Evaluation of the Suitability of Natural Flower Dyes Spectra on Dye-Sensitized Solar Cell (DSSC) Containing TiO2 and I-/I-3 With Respect to Stability and Efficiency

Different dye-sensitized solar cells (DSSC) were fabricated using TiO2 as a semiconductor with two groups of flowers, nursery flowers from Duhok city and mountain flowers. The ultraviolet (UV)-visible absorption spectra of these dyes were examined. Current density–voltage (J–V) characteristics for each fabricated cell were recorded at ambient temperature. The stability of these cells has been examined through annealing time. The mountain flower dyes have been collected from Gara Mountain at a height about 2500 m above sea level for which they are believed to have resistivity to UV light. The efficiencies of the cells mostly depend on the visible absorption of the dye while their stabilities depend on the ultraviolet absorption of these dyes. As a compromise, the nursery flower dyes efficiencies were between 1.88% and 4.19%, while mountain flowers were between 1.83% and 4.82%. Although we get low efficiency in comparison to other workers, however, in this study we are after getting the best stability of natural flower dyes. The current findings are substantially added to our understanding of the stability of DSSC. Some suitable stable natural flowers were selected for our new program of DSSC research. The statistical P values show that the efficiencies of mountain flower dyes are much more stable than that of nursery flower dyes.

complexes [3] in late 1970-1990 as in a dye-sensitized solar cell (DSSC) of Grätzel. He had used an organometallic complex similar to chlorophyll which he calls photosensitized solar cell, DSSC. Recently, the improvement of the efficiency of dyed monocrystalline silicon solar cells by covering them with natural plant pigments has been done in our lab [4]. In the late 1960s, the researchers generated electricity from illuminated organic dyes in electrochemical cells [5]. Later, many years of unremitting efforts in improving the efficiency of DSSC were spent since the amount of energy provided by the sun in one hour is huge and much more the consumption by the human in one year. Then in 1991, Grätzel made a breakthrough when he introduced a titanium dioxide semiconductor to increase the light-harvesting getting 7% efficiency and called these cells DSSC [6]- [8]. The field is growing fast, and the cells of 11% efficiency are reachable [9], [10]. In 2013, Grätzel announced a 15.0% efficiency of Solid State DSSC [11].
Economic as well as nontoxic and environmentally friendly aspects of DSSC make it one of the most promising types of photovoltaic cells for the conversion of light energy to electricity.
The DSSC which is the subject of this research was experimented by O'Regan and Grätzel [12]. It consists of an organic-based photovoltaic cell as a dye-sensitized semiconductor material, often titanium dioxide TiO 2 , where the dye molecule attached to the semiconductor acts as a light absorber.
Under illumination, many processes play an important role in converting the light energy to electricity. Since the dye molecule (D) is sensitive to the light, it is easily excited by light energy resulting in an unstable sensitizer molecule (D * ) in the lowest unoccupied molecular orbital (LUMO) band as shown in reaction (1) and Fig 1. To be stable and in an ideal condition, a photocurrent electron is released by the excitation molecule which injects into the conduction band (CB) in the semiconductor yielding a sensitizer hole (D+) as shown in reaction formula (2). Next, the photoelectrons in CB as in reaction (3) reach the counter electrode (CE) through an external circuit as illustrated in reaction (4). With the help of the electrons that come from the CE, dye at an excited state is reducing some of its oxidation energy in the electrolyte. This led to Red according to reaction (5). Finally, as in the reaction formula (6), D+ This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ Fig. 1. Construction of DSSC cell with dye molecules based on TiO 2 semiconductor and I -/I 3 electrolyte. The dye absorbs light radiation and releases photo-excited electrons from HOMO to LUMO at V 1 , which is supposed to be injected into the conduction band (CB) in the semiconductor at V 2 ; at V 3 the excited electron reaches the anode through an external circuit, yielding a sensitizer hole which is regenerated by electron donation from iodide in the electrolyte at V 4 , now this iodide must be regenerated by reduction process of triiodide on the counter electrode at V 5 on the cathode. Recombination at the oxide/dye interface at V 6 and electrolyte/oxide interface at V 7 besides a relaxation of the excited dye to its ground state at V 8 also occurs.
is regenerated by Red. However, the disadvantage that occurs in these processes is that they companied with electron-hole recombination processes as shown in reaction formulae (6)-(8) that significantly decreased the efficiency of solar cells [13], [14].
Ox + e − (CE) → Red − (5) In this study, the DSSC was fabricated with natural plant dyes. The extraction process was simple and it is done with primitive tools. This study did not use Ruthenium-based, organometallic material, or any other artificial dyes which are expensive and mostly highly toxic.

A. Synthesis of Natural Dye Sensitizer
This study seeks simplicity and relying on what is available in nature to get electricity. Generally speaking, all-natural dyes could be used in DSSC as sensitizers, but not all of them provide us the same stability and efficiency in converting light to electricity. For this, two groups of flowers are selected in this study for the comparison process. The first group consists of natural dyes from different special flowers at the top of Gara Mountain which is about 2500 m above sea level in the Kurdistan region, north of Iraq. They were used as sensitizers and sunlight harvesters. They are considered having a high resistivity to UV light. The second group of flowers are selected from a nursery at Duhok city (maybe some of them also originally from mountain). The extraction process was done in primitive ways either by squishing the flower or by using different solvents to extract the dyes. The process is done using the same weight of flowers and the same volume of different solvent that does not affect the color of the flower or cause degradation of pigments. However, it was difficult to match them because not all of them are soluble in the same solvent. As we are after the UV-visible spectra, the dyes were subjected to 6800 UV/Visible spectrophotometer (JENWAY, China, QA in the UK) to understand the light response of these different dyes. It is believed that these wavelengths will serve us in achieving two goals. The first goal which is the main goal from this study is that the UV spectra wavelength is thought to support stability, while the visible spectra wavelength is believed will support efficiency as it is the second goal. For example, the collection of some of these dyes' spectra is shown in Fig. 2.
Despite using the same weight of flowers in the same volume of different appropriate solvents, the spectra of theses give us different optical densities (OD) as a function of different wavelengths beside different concentrations of dyes in each flower, see the following equation: where ε is the extinction coefficient of dye molecules in each flower, L is the path length of the cell in cm, and C is the concentration of dye in molar. The maximum absorbance for some of them reached up to 4 OD, while for the others, it reaches several tenths of the unit as shown in Figs. 2 and 3.
The code, the name, and the photo of each flower are illustrated in Appendix 1 [15]- [17]. The absorbance peaks at UV and visible regions for each flower are recorded in Table I.

B. Materials
Transparent titania paste, TiO 2 with 20 nm average diameter of the nanoparticle (paste A), Titania paste, reflector TiO 2 with  150-250 nm average particle size (paste B) were used as working electrode layers, EL-HPE high-performance electrolyte was used as the transport for redox mediator between two electrodes, platinum paste, Pt with a highly transparent layer and high electrocatalytic activity to provide good transparency and high catalytic activity for the electrochemical reduction of I 3 to Iwhich is used as a counter electrode layer, fluorine-doped tin oxide coated glass slide FTO of surface resistivity equal to ∼7 Ω/sq, and 2.2 mm thickness, 80%-82% (visible) transmittance was used as a conducting glass, act as substrates depositing for both TiO 2 films and platinum electrodes. All these material were purchased from Sigma Aldrich, USA, and used without any further purification.

C. Preparation of the Working Electrodes
Two types of TiO 2 paste-transparence TiO 2 (paste A) and reflector microcrystalline TiO 2 (paste B) were deposited by the doctor blade method manually, forming respectively the transparent and the light-scattering layers of the photoanode electrode.
During the sintering process of the TiO 2 layer, iron cations should be removed from the surface for it enhances charge recombination in photocells [18] and quenches the dye-sensitized photocurrent (Jsc) effectively, beside the Jsc decreased by 30% upon the addition of 400 ppm Fe 2 O 3 [19], [20].
To prepare the DSSC working electrodes, the FTO conductive glass was cut in area 2 × 2 cm 2 and was washed with a detergent solution using an ultrasonic bath for 20 min and then rinsed with distilled water and ethanol to remove organic pollutants and iron contamination before use. Paste A was coated on the FTO glass plates by the doctor blade method. To avoid surface irregularity, the glass is kept in a clean box for 40 min relaxation time. Then dried for 6 min at 125°C [21]. Then one layer of the same area and 10-15 μm thickness of light-scattering anatase TiO 2 film (paste B) were coated by the doctor blade method. Two different geometrical sizes, 0.28 and 0.19 cm 2 , of the TiO 2 electrodes were used; however, the real size was 0.04 cm 2 . After fabricated the cells, we get a different arbitrary geometrical area of our devices. So, a suitable mask of 0.04 cm 2 area for each device is used to let the light fall only on a certain area which is the area of the mask during measuring the energy conversion efficiency under illumination and this is the active area. In other words, the active area is defined by the mask area, not the TiO 2 area, and 20-24 μm thickness. The electrode coated with the TiO 2 pastes were gradually heated under an airflow at 325°C for 5 min, at 375°C for 5 min, and at 450°C for 15 min, and finally, at 500°C for 15 min [21]. After cooling to 80°C, the TiO 2 electrode was immersed into flower dyes for about 20 h at room temperature.

D. Preparation of the Pt-Counter Electrode
First of all, two holes of 1-mm diameter on the FTO surface were drilled by a diamond drill. The FTO conductive glass was washed with ethanol and distilled water. Based on the doctor blade method, a platinum counter electrode was prepared by coating the FTO glass substrate with platinum paste. First, dry at 80°C, and then sintered at 450°C for 30 min with a thickness equal to the thickness of the scotch tape.

E. Electrolyte
EL-HPE High-Performance Electrolyte, Sigma Aldrich, USA, is used in this study because it has a low viscosity, high ionic conductivity liquid electrolyte (16.5-18.5 mS/cm at 20°C), and, therefore, provides the highest performance [22].

F. Assembly of the DSSC
The working TiO 2 electrode and Pt-counter electrode were assembled into a sandwich-like structure (see Fig. 4). A hot-melt sealant was put between two electrodes, with inner dimensions 2-mm larger than the TiO 2 layer. The electrolyte was injected between the two electrodes through the holes in the back of the CE. Then, the two holes were sealed by silver tape and silicon glue to prevent evaporation.

G. Apparatuses and Instruments
In this study four prop computerized graphical interface Source-Measure Unit (SMU) 2450 -KEITHLEY is used to provide J-V curves. The intensity of the beam of light is measured by Voltcraft PL-110SM Solar Light Meter.

H. Photovoltaic Outdoor Measurements
In this study, we considered outdoor photovoltaic measurements at noon with light intensity between 700 and 800 W/m 2 . J-V characteristics were obtained by using a Keithley Model 2450 SourceMeter SMU which can source and measure both current and voltage simultaneously. The voltage step was 10 mV and the delay time of photocurrent was 40 ms.

III. RESULTS AND DISCUSSION
The absorption spectra and the J-V curves of all cells were plotted. Many cell parameters such as open-circuit voltage (V oc ), short circuit current density (J sc ), maximum power point (P max ), fill factor (FF), efficiency (η), and wavelength at the maximum optical density (λ max ) are measured according to the following mathematical equations: where J SC is the current per unit active area A of the device when the open-circuit voltage across the device is zero, I SC is the maximum current available from a solar cell, and this occurs at zero voltage.
where P max is the maximum generated power, P in is the electrical input power from the light source, P theo is the theoretical power maximum of a solar cell, V MP and I MP are potential and current of the I-V curve when the generated power is at the maximum, V OC is the maximum voltage available from a solar cell, and this occurs at zero current.

A. Optical Properties of Dyes
The spectra of the Gara mountain flower dyes and nursery Duhok city flower dyes are plotted in Figs. 6(a) and (b) and 7, respectively. The absorption peaks of the spectra for all extracts were tabulated in Table I.

B. Photovoltaic Performance of the DSSC
To investigate the effect of the dyes on the efficiency of DSSC, all the cells were treated and fabricated under the same      Table I while the IV  characteristics are tabulated in Table II.  Table I while the IV  characteristics are tabulated in Table II. J-V sweeps show that when a TiO 2 working electrode was used without dye, the V oc was only 0.02 V, the J sc was 0.41 mA.cm −2 , and the η achieved was only 0.02%, as described in Table II. The reason behind such low efficiency comes from the absence of the dynamic parameter, dye-sensitized molecules. Indeed, the functioning of DSSC significantly depends on the role of dye sensitizer molecule (in UV and visible absorbance) on the surface of semiconductor materials such as TiO 2 , SnO 2 , Nb 2 O 5 [23], that is, the dye naturally has an ability to absorb light, paving the way of working as an electron pump to excite an electron to the LUMO where it will be injected into the CB of the semiconductor, receiving an electron from the charge mediator such as iodide/triiodide redox couple, and repeat the cycle [24], as has been mentioned earlier in reactions (1)-(8) and Fig. 1.
On the other hand, the results show that the efficiency of the DSSC with dye-covered TiO 2 electrode was increased significantly in both flower groups. In the Gara mountain group, the efficiency is ranged from 1.83% for AD dye to 4.82 for TT dye while in nursery city flower the efficiency is ranged from 1.88% for FF dye to 4.19 for SS dye. Although the nursery flowers have a better spectrum than the Gara flowers, the results show that Gara flower efficiencies are higher as in Table II. J-V characteristic curves for the first readings (T = 0) of all samples show that the highest short circuit current density J varies from 2.21 mA/cm 2 for YY dye to 5.21 mA/cm 2 for TT dye in Gara mountain flowers, and from 3.54 mA/cm 2 for FF dye to 7.33 mA/cm 2 for SS dye in nursery flowers at Duhok city. The open-circuit voltage varies from 0.49 volt for WW yellowish color dye to 0.68 volts for ZZ dye in Gara mountain flowers, and from 0.44 volt for HH dye to 0.54 volt for AA dye in nursery flowers at Duhok city as shown in Figs. 9 and 10.

C. Statistical Analysis
The study has gone some way towards enhancing our understanding of the stability of cell efficiency. The two groups of flowers have been examined in the same period, T1 and T2, to know the differences in their stability at certain periods. Nine from the mountain group and eight from the city nursery group.
T-test paired two samples for means are applied in the range of visible band absorption because it is this band that effects   Table IV. Since they are less than 0.05, this means that their efficiencies are not stable. The reason behind the stability of efficiencies in Gara mountain flowers is counted to the existence of UV radiation at the top of the mountain where the flowers were adapted to grow in. It is obvious that the UV intensity is much more at a higher altitude, so, whatever plant grows there should have some resistivity to change to stay alive as much as possible. Regarding this fact, it is believed that there is a significant relationship between the mountain plants which have absorbance peaks at UV band and the stability of their efficiencies as described in Tables III and IV.

IV. CONCLUSION
In our previous work, we focus on improving the efficiency of Si solar cells using natural flower dyes [4], while in this work we are after searching for the desired natural dyes that give the best stability to the cell as this is the main goal of our study. For this, many natural flower dyes have been used but only some of them are selected.
In DSSC, when we test the efficiencies at different times, T1 and T2, indeed we are measuring the contribution of the dye to pumping the electrons and increasing the efficiency, but in this type of DSSC, the efficiencies are low; maybe there is no cooperation between the excited dye electrons to promote the electrons from the valance band to CB of the TiO 2 semiconductor.
Although the present study hasn't been done for a long period, it makes several noteworthy contributions to the role of both spectra and environmental adaptation. The existence of UV peaks in dyes spectra means that the molecules of dyes have some resistivity for degradation as they can absorb the UV radiation and protecting the chromatic bond in a dye from broken and hence high stability as well. Accordingly, the dyes with high UV optical density absorbance peaks should be more stable than those with low UV optical density absorbance peaks. This is valid for FF, HH, and AA in nursery flowers. However, for mountain flowers, the results add some other fact to this idea. All Gara flower dyes having high peaks, whether in UV or visible region, have high stability. This is because the flower that grows on the top of a mountain adapts itself to be alive with the existence of UV no matter how much the absorbance rate of UV. This adaptation enables the flower to get much more resistivity for degradation and hence supports more stability of the efficiency under severe conditions. Statistically, this fact has been proved as there is no significant difference according to P-value for the efficiencies during a different period for all Gara mountain flowers that have high optical density whether in the UV region or visible region. So, the above two points are the reason why one should prefer mountain flower dyes over local flower dyes to get more stability of the DSSC. The small differences between the efficiencies at different periods are maybe due to the changing of the electrolyte concentration.
Regarding the efficiencies for all outdoor experiments that have been done at Duhok city (not on the mountain) which is only 410 m above sea level, it is found that all flowers dyes with most of their peaks at UV region have fewer efficiencies, like AD, YY, and XX in Gara Mountain flower and FF, HH, RR, and AA in nursery flowers, as shown in Fig 11. This is because of the low rate of UV radiation at the place of the experiment (Duhok city). In other words, it is believed that more UV radiation for such flowers might support more efficiency.
All flowers which have obvious peaks at the visible region in their spectra as in Fig. 12(a) and (b) give more efficiencies, like ZZ, VV, and TT in Gara mountain flowers and BB, II, CC, and SS in Duhok nursery flowers. This is because the rate of visible light is more than UV light at Duhok city.
It is obvious that the efficiencies were low in comparison to other workers. This might be due to the other components of these cells rather than the dye which we used, such as the lack of technology in the fabrication process besides the design of the cell which used only TiO 2 semiconductor as a working electrode, the photoelectrode, also it is not chemically treated using different precursor such as titanium (III) chloride (TiCl3), titanium tetrachloride (TiCl4), indium (III) and so on, that have been scientifically proven to help accelerate the electrons diffusion in TiO 2 film and preventing the recombination process [25]. Also, getting a certain thickness is very difficult especially with the manual doctor blading method that we depended on in this study; variation in several microns is huge when it comes to photovoltaic technology and has a significant effect on cell performance.
The idea of our research is to find good, stable, cheap, and easy for manufacturing solar cells with many natural dyes. The suitable natural dye which promotes electron from HOMO to LUMO and then to CB of the semiconductor of solar cell and increase the efficiency is our goal of this research. the next step is to find the structure of these dyes and with the help of the other parameters we may find the suitable new solar cell for industrial application.
APPENDIX 1 SCIENTIFIC NAME [15]- [17], THE COD, AND THE PHOTO OF  EACH FLOWER IN GARA MOUNTAIN AND NURSERY GROUPS  FLOWERS