Simulation on the Performance of Dye Solar Cell Incorporated with TiO2 Passivation Layer

Dye Solar Cell (DSC) has started to gain interest in the recent years for practical application because of its ecofriendly, low cost, and easy fabrication. However, its efficiency is still not as competitive as the conventional silicon based solar cell. One of the research efforts to improve the efficiency of DSC is to use the passivation layer in between the photoelectrode material and the conductive oxide substrate. Thus, the objective of this simulation study is to investigate the effect of passivation layer on the performance of DSC. Properties from literatures which are based on physical work were captured as the input for the simulation using process, ATHENA, and device, ATLAS, simulator. Results have shown that the addition of two-20 nm TiO 2 passivation layers on DSC can enhance the efficiency by 11% as the result of less recombination, higher electron mobility, and longer electron lifetime.


Introduction
Dye Solar Cell (DSC) is the third-generation solar cell that offers ecofriendly, low cost, and easy processing [1].As shown in Figure 1(a), DSC is made up of photoelectrochemical cell consisting of dye-sensitized TiO 2 layer, electrolyte, and platinum (Pt) film in between two transparent conducting oxide (TCO) glasses.The material used for fabricating DSC is abundant and belongs to nontoxic material.The procedure to build the DSC is also simple and easy.But behind all these advantages, there are a number of problems that still need to be addressed in order to realize DSC practically.One of them is its efficiency, which can be described by the following equation: where  sc is the short circuit density,  oc is the open circuit voltage, FF is the fill factor, and   is the intensity of the incident light.
In the presence of sunlight, the photons strike the dyes with enough energy to create an excited state of the dye, resulting in generating electron which can be injected directly into the conduction band of TiO 2 as shown in Figure 1(b).Regeneration process in dye happens when the electrons from the electrolyte are restored.The process is cycled, causing the generation of electricity.In reality, the generated electrons could travel in the reverse direction and recombine with the oxidized dyes [2] before regeneration process takes place.The dynamic competition between the electrons generation and recombination has been found to be the limitation that restricts the development of higher efficiency DSCs.Recombination mainly occurs at the interface between DSC elements such as TiO 2 /electrolyte and the transparent conducting oxide (TCO)/electrolyte.Recombination may cause significant losses in DSC.In order to improve the performance of DSC, reducing the recombination is essential [3].It has been reported that the back reaction process of electrons to the electrolyte is more prevalent at the TCO/electrolyte interface than at the photoelectrode film/electrolyte interface.This condition highlights the potential of utilizing two thin films of TiO 2 layer above (marked as A) and below (marked as B in Figure 1(a)) the photoelectrode TiO 2 layer which later proved to have increased the efficiency [4].One of the methods to be employed to control the recombination process in DSC is through the use of a thin layer at the TCO/electrolyte contacts.This layer is the single TiO 2 passivation layer, indicated as A in Figure 1(a).Several works have reported the use of a thin layer of other materials such as ZnO and Nb 2 O 5 instead of TiO 2 between the photoelectrode and the TCO substrate [5][6][7][8][9][10].They also reported that adding TiO 2 passivation layer may affect the performance as well as the efficiency of DSC.
Adding TiO 2 passivation layer is expected to reduce the possibility of recombination on electrolyte/TCO interface.On the other hand, the addition of TiO 2 passivation layer can reduce conductivity of electrolyte that will lead to the decrease in short circuit density ( sc ), as described in (1), as well as the efficiency.Moreover, adding TiO 2 passivation layer may result in huge electron trapping in DSC which may reduce the electron lifetime.Thus, striking a balance between these two factors may prove beneficial to the performance of DSC.It is believed that optimizing the thickness of TiO 2 passivation layer will improve the efficiency as well as DSC performance.
In order to understand how the presence of TiO 2 passivation layer can reduce the recombination process leading to an improvement in DSC performance, one needs to analyze the kinetics of the components within DSC system.If recombination is reduced, more electrons can flow to the cathode, generating more electricity in DSC.Recombination process in DSC may affect recombination rate through where  is recombination rate,  is constant of recombination, and  is electron concentration [11].Changing the amount of electron concentration in DSC will cause a change in the electron lifetime.This is because the electron lifetime is inversely proportional to the recombination rate as described in where  is electron lifetime and Δ is the change in electron concentration [12].Thus, smaller recombination rate would mean longer electron lifetime, whilst longer electron lifetime may cause higher  sc , leading to higher performance of DSC.Another parameter related to the recombination is recombination resistance.Electron lifetime affected the recombination resistance through where  is temperature,  is area,  is electron charge, and  is the thickness [13].The thickness, , is referring to the thickness of the photoelectrode material, namely, TiO 2 film.When (3) is equated to (4), it explained the inversely proportional relationship between  and  rec .Higher recombination resistance means less occurrence of recombination process in DSC.Meanwhile, there is also another parameter called transport resistant that is affected by the thickness as shown in where  is diffusion coefficient [13].The thicker the layer is, the higher the transport resistance would be.On the other hand, higher diffusion coefficient will increase the electron mobility as described in where  is electron mobility [14].From (4), there is the need for a thicker thickness, , of the film in order to increase the transport resistance which can then reduce the recombination process.In DSC system, it would mean a layer of film at the TCO/electrolyte interface where the occurrence of recombination is more severe.Experimental work by Eskandar et al. showed that adding the passivation layer can improve the performance of DSC [15].Recombination is reduced as indicated by the higher recombination resistance for DSC with TiO 2 passivation layer.Moreover, lower recombination may increase the electron lifetime.It is also confirmed by Waita et al. [16] that electron lifetime in DSC increased as the effect of adding TiO 2 passivation layer.Several researchers have successfully improved the performance of DSC by using passivation layer [17][18][19][20].However, another finding [15] showed that TiO 2 passivation layer can cause the electrons to be trapped so that they are unable to flow to the cathode, resulting in low  sc .Thus, with these contradicting effects, there is the need to optimize the thickness of the TiO 2 passivation layer in order to achieve longer electron lifetime with minimal electron trapping and higher diffusion coefficient for faster electron mobility but also lower recombination.
The work here involved simulated analysis on the performance of DSC by adding TiO 2 passivation layer of different thickness at the TCO/electrolyte interface and adopting this optimum thickness for the second passivation layer at the TiO 2 /electrolyte interface.Properties of the components of DSC were extracted from literatures which are based on physical work and were used as the input for the simulated environment.Michael et al. [20] had earlier reported a simulation work on Si-based solar cell using ATLAS by Silvaco.Simulation study allows for prediction of the behavior of the components in the system where, in this case, it can lead to a better understanding of the kinetics within DSC.This can be acquired without running a series of costly experimental works.The predicted behavior based on optimized simulated parameters can then be validated directly using the physical work.Simulation result also provides detail that may not be experimentally measurable using the current technology.

Simulation Model
ATHENA and ATLAS simulation software by Silvaco were used to simulate DSC performance.ATLAS is a device simulator whilst ATHENA is a group of process simulations.By using this software, virtual fabrication of DSC will be built.ATLAS calculates the electrical parameters by simulating the electron transport on a two-dimensional mesh.It will give the output of extracted electrical characteristics of the simulated model.The most suitable model from several models offered in ATHENA needs to be chosen in order to represent DSC system.In this model, manipulation of material, physical structure, and dimension can be carried out.
ATLAS offers several selections of material models that can be employed in the simulation, namely, mobility, recombination, statistics, impact ionization, tunneling model, energy transport, and heat flow equation [33].These models can be endorsed for the entire device or the specific region only.Statistic model is about acceptor and donor ion density.Impact ionization is more about electron injection when recombination happened.Tunneling model is more suitable for semiconductor device with occupied tunnel in it.Energy transport is more about energy used for carrier transport.Heat flow equation is used for defining heat flow.Among those categories, mobility and recombination models are the nearby model for DSC simulation in which both processes occurred in DSC.
Mobility models in ATLAS are more suitable for silicon and GaAs based semiconductor device with limited conditions whereas DSC will not be using those materials.Instead, the recombination model was selected since, for DSC, the simulation will focus on recombination process.Since DSC utilized TiO 2 that belongs to wide band gap material then SRH (Shockley-Read-Hall) model was adopted as it was more relevant to the recombination process in DSC.The net recombination rate for SRH recombination is given by where  is electron density,  is hole density,   is electron in intrinsic level,   and   are the minority carrier lifetimes for electrons and holes, and it is assumed that the trap level coincides with the intrinsic level [33].Equation ( 7) is derived from (2).
As the input for the simulation of DSC here in ATLAS and ATHENA, properties, namely, affinity, energy gap, and permittivity, are extracted from literatures.Affinity is the amount of energy obtained when an electron is moving whereas energy gap is the amount of energy required to excite it whilst permittivity is its ability to store energy.DSC was then designed in ATHENA based on the data properties shown in Table 1.In this simulation, working temperature for DSC is set at 300 K with 1.5 AM full sunlight condition.The electrolyte used is iodine-based electrolyte.To fulfill the need of simulated light source, input of an optical file consisting of  values is extracted from UV/VIS spectroscopy result.This optical file contains the refractive index, , and extinction coefficient, , for dye N719-sensitized TiO 2 .It is important for the simulation of DSC as it defines the condition of light passing through the DSC.
In this simulation, the input and structure of DSC are shown in Figure 2. DSC is designed in a sandwich-like structure, consisting of a thin layer of metal oxide layer as the passivation layer and TiO 2 nanoparticle with absorbed N719 dye, represented by the red region.The electrolyte used is iodine, which is displayed by the green region.The performance in the form of electrical data was extracted for the simulated DSC system and then compared and verified with the results of the closest related experimental work.To investigate the effects of TiO 2 passivation layer, simulations of DSC without any passivation layer and also with TiO 2 passivation layer in DSC as illustrated in Figure 1(a) are compiled.Simulations have the same structure as shown in Figure 2. Properties of TiO 2 used as photoanode and passivation layer are shown in Table 1.

International Journal of Photoenergy
In DSC, electron will be injected from the dye to TiO 2 photoanode.Then, electron will travel to the cathode through passivation layer.In order to get higher electron injection, the conduction band (CB) of material used as the passivation layer needs to be closer to the CB of TiO 2 , which is the photoanode.As shown in Figure 3, CB of TiO 2 , Nb 2 O 5 , and ZnO are located below the CB of N719 dye.It means that those materials can be used as the passivation layer.So as to comprehend the effects of material used as the passivation layer, simulations are conducted by adjusting the properties of material used as passivation layer.Simulation with the same structure as shown in Figure 2 is conducted and compared with simulation which has different material of additional layer as passivation layer, namely, TiO 2 , Nb 2 O 5 , and ZnO.Properties of materials used as passivation layer are shown in Table 1.By varying the thickness of TiO 2 passivation layer, simulations are also compiled to study the effects of thickness of TiO 2 passivation layer.TiO 2 passivation layer is varied with its thickness from 10, 20, and 50 to 100 nm.Properties of TiO 2 used as photoanode and passivation layer are as shown in Table 1.Structure used in simulation is as shown in Figure 2. Simulated results obtained are compared with the experimental result which has the closest condition to the simulated DSC.Lastly, DSC is simulated with two-20 nm thick TiO 2 passivation layers with position as shown as A and B in Figure 1.
To measure the accuracy of simulation, calculation of error is also conducted based on

Results and Discussion
3.1.Effects of Passivation Layer.By adding TiO 2 passivation layer, the thickness of TiO 2 photoelectrode will increase.As shown in (4), there is an inversely proportional relationship between  rec and thickness, . rec is higher for simulated DSC without any additional passivation layer compared to  rec for simulated DSC with additional TiO 2 passivation layer.This implies that recombination is easily occurring in simulated DSC without any additional passivation layer compared to the one with the additional TiO 2 passivation layer.Meanwhile, there is proportional relation between   and thickness as shown by (5).  is lower for simulated DSC with additional TiO 2 layer as passivation layer compared to   for simulated DSC without any additional passivation layer.
It means that electron transport is easier in simulated DSC with additional TiO 2 layer as passivation layer compared to electron in simulated DSC without any additional passivation layer.This condition indicates that the charge transfer remains more efficient and results in higher  sc in simulated DSC with additional TiO 2 layer as passivation layer compared to that in simulated DSC without any additional passivation layer.From Table 2, it can be seen that the simulated DSC with TiO 2 passivation layer has higher  sc compared to simulated DSC without additional layer. sc is 11.649 mA/cm 2 for simulated DSC with additional TiO 2 layer as passivation layer and 11.364 mA/cm 2 for simulated DSC without any additional passivation layer.This higher  sc results in higher efficiency of DSC with additional TiO 2 passivation layer compared to DSC without any passivation layer due to less recombination.Meanwhile,  oc values of them both are close to each other.It demonstrated that the additional passivation layer can increase the performance of DSC. 2 and also being compared with the closest experimental works.As presented in Table 2,  sc of simulated DSC using TiO 2 , Nb 2 O 5 , and ZnO as passivation layer are 11.649 mA/cm 2 , 7.952 mA/cm 2 , and 7.972 mA/cm 2 , respectively.These three materials are selected since they have close conduction band with TiO 2 as can be seen in Figure 3.It can be observed that simulated DSC with TiO 2 passivation layer has the higher  sc compared to the simulated DSC with Nb 2 O 5 passivation layer and simulated DSC with ZnO passivation layer.This is attributed to the use of the same material of the passivation layer as the TiO 2 photoanode where interparticle connectivity will be very much improved with the CB at the same position in the energy level.

Effects of Passivation Layer Material. The performance of simulated DSC is shown in Table
As can be seen in Table 2, the efficiency of simulated DSC using TiO 2 , Nb 2 O 5 , and ZnO as passivation layer is 3.428%, 3.347%, and 2.825%, respectively, making the efficiency of the simulated DSC with TiO 2 passivation layer the highest.As for the simulated DSC with ZnO passivation layer, its lowest efficiency is believed to be due to the position of its CB above the CB of TiO 2 photoanode.The difference in the position of CB will create a boundary for electrons to be injected from dye N719 to the cathode through the passivation layer.On the other hand, the simulated DSC using Nb 2 O 5 passivation layer shows higher efficiency compared to the simulated DSC with ZnO passivation layer but still lower one compared to the simulated DSC using TiO 2 passivation layer.This is due to the occurrence of recombination with the electrolyte because the CB of Nb 2 O 5 is close to the CB of the electrolyte as shown in Figure 3.
From this simulation work, it can be predicted that, using appropriate material for the additional passivation layer in DSC, the electron loss from the photoanode to the dye through recombination process can be prevented.Besides that, it is also suggested that the material chosen as passivation layer has near or the same energy level with the photoanode in order to improve the electron mobility.

Thickness of TiO 2 Passivation
Layer.Based on data of diffusion coefficient and electron lifetime from Eskandar et al. [15], related to (4) and ( 5), simulated recombination resistance and transport resistance were calculated.As can be seen in Figure 4, the recombination resistance reached optimum when the passivation layer thickness is 20 nm.This implies that the occurrence of recombination is low when 20 nm thickness of passivation layer is used.Furthermore, transport resistance touched the minimum when the passivation layer thickness is 20 nm which translates better electron mobility in such condition.The result of simulated DSC is shown in Table 3 together with the outcome of the experimental work by Eskandar et al [15].Values of  sc shown in Table 3 seem to be decreasing with increasing passivation layer as the amount of electron density decreases.Result of simulated  sc for DSC without TiO 2 passivation layer is 11.364 mA/cm 2 , whilst those with 10 nm, 20 nm, 50 nm, and 100 nm of TiO 2 passivation layer are 10.28 mA/cm 2 , 11.649 mA/cm 2 , 11.638 mA/cm 2 , and 11.813 mA/cm 2 , respectively.It can be seen that  sc is maximum for 100 nm of TiO 2 passivation layer.High  sc implies high moving electron condition where its lifetime is longer in DSC with thicker TiO 2 passivation layer.
Values of  oc for simulated DSC were found to be little fluctuating as shown in Table 3.  oc for DSC without TiO 2 passivation layer is 0.651 V, whilst, for DSC with 10, 20, 50, and 100 nm of TiO 2 passivation layer, the values are 0.65, 0.694, 0.669, and 0.657 V, respectively.From that simulation result, the highest  oc is reached when DSC is added with 20 nm TiO 2 passivation layer.This 20 nm passivation layer is considered to be the optimal thickness for maximum suppression of recombination.Fill factor (FF) for DSC without TiO 2 passivation layer is 0.439, whilst, for DSC with 10, 20, 50, and 100 nm of It is observed that the FF value is optimum when DSC is added by 20 nm TiO 2 passivation layer.It is the result of less occurrence of recombination because of sufficient barrier provided by 20 nm of TiO 2 passivation layer.This simulation result is in line with the simulation of resistance and also agreeable with the experimental result obtained by Eskandar et al., where recombination resistance peaks when passivation layer thickness is 20 nm, at a point where the transport resistance is minimal.As shown in Table 3 and Figure 5, efficiency of DSC without TiO 2 passivation layer is 3.247%, whilst DSCs with 10, 20, 50, and 100 nm of TiO 2 passivation layer are 3.408%, 3.428%, 3.34%, and 3.341%, respectively.The highest efficiency of simulated DSC is achieved by the DSC with 20 nm TiO 2 passivation layer.
As illustrated in Figure 5, efficiency of DSC with added TiO 2 passivation layer increased by 5.28% compared to the one without any passivation layer.This is attributed to the suppression of recombination by the passivation layer added in between electrolyte and TCO [34] where recombination is more prevalent.The additional TiO 2 passivation layer is also believed to have extended the electron lifetime for such appropriate thickness of passivation layer as also confirmed by Mohamed et al. [4].Thicker passivation layer may lead DSC to have electron trapping which can reduce the efficiency of DSC.Both simulation and experimental result are illustrated in Figure 5 where efficiency was found to be optimum for DSC with 20 nm of TiO 2 passivation layer.
In Table 4,   and  Rec sim are the simulation result for   and  Rec whilst   and  Rec exp are experimental data extracted from Eskandar et al. work [15].The maximum error is 3.237% for RMSE, 1.1% for MBE, and 2.96% for MAE as shown in Table 4.The errors are acceptable since they are still less than 4%.

Effects of Two-TiO 2 Passivation
Layers.From the previous discussion about effects of passivation layer, TiO 2 passivation layer, and thickness of TiO 2 passivation layer, simulation on DSC using two-TiO 2 passivation layers was As can be seen in Table 5,  sc for simulated DSC without any passivation layer, 20 nm TiO 2 passivation layer, and two-20 nm TiO 2 passivation layers are 11.364 mA/cm 2 , 11.649 mA/cm 2 , and 11.678 mA/cm 2 , respectively.It can be seen that simulated DSC with two-20 nm TiO 2 passivation layers has the highest  sc among others which also indicates lower recombination occurred in simulated DSC.
oc for simulated DSC without any passivation layer, 20 nm TiO 2 passivation layer, and two-20 nm TiO 2 passivation layers are 0.651 V, 0.694 V, and 0.692 V, respectively.From Table 5, it can be seen that  oc for simulated DSC with additional TiO 2 passivation layer has higher  oc than simulated DSC without any passivation layer.It shows that, by adding passivation layer, electron will go faster which results in higher  oc .
It is proven that those passivation layers suppress recombination in interface of TCO/electrolyte and TiO 2 /electrolyte (as shown in positions A and B in Figure 1).Lower recombination will lead to longer electron lifetime and higher electron mobility.It is confirmed that adding two-20 nm TiO 2 passivation layers can enhance the performance of simulated DSC.
Passivation layer is added in DSC as light scattering layer with the material the same as the photoelectrode material which is TiO 2 .The presence of this layer can help to suppress the recombination since it is located between TCO/electrolyte and TiO 2 /electrolyte where recombination is more prevalent.Less recombination happens in DSC which can enhance the performance of DSC.

Conclusion and Future Works
Material used and thickness of TiO 2 passivation layer were found to affect the performance of DSC.Simulations have been compared and validated with other experimental works that have error percentage less than 5%.Result showed that adding two-20 nm TiO 2 passivation layers can improve the efficiency of DSC performance by 11% as the result of less recombination, higher electron mobility, and longer electron lifetime in DSC with TiO 2 passivation layer.
Simulation of DSC performance showed that adding TiO 2 passivation layer can get DSC to reach the highest efficiency but practically, the addition of TiO 2 passivation layer may also be affected by the fabrication procedure.For future work, it is suggested that the fabrication of DSC should include additional TiO 2 passivation layer and also other optimum parameters of DSC such as 10 m thickness of TiO 2 [35] and 1 M of iodide electrolyte concentration [36] to get better performance of DSC.

Figure 5 :
Figure 5: Efficiency of DSC with various thicknesses of TiO 2 passivation layer.

Table 2 :
Performance of DSC with various materials as passivation layer.

Table 3 :
Performance of DSC with various thicknesses of TiO 2 passivation layer.

Table 5 :
Performance of simulated DSC without and with single and two TiO 2 passivation layers.