A Visible-NIR Responsive Dye-Sensitized Solar Cell Based on Diatom Frustules and Cosensitization of Photopigments from Diatom and Purple Bacteria

Diatoms exhibit high solar energy harvesting efficiency due to their remarkably organized, hierarchical micro/nanoporous, lighttrapping, and scattering frustules. At present, few studies focus on cosensitization of natural near-infrared dye to expand the spectral response of dye-sensitized solar cells. In this study, the diatom frustule-TiO2 (12 : 5) composite film was prepared and assembled it on the TiO2 electrode. Compared to the single TiO2 layer film, diatom frustule-TiO2 (12 : 5) composite film sensitized by diatom’s dye showed the conversion efficiency of 0.719%. To expand the light-harvesting response to near-infrared region spectra, the cosensitized dyes were used to fabricate the visible-near-infrared responsive dye-sensitized solar cells. 1e cosensitization diatom frustule-TiO2 (12 : 5) composite film exhibited two distinct absorption bands in the near-infrared region and reached a higher conversion efficiency of 1.321%, which was approximately 1.4 or 1.7 folds higher than that of cosensitization double-TiO2 film or single TiO2 layer film, respectively, and approximately 3.7 or 1.7 folds higher than that of the single TiO2 layer film sensitized by diatom dye or purple bacterial dye, respectively. 1e results showed that the combination between diatom frustule-TiO2 with cosensitization natural dyes could significantly improve the photoelectric performance of visible-near-infrared responsive dye-sensitized solar cells.


Introduction
Dye-sensitized solar cells (DSSC) have attracted considerable attention due to their ease of fabrication, low production costs, and architectural and environmental compatibility compared to silicon-based solar cells [1][2][3][4]. Currently, the dyes used in DSSC mainly fous on synthetic dyes and some plant natural dyes. Synthetic dyes such as N719 and N3 ruthenium are capable of yielding higher conversion efficiencies [5][6][7]. However, their fabrication is complicated, and some are toxic [8]. Natural dyes are abundant, easily available, cost-effective, and eco-friendly [9,10]; however, the conversion efficiency of DSSC was yet to be improved. In general, a challenge in DSSC is the expansion of absorbance range of photoactive layers. To the best of our knowledge, up to 50% of solar spectra are larger than 700 nm, which are located in the range of the infrared radiation [11]; however, the absorption maxima in most of the synthetic or natural dyes in DSSC are within 700 nm [10], and their conversion efficiency in the available DSSCs may be compromised due to limited sunlight absorption of dyes. erefore, more interests are focused on exploring the nearinfrared absorbance dyes or spectral complementary cosensitization dyes, such as synthesized FT89 NIR dye enhanced photocurrent conversion efficiency by 6% compared to N749 benchmark [12]; bacteriochlorophyll c (BChl c) from green sulfur bacterium showed a photocurrent conversion efficiency of 0.1% at 600-800 nm [13]. Cosensitizers (N719, black dye, SQ1, NI5, and porphyrin) could enhance the photoelectric performance of DSSC [14][15][16]. However, few studies focus on natural NIR dye or cosensitized natural NIR dye. e strategy for improving the structure and composition of TiO 2 electrodes is an additional effective approach to enhance the performance of DSSC, such as a double-layer TiO 2 film consisting of transparent nanocrystalline and microcrystalline light-scattering anatase particles was used for photocurrent enhancement of DSSC [17]. Mohammadpour and Janfaza coated nanofibrous as the scattering layer on the TiO 2 film, enhancing photocurrent by 11% [18].
ese results showed that scattering double-layer electrodes exhibited a power performance. Novel structure, such as TiO 2 nanowires or nanotubes, was also used to enhance the performance of DSSC [19,20].
Diatom frustule could be a promising photoelectric device [21][22][23][24]. Previous studies showed that diatom frustule-TiO 2 composites as the working electrode could enhance DSSC efficiency. Toster et al. coated frustules with titania nanoparticles to improve the conversion efficiency of DSSC [22]. Huang et al. made diatom frustule-TiO 2 composite electrodes with a multilayer structural design, enhancing efficiency by 38% [23]. In the above studies, sensitizers used in DSSC are N719 synthetic dyes. It remains unclear whether the diatom frustules could improve photoelectric performance of TiO 2 electrodes sensitized by natural dyes or other dyes.
Our previous study showed that pigment extracts from purple bacteria exhibited the maximal absorption at ∼770 nm (bacteriochlorophyll a) in the NIR region and at 400∼550 nm in the visible region, and diatom pigment extract exhibited the maximal absorption at ∼660 nm (chlorophyll a) [25,26]. e aim of this study was to design diatom frustule-TiO 2 composite and cosensitize it by two near-infrared spectral complementary natural dyes from purple bacteria and diatom with wide spectrum (400 nm∼770 nm) in order to improve the photoelectric conversion efficiency of DSSC.
is study showed that pigment extracts from purple bacteria exhibited the characteristic maximal absorption in the NIR region (at ∼770 nm, bacteriochlorophyll a) and in the visible region (at 400∼550 nm), and the pigment extract from the diatom exhibited the maximal absorption in the visible region (at ∼660 nm, chlorophyll a). e absorption spectra of the three pigment mixtures displayed an expanding absorption region from Vis to NIR (400∼500 nm, ∼660 nm, and ∼770 nm). e aim of this paper was to prepare diatom frustule-TiO 2 composite and cosensitize it by two near-infrared spectral complementary natural dyes from purple bacteria and diatom. e results showed that the combination utilization of diatom frustules and cosensitized dyes greatly enhanced the photoelectric performance of visible-NIR responsive DSSC.
is work highlighted the trends in using the biomolecular photosensitizers with Vis-NIR absorbance or spectral complementary cosensitization characteristics for the dyesensitized solar cell and also demonstrated that natural pigment extracts from diatom and purple bacteria and diatom frustules have the potential to be used as materials for fabricating low-cost and eco-friendly dye-sensitized solar cells.

Extraction of the Pigment and Diatom Frustules.
Pigment was extracted from diatom NA56 following the method as described as follows. In brief, 10 g of wet cells was extracted in 100 mL of acetone and under ultrasonic treatment (pulse 4 s, stop 6 s, 400 W, 90 times) of ice bath for 15 min in dim light. Extract was fractioned by centrifuging, and the supernatant solvent was dried under nitrogen. e dried extract suspended in ethanol was the pigment extract of NA56. e pigment of purple bacteria CQV97 was extracted with an acetone/methanol mixture (7 : 2, V/V) as described previously [27]. In brief, 5 g of wet cells was extracted in 60 mL of the acetone/methanol mixture and under the conditions similar to those used for diatom. e dried extract suspended in ethanol was the pigment extract of CQV97. Pigment extracts were stored at −20°C under dark conditions. e extract of frustules was based on Gell's methods [28]. Deal diatoms with hydrogen peroxide (30 wt%) and hydrochloric acid (2 mol·L −1 ) mixture (V/V � 1 : 1) for 72 h under dark conditions in order to remove carbonates and organic matter. e pellet was rinsed with ethanol two times and widely washed with distilled water to get neutral pH. Diatom frustules were dried for 3 h at 130°C.

Preparation of the TiO 2 Film and Diatom
Frustule-TiO 2 Film. A TiO 2 colloid was prepared as described previously [29]. Tetrabutyltitanate (10 mL) was rapidly added to distilled water (100 mL), and a white precipitate formed immediately. e precipitate was filtered using a glass frit and washed with distilled water. Under vigorous stirring, the filter cake was added to aqueous solution (150 mL) containing 1 mL nitric acid and 10 mL acetic acid at 80°C until the slurry became a translucent blue-white liquid. e bluewhite liquid was autoclaved at 200°C for 12 h to form a milky, white slurry. e resultant slurry was concentrated to 25% of its original volume, and then PEG-20000 (0.80 g) and a few drops of the Triton X-100 emulsification reagent were added to form a TiO 2 colloid, labeled T 0 . e TiO 2 colloid was coated on the fluorine-doped tin oxide-coated glass plate using a doctor scraping method followed by sintering at 450°C for 30 min in air. TiO 2 film was immersed in 50 mM TiCl 4 aqueous solution at 70°C for 30 min and sintered at 450°C for 30 min in air again. e film was labeled T, and the thickness and area of it were controlled to be about 8 μm and 0.1 cm 2 , respectively. e procedures used for the diatom frustule-TiO 2 colloid were similar to those for the TiO 2 colloid. 0.40 g, 0.80 g, 1.20 g, 1.60 g, and 0.80 g diatom frustules were added to distilled water (20 mL), respectively, and sonicated for 30 min. Subsequently, tetrabutyltitanate (2 mL) was rapidly added (except for the fifth group) to distilled water containing diatom frustules, respectively, and followed procedures were the same with preparation of TiO 2 colloid. e colloids were labeled F 1 , F 2 , F 3 , F 4 , and F 0 in turn. en, T 0 (without diatom frustules), F 1 , F 2 , F 3 , F 4 , and F 0 (without TiO 2 ) colloids were individually coated on the T film. After sintering at 450°C for 30 min in air, double-layer films were obtained and labeled TT (cosensitization double TiO 2 film), TF1(diatom frustule-TiO 2 (4 : 5) composite film), TF2 (diatom frustule-TiO 2 (8 : 5) composite film), TF3 (diatom frustule-TiO 2 (12 : 5) composite film), TF4 (diatom frustule-TiO 2 (6 : 5) composite film), and TF0 (diatom frustules without the TiO 2 film) in sequence.

Different Concentrations of the Pigment Extract-Sensitized TiO 2 Film.
e concentration of NA56 or CQV97 pigment extracts was determined by the amount of Chl a or BChl a, respectively. e single TiO 2 layer films (T) were placed in five concentrations of NA56 (with Chl a concentration of 19, 38, 155, 310, and 620 μg·mL −1 ) and CQV97 (with BChl a concentration of 9, 36, 144, 288, and 576 μg·mL −1 ) pigment extracts at 4°C for 24 h, respectively.

Cosensitization.
e T films cosensitized by NA56 and CQV97 pigment extracts were prepared following two methods. (i) e cocktail process, in which the T films were immersed into the mixture of NA56 and CQV97 pigment extracts (V/V, 1 : 1, at the concentration optimized) at 4°C for 24 h. (ii) e step-by-step process, in which the T films were first immersed in the NA56 pigment extract for 0 h, 2 h, 6 h, 12 h, 18 h, 22 h, and 24 h, respectively, and subsequently transferred into the CQV97 pigment extract and incubated for 24 h, 22 h, 18 h, 12 h, 6 h, 2 h, and 0 h in sequence.

Diatom Frustule-TiO 2 Film Sensitized by the NA56 Pigment Extract.
e TT, TF1, TF2, TF3, TF4, and TF0 films were immersed in the NA56 pigment extract at 4°C for 24 h. TF3 with a better photoelectric performance and its thickness were next optimized. Different amounts of the F 3 colloid were coated on the T film (8 μm) and sintered at 450°C for 30 min, and TF3 films were controlled with the thickness of 12 μm, 14 μm, 16 μm, and 18 μm, respectively. Subsequently, they were immersed in NA56 pigment extracts at 4°C for 24 h.

Fabrication of DSSC.
e pigment-sensitized T and TF electrodes and Pt counter electrode were clipped together, and cyanoacrylate adhesive was used as sealant. e composition of the electrolyte was 0.1 M I 2 , 0.1 M LiI, 0.6 M tetrabutylammonium iodide, and 0.5 M 4-tertbutylpyridine in acetonitrile.

Characterization and Measurement.
e absorption spectra of pigment extracts and dye-sensitized electrodes were recorded from 300 to 900 nm using the UV-3200PCS spectrophotometer. e morphologies of diatom frustules and diatom frustule-TiO 2 composites were observed by a field emission scanning electron microscope (FESEM) (S-4800, HITACHI). e photoelectric test of DSSC was performed by measuring photocurrent-photovoltage (I-V) characteristic curves at room temperature, white light of 100 mW·cm −2 ·t (AM1.5) irradiated from a solar simulator, and a CHI660D electrochemical measurement system. According to the I-V curves, the relevant parameters were measured, and the efficiency was calculated. e filling factor (FF) and conversion efficiency (η) of DSSC were calculated from the I-V curve, and equation FF � P opt /(I sc × V oc ) � (I opt × V opt )/ (I sc × V oc ), η � P opt /P in � (FF × I sc × V oc )/P in . In the above equation, I sc represents the short-circuit current, I opt represents the optimal photocurrent, V oc represents the opencircuit voltage, V opt represents the optimal photovoltage, P opt represents the optimal output power of the battery, and P in represents the sunlight input power. Figure 1(a) shows the absorption spectra of NA56 and CQV97 pigment extracts dissolved in ethanol. NA56 pigment extract mainly consisting of Ch a and Car exhibits absorption peaks in the visible region at approximately 410 nm (Ch a), 666 nm (Ch a), and 450 nm (Car). CQV97 pigment extract mainly consisting of BCh a and Car shows absorption peaks in the UV-NIR region at approximately 368 nm (BCh a), 598 nm (BCh a), 774 nm (BCh a), and 475 nm (Car). Cosensitizer of NA56 and CQV97 pigment extracts exhibits a wider spectral response range than the individual pigment extract. Figure 1(b) shows the absorption spectra of pigments adsorbed on the TiO 2 film. Compared with Figure 1(a), the characteristic absorption peaks in Figure 1(b) were redshifted. e absorption spectra of the cosensitizer on the TiO 2 film were expanded within the Vis to NIR region and displayed both characteristic absorption peaks of the pigment extract from NA56 and CQV97.

Characterization of Pigment Extracts.
is might be an important reasonable reason for improving photoelectric properties of the cosensitizer on the TiO 2 film. Figures 2(a) and 2(b), η and I sc increase from 19 to 155 μg·mL −1 of Chl a and 9 to 144 μg·mL −1 of BChl a, respectively, but decrease with further increase in   concentration. Dye in the film will build up with increasing concentration, and an increase in η and I sc would be expected. Dye aggregation, however, becomes more serious in higher concentration, which may lead to intermolecular quenching of photo-excited states or molecules residing in the system that are not functionally attached to the TiO 2 surface, causing the decrease of η and I sc [30]. And V oc fluctuates slightly with the change of concentration. e highest η and I sc values of NA56 and CQV97 pigment extracts are 0.368%, 1.139 mA·cm −2 and 0.417%, 0.950 mA·cm −2 at the concentration 155 μg·mL −1 (Chl a) and 144 μg·mL −1 (BChl a), respectively.

Cosensitization of NA56 and CQV97 Pigment Extracts.
e photovoltaic parameters of the cosensitizer sensitized by the T film are shown in Table 1. A slight increment of conversion efficiency (0.432%) is observed in the pigment extract mixture-cosensitized DSSC compared to individual pigment extract-sensitized DSSC (0.420%, 0.372%).
Remarkably, the photoelectric performance is greatly enhanced by the cosensitization through stepwise procedure. e maximum values of η and I sc reach up to 0.795% and 1.913 mA·cm −2 when T films are immersed in the NA56 pigment extract for 12 h and then in the CQV97 pigment extract for 12 h. η exceeds NA56 and CQV97 extract-sensitized DSSC (0.372%, 0.420%) by 114% and 89%, respectively. Such increase is possibly attributed to higher lightharvesting capacities of the cosensitizer, which is in agreement with the UV-Vis spectra (see Figure 1) of individual pigment extracts and cosensitizer. After absorption on the TiO 2 film, both individual pigment extracts and cosensitizer show characteristic absorption spectra with redshift when compared to their absorption spectrum of dye solution, and cosensitizer demonstrates absorption bands in the visible region at ∼400 nm and ∼670 nm and the NIR region at ∼770 nm, with a highly broadened spectral response range. Furthermore, cosensitizer exhibits a great potential to enhance light-harvesting capacities if compared to individual dyes. As shown in Table 1, DSSC cosensitized by stepwise procedure are more efficient than the mixturesensitized DSSC. When NA56 and CQV97 extracts are mixed, unfavorable intermolecular interaction such as dye aggregation could occur which restricts the utilization of original properties of individual dyes for efficient photoenergy conversion process [31]. Stepwise cosensitization could retard the charge recombination and decrease aggregation of the dye absorbed on the TiO 2 film to further improve the device performance [32]. Table 2, DSSC with the double-layer film are more efficient (except for TF0) than the single layer. e conversion efficiency of the NA56 pigment extract-sensitized TT, TF1, TF2, TF3, and TF4 film is 0.520%, 0.641%, 0.706%, 0.719%, and 0.522% in sequence, exceeding that sensitized T film by 41%, 74%, 92%, 95%, and 42%, respectively. e photocurrent of DSSC firstly increases and then decreases with the increasing diatom frustules mass ratio of the second layer ( Figure 3); a similar trend is seen in their conversion efficiency. However, under the same conditions, V oc slightly changes.

e Photoelectric Performance of the NA56 Pigment Extract-Sensitized Diatom Frustule-TiO 2 Film. As shown in
e maximum values of η and I sc reach up to 0.719% (exceeding T and TT films by 95% and 38%) and 2.006 mA·cm −2 (exceeding T and TT films by 76% and 41%) when the mass ratio of diatom frustules/TiO 2 of the second layer is 12 : 5 (TF3 film). e improvement for the cell incorporating diatom frustules is possibly attributed to enhanced light scattering and trapping. When light strikes diatom frustules, the pores cause multiple reflections thereby increasing the probability for photons to be absorbed by the dye and promoting injection into the semiconductor [22,23]. As shown in Figure 4, after the NA56 pigment extract is absorbed on TiO 2 and frustule-TiO 2 film, both of them show absorption spectra with redshift as compared to that of dye solution. And the spectral response range on the frustule-TiO 2 film is wider than that on the TiO 2 film, which may be a reason why the diatom frustules incorporated into TiO 2 improved the photoelectric performance of DSSC.
Diatom frustules consist of two halves with many small holes on their surface, approximately 5 μm in width and 20 μm in length ( Figure 5(a)). e small holes of the external surface with diameter of approximately 200 nm ( Figure 5(b)) and internal surface of frustules ( Figure 5(c)) have a honeycomb structure with smaller holes (with diameter of approximately 50 nm). e results above were basically consistent with the previous study [33]. e SEM images (Figures 5(d) and 5(e)) of diatom frustule-TiO 2 materials show that TiO 2 nanoparticles successfully coated the surface of diatom frustules, and the diameter of TiO 2 particles is about 20 nm. TF3 thickness is optimized by attaching the second layer diatom frustule-TiO 2 film thickness (the first layer is T film with 8 μm thickness). e correlation between thickness and photoelectric performance parameters is shown in Figure 6 η and I sc increase within the TF3 film thickness range between 12 and 14 μm but decrease with further increase in thickness. e charge recombination between electrons injected from the excited dye to the conduction band of TiO 2 and the I3 ions in the electrolyte will, however, become more serious in thicker films, which leads to the decrease of η and I sc [34]. However, decreased V oc is attributed to the charge recombination and mass transport limitations in the thicker film.

e Photoelectric Performance of Different Dye-Sensitized TT and TF3 Film.
e above results show that diatom frustules incorporated into TiO 2 enhanced the photoelectric performance of the NA56 pigment extract-sensitized solar cell. Herein, to test whether the frustules can improve photoelectric performance of DSSC sensitized by other dyes, we further investigate the photovoltaic parameters of the CQV97 pigment extract, cosensitizer (NA56 and CQV97 pigment extracts), and N719-sensitized TT and TF3 films. As shown in Table 3, CQV97 pigment extract, cosensitizer (NA56 and CQV97 pigment extracts), and N719-sensitized TF3 films show conversion efficiency of 0.976%, 1.321%, and Journal of Chemistry 5.844%, enhancing efficiency by 42%, 32%, and 34% compared with that of the sensitized TT film, demonstrating diatom frustule-TiO 2 composites made in this work possibly improve photoelectric performance of solar cells sensitized by other dyes. is result was basically consistent with the previous study [23]. e conversion efficiency of cosensitizer (NA56 and CQV97 pigment extracts) sensitized by the TF3 film reaches up to 1.321%, exceeding that (0.998%, 0.520%, and 0.687%) of cosensitizer and individual dye-sensitized TT film by 32%, 154%, and 92%, respectively; furthermore, it enhances efficiency by 259% and 217% when compared with that (0.368% and 0.417%) of the individual dye-sensitized T film, respectively. It is more efficient than most of the reported chlorophyll-and carotenoid-sensitized solar cells.     8

Journal of Chemistry
Our results further demonstrated that combining cosensitizer with TiO 2 electrode incorporating diatom frustules greatly improved the photoelectric performance of DSSC.

Conclusion
In conclusion, our results showed that cosensitizer-(NA56 and CQV97 pigment extracts-) sensitized TiO 2 film exhibited absorption bands in the visible region at ∼400 nm and ∼670 nm and the NIR region at ∼770 nm, expanding absorption spectra of individual dyes and yielding conversion efficiency of 0.795%, exceeding that of NA56 and CQV97 pigment extract-sensitized TiO 2 film by 114% and 90%, respectively. e NA56 pigment extract, CQV97 pigment extract, cosensitizer, and N719-sensitized solar cells incorporating diatom frustules (TF3 films) enhanced efficiency by 38%, 42%, 32%, and 34% compared with that of sensitized TT films, respectively. e conversion efficiency was further promoted (up to 1.321%) when a cosensitizersensitized TF3 film was created. It was more efficient than most of the available chlorophyll-and carotenoid-sensitized solar cells. Pigments of diatom and purple bacteria are easily extracted and eco-friendly. Diatom frustule is a structure designed by nature with the ability of trapping and scattering light. Apparently, the above results show that natural pigment extracts of diatom and purple bacteria and diatom frustules have the potential to be used as materials for fabricating low-cost and eco-friendly dye-sensitized solar cells.

Data Availability
e data used to support the findings of this study are enclosed within the article. Additional data are accessible from the corresponding author upon request.

Conflicts of Interest
e authors declare that there are no conflicts of interest regarding the publication of this paper.