Influence of Different Spherical Binary Plasmonic NPs on HTM Layer in Methyl Ammonium Lead Triiodide Solar Cell

© 2019 The Authors. This article is licensed under a Creative Commons Attribution 4.0 License Abstract. Methylammonium lead triiodide perovskite solar cells have attracted huge research interest. Its optoelectronic properties are competing with those of silicon wafers. It is a hybrid absorber with a direct band gap of about 1.53 eV with good light-absorption capability appropriate for optoelectronic applications. A typical perovskite solar cell HTML layer rarely incorporates ZnO or Cu2O or TiO2 nanoparticles to increase charge carrier transport. These ZnO, Cu2O, TiO2 nanoparticles can be introduced into the HTM layer to modify its PSCs efficiency and performance. These nanoparticles are direct band gap binary semiconductors with a wide band gap energy range of 2.17 eV to 3.37 eV respectively which can lead to higher transport mobility and enhanced HTM nanostructured layer. In this paper, two model solar cell having a ITO/TiO2/CH3NH3PbI3/P3HT/Ag and ITO/TiO2/Ag:CH3NH3PbI3/P3HT/Ag structures were proposed, geometrically modelled and simulated using SCAPS-1D software. Their HTM layer (composed of P3HT) was doped with ZnO, Cu2O, and TiO2 nanoparticles respectively to determine their influence on PCEs of this solar cells. It was revealed that starting from undoped P3HT layer all through the Cu2O, ZnO to TIO2 doped layers, efficiency reduced from 13.123 % and 9.071% respectively; fill factor (FF) also reduced from 69.4% to 48.9 % for the doped CH3NH3PbI3 perovskite solar cell while efficiency of doped CH3NH3PbI3 perovskite solar cell reduced from 13.033 % and 9.091%, the fill factor (FF) also reduced from 66.4% to 52.9 % respectively. It was noted that the solar cell employing P3HT undoped layer had the best performance and concluded that introducing nanoparticles onto P3HT layer has a negative impact on the performance of CH3NH3PbI3 perovskite solar cell.


INTRODUCTION
Modern thin-film solar cells have reduced material consumption [2] and fabrication costs. However, a major limitation facing hybrid perovskite thin-film solar cells is its poor photon absorption [1]. It is documented that light-trapping can be increased by increasing the optical path length inside a film, but how to implement it remains a mirage [4]. Plasmonic structures can modify the excitations of localized surface plasmon to improve photon absorption [5]. Surface plasmon is a collective oscillation of free exciting electrons of metallic nanoparticles [5,7,9] which can be used to enhance optical absorption through scattering and near-field concentration photons depending on particle shape [11], size [15], inter-particle distance [22], optical resonance [25], material nature [34] and type of coupled systems avail-able [29]. However, plasmon coupling only occurs when closely spaced nanoparticles have their associated electron oscillations influencing its local field affect electron oscillations of neighbouring particles [28,31,33]. Many noble metallic nanoparticles have been tested for plasmonic coupling. These include gold (Au), copper (Cu), silver (Ag) and aluminium (Al) nanoparticles [9,12,18]. Their dipole and quadrupole plasmon resonances can be described qualitatively using spherical nanoparticles [20,26]. They have already shown a red-shift in total absorption flux enhancement and tunable localized surface plasmon resonance. The role of near-field coupling, resonance, scattering and transmission of light have been investigated too using Ag nanoparticles [6,11,27]. An enhancement factor of 2.3 in external quantum efficiency at 1100 nm wavelength, a sevenfold enhancement of light absorption and a 16-fold enhancement for 1250 nm using Ag nanoparticles have been reported by some publications [14,18,25,33]. This paper focuses on the doping the HTM layer and its influence on the efficiency of methylammonium lead triiodide solar cells as simulated using SCAPS-1D software at AM1.5G solar radiation that employs FDTD, FEM, and FIT methods.
SCAPS-1D simulation software. SCAPS-1D is a onedimensional solar cell simulation software that employs three coupled differential equations. The first equation is the Poisson's equation given as (1): , where ψ is electrostatic potential, q is electron charge, p is free holes, n is free electrons, pt is trapped holes, nt is trapped electrons, Na − is ionized acceptor-like doping concentration, and Nd + is ionized donor-like doping concentration [24,30].
The second equation (2) is the continuity equation for holes as [2,5,11]: (2) while the 3 equations is the continuity equation for electrons as: where G is generation rate, ξ is permittivity while D is the diffusion coefficient.
Computer Simulation Technology. Computer Simulation Technology (CST) is a 3D simulation software used to numerically calculate optical properties required for many applications that include plasmonic solar cells, electromagnet metamaterials and antennas when exposed to an electromagnetic field. It is a 3D EM solver that solves Maxwell's equation in the time domain with Finite integration method (FIT) and frequency domain with the Finite element method (FEM) by incorporating FDTD and FDFD techniques. CST software is a Multiphysics software that employs FIT, FEM, FDTD and FDFD to simulate designed plasmonic nanostructures of various PV devices. CST is a 3D simulation software used in order to attain accurate simulation results under the absorption profile and internal quantum efficiency as a function spatial position, photon density and frequency. It employs FDTD technique to evaluate electron transport, series resistance and fill factor enhancement, distribution profile of the photons absorbed and also electron-hole pair recombination rate. It is a time domain technique used to solve Maxwell's equations in differential form over a grid-based domain in a single simulation by calculating electric field, E, and magnetic field, B for time irrespective of the designed 1D, 2D and 3D models expressed as [14,15,23]: It reveals nano-structural shapes, material types, dielectric environments, array pitches and particle locations for the source of light scattering. Its optical computational power provides the final power conversion efficiency of plasmonic solar cells. FEM available in CST software and it is used to calculate solutions to partial differential and integral Maxwell's equations. It is suitable for simulating irregular shaped geometrical models for optical devices as it provides information even of large dtime and frequency domains small elements in regions where fields may abruptly change. It can also simulate larger elements in less important and unexpected electromagnet regions.

METHODOLOGY
Device Structures. A plasmonic model solar cell was designed containing silver nanoparticle doped and undoped methylammonium lead triiodide (CH3NH3PbI3) as a perovskite absorber layer. The solar cell had five layers as a glass cover (protection purposes), an anode (transparent conducting film), substrate (glass), n-type compact layer, absorber layer (Ag:CH3NH3PbI3 and/or CH3NH3PbI3), P3HT (p-type hole transport layer) and a cathode (silver) designed according to [35] as shown in Figure 1. It was modeled as shown in Figure 2. All the models were considered to be solid-state planar heterojunction p-i-n solar cells with low p-typedoped CH3NH3PbI3 sandwiched between the ntype ETM (compact TiO2) and p-type HTM (P3HT) layers according to [15].
Simulations were performed using light that propagated along z direction, through TiO2, CH3NH3PbI3 and entering P3HT hole transporting layer in that order. The same design was maintained for the one with or without the silver nanoparticle in its perovskite layer. The p-type hole transport (P3HT) layer also modelled as doped with ZnO, Cu2O and TiO2 nanoparticles separately and simulated using CST (microwave studio). Numerical simulation using SCAPS-1D simulation software was carried out to analysed the electrical parameters and the PCEs of the resulting solar cells. Table 1 and Table 2 shows the model parameter of the solar cells that were simulated.  Device Simulation. Two simulation software were used in this work. The SCAPS-1D simulation software was used to numerically simulate photovoltaic cell analysis while other parameters were simulated using Computer Simulation Technology (CST) software. These are easy to use software's available from authorized vendors.
The solar cell models in Table 1 and Table 2 were simulated.
Choice of simulation parameters. The material optical parameters used in SCAPS-1D simulation were selectively picked from published articles by [35] and a number of supporting optoelectronic theories. The other optical constant and absorption coefficients for both CH3NH3PbI3 and Ag:CH3NH3PbI3 were obtained from [16]. A summary of some of these optical parameters are as tabulated in tables 1. It was adopted that the defect density of 1 × 10 10 cm −3 , electron hole and thermal velocity of 1.0 ×10 7 cm/s, donor density, (ND) of zero (0), defect reference energy level was taken above Ev while energy level reference to 0.7e V respectively. Finally, energetic distribution and reference point for defect energy level (Et) were assumed to be single and above the highest Ev respectively when simulating the active layers (CH3NH3PbI3, Ag:CH3NH3PbI3) and perovskite/P3HT interface layers. Numerical simulation was finally performed. Limitation on Simulations. All simulations were limited to either thickness, defect at hole transporting (HTM) layer, the density of states (DOS) and different nanoparticles on P3HT host HTM layer. As a result, thermal velocity of electrons and holes were taken as equal at 1 × 10 7 cm/s while an illumination of 1000W/m 2 , temperature of 25°C, and an air mass of 1.5G were adopted and the simulation software was limited to by SCAPS-1D simulation software to determine to shortcircuit current density, Jsc, open-circuit voltage, Voc, fill factor, FF, and power conversion efficiency, η, for different doped HTM layer. All other simulations were subjected to CST (microwave studio) software.

RESULTS AND DISCUSSIONS
Influence of CH 3 NH 3 PbI 3 and Ag:CH 3 NH 3 PbI 3 layer thickness. Figure 3 shows the variation of thickness of CH3NH3PbI3 and Ag:CH3NH3PbI3 absorber layers with respect to open-circuit voltage. It can be noted that the open-circuit voltage of CH3NH3PbI3 layer increases gradually as thickness increases to a certain peak point. This in effect implies that the efficiency increases up to a certain value which can be considered as the optimum at a specific thickness. Therefore, with increasing thickness, the short circuit current (Jsc) increases and therefore a thicker absorber layer will absorb more photons which in turn, will relatively create more electron-hole pairs resulting in higher open-circuit voltage values. From figure 3 also, it can be observed that a thickness of 320-440 nm is appropriate for Ag:CH3NH3PbI3 perovskite layer for optimal photon absorption based on the AM 1.5G simulation radiation. The optimum thickness that recorded the highest VOC for both Ag:CH3NH3PbI3 and CH3NH3PbI3 layers was 420 nm. The gradual increase in VOC from 280-420 nm was attributed to the plasmonic contributions due to silver nanoparticles with a maximum localized surface plasmonic effect attained at 420 nm. Beyond this thickness, chances of recombination of electron-holes increase as the charge carriers traverse the longer distance in thick films and therefore diffusion is hampered. Thickness is a major parameter that plays a vital role in the overall performance of the solar cell. It could be the thickness for the absorber layer, HTM layer or the back contact. However, the efficiency of a perovskite solar cell depends largely on its response to the solar spectrum as influenced by thickness.

Figure 4 -Influence of thickness on open-circuit voltage perovskite layer
The thickness of Ag:CH3NH3PbI3 absorber layer was varied from 280 nm to 500 nm while that of the HTM layer was varied between 28 nm to 50 nm. The simulated VOC was tabulated as shown in figure 1 using  (5): , where I0 and IL are dark saturation current and light-generated current, while kT/q and A are the thermal voltage and the photodiode ideality factor [31].
It also implies that the fill factor of this solar cell will decrease as thickness increase and the consequence will be that the cell internally consumes power reducing efficiency after a certain peak thickness. Similar observations were re-ported by the experimental result by [16], who concluded that recombination expedites with increasing thickness.

Influence of Defect State at HTM / perovskite Interface
Layer. Methyl ammonium lead triiodide films have a number of defects. These include point defects. A defect layer on the Ag:CH3NH3PbI3/ HTM interface layer was considered during simulations which took into account the interface recombinations. The simulation parameters used are shown in Table 4. Integrated total density 1/cm 2 1.0 × 10 +12 Figure 5 shows the effect of interface defect density versus efficiency curves for three different nanoparticle dopants on the HTM layer. From figure 5, it can be noted that there was a negligible effect on efficiency above defect density of 1.12 × 10 12 cm −3 . When defect density went below 1.01 × 10 12 cm −3 , a notable decrease in PCEs was recorded. With an increase in defect density, the recombination rate increases and as a result efficiency decreases. It was noted that doping P3HT layers with binary impurities reduced its potential as a HTM layers for hybrid perovskite solar cells. Such attempts have not been carried out through attempts to practically implemented this are on progress. HTM layer can be deposited by sputtering [10], copper oxidation [12,15], spin coating, atomic layer deposition and even more expensive techniques like molecular beam epitaxial technique.
Influence of Density of State of perovskite Layer. Figure 6 shows the variation of density of state (DOS) with thickness.

Figure 6 -Defect density of the perovskite layer versus thickness
The effect of DOS of the CH3NH3PbI3 perovskite absorber layer was simulated with respect to thickness without accounting on the influence of the doped HTM layers. It can be observed from figure 6 that the density of state (Nv) varied from 13.76 × 10 19 cm −3 to 7.5 × 10 19 cm −3 for CH3NH3PbI3 layer and from 9.75 × 10 19 cm −3 to 7.79 × 10 19 cm −3 for Ag:CH3NH3PbI3 layer. This suggested that as thickness increases, the rate of recombination increases and this introduces multiple parasitic capacitances that negative affect solar cell efficiency.
Influence of Valance band effective density of state on efficiency. Figure 7, 8 show the curves of valance band effective density of state versus efficiency for CH3NH3PbI3 and Ag:CH3NH3PbI3 layers. From the curves in figure 6, it can be noted that in general, the cell efficiency of CH3NH3PbI3 absorber layer decreases with the increase in valence band effective density (Nv). As the number of holes increases in the CH3NH3PbI3 absorber layer. Their possibility of taking part in reverse saturation current once the solar cell circuit is completed also increases. Consequently, the open-circuit voltage declines which leads to low electric conversion efficiency and hence poor solar cell performance.
Since low electron affinity of the HTM layer usually has a great significance on charge carrier mobility low affinity safeguards high-hole mobility. As a consequence, carrier mobility in P3HT as a hole transporting layer increases it as a reliable hole transport material (HTM) for hybrid perovskite heterojunction devices. Influence of silver Back Contacts. Simulations were carried out using the silver metal paste as a prospective back contact for methyl ammonium lead triiodide perovskite solar cells to determine the effect of doping P3HT as its hole transport layer. Parameters used in simulating the back contact are as shown in Table 5 while Table 4 illustrates the simulation result of efficiency. It was ob- served that the performance of the solar cells decreased depending on the type of nanoparticle material used. Cu2O dopant had the highest efficiency while TiO2 dopant had the least efficiency. This showed that doped P3HT was energetically unfavourable for holes to travel towards the silver electrode with an opposing electric field within or close to HTM since the back contact becomes negative. The decrease in efficiency was also attributed to the back contact as shown in Table 6. When compared to Au contact used elsewhere [17,18,21,28], it was anticipated that the lower work function of silver metal contact was attributed to the lower efficiency. It was recommended that Au should be given a priority when developing hybrid perovskite solar cells.
Numerical Analysis open-circuit voltage. Figure 4 shows  Figure 4.

Figure 9 -Influence of HTM thickness on open-circuit voltage
It was observed from Figure 9     The HTM layer structure suggested here is different from that proposed by other researches where Ag:CH3NH3PbI3 absorber perovskite layer instead of CH3NH3PbI3 absorber layer. In all model structures, Cu2O, ZnO, and TIO2 were considered as dopant in HTM layer. It was concluded that Cu2O ensures the highest performance among dopant nanoparticles slightly lower than P3HT layer. This was attributed to the properties of Cu2O since it is a 2.17 eV direct bandgap p-type binary inorganic absorber solar cell material [23]. Starting from P3HT layer through Cu2O, ZnO to TIO2 doped layers, the efficiency of 13.123 % and 9.071% FF range of 69.4% to 48.9 % for CH3NH3PbI3 perovskite solar cell and 13.033 % and 9.091% FF range of 66.4% to 52.9 % CH3NH3PbI3 perovskite solar cell.

CONCLUSION
As compared to other simulated perovskite solar cell efficiencies, the I-V characteristics obtained in this work reflect the performance outcome expected by this perovskite layer of 450 nm thickness exhibited by a valance band density of states of 3.2 × 10 18 cm −3 and interface defect density of 1.12 × 10 12 cm −3 respectively. ITO/ TiO2/ CH3NH3PbI3/P3HT/Ag perovskite structure gave the best performance among all the PSCs simulated. From the findings of this study, it was revealed that among all the simulated solar cells, the solar cell employing P3HT undoped HTM layer had the best performance. Open-circuit voltage and short circuit current changes were significant. This finding shows that we can modify the performance of a hybrid solar cell by modifying its organic HTM layer using binary inorganic plasmonic nanoparticles. It was therefore concluded that these findings can be used to justify the model in this paper as a potential alternative way to developing conventional hybrid perovskite solar cells. It was recommended that experimental investigation was required to determine its viability. It was concluded that binary inorganic nanoparticles hamper the performance of P3HT HTM layer for solar cell applications.