# The role of structural parameters on efficiency and transparency of semi-transparent non-fullerene organic solar cell

This work has studied ITO/ZnO/PBDB-T:ITIC/MoO3/Ag/MoO3 structure with different active layer thicknesses and various D/M/D thicknesses. The thicknesses of active layers are 53, 59, 72, 91, 100, 114, 143 nm, and DMD (Top contact) thickness are 10 (nm)/dm/30(nm) with dm: 4, 6, 8, 10, 12, 14, 16 nm. We first examined our model’s accuracy in comparison with experimental data reported for the device in Ref.38, in which the D/M/D thickness is fixed to (6 nm/10 nm/40 nm), and active layer thickness varies from 53 to 143 nm. As shown in Figs. 2, 3, 4, the obtained results are in very good agreement with the experimental results. The characteristics parameters are listed in Table 1. To model the ST-OSC’s performances, one has to know the absorption coefficient of the devices with the structures including different layers, especially for different thicknesses of active layers. For this purpose, we calculated the absorption coefficient as a function of the active layer thickness and fitted it to the absorption coefficient reported experimentally. The fitted relation is presented in Eq. (3), which can predict the absorption coefficient for any thickness of the active layer.

$$\alpha_{L} \approx \left( {\frac{{\left( { – 37.12 + 100\lambda } \right) + \left( {0.97 – 3e6\lambda + 2.85e12\lambda^{2} } \right)L}}{{\left( { – 37.12 + 100\lambda } \right) + \left( {0.97 – 3e6\lambda + 2.85e12\lambda^{2} } \right)L_{0} }}} \right) \times \alpha_{{L_{0} }} ;$$

(3)

where $$\lambda$$ is the wavelength, and L is the active layer thicknesses. Knowing the absorption coefficient, $$\alpha_{{L_{0} }}$$, for an active layer thickness, L0, which is reported experimentally, one can find the absorption coefficient for any other thicknesses at different wavelengths.

Figure 2 shows the J-V curve of the devices with structure of ITO/ ZnO/ PBDB-T: ITIC/ MoO3/ Ag/ MoO3, with fixed thickness of D/M/D in 6 nm/10 nm/49 nm, and PBDB-T: ITIC active layer thickness is: a) 53, b) 59, c) 72, d) 91, e) 100, f) 114, g) 143 nm. It is clear from the figures, that there is very good agreement between our model and experimental results. This figure indicates that all devices have the same Voc, which is close to 0.85 V. This value is ~ 0.2 V higher than devices containing traditional fullerene acceptors, due to the high LUMO of non-fullerene acceptors29. This is a major factor that helps improve this material system’s photovoltaic performance. As the active layer thickness was increasing, the Jsc also is increasing.

The transmittance spectrum of the ST-OSC for different active layer thicknesses is calculated and compared with those experimental data. As an example, the transmittance spectrum of the ST-OSC with an active layer of 100 nm is presented in Fig. 3a. The figure shows the experimental transmittance for whole devices, besides, the calculated transmittance of the: MoO3/Ag/MoO3 anode, ITO and ZnO compact layer, the active layer, and the whole device. Moreover, for a better understanding of the device’s semi-transparency, AM1.5 spectral irradiance, $${S}_{AM1.5}\left(\lambda \right)$$, and $${S}_{AM1.5}\left(\lambda \right)*V\left(\lambda \right)$$ are demonstrated. It can be seen that there is a good agreement between the obtained transparency for the device and the experimental data. Also, the error bar is included which shows the model’s accuracy. In the wavelengths of FWHM of $${S}_{AM1.5}\left(\lambda \right)*V\left(\lambda \right)$$, the ITO and ZnO compact layer has more than 85% transparency, and the MoO3/Ag/MoO3 anode transparency is about 60% to 75%, and active layer transparency is about 35% ~ 50%.

In Fig. 3b, the calculated transmittance of the ST-OSC for different active layers thickness is presented. As shown in the figure, the transmittance of the ST-OSC with thin active layers thickness (53–72 nm) is higher than 25% at all wavelengths of FWHM of $${S}_{AM1.5}\left(\lambda \right)*V\left(\lambda \right)$$, which makes it much suitable for widow application. By exceeding the increment of the active layer thickness, the transmittance of the ST-OSC decreases, whereas, for a longer wavelength, it decreases to less than 25%. However, the AVT of the solar cells in the visible region (370–740 nm) of the devices with active layer thickness thinner than 100 nm is higher than 25% and still suitable for widow application.

In Table. 2, we compared the parameters of the solar cell such as short-circuit current, open-circuit voltage, FF, PCE, AVT, and resistances of the modeled devices with experimental data31. In Table.2, Th. represents the calculated data, and Exp. represents the experimental data reported by 31.

As expected and depicted in Table 2 and Fig. 4, with increasing the thickness of the active layer, the photo-absorption increases, and consequently PCE increases. Then, PCE has been decreasing as FF decreases because of the explained recombination effects. FF almost declined with increasing active layer thickness, which is associated with the decrease in shunt resistance values39, 39. Although the Jsc is highest for the active layer with a thickness of 143 nm, the FF is low and is 49.6%. The optimum PCE is obtained at the active layer thickness of 100 nm with a maximum PCE value of 9.32%.

Unlike the PCE, the AVT decreases with increasing active layer thickness. As shown in Fig. 4, and Table 2, for the purposed structures, the devices with PCE of more than 5% and AVT of more than 32% are achievable. The mismatch between experimental and theoretical AVT with increasing the active layer thickness comes from the fitted absorption coefficient (Eq. 3) which has a very low deviation from experimental for thinner active layers.

It is well known that organic thin-film solar cells act as multilayer optical cavities in which the distribution of the optical field is governed by the effect of optical interference, due to the reflection of the incident light at the layer interfaces 41. In the studied devices, the D/M/D top contact which includes 3 layers can be an important multi-layer for optical interference. On the other hand, as the model results for studied structures (Table. 2) are in very good agreement with the experimental data, so the model can be applied to the same structures with different Ag thicknesses in the D/M/D layers. For this purpose, all reported devices in Table 2 have been studied using different metal thicknesses in the D/M/D layer. The considered D/M/D layers has the thickness of MoO3(10 nm)/Ag(dm)/MoO3(30 nm) with dm = 4, 6, 8, 10, 12, 14, 16 (nm). Using the theoretical model previously explained, all performance parameters such as the J-V curve, EQE, T, AVT, and color coordinates are calculated. As an example, for the devices with the active layer thickness of 53 nm, and dm = 4, 6, 8, 10, 12, 14, 16 (nm), the performance parameters are presented in Fig. 5. As shown in Fig. 5a, with increasing the metal thickness, JSC increases, but the VOC does not change. In these devices, the exciton generation rate depends on the optical field intensity which is located close to the anode/active layer interface when light enters through the D/M/D electrode under top illumination42. So, the metal thickness ‘dm’ changes can dominantly affect the JSC values. Figure 5b shows the EQE of the devices as a function of wavelengths, in which the highest EQE value belongs to thick metal layers, dm = 16 nm. For all devices, the transmittance is higher than 25% for most visible wavelengths (Fig. 5c). Finally, Fig. 5d shows the AVT of the devices as a function of metal thickness. The AVT of all devices is higher than 36% and the maximum AVT is obtained for dm = 6 nm. So, all devices can be used in the windows application.

All performance parameters for the device are presented in Table 3. As depicted in the table, all devices are practical for the window application with different PCE, while the highest value for PCE is for dm = 16 nm. Also, the highest AVT was achieved for dm = 6 nm, with 3.11% PCE.

In Fig. 6, we have presented the J-V curves of the devices with different active layer thicknesses and various dm thicknesses. The figure shows that any change in dm thickness doesn’t change the VOC values. The JSC increases with increasing dm and reaches 14.88 mA/cm2 for the sample with an active layer thickness of dm = 16 nm, where the fill factor is the lowest valve (46%-48%) in comparison with other samples. From the point of view of the fill factor, the sample with dm = 16 nm and the active layer of 59 nm has the maximum FF, 70.49%.

In Fig. 7, the devices’ JSC, FF, and PCE are presented as a function of ‘dm’ thickness and for various active layer thicknesses. As shown in the figure, for any fixed active layer thickness, with increasing ‘dm’, the JSC is increasing slightly, and the FF is almost constant, so, the PCE increases slightly. With increasing the thickness of active layers for any fixed ‘dm’, the JSC is increasing, and the FF hasn’t any certain functionality, then, the PCE increases and reaches a maximum, decreases. The maximum value of PCE happens for samples with an active layer of about 100 nm. For devices with a thicker active layer (more than 100 nm), the FF was lower due to the increase in series resistance, which could be due to the distorted distribution of exciton generation within the active layer (most excitons are generated near the anode/active layer interface) and subsequent carrier transport towards respective electrodes.

To show the applicability of the studied devices in the windows application, the AVT values of the devices are shown in Fig. 8. By increasing the active layer thicknesses, the AVT is increasing, then decreases almost linearly. The maximum value for the AVT belongs to the ST-OSC with the active layer thicknesses of 59 nm (see Fig. 7c). As shown in the figure, for a fixed active layer thickness, the AVT value has a maximum at dm = 6 nm, then with increasing the dm value, the AVT decreases. The figure shows that all devices with different dm and active layer thicknesses thicker than 114 nm deserve semitransparent solar cell conditions.

The CIE color space, including the coordinates of ST-OSC consisting of different active layer thicknesses and different ‘dm’, is shown in Fig. 9. The color coordinates of translucent OSCs with an active layer thickness of about 90- 100 nm are located close to the color point or so-called “white dot” in the CIE chromaticity diagram. Proximity indicates that there is a good achromatic or neutral color sensations when looking through devices under AM1.5G illumination. Hence, these devices can transmit high-quality light with near white sensation to the human eye without changing the original color of an object. However, as the thickness of the active layer changes, the color coordinates move in different directions from the white dot. Also, the coordinates are sensitive to dm values, and both coordinates x and y increase with increasing dm (see the inset of Fig. 9). For a device with the best PCE and AVT, the thickness of the active layer is 100 nm and the color coordinates are slightly away from the achromatic point, however, the device does not alter the transmitted light by a large extent.