Introduction

Dye-sensitized solar cells (DSSC) have gained worldwide attention for many years due to low production cost and environmentally friendly operation. The operation principle of DSSC is similar to the photosynthesis, a natural process. Here, the device is capable of generating energy by converting the absorbed Sunlight into electrical energy. Generally, a DSSC is composed of a mesoporous metal oxide semiconductor, a dye sensitizer, an electrolyte containing iodide and triiodide ions and a counter electrode1. In DSSC, sensitizers play a key role in harvesting the sunlight and then transforming it into electrical energy. Numerous metal complexes and organic dyes have been synthesized and utilized as sensitizers. Yet, ruthenium based synthetic organic dyes are found to be effective sensitizers. By far, the highest efficiency of over 11% has been reported for DSSCs sensitized by Ru based N719 dye2. However, the preparation routes for metal complexes are often based on multi-step procedures involving tedious and expensive chromatographic purification procedures3. It is anticipated that replacing synthetic organic dyes with natural pigments, such as chlorophyll and anthocyanin, could resolve the above limitations as they can be easily extracted from the fruits, leaves, roots and flowers of plants.

Generally, many of the plant parts contain chlorophyll and anthocyanin pigments. Chlorophyll is the most abundant pigment in green plants and each chlorophyll molecule possesses a Mg2+ ion surrounded by four pyrrole rings, one of which is bonded to a phytol tail4. Chlorophyll molecules are described as photoreceptors due to their light absorbing property. There are two types of chlorophyll, namely chlorophyll a and chlorophyll b, which differ in their structures at the C3 position of one of the pyrrole rings. The C3 position of the said pyrrole ring in chlorophyll b contains a formyl (–CHO) side chain whereas a methyl (–CH3) group is present at the same position in chlorophyll a5. Due to the presence of different substituents, the chlorophyll a and chlorophyll b molecules exhibit varied light absorption properties. Hence, chlorophyll absorbs light in a wide wavelength range corresponding to blue, red and violet regions of the visible spectrum6. Anthocyanin is another pigment responsible for the variety of colours in petals of flowers and fruits. Employing anthocyanin as dye for DSSC leads to absorption of light in the blue-green region of the solar spectrum7 and the carbonyl and hydroxyl groups present in anthocyanin molecules demonstrate efficient anchoring to the TiO2 surface (photoelectrode), thus enabling an effective electron injection mechanism in DSSC8.

Commonly, natural dyes contain either chlorophyll or anthocyanin and rarely contain both. Studies focusing on enhancing PV performance of the fabricated devices with dye mixtures comprising of both pigments have been reported. Wongcharee et al.9 fabricated three different types of DSSCs with natural dyes extracted from rosella and blue pea and a mixture of the said dye extracts. The light absorption spectrum of the dye mixture showed peaks corresponding to the individual natural pigments present in rosella and blue pea. However, the dye mixture adsorbed on TiO2 did not exhibit synergistic light absorption and photosensitization effect compared to the individual constituent dyes9. In a different study, Sengupta et al.10 has reported that a mixture of chlorophyll and betalain dyes, extracted from fresh spinach leaves and beetroots respectively, achieves photovoltaic performance of the device upto 0.29% due to light absorption in a wider range of solar spectrum10. In a separate study, Park et al.11 has demonstrated mixing the dyes, extracted from the flowers of Gardenia Jasminoide Elli with two different colours (yellow and blue), results in widening the light absorption wavelength compared to the individual dyes, thus improving the PV performance of the fabricated device11.

For the present study, we have chosen the flower of Mussaenda erythrophylla, which contains both chlorophyll and anthocyanin, as a natural sensitizer for DSSCs. Mussaenda erythrophylla is a rambling shrub, grows best in warmly temperate or subtropical regions and is semi deciduous in cooler parts12. The flower of the plant possesses much smaller flowers in the centre and is available in a wide variety of colours, including red, rose, white and pale pink. The flowers bloom for most of spring through summer. Since this plant is widely found in Sri Lanka, the suitability of the dye extracted from the flowers of the plant in DSSC application was investigated. The study involved extraction of the dye from the flowers of Mussaenda erythrophylla, its phytochemical analysis, optical, structural and electrochemical characterizations and evaluation of PV performance of the corresponding devices.

Experimental

Materials

Absolute ethanol (> 99%), Triton X-100 (laboratory grade), di-tetrabutylammonium cis-bis (isothiocyanato) bis (2,2’-bipyridyl-4,4’-dicarboxylato) ruthenium (II) dye (N-719, 95%), acetonitrile (gradient grade), tert-butyl alcohol (≥ 99.7%) and titanium dioxide nanopowder (21 nm primary particle size, ≥ 99.5% trace metals basis) were purchased from Sigma–Aldrich, Oslo, Norway. Acetylacetone (≥ 99.5%) was purchased from Fluka Analytical, Munich, Germany. All the materials were used without further purification unless otherwise stated.

Characterization

The optical absorbance spectra were recorded using Shimadzu 1800 scanning double beam UV–Visible spectrophotometer. The structural properties of the dye coated films were studied by fourier transform infrared spectroscopy (Thermo Scientific™ Nicolet™ iS5 FTIR spectrometer) and atomic force microscopy (AFM-Park XE7, in 1 × 1 µm scan area). pH of the prepared dye solutions was measured using Eutech pH 700 m. The photovoltaic performance of the fabricated devices with effective area of 0.25 cm2 was studied using Keithley-2400 source measurement unit under simulated irradiation by 150 W Xe lamp with an intensity of 100 mWcm−2 and AM 1.5 filter (Peccell-PEC-L12, Kanagawa, Japan). All electrochemical studies were carried out using Bio Logic SP-150 potentiostat.

Methodology

Natural dye extraction

The natural dye was extracted from the flower of Mussaenda erythrophylla. Initially, the fresh flowers of Mussaenda erythrophylla were collected from the Thirunelvely premises of University of Jaffna, Jaffna, Sri Lanka and washed with tap water followed by de-ionized water (DI-water) to remove the dust particles (Step 1). Then, the cleaned flower petals were allowed to dry at a muffle furnace for 2 days at 60 °C (Step 2). The dried petals were crushed into powder using an electrical blender (Step 3). 1 g of the dried petal powder was soaked in 10 mL of two different solvents (ethanol and DI-water) separately for 24 h. Then, the solutions were filtered separately using Whatman No.1 filter paper (Step 4). The individual filtrates (natural dye samples) were stored in airtight containers at room temperature and used as sensitizers without further purification (Step 5) (Fig. 1). All procedures involving plant materials were performed in compliance with the relevant ethical standards and institutional and/or national guidelines/regulations/legislation.

Figure 1
figure 1

Schematic representation of the procedures followed during extraction of the natural dye.

Full size image

Phytochemical analysis

The natural dye samples, extracted separately in ethanol and DI-water, were heated at 60 °C to remove the unwanted volatile substances and the resultant residues were stored appropriately. The said residues of 162 mg from ethanol extract and 160 mg from DI-water extract were again dissolved separately in 20 mL of the respective solvents and the phytochemical analysis was carried out by adopting the standard procedures (Table S1) stipulated elsewhere13,14. The results of phytochemical analysis are given in Table 1 and Fig. S1.

Table 1 Phytochemical analysis of Mussaenda erythrophylla flower extracted in ethanol and DI-water.
Full size table

Device fabrication

The Fluorine doped Tin Oxide (FTO) coated glass substrates with surface resistivity of 7.5 Ωcm–2 were cleaned in ultrasonic bath for 10 min by treating sequentially with soap water, DI-water and ethanol. 100 mg of P25-TiO2 was ground with DI-water, acetylacetone and Triton X-100 using an agate motor with pestle to make semisolid TiO2 paste. Subsequently, the prepared TiO2 paste was coated on the previously cleaned FTO glass by doctor blading method and calcinated at 500 °C for 30 min to obtain TiO2 coated thin films. Then, the said films were soaked separately in the natural dye samples extracted in ethanol and DI-water for 12 h. Afterwards, the dye coated thin films were rinsed with the respective solvents and dried. The platinum (Pt) coated glass substrate was assembled with the individual dye-coated photoanode as the counter electrode. Finally, a small amount of \({I}^{-}\)/\({I}_{3}^{-}\) electrolyte was injected in between the dye coated photoanode and Pt counter electrode to complete the DSSC fabrication.

Results and discussion

Phytochemical analysis

The dried flower petal powder of Mussaenda erythrophylla was soaked in ethanol and DI-water separately and the corresponding natural dye samples were extracted and subjected to quantitative phytochemical analysis which revealed the presence of phytochemicals as displayed in Table 1.

As per Table 1, both (ethanol and DI-water) extracts of Mussaenda erythrophylla flower contain terpenoids, glycosides, alkaloids, phenols, quinones, and coumarins in different quantities. Also, the above phytochemical analysis reveals that flavonoids are absent in the DI-water extract although presence of flavonoids in the flowers of Mussaenda erythrophylla has been reported in the literature15. Anthocyanin is one of the pigments responsible for the colour of flower petals and belongs to the family of flavonoids16. It consists of phenolic-OH group (hydroxyl group bonded directly to an aromatic hydrocarbon ring) and frequently occurs in plants as glycosides (bound to sugar groups)17. Since the phytochemical analysis of the DI-water extract of Mussaenda erythrophylla flower demonstrates the presence of phenols and glycosides, it could be assumed that modified anthocyanin pigments may be found in the said extract.

Optical characterization

UV–Visible spectroscopic analysis

The optical properties of both dye extracts and the corresponding dye coated TiO2 films were analyzed by UV–Visible spectroscopy. Prominent optical properties were observed for the dye extracted in ethanol compared to DI-water. The peak observed at 533 nm for the ethanol extract as shown in Fig. 2 corresponds to the light absorption by anthocyanin molecules18 which is in good agreement with the phytochemical study. In addition, an intense absorption peak at 665 nm corresponding to the n to π* transition in chlorophyll a molecule was observed for the ethanol extract19,20. The observed absorption peak of chlorophyll a is dominant over the absorption peak of anthocyanin. Overall, the UV–Visible spectrum of the ethanol extract confirms that chlorophyll a and anthocyanin pigments are successfully extracted in ethanol from the Mussaenda erythrophylla flower petals.

Figure 2
figure 2

UV–Visible spectra of the dye extracted from Mussaenda erythrophylla flower using ethanol and DI- water.

Full size image

But, the DI-water extract did not show any peak in the regions corresponding to anthocyanin and chlorophyll a absorptions. Hence, it is expected that the natural dye extracted from Mussaenda erythrophylla flower in ethanol would exhibit better PV performance than the dye extracted in DI-water.

Dye stability and sensitization

A dye with long term stability is preferred in DSSC applications. Usually, the stability of natural dyes decreases with time as they undergo biodegradation and photo-oxidation. Hence, stabilities of the natural dye samples extracted from Mussaenda erythrophylla flower in ethanol and DI-water were determined in the present study by keeping the respective dye solutions in the dark at room temperature and measuring their light absorbance intensities, using UV–Visible spectrophotometer, periodically for 50 days (Fig. 3).

Figure 3
figure 3

Time dependent UV–Visible spectra of natural dye extracted from Mussaenda erythrophylla flower in (a) ethanol (b) DI-water.

Full size image

The time dependent UV–Visible spectrum of the dye extracted in ethanol did not show any significant change with time, except a reduction in intensity of the peak at 665 nm corresponding to chlorophyll a. In contrast, the said spectrum of the dye extracted in DI-water exhibited increased light scattering with time which may be due to fungal growth. Hence, the stability study suggests that the dye extracted in ethanol is more stable than the dye extracted in DI-water.

When employing a dye sensitizer in photovoltaics, optimizing its sensitization duration is essential. In order to accomplish this, the TiO2 coated films were dipped in ethanol and DI-water extracts of Mussaenda erythrophylla flower separately and their light absorption properties were studied at regular time intervals using UV–Visible spectrophotometer.

As illustrated in Fig. 4, a monolayer of dye molecules formed on the TiO2 surface upon 12 h of sensitization in both extracts. Interestingly, TiO2 films coated with the dye samples extracted in ethanol and DI-water showed similar peaks in the UV–Visible spectra. Further, greater light absorption capabilities were exhibited by both dye coated TiO2 films compared to the bare TiO2 film. In addition, UV–Visible spectroscopic measurements of the dye solutions, taken before and after dipping the TiO2 films in the said dye solutions, demonstrated a reduction in their light absorption capabilities, thus confirming that dye molecules had been adsorbed on the TiO2 surface while dipping (Fig. S2).

Figure 4
figure 4

Time dependent UV–Visible spectra of TiO2 films dipped in the dye extracted from Mussaenda erythrophylla flower in (a) ethanol (b) DI-water.

Full size image

Structural characterization

FTIR spectroscopic analysis

It is noteworthy to mention here that strong anchoring between the functionalized groups of the dye sensitizers and the surface of the TiO2 nanoparticles is vital for injection of electrons from the excited state (ES) energy level of the dye molecule into the conduction band (CB) of TiO2 molecule, thus producing highly efficient DSSCs. The FTIR spectroscopic analysis is widely used to investigate such anchoring in dye coated TiO2 films. In the present study, bare and dye coated TiO2 nanoparticles were detached from the corresponding FTO glasses and subjected to FTIR measurements within the 400 to 4000 cm–1 wavenumber range and the results are illustrated in Fig. 5.

Figure 5
figure 5

FTIR spectra of (a) bare TiO2 (b) TiO2 sensitized with dye extracted in DI-water and (c) TiO2 sensitized with dye extracted in ethanol.

Full size image

The FTIR spectra of dye coated TiO2 films indicate the presence of the same sensitizers in both dye extracts with a few shifts in wavenumber and trivial deviations in the transmittance level (Fig. 5).

As displayed in Table 2, the IR spectrum of the dye extracted in ethanol showed peaks at 3343, 2922, 2845, 1353 and 1065 cm−1 corresponding to stretching vibrations of O–H, C–H (asymmetric), C–H (symmetric), C–N and C–O groups, respectively while the C–N and C–H (symmetric) stretching vibrations were not observed in the IR spectrum of the dye extracted in DI-water. It should be noted that the porphyrin ring in chlorophyll a molecule is responsible for the appearance of peak corresponding to C–N stretching vibration in the IR spectrum; hence it is confirmed that the dye extracted in ethanol contains chlorophyll a molecules which is in consistent with the findings of UV–Visible spectroscopic analysis (Fig. 2).

Table 2 IR absorptions of functional groups in Mussaenda erythrophylla flower dye adsorbed on TiO2 films.
Full size table

As illustrated in Fig. 6, chlorophyll a molecule attaches to the TiO2 surface through the carbonyl group of porphyrin ring by C=O⋅ ⋅ ⋅ TiO2 coordination25 and anthocyanin molecule anchors to the TiO2 molecule with its carbonyl and hydroxyl groups26. However, the presence of such anchoring between the TiO2 surface and the sensitizers (chlorophyll a and anthocyanin) could not be confirmed from the obtained IR spectral data.

Figure 6
figure 6

Bonding of TiO2 molecule with (a) chlorophyll a molecule and (b) anthocyanin molecule.

Full size image

Atomic force microscopic (AFM) analysis

The distribution of dye molecules on the TiO2 surface influences the PV performance of the DSSC. In the present study, distribution of dye in the prepared dye coated TiO2 films was analyzed and compared with the bare TiO2 film by AFM.

As shown in Fig. 7, a high degree of roughness was observed on the surface of bare TiO2 film. The root mean squares of roughness of bare TiO2, TiO2 sensitized with dye extracted in DI-water and TiO2 sensitized with dye extracted in ethanol were found to be 16.22, 10.46 and 9.46 nm, respectively. The reduction in TiO2 surface roughness upon sensitization with the dye could be attributed to filling of pores on the TiO2 surface by the said dye molecules. Also, the uniform distribution of colour in the images of dye sensitized TiO2 films indicates that the dye molecules are uniformly dispersed over the entire TiO2 surface.

Figure 7
figure 7

2D and 3D Topographic images of (a) bare TiO2 film (b) TiO2 film sensitized with dye extracted in DI-water and (c) TiO2 film sensitized with dye extracted in ethanol.

Full size image

Electrochemical studies

Since the dye extracted from Mussaenda erythrophylla flower in ethanol demonstrated more suitable optical and structural characteristics for PV application than the dye extracted in DI-water, the cyclic voltammetry (CV) performance of the DSSC fabricated with the said dye sensitizer extracted in ethanol was analyzed. The CV was measured by employing a nickel foam dipped in the dye extracted in ethanol, Pt electrode, Ag(s)/AgCl(s) electrode, and 2 M KOH(aq) as working electrode, counter electrode, reference electrode and electrolyte respectively27 and CV of bare Ni foam was also measured for comparison.

As shown in Fig. 8a, the oxidation \(({E}_{oxd}^{onset}\)) onset potential of the ethanol extract was determined from the intersection of tangent between the rising current and the baseline charging current of the respective CV curve. The calculated value for \({E}_{oxd}^{onset}\) was 0.49 eV. Further, energies of the highest occupied molecular orbital (\({E}_{HOMO}\)) and the lowest unoccupied molecular orbital (\({E}_{LUMO}\)) were calculated using the following equations as reported elsewhere, where a correction factor of 4.4 eV was incorporated for the Ag(s) /AgCl(s) reference electrode28,29,30 and the optical band gap (\({E}_{g}^{opt}\) = 2.20 eV) was obtained from Tauc plot (Fig. S3).

Figure 8
figure 8

(a) Cyclic voltammograms of Ni foam with and without the dye in ethanol extract (b) Nyquist plots (c) Bode plots and (d) Transient photocurrent—Time (Jphoto–t) profile of DSSCs with and without sensitization by the dye in ethanol extract.

Full size image
$${E}_{HOMO}=-\left({E}_{oxd}^{onset}\right)+\left(-4.4\right),$$
$${{E}_{LUMO}=E}_{HOMO}+{E}_{g}^{opt}.$$

The calculated values for \({E}_{HOMO}\) and \({E}_{LUMO}\) were – 4.89 and – 2.69 eV respectively. Since the literature value for conduction band (CB) energy of TiO2 is nearly – 4.3 eV31, obviously the energy of LUMO of the dye extracted from Mussaenda erythrophylla flower in ethanol (– 2.69 eV) is greater than the CB energy of TiO2. Therefore, injection of electrons from the photoexcited dye molecule to the CB of TiO2 molecule is possible. Moreover, regeneration of the oxidized dye sensitizer during the operation of DSSC is possible when HOMO of the dye is lower than redox potential of the iodide/triiodide couple. In this study, energy in the HOMO (– 4.89 eV) of ethanol extract is slightly lower than the redox potential of the iodide/triiodide couple (– 4.8 eV)32 compare to general metal complex sensitizers. This may be the reason it can’t achieve the PCE of the device comparable to the synthetic metal complex-based dyes.

Further, the charge transport properties at the interface of the fabricated DSSC were evaluated by electrochemical impedance spectroscopy (EIS)33,34,35. The EIS measurements were recorded at the frequency range from 10–2 to 106 Hz with 0.7 V applied bias voltage under dark condition. The impedance at each interface was determined, after fitting the EIS data in the ZView software, in terms of an appropriate equivalent circuit which is inserted in Fig. 8b. The device, sensitized by the dye in ethanol extract, exhibited series resistance at the FTO/TiO2 interface (Rs), charge transport resistance at the electrolyte/counter electrode interface (Rct), and recombination resistance at the TiO2/dye/electrolyte interface (Rrec) along with the time taken for recombination (\({\mathrm{T}}_{{\varvec{e}}}\)) which were quantified as 14.63 Ω, 1.60 Ω, 85.52 Ω and 54.88 ms. Moreover, the device without the dye sensitizer demonstrated Rrec and \({\mathrm{T}}_{{\varvec{e}}}\) values of 2.29 Ω and 42.78 ms respectively (inserted Fig. 8b and c). The very low Rrec value for the device without dye sensitizer indicates a high rate of charge recombination.

To understand the photovoltaic response stability of the device sensitized by the natural dye in ethanol extract with time, the transient photocurrent—time (\({\mathrm{J}}_{\mathrm{photo}}-\mathrm{t})\) profile was recorded and is displayed in Fig. 8d. The photocurrent was provided at regular time intervals in response to a light on–off sequence produced by opening and closing a mechanical shutter manually that blocks the light beam (1 sun illumination). The curve representing the DSSC with the natural dye sensitizer showed the stability of photocurrent with no significant loss or decay during the illumination period of 60 s compared to the device without sensitizer. This shows that the dye regeneration process is very fast and analogous to injection of charge carriers to the CB of TiO2 by the excited dye molecules. All the electrochemical studies suggest that this natural dye is a potential candidate to be employed as a sensitizer in DSSC application.

PV performance

Influence of solvents employed for dye extraction

The PV performances of the optimized photoelectrodes were analyzed under the illumination intensity of 100 mWcm−2 with Air Mass (AM) 1.5 filter and effective device area of 0.25 cm2. Figure 9 shows J-V curves of the P25-TiO2 based DSSCs sensitized with the natural dye extracted from Mussaenda erythrophylla flower in ethanol and DI-water and the control device. The power conversion efficiency (η) expression can be written as follows:

Figure 9
figure 9

Photovoltaic performances of DSSCs with photoanodes sensitized by the natural dye extracted in ethanol and DI-water under illumination intensity of 100 mWcm–2 with AM 1.5 filter.

Full size image
$$\eta =\frac{{J}_{sc} \times {V}_{oc} \times FF}{{P}_{in}} \times 100 \%.$$

The short circuit current density, \({J}_{SC}\) of the device, is defined as the current density passes through the device when the applied voltage is zero. Open circuit voltage, \({V}_{OC}\), is the voltage developed in the device when the current passes through the device is zero. \({P}_{in}\) is the intensity of light irradiation and FF is the fill factor of the device which is defined as

$$FF = \frac{{J}_{m} \times {V}_{m} }{{J}_{sc} \times {V}_{oc}},$$

where \({J}_{m}\) is the current density at peak power and \({V}_{m}\) is the voltage at peak power. The values of all photovoltaic parameters obtained from the J-V curves are summarized in Table 3.

Table 3 Photovoltaic parameters of DSSCs with photoanodes sensitized by the natural dye extracted in ethanol and DI-water under illumination intensity of 100 mWcm–2 with AM 1.5 filter.
Full size table

As per Table 3, relatively higher JSC, VOC, FF and η values of 0.72 mAcm–2, 0.58 V, 0.64 and 0.27% respectively were observed for the device sensitized by the natural dye extracted in ethanol while the same device structure sensitized by N719 dye (a Ru-based dye) exhibited the PCE (η) of 5.15% (Fig. S4). It was noted that the VOC and FF had shown slight enhancements while the JSC had demonstrated a significant improvement in the device with photoanode sensitized by the dye in ethanol compared to the photoanode with the dye in DI-water sensitization. A two-fold increment in the JSC (from 0.37 to 0.72 mAcm–2) and the efficiency (from 0.12 to 0.27%) was observed for the device with the photoanode sensitized by the dye in ethanol extract compared to the device sensitized by the dye in DI-water extract. Moreover, the control device (without sensitizer) exhibited poor photovoltaic performance which was 14 times lesser than the PV performance of the best device in the present study. The following factors may have contributed to the said observations: firstly, better solubility of natural pigments (chlorophyll a and anthocyanin) found in Mussaenda erythrophylla flower in ethanol which could have prevented aggregation of the pigment molecules and led to their uniform dispersion on the TiO2 surface as evidenced by the AFM studies; secondly, more light harvest in the visible region of solar spectrum by the pigment molecules extracted in ethanol as shown in Fig. 2 that resulted in increased electron transfer from the LUMO of the excited dye to the TiO2 conduction band. Furthermore, the observed photovoltaic performances of the said devices were found to be retained for 24 h (Fig. S5 and Table S2).

Influence of dye concentration and pH

The influence of dye concentration and pH of the dye solution on PCE of the fabricated devices were also studied. Different concentrations of the dye solution were prepared by serially diluting the dye in ethanol extract with ethanol, keeping the total volume as 12 mL (Table 4).

Table 4 Photovoltaic parameters of DSSCs with photoanodes sensitized by different concentrations of the dye under illumination intensity of 100 mWcm–2 with AM 1.5 filter.
Full size table

The optical study of the prepared dye solutions confirmed that light absorption increases with increase in the dye concentration (Fig. S6). However, the PV performance of the corresponding devices was only slightly improved with the change in dye concentration (Fig. 10). When the natural dye extracted in ethanol was diluted to half of its initial concentration, a slight improvement in the PCE of the corresponding device was observed.

Figure 10
figure 10

Photovoltaic performances of DSSCs with photoanodes sensitized by different concentrations of the dye under illumination intensity of 100 mWcm–2 with AM 1.5 filter.

Full size image

To investigate the influence of pH on the optical properties of the extracted natural dye, different volumes of 0.1 M HCl were separately added to the optimized dye solution (dye: ethanol = 6:6) and the resultant dye solutions were analyzed by UV–Visible spectroscopy. Subsequently, the TiO2 films were dipped in the said dye solutions at varied pH for 12 h, the corresponding DSSCs were fabricated, and their PV performances were evaluated.

Figure 11 shows the UV–Visible absorption spectra of the dye solutions at four different pHs. The absorption spectrum of a dye reflects its optical transition probability between the ground state, the excited state and the solar energy range absorbed by the dye. The pH of the optimized dye solution was found to be 5.78. When the pH of the said dye solution was decreased step by step, the intensity of the peak corresponding to chlorophyll a showed a downward trend whereas the intensity of the peak responsible for anthocyanin exhibited the opposite trend in the UV–Visible spectra. It has been reported that exposure of chlorophyll a molecules to weak acids, oxygen or light accelerates their oxidation potential and results in the formation of numerous degradation products36 which could be attributed to its decreased peak intensity at low pH (acidic condition). Moreover, the colour of anthocyanin is sensitive to pH37 due to transformation of its molecular structure at different pH. In the present study, the natural dye extracted from Mussaenda erythrophylla flower in ethanol showed an absorption peak near 530 nm, corresponding to anthocyanin, at pH 5.78 and it became more intense and broader when the pH of the said dye solution was decreased to 2.00. It has been reported in the literature that protonation of anthocyanin molecules leading to the formation of red coloured flavylium salt occurs under acidic condition (inserted image in Fig. 11)8. As the optical properties of the dye solution varied with pH, devices were fabricated by employing the said dye solution at four different pH and their PV performances were investigated.

Figure 11
figure 11

UV–Visible spectra of the dye solutions at different pH.

Full size image

As illustrated in Fig. 12 and Table 5, the optimum efficiency of 0.41% with the JSC of 0.98 mAcm–2 and VOC of 0.60 V was observed for the device fabricated with the natural dye extracted from Mussaenda erythrophylla flower in ethanol at pH 5.00. The investigation on the influence of pH on device performance revealed that when the pH was reduced from 5.78 to 5.00, the JSC and PCE increased from 0.82 mAcm–2 and 0.34% to 0.98 mAcm–2 and 0.41% respectively. Hence, it is proposed that a small quantity of flavylium ions, the stable form of anthocyanin, are formed at pH 5.00, get firmly attached to the TiO2 and thereby enhance the PCE3. However, further decrease in pH reduces the device performance. It could be concluded that even though the natural dye extracted from the flowers of Mussaenda erythrophylla contains both chlorophyll a and anthocyanin, PCE of the device is largely influenced by the light absorption ability of chlorophyll a. Table 6 summarizes PV performances of DSSCs sensitized by natural dyes extracted from different plants.

Figure 12
figure 12

Photovoltaic performances of DSSCs with photoanodes sensitized by the natural dye solution at different pH under illumination intensity of 100 mWcm–2 with AM 1.5 filter.

Full size image
Table 5 Photovoltaic parameters of DSSCs with photoanodes sensitized by the natural dye solution at different pH under illumination intensity of 100 mWcm–2 with AM 1.5 filter.
Full size table
Table 6 A comparison of PV performances of DSSCs with natural dye sensitizers extracted from different plants.
Full size table

The above comparison has revealed that the ethanol extract of Mussaenda erythrophylla flower is a promising candidate to serve as sensitizer in the DSSC application.

Conclusion

A natural dye, extracted from the flowers of Mussaenda erythrophylla in ethanol and DI-water separately, was employed as a photosensitizer in DSSCs. The phytochemical analyses of the said dye confirmed the presence of many phytochemicals. The optical, structural and electrochemical characterization of the said dye and the dye coated TiO2 films revealed the presence of anthocyanin and chlorophyll a in the ethanol extract only. The DSSC fabricated with the optimized natural dye sensitizer in ethanol at pH 5.00 demonstrated the best PCE of 0.41% with 0.98 mAcm–2 of JSC.