-blue) are in equilibrium. After adsorption onto TiO 2 , electrodes became red-blue colored, showing that quinonoid is the main form when conjugated to the semiconductor [31]. The quinonoid form has C=O groups, whereas the flavylium does not.
After union to the TiO 2 , some signals coming from phenol groups disappeared: in particular, signals at 1454 cm −1 and 1421 cm −1 (assigned to C-C=C vibration), in addition to the signal at 1250 cm −1 (from C-O vibration) [56]. These observations proved that DEL3GLU could be adsorbed to the titania through using the phenolic -OH moieties.
The strong signal at 1060 cm −1 is split into two weak contributions at 1130 and 1000 cm −1 after adsorption onto TiO 2 . The mentioned signal could be explained due to the C-O vibration of bonds from primary alcohol groups, such as those found in glucose and the stretching of C-O-C bonds of carbohydrates or flavylium [57,58]. These facts involved the -OH groups from glucose or the flavylium ion in coordination with titanium.
Finally, a signal related to the stretching Ti-O-Ti is detected at 665 cm −1 . The peak intensity at 1647 cm −1 ratio to the peak intensity at 665 cm −1 decreased after the adsorption of DEL3GLU to the semiconductor, confirming the existence of intermolecular interactions between anthocyanin and the semiconductor.

Redox Characterization
From cyclic voltammetry, profiles registered at v = 0.05 Vs −1 for Au-pc in the supporting electrolyte (ethanol/0.1 M NaClO 4 in Milli-Q water 50/50) displayed an anodic contribution at 0.88 V with a cathodic counter-peak at 0.52 V. The potential of the pseudoreference electrode is 0.238 V vs. SHE. The potentials in the text are referred to as this value.
Voltametric profiles obtained for Au-pc in an anthocyanin solution dissolved in the supporting electrolyte showed the presence of an anodic peak at 0.98 V with a cathodic contribution at 0.52 V (Figure 3). The -OH groups from the anthocyanin adsorb to the electrode surface first, and the oxidation of this group then occurs at 0.98 V [43].
An oxidation potential of ca. 1 V is high enough to assure electronic transfer inside DSSCs once assembled through using such sensitizer in the photoelectrode.
Voltametric profiles obtained for Au-pc in an anthocyanin solution dissolved in the supporting electrolyte showed the presence of an anodic peak at 0.98 V with a cathodic contribution at 0.52 V (Figure 3). The -OH groups from the anthocyanin adsorb to the electrode surface first, and the oxidation of this group then occurs at 0.98 V [43].
An oxidation potential of ca. 1 V is high enough to assure electronic transfer inside DSSCs once assembled through using such sensitizer in the photoelectrode.

Thermal Stability
The stability of DEL3GLU extracted with ethanol and water was evaluated by heating the solution at temperatures within the range of 40 to 80 °C in the presence and absence of NP100 (Figure 4). For this reason, changes in the absorption values for the maxima were recorded.
In all evaluated cases, the stability of the solutions did not show significant changes. Only a slight increase was detected at 50-55 °C and could be explained due to the co-pigmentation effect. The flavylium cation is a planar chromophore. Under the influence of the temperature, these planar structures stack with other, similar ones, resulting

Thermal Stability
The stability of DEL3GLU extracted with ethanol and water was evaluated by heating the solution at temperatures within the range of 40 to 80 • C in the presence and absence of NP100 (Figure 4). For this reason, changes in the absorption values for the maxima were recorded. Solar 2023, 3, FOR PEER REVIEW 8 in increased stability instead of decomposition due to a nucleophilic attack of water, increasing the color intensity and the absorbance values [59].

TG and DSC Evaluation of the Samples
To evaluate DEL3GLU and TiO2 adsorption, in addition to the stability of their union, TG and DSC techniques were carried out. Samples with TiO2 were obtained from FTO/TiO2 electrodes used to assemble the cells, and the others were used as is. Figure 5 shows the TG and DSC thermograms. The TG signal's derivative is displayed for a better visualization of the events. From 5a, it is possible to observe weight loss during the heat treatment. Each sampling event with mass percentages and associated temperature ranges is described in Table 2. Four events were observed in the thermal decomposition of the pigment: the first, between 0 °C and 100 °C, with a mass loss of 0.5% associated with the evaporation of the solvent (water); the second, which seems to be two overlapping events (one between 100 °C and 160 °C and the other between 160 °C and 250 °C), with a 43% mass loss; the third, between 250 °C and 400 °C, with a 17% mass loss; and the fourth event, between 400 °C and 600 °C, with a 25% mass loss. Therefore, there is a 85% mass loss associated with the thermal decomposition of the pigment. Regarding TiO2, it can be observed that there is no associated mass loss event in the range studied. While the DEL3GLU sample supported on TiO2 shows the same type of events as In all evaluated cases, the stability of the solutions did not show significant changes. Only a slight increase was detected at 50-55 • C and could be explained due to the copigmentation effect. The flavylium cation is a planar chromophore. Under the influence of the temperature, these planar structures stack with other, similar ones, resulting in increased stability instead of decomposition due to a nucleophilic attack of water, increasing the color intensity and the absorbance values [59].

TG and DSC Evaluation of the Samples
To evaluate DEL3GLU and TiO 2 adsorption, in addition to the stability of their union, TG and DSC techniques were carried out. Samples with TiO 2 were obtained from FTO/TiO 2 electrodes used to assemble the cells, and the others were used as is. Figure 5 shows the TG and DSC thermograms. The TG signal's derivative is displayed for a better visualization of the events. From 5a, it is possible to observe weight loss during the heat treatment. Each sampling event with mass percentages and associated temperature ranges is described in Table 2. Four events were observed in the thermal decomposition of the pigment: the first, between 0 • C and 100 • C, with a mass loss of 0.5% associated with the evaporation of the solvent (water); the second, which seems to be two overlapping events (one between 100 • C and 160 • C and the other between 160 • C and 250 • C), with a 43% mass loss; the third, between 250 • C and 400 • C, with a 17% mass loss; and the fourth event, between 400 • C and 600 • C, with a 25% mass loss. Therefore, there is a 85% mass loss associated with the thermal decomposition of the pigment. Regarding TiO 2 , it can be observed that there is no associated mass loss event in the range studied. While the DEL3GLU sample supported on TiO 2 shows the same type of events as DEL3GLU alone, the second and fourth events, however, are lower, as is possible to observe in the derivative thermogravimetry (DTG) plot in the inset figure and in Table 2. From Figure 5b, it is possible to observe the DSC signal for the same samples. In the TiO 2 sample, as in TG, no signal was observed. Meanwhile, for DEL3GLU it is possible to observe signals (exo and endo) in the same range of temperature indicated in the TG thermogram. In the range of 100-260 • C it is possible to distinguish a first endothermic and a second exothermic peak. In the fourth event it is possible to observe a bigger exothermic peak. Solar 2023, 3, FOR PEER REVIEW 9 the second event, 43% of the mass is lost, this percentage of loss could be associated with the rupture of the glycosidic bond, resulting in aglycone, which has a very unstable structure, producing the opening of the oxygenated hetero-cycle and thus losing a compound derived from delphinidin. A similar decomposition was observed in anthocyanins from black soybean, where TG and high-performance liquid chromatography (HPLC) studies were able to associate the decomposition in the range of 100 to 400 °C of sugar with the cyanin-derived molecule and its decomposition. In the last event, a loss of 25% can be observed, which would correspond to the decomposition of sugar; however, this value is slightly lower than the theoretical value (35%) [60]. This difference may be due to the difficulty of determining each reaction's start and end points since, as observed in the DTG in the range of 100 to 600 °C, there is an excellent overlapping of peaks, indicating the coexistence of decomposition reactions simultaneously. The prominent final exothermic peak observed in DSC could confirm the above, in addition to showing the difficulty in determining the start of the reaction [61,62].
Regarding the DEL3GLU supported on TiO2, the same behavior and similar temperature ranges can be observed. Due to the difficulty of precisely knowing the proportion of anthocyanin adsorbed onto TiO2, the observed variations in DSC and TG can be assigned to a mass effect.    -240  11  3  240-450  14  TiO2 ---  Considering the structure of delphinidin-3-glucoside as a starting compound, it is possible to postulate some decomposition path. Therefore, if we take into account that, in the second event, 43% of the mass is lost, this percentage of loss could be associated with the rupture of the glycosidic bond, resulting in aglycone, which has a very unstable structure, producing the opening of the oxygenated hetero-cycle and thus losing a compound derived from delphinidin. A similar decomposition was observed in anthocyanins from black soybean, where TG and high-performance liquid chromatography (HPLC) studies were able to associate the decomposition in the range of 100 to 400 • C of sugar with the cyaninderived molecule and its decomposition. In the last event, a loss of 25% can be observed, which would correspond to the decomposition of sugar; however, this value is slightly lower than the theoretical value (35%) [60]. This difference may be due to the difficulty of determining each reaction's start and end points since, as observed in the DTG in the range of 100 to 600 • C, there is an excellent overlapping of peaks, indicating the coexistence of decomposition reactions simultaneously. The prominent final exothermic peak observed in DSC could confirm the above, in addition to showing the difficulty in determining the start of the reaction [61,62].

DSSC Characterization
Regarding the DEL3GLU supported on TiO 2 , the same behavior and similar temperature ranges can be observed. Due to the difficulty of precisely knowing the proportion of anthocyanin adsorbed onto TiO 2 , the observed variations in DSC and TG can be assigned to a mass effect.

DSSC Characterization
Once built up, the cells were evaluated via the recording of their J vs. V profiles and use of EIS. The results depend on the sensitizers applied: DEL3GLU or co-mixed with silver nanoparticles (Table 3, Figure 6). Figure 6 shows the current density profiles vs. the applied potential for DSSCs sensitized with different DEL3GLU extracts or after adding silver nanoparticles sequentially. The conversion efficiency is raised in the presence of NP100; thus, a relationship between surface coverage and electron exchange can be corroborated. It is meaningful to mention that DSSCs sensitized only in the presence of NP100 reached efficiency values of 0.013%. This being the case, co-adsorption interestingly increases the measured PCE values. The best efficiency is reached when the photoanode's surface is fully covered (see an example of sensitization on Scheme 2).
Once the anthocyanins adsorb onto the FTO/TiO 2 electrode, some surface defects will remain. Using silver nanoparticles as co-sensitizers covers the remaining bare spots on the FTO/TiO 2 electrode and improves the electron transference from DEL3GLU to the semiconductor, increasing cell performance. Consequently, the current density of DSSCs is raised in the presence of NP100 [63]. Table 3. Photovoltaic properties for DCC cells assembled using different anthocyanin concentrations or extraction solvents. Measured under an intensity of 100 mW cm −2 , AM 1.5 G. Electrodes' active areas: 0.36 cm 2 . Jsc is the current density, Voc is the open-circuit potential, FF is the fill factor, and η is the conversion efficiency. Results come from at least three independent experiments performed for each sensitizer. The estimated error is ± 0.005%.    Cells sensitized with DEL3GLU extracted with ethanol (black line), DEL3GLU extracted using Milli-Q water (blue line), and those co-sensitized using a sequential approach (DEL3GLU first extracted from ethanol and then silver nanoparticles, NP100, red line). Solar 2023, 3, FOR PEER REVIEW 10 Once built up, the cells were evaluated via the recording of their J vs. V profiles and use of EIS. The results depend on the sensitizers applied: DEL3GLU or co-mixed with silver nanoparticles (Table 3, Figure 6). Figure 6 shows the current density profiles vs. the applied potential for DSSCs sensitized with different DEL3GLU extracts or after adding silver nanoparticles sequentially. The conversion efficiency is raised in the presence of NP100; thus, a relationship between surface coverage and electron exchange can be corroborated. It is meaningful to mention that DSSCs sensitized only in the presence of NP100 reached efficiency values of 0.013%. This being the case, co-adsorption interestingly increases the measured PCE values. The best efficiency is reached when the photoanode's surface is fully covered (see an example of sensitization on Scheme 2).

Scheme 2.
Photoanode sensitized with DEL3GLU after overnight immersion in the solution. Table 3. Photovoltaic properties for DCC cells assembled using different anthocyanin concentrations or extraction solvents. Measured under an intensity of 100 mW cm −2 , AM 1.5 G. Electrodes' active areas: 0.36 cm 2 . Jsc is the current density, Voc is the open-circuit potential, FF is the fill factor, and η is the conversion efficiency. Results come from at least three independent experiments performed for each sensitizer. The estimated error is ± 0.005%. More details about the electronic processes occurring at the DSSCs are obtained by applying EIS techniques ( Figure 7). As previously reported, the impedance registers are fitted via the use of equivalent circuits integrated for several elements [64][65][66][67]. Solar 2023, 3, FOR PEER REVIEW 11 Milli-Q water (blue line), and those co-sensitized using a sequential approach (DEL3GLU first extracted from ethanol and then silver nanoparticles, NP100, red line).

DEL3GLU in
Once the anthocyanins adsorb onto the FTO/TiO2 electrode, some surface defects will remain. Using silver nanoparticles as co-sensitizers covers the remaining bare spots on the FTO/TiO2 electrode and improves the electron transference from DEL3GLU to the semiconductor, increasing cell performance. Consequently, the current density of DSSCs is raised in the presence of NP100 [63].
More details about the electronic processes occurring at the DSSCs are obtained by applying EIS techniques ( Figure 7). As previously reported, the impedance registers are fitted via the use of equivalent circuits integrated for several elements [64][65][66][67]. Some particular parameters are useful in understanding the calculated efficiencies (displayed in Table 4): Rce, the resistance associated with the regeneration of I3 − into I − at the counter-electrode.  Some particular parameters are useful in understanding the calculated efficiencies (displayed in Table 4): Table 4. Determined EIS parameters for DSSCs assembled with DEL3GLU (extracted with ethanol) and also co-sensitized with NP100. Measured at 0.5 V in darkness. Results arise from at least three independent experiments performed for each sensitizer. Rce, the resistance associated with the regeneration of I 3 − into I − at the counter-electrode.

EIS
Rct, the charge transport resistance for the recombination of electrons at the TiO 2 /dye/ electrolyte interfaces.
Rt, the electron transport resistance in the photoanode. Γt, the time constant for the transport of the injected electrons that diffuse through the nanoparticle network. Calculated as Γt = Rt × Cµ (Cµ, the chemical capacitance at the TiO 2 /dye/electrolyte interface associated with the variation in the electron density and the displacement of the Fermi level).
Γrec, the recombination time that reflects the lifetime of an electron in the photoanode. Calculated as Γrec = Rct × Cµ.
Recombination implies a process of injecting electrons from the TiO 2 into the electrolyte. Then, once generated, the electrons in the dye would follow an undesired path through the mesoporous titania towards the FTO/TiO 2 interface.
High recombination times and time constants for transporting the injected electrons explain good performance efficiency.
For the analyzed samples, the electron transport times increased in the presence of the nanoparticles, affecting the cell's performance [63,68,69]. Incorporating nanoparticles improves the surface coverage; the electron mobility increases, and the electron transfer increases.
The ratio between the Rct and Rt resistances helps in understanding efficiency values. In both cases (i.e., in the presence or absence of NP100), ratios are high; thus, when electrons are generated, they follow transference across the semiconductor instead of recombining with the electrolyte.
The resistance connected to iodide regeneration at the counter-electrode (Rce) must also be contemplated in order to explain the calculated η values. The adsorption of molecules onto the Pt surface increases the Rce values. In our case, cells assembled with silver nanoparticles displayed higher Rce values. Once NP100 are adsorbed to the counterelectrode, the area of the electrode increases, resulting in a higher surface extent available and necessary for the I 3 transformation into iodide; however, electron conduction through silver is worse than that through platinum, increasing Rce.
In summary, anthocyanins, such as delphinidin-3-glucoside, fulfill the characteristics necessary for them to be utilized as sensitizers in DSSCs, in particular, joined to silver nanoparticles.

Conclusions
The adsorption of delphinidin-3-glucoside extracted from blueberries (Southern Highbush type) and silver rounded nanoparticles to FTO/TiO 2 electrodes was proven from absorbance spectra recorded via the use of a reflection optic probe. FTIR measurements showed molecular interactions between the semiconductor and the delphinidin-3-glucoside through the phenolic -OH moieties of the anthocyanin. DSC and TG also confirmed the adsorption of anthocyanins to the semiconductor.
Additionally, cyclic voltametric measurements showed an anodic contribution at 0.98 V raised from the oxidation of -OH groups of the delphinidin-3-glucoside, enough to assure the electronic transfer inside DSSCs via the use of a sensitizer in the photoelectrode. Additionally, regarding its application as a light harvester in a device exposed to the sun and, therefore, its capacity to warm up, the stability of delphinidin-3-glucoside was evaluated by heating the solutions to 80 • C without any significant change.
Once assembled in DSSCs, the photoelectrodes sensitized with delphinidin or cosensitized with nanoparticles showed high recombination times and time constants for transporting the injected electrons; however, power conversion efficiencies were higher in the presence of NP100, probably due to obtaining a better surface coverage.
The exposed results indicated the possibility of applying delphinidin-3-glucoside extracted from blueberries following a straightforward path as potential impregnation dyes in DSSCs, particularly in the presence of co-adsorbents, such as silver nanoparticles, offering a technological application of discarded and not commercialized fruits for countries where agricultural production provides an abundant origin of extraction.