A comparative study of titanium dioxide preparation methods in solar cells based on the TiO2 semiconducting polymer heterojunction

Article history: Received 29 April 2020 Received in revised form 30 July 2020 Accepted 12 August 2020 The influence of the morphology and thickness of titanium dioxide (TiO2) films in solar cells based on dioxide materials-semiconducting polymer was investigated. These parameters can significantly impact the mechanism of charge generation in double-layer solar cells. TiO2 films were deposited on glass surfaces using two different techniques. Atomic force microscopy examination revealed a crystalline structure in the thick TiO2 film produced using the doctor blade technique, whereas an amorphous structure and thin layer were obtained via the spin-coating technique. Under light conditions, the best performance between the two solar cells was obtained with the amorphous TiO2thin film, which displayed a higher short circuit current density and power conversion efficiency (Jsc=1mA/cm2, μ=0.24%) than the nanocrystallineTiO2 solar cell (Jsc=0. 23mA/cm2, μ=0.1%). This is explained by the improved efficiencies of hole-electron pair separation and charge collection through reductions in the resistance across the bulk region and in charge recombination at the interfacial layer in the amorphous solar cell. However, the high rectification ratio and high turn-on voltage under dark conditions, and forward bias current in the nanocrystalline solar cell led to an increase of 0.3 V in the open circuit voltage in comparison with the amorphous device. These results confirm that, for double-layer solar cells, an amorphous thin film performs better than a thick nanocrystalline TiO2 layer, in contrast to previous observations for dye-sensitized solar cells.


Introduction
*In the last two decades, solar cells based on the TiO2 semiconducting polymer heterojunction have attracted significant attention because of their high power conversion efficiencies compared with other types of solar cells (Thu et al., 2018). Several techniques have been applied to improve the performance of polymer-metal oxide solar cells, exploiting the characteristics of their components. Metal oxides semiconductors (MOSs) have been widely studied as the electron transport layer, including CeO2, TiO2, ZnO, and SnO2 (Vittal and Ho, 2017). Their high electron mobilities, good pore filling, and wideband gaps make MOSs very attractive for electronic and optoelectronic applications such as solar cells and sensors. They can be used as a layer deposited directly onto a transparent conductive oxide (TCO) substrate or mixed with a hole conductor to increase the solar cell's efficiency (Shirkavand et al., 2019). Additionally, these materials have a sufficient offset of the lowest occupied molecular orbital (LUMO) of the hole transport material (HTM) to provide enough energy to separate the electron-hole pair at the interfacial layer (Cheema and Joya, 2018). O'Regan and Grätzel (1991) fabricated a high-efficiency solar cell using a nanocrystalline nc-TiO2 electron transport layer because of its high chemical and optical stability, nontoxicity, low cost, and corrosion resistance in comparison to other types of MOs. Additionally, good light harvesting and increased photoinduced charge separation were accomplished by optimizing the morphology and thickness of the nc-TiO2.
These physical attributes play important roles in improving the efficiency of dye-sensitized solar cells (DSSCs) by increasing the absorption of dye at the surface of the TiO2 and interfacial layer using nc-TiO2. In the absence of a dye, however, that will be a disadvantage in making solar cells because there is increased aggregation at the interface between the hole transport layer and TiO2.In this article, we report our study of the fabrication, characteristics, and structural properties of TiO2layers produced using two different deposition methods for solar cell application.
We confirm that amorphous TiO2/polymer solar cells show better performance in comparison to nc-TiO2 solar cells.

Sintering of titanium dioxide films
Two devices were fabricated using an active layer sandwiched between the bottom and top electrodes. The first device had a thin layer of TiO2 coated on a pre-cleaned fluorine-doped conducting tin oxide (FTO) glass plate (sheet glass, 8 Ω/sq, Solaronix com). The TiO2solution (5mL, Solaronix com) was dropped onto the bottom electrode and, within a few seconds, was spun at 1000 rpm for 60 s. For the second device, a thick layer of nc-TiO2was produced by depositing TiO2 sol gel (50μL, Solaronix com) on the edge of the bottom substrate using a micropipette; then, a cleaned glass rod was used to spread the sol gel over the electrode substrate area defined by tape (Shirkavand et al., 2019). This film was left to dry in air for 10 min until its milky color disappeared. The tape was then removed, and the substrate placed on a hot plate preheated between 75 0 Cand 100 0 C (limited by the accuracy of the dial setting) for 30 min. For both devices, the left on the hot plate for 110 min, during which time the temperature was increased from ~75 0 C to 450 0 C in 4 steps. Finally, heating was stopped and the temperature allowed to fall naturally to ~70 0 C, where upon the substrate was recovered.

Preparation of poly (3-hexylthiophene) (P3HT) Film
Poly (3-hexylthiophene) was chosen as the organic semiconductor for the second part of device components because it h has been widely used in polymer-based electronics such as transistors and solar cells (Yang et al., 2017). The P3HT solution for spin coating was prepared by weighing regioregular poly (3-hexylthiophene-2,5-diyl) (0.1g, Sigma-Aldrich) into a 5 mL volumetric flask, and then topping up with chloroform (Aldrich Chemical, UK). To enhance the solubility of P3HT in chloroform, the volumetric flask was placed in a warm ultrasonic bath at ~50℃ and sonicated for 30-60 min. Undissolved particles were removed by passing the solution via a syringe through a 0.2μm PTFE filter into a clean vial. To prepare the P3HT layers in the two devices, each substrate was placed on the vacuum chuck of a photo resist spinner (Model 4000, Electronic Micro Systems Ltd.). A pre-cleaned glass pipette was used to drop the P3HT solution (5mL) onto either the nc-TiO2or thin TiO2 layer. After several seconds, the substrate was spun at 1000 rpm for the 60s.

Device characterization
The substrates were removed and stored in the dark until gold electrodes were deposited onto the P3HT layers. The current-voltage characteristics of the solar cells were evaluated using a Keithley model 4200 source-measure unit, which was capable of applying negative and positive voltages in the range of +1 to -1V to the device and measuring currents in the range of 1pA to 1mA. The optical responses of the solar cells were obtained using a tungsten lamp. The light was focused onto the transparent FTO electrode with an incident power of 72mW/cm 2 , constrained by an aperture to fall on an area coinciding with one of the gold top electrodes. Fig. 1 shows the semilog plots of current density (J) versus voltage (V) applied for the P3HT/amorphousTiO2 solar cell when the voltage was swept from−1 to 1 V under different conditions. In the dark, the current is higher when the bias was swept from 0 to 1 V compared to the opposite sweep direction. The rectification ratio is defined as the forward current IF divided by the reverse current IS at a ±1 V bias. As can be seen from Fig. 1, the J-V characteristics of the P3HT/amorphous TiO2 solar cell display rectification with a turn-on voltage for the solar cell of~0.75V. Additionally, the rectification ratio is estimated to be ~10 2 at ±1V. Under illumination, the rectification ratio nearly disappears, and the current increases from 7×10 6 to 1mA/cm 2 at −1V. This is attributed to the change in conductivity of the bulk region at the interface inside the solar cell. For the region between −1 to 0 V (under the reverse bias condition), the photo conductivity mode governs the increase of the photocurrent. This depends on the generation of free charges through the absorption of the incident of photos and their transport under the influence of the electric field and bulk-region capacitance. In contrast, there was a small increase in forward bias current under illumination, from the turn voltage to +1V related to the small enhancement of majority charge carriers, which was already high in the dark.

Results and discussion
where PL is light intensity falling on the device, and Jmax and Vmax are the current and voltage corresponding. Using the data obtained from Fig. 2 (Jmax=0.571mA/cm 2 and Vmax=0.3V), the values of μ, and FF were determined as 0.24% and 28 %, respectively. The performance of the P3HT/nanocrystalline TiO2solar cell was next investigated. Fig. 3 shows the semilog plots of J versus V for this device under dark and light conditions when the voltage was swept from 1 to −1V. The reverse bias current (from 0 to −1V) is lower compared to the opposite sweep direction; however, in the negative region, the current is nearly constant with increasing applied voltage. In contrast, the forward current of the nc-TiO2 solar cell increases with the applied voltage and is lower than that in the amorphous solar cell. Thus, a higher rectification of ~10 3 at ±1 V is observed in the nc-TiO2 solar cells, arising from the difference in the energy bands of the materials used to fabricate the cell. These results agree with those of Hwang and Lee (2008), who observed a similar high rectification ratio due to the best crystallization in p-n junction diode. Additionally, the turn-on voltage of the nc-TiO2 solar cell is 0.635V higher than the amorphous solar cells. On the other hand, when the device was under illumination, the reverse current significantly increases in revere current more than four orders of magnitude, whereas the current in the forward direction is almost unchanged for the region between 0.85 and 1 V, under the effect of photoconductivity. To compare the efficiencies of amorphous solar cells, the J-V characteristics of the P3HT/nc-TiO2solar cell were also studied under dark and light conditions (Fig. 4) in the forward bias region of 0 to 1V. This solar cell exhibits a JSC value of 0.23mA/cm 2 , which is substantially lower than that of the amorphous TiO2 solar cell (1mA/cm 2 ). Thus, the nc-TiO₂-based device produces a low μ of 0.1%, even though its VOC and FF values are 0.8 V and 40%, respectively.
of the voltage characteristics -Current density Under dark conditions, the J-V characteristics of the amorphous and nanocrystalline TiO2 solar cells exhibit exponential behavior at low voltage and nonlinear behavior at high voltage. Thus, at low voltage, both cells exhibit typical diode-like behavior in the dark, such that the forward current is solely controlled by thermionic emission theory (Bartolomeo, 2016). According to this theory, the relationship between the voltage and dark current density is given by Eq. 3 and 4: where n is the ideality factor, k is the Boltzmann constant, q is the electronic charge, A* is the Richardson constant, T is the temperature, and Vb is the potential barrier height. From Eq. 3, the ideality factor can be determined from the slope of the exponential regime of the semi-logarithmic plot of the dark J-V characteristics, and can be expressed as: If n is equal to unity (n=1), the device behaves as an ideal diode in which there is no trapping or recombination of charge carriers, nor the presence of high series resistance or interfacial layers. For example, Kim observed that the ideality factor of an Au/Al2O3 device was ~ n=2 for an Al2O3 thickness up to 4.9nm, but increased to n=2.9 at an Al2O3 thickness of 8.3 nm . Using equation 5, we calculated the ideality factors for the prepared amorphous and nc-TiO2 solar cell devices, obtain in values of 1.58 and 3, respectively. These results may be attributed to drop bias voltage cross the series resistance of the bulk region in the device and interfacial layer or depletion region between the P3HT and nc-TiO2. In addition, the thickness of the bulk region and the large interfacial layer of the junction in the nc-TiO2 solar cell increases the ideality factor compared to the other device. This indicates to increase in the electron/hole recombination dominated at a high interfacial layer between P3HT and nc-TiO2 layers. However, our previous report showed that a potential interfacial barrier might be formed between the P3HT and nc-TiO2, which was interpreted as the origin of the observed rectification (Al-Dmour and Taylor, 2009). Rectification depends on the energy level alignment between the P3HT and nc-TiO2. And/or the work functions of the bottom and top electrodes. When V>0 is applied to the device, the Au electrode is forward-biased, and the current characteristics are dominated only by the series resistance of the bulk region, with decreased shunt resistance at the interface. Thus, the results here imply that a thicker layer of TiO2 reduces the dark current and gives rise to high rectification. Fig. 5 shows the tapping-mode atomic force microscopy (AFM) images of the two TiO2 films. There is a significant difference in their topographies, such as their surface roughness, grain formation, and thickness. In the amorphous TiO2 film, the surface is nearly flat with low roughness and a thickness of less than 100nm. In contrast, greater differences are seen in the nc-TiO2 film, which has a root-mean-square (RMS) roughness of 20.5 nm, a mean particle height of 86 nm, and a thickness of 2μm. These findings confirm that, for the nc-TiO2, inclusions of anatase TiO2 phases appear when the film thickness is larger, in comparison with the amorphous TiO2 film. Several studies have examined the surface properties and thicknesses of TiO2 films to improve the power conversion efficiencies of metal-oxidebased solar cells (Liu et al., 2018). Here, we find that the thinner layer and flatter, smoother surface of the amorphous TiO2 are much better for producing a photocurrent than the nanocrystalline TiO₂ having a rough surface and large interfacial layer at the junction in solar cells. The power conversion efficiency of a solar cell depends on the efficiencies of five processes, which are described in Eq. 6 as: where is the efficiency of light absorption in the active layer; is the efficiency of exciton migration to the interface; is the efficiency of charge dissociation at the interface and transport to the electrodes; is the efficiency of charge collection; and R is the reflectivity of the substrate at the air interface. The highest efficiencies among these five processes are charge dissociation and charge collection for the thin amorphous solar cell, whereas the efficiency of light absorption is highest for the thick nc-TiO2 solar cell. For the amorphous TiO2 solar cell, there is enough separation energy stemming from the work function difference of the two electrodes and semiconductor in addition to the interfacial layer at the junction to separate the electron-hole pairs. Thus, they require less time to flow electron and holes out of the device as current. In contrast, in the nanocrystalline solar cell, the contribution of the work function difference between the electrodes in charge separation is limited, and coupled with the increase in the series resistance of the bulk region, increased charge carrier recombination is observed. Thus, we have observed the power conversion efficiency of the amorphous solar cell is enhanced. Fig. 6 shows the power characteristics of the two devices as a function of voltage. The results confirm that the amorphous TiO2 solar cell produces a high maximum electric power of 0.18mW/cm 2 as compared to 0.06mW/cm 2 for the nanocrystalline solar cell. On the other hand, the range of operation of solar cells starts from zero voltage to open circuit voltage, which was small in amorphous TiO2 solar cells. That confirms the difference in interfacial layers, which affect the hole-electron separation. The observed results confirm the importance of the thickness and morphology of the TiO2 on the performance of a solar cell in agreement with the results reported by other research groups. For example, Boroomandnia et al. (2015) studied different types of TiO2 films and their effects on the performance of solar cells. It was found higher densities of TiO2 solar cells produce lower power conversion efficiency than TiO2 nano particles solar cells (Boroomandnia et al., 2015). That is attributed to the difference in the interface area in tow devices. Finally, Al-Attafi et al. (2018) used a similar approach to our work to study the effect of amorphous TiO2 on DSSC performance. A detrimental effect on DSSC performance was investigated with the amorphous material in contrast to the observations here in for double layer solar cells (Al-Attafi et al., 2018).

Conclusion
Amorphous TiO2 solar cells offered better performance than nanocrystalline TiO2 solar cells under light conditions. The amorphous TiO2 solar cells generated a high power conversion efficiency (of 0.24% and a short current density Jsc of 1mA/cm 2 , which were due to the improved efficiencies of hole-electron pair separation and charge collection through the reductions of resistance across the bulk region and charge recombination at the interfacial layer. The dark current results emphasized the effect of the TiO2 structure in terms of the high rectification ratio and low reversed bias current observed in the nanocrystalline solar cell.