Improved PEDOT:PSS/c-Si hybrid solar cell using inverted structure and effective passivation

The PEDOT:PSS is often used as the window layer in the normal structured PEDOT:PSS/c-Si hybrid solar cell (HSC), leading to significantly reduced response, especially in red and near-infrared region. By depositing the PEDOT:PSS on the rear side of the c-Si wafer, we developed an inverted structured HSC with much higher solar cell response in the red and near-infrared spectrum. Passivating the other side with hydrogenated amorphous silicon (a-Si:H) before electrode deposition, the minority carrier lifetime has been significantly increased and the power conversion efficiency (PCE) of the inverted HSC is improved to as high as 16.1% with an open-circuit voltage (Voc) of 634 mV, fill factor (FF) of 70.5%, and short-circuit current density (Jsc) of 36.2 mA cm−2, an improvement of 33% over the control device. The improvements are ascribed to inverted configuration and a-Si:H passivation, which can increase photon carrier generation and reduce carrier recombination, respectively. Both of them will benefit the photovoltaic performance and should be considered as effective design strategies to improve the performance of organic/c-Si HSCs.

(PCE) of the inverted HSC can reach as high as 16.2%, an improvement of 33% over the control device. Further device analysis based on carrier lifetime measurement show that the PCE of the inverted PEDOT:PSS/c-Si HSC can potentially exceed 20%.

Results and Discussion
To investigate the effects of a-Si:H passivation and back PEDOT:PSS on the performance of HSC, two modified HSCs have been proposed and fabricated as shown in Fig. 1a,b and listed in Table 1. Figure 1c,d presents the energy level diagram and carrier transport process in the Grid Ag/ITO/a-Si:H/c-Si/PEDOT/Ag solar cell. As seen in Fig. 1c, the levels of PEDOT are well aligned with the bands of n-type c-Si, which allows holes in silicon to flow into PEDOT unimpeded and reflect electron back into the bulk and reduce the surface recombination (Fig. 1d). Additionally, the insertion of n-type a-Si:H on the cathode electrode not only decreased the contact resistance, but also formed a electric field, which can facilitate electron collection as well as repel the hole into the bulk to reduce surface recomnication. These characters of this device configuration could increase photogenerated carriers collection efficiency and reduce dark current or J 0 and carrier recombination, which leads to higher V OC .
The PEDOT:PSS/c-Si interface plays a key role in the cell performance 9,12,15 . Figure 2a,b show the top-view SEM images of pyramids structured c-Si wafer surface with and without PEDOT:PSS, respectively. It is seen that the bare c-Si wafer surface consists of pyramids with different size in the range of 5~10 μ m. When the PEDOT:PSS infiltrates the gaps between the pyramids, it forms a uniform layer above the textured Si wafers. It is clearly seen that the c-Si pyramids are completely coated by PEDOT:PSS (Fig. 2c). Figure 2d shows cross-sectional image of the pyramid surface covered by the PEDOT:PSS layer. The hills and valleys of the Si pyramids are fully covered with about 100 nm PEDOT:PSS. The better coverage on the valleys is due to chemical etching, the larger pyramid size and better adhesion to hydrophobic c-Si surface comparing to the poorer coverage reported 23 . The good coverage with PEDOT:PSS coating is of benefit to reduce the surface recombination and charge collection 30 .  To investigate the effect of a-Si:H on the surface passivation, a normal HSC with an a-Si:H passivation layer has been prepared. In Fig. 3a, the J-V curve of the champion PEDOT:PSS/c-Si HSC with and without the a-Si:H layer has been compared, with key J-V parameters listed in Table 2. It shows that the best cell without the a-Si:H film exhibits J sc of 26.8 mA cm −2 , V oc of 548 mV and FF of 56.5%, yielding a PCE of 8.3%. The lower V oc derives from the serious recombination losses at the fully metalized rear electrode, the direct contact between the metal electrode and the doped c-Si apparently forms a Schottky-barrier, as commonly observed at the metal-semiconductor interface 31 , leading to high contact resistance and thus an inferior FF and J sc . When the c-Si rear surface is passivated by intrinsic a-Si:H film and back surface field is constructed using highly doped n-type a-Si:H layer, the PCE of normal HSC is improved to 12.1% with J sc of 29.7 mA cm −2 , V oc of 620 mV and FF of 65.8%. The substantial efficiency increase is attributed to the high quality passivation of the intrinsic a-Si:H thin film and n-type a-Si:H as back surface field. The intrinsic a-Si:H film is known to provide effective passivate to the surface defects between the c-Si and a-Si:H(n) layer, resulting in reduced recombination rate at the c-Si surface via a downward band bending for reflecting holes (minority carrier) 32,33 . In addition, the highly doped a-Si:H(n) film generates the strong built-in back surface electric field (BSF) and improves the electronic contact at the back  surface, facilitating the separation and diffusion of the minority carriers at the front heterojunction, therefore significantly enhances the HSC cell performance.
As the a-Si:H could passivate the c-Si surface as well as form a transparent electrode with ITO, the inverted HSC with transparent ITO electrode and PEDOT:PSS back electrode has been fabricated and characterized. The ITO and a-Si:H thin film serves as window layer for the inverted PEDOT:PSS/c-Si HSC, the intrinsic a-Si:H is employed as the passivation layer, and the highly doped n-type a-Si:H forms a front surface field. Figure 3a compared the J-V curves of the best normal and inverted types of PEDOT:PSS/c-Si HSC passivated using the a-Si:H thin films. Surprisingly, the inverted PEDOT:PSS/c-Si device exhibits much better photovoltaic performance with J sc 36.2 mA cm −2 , V oc 634 mV and FF 70.2%, giving a PCE of 16.1%. It is obvious that all key parameters are dramatically increased for the inverted HSC with PCE increased by as much as 33% compared to the normal-type HSC. In addition, as the PEDOT:PSS is hygroscopic in nature, it may result in degradation of the HSC performance 23 . However, in the inverted HSC, the PEDOT:PSS layer is fully covered by the metal electrode, isolating it from direct contact with air. Therefore, it may make the inverted HSC more stable than the normal structured HSC.
The dark J-V characteristic is often a good indicator for the solar cell performance. Figure 3b shows the J-V curves measured in dark for the normal HSC with and without a-Si:H passivation layers to compare with the inverted HSC. It is known that to attain a high V oc , the reverse saturation current density J 0 must be reduced because the leakage current leads to reduced V oc . In theory, the dark J-V curves can be simulated using the diode Equation (1) where n the diode ideality factor, k the Boltzmann constant, T the absolute temperature and e is charge of an electron. Figure 3b shows J dark -V curves of the normal HSC with and without the a-Si:H passivation layers compared with the inverted HSC measured in the dark. The J dark -V response in region A is affected by the shunt behavior, while the voltage ranges B and C rely more on exponential diode behavior and the series resistance, respectively. The reverse saturation current density J 0 and the diode ideality factor (n) can be determined using numerical fit to the J dark -V curve in the region B. Based on the least square fitting of the J dark -V characteristic curves, the diode ideality factor (n) and reverse saturation current density (J 0 ) of the PEDOT:PSS/c-Si heterojunction solar cells are extracted, as summarized in Table 3. The n values of the normal HSC without a-Si:H passivation layer are higher than what with passivation, which is most likely attributed to the injection-dependent recombination at the entirely metalized contact. The inverted HSC with a-Si:H solar cell shows the smallest n value, indicating the best passivation quality of the a-Si:H layer with good p-n junction. In additional, the J 0 values displayed a similar tendency with the inverted HSC showing the smallest J 0 value and the highest V oc . Figure 4a shows the external quantum efficiency (EQE) characteristics of the HSC including both normal and inverted structures with and without the a-Si:H passivation. The J sc was confirmed by integrating the EQE with a standard AM 1.5 spectrum. The EQE integrated J sc of the normal HSC with the a-Si:H layers increased by 2.9 mA cm −2 compared with the normal device without a-Si:H. The high EQE and J sc are ascribed to higher passivation quality using the intrinsic a-Si:H and n-type a-Si:H as BSF, leading to reduced recombination at the back surface and enhanced carrier collection. Meanwhile, EQE of the inverted HSC with a-Si:H layers is significantly higher than that of normal device. It is attributed to the EQE gain from increased light trapping in spectrum 500-1200 nm. Figure 4b shows the reflectivity spectra of the HSC. It is clear that the reflectivity of the inverted HSC is lower than that of the normal device in 400-800 nm region because of the 80 nm thick ITO film in the inverted device serving not only as a transparent conductive coating but also an antireflection layer for improved light utilization. However, the EQE below 500 nm is low due to the absorption of a-Si:H. We therefore control the   bilayer thickness to <15 nm. As the photon flux in this range is far lower than that in the red spectrum, even though the EQE below 500 nm is low, its contribution to J sc is very little as shown in Figure S2.
The reflectivity spectra of all three samples are similar in 800-1100 region. However, there are huge differences in EQE results among inverted and normal structured HSCs, which should be ascribed to the parasitic absorption in PEDOT:PSS layer reducing the photo response in red and NIR region. As the photon flux at 600-1000 nm in the AM1.5 spectrum is more abundant than in the UV region, the EQE enhancement to J sc in the long wavelength range is more pronounced, as seen in Figure S2. The EQE results clearly disclose the advantage of inverted HSCs in red and NIR photon utilization, which can effectively increase the photocurrent.
To elucidate the potential advantages of passivation of the c-Si using a-Si:H and PEDOT:PSS, we measured minority carrier lifetime (MCL) with transient photoconductance decay (PCD) using a WCT-120 lifetime tester from Sinton Instruments 35,36 . The MCL is up to 112 μ s for PEDOT:PSS/c-Si/a-Si:H heterojunctions at an injection level of 1 × 10 15 cm −3 , as seen in Fig. 5a, ten times larger than device without a-Si:H passivation. Based on the injection-dependent lifetime data, the reverse saturation current density J 0 is calculated by fitting the reciprocal effective lifetime 1/τ eff versus the excess carrier concentration Δ n using the Equation (2) where W is the thickness of wafer, q the elementary charge, and n i = 9.63 × 10 9 cm −3 is the intrinsic carrier concentration of silicon at absolute temperature (298 K), N dop = 3 × 10 15 cm −3 37 . Note that the influence of bulk minority carrier lifetime (τ bulk ) on the J 0 value is negligible because the value of τ bulk is much higher than τ eff . An effective surface recombination velocity S a-Si:H is measured in the range between 3.5 and 7.5 cm/s (excess carrier density in the range between 10 14 and 10 16 cm −3 ) for the wafer passivated with a-Si:H, as shown in supporting information ( Figure S3 and Table S1) Fig. 5b shows the J 0 -related contribution of Eq. (2) as a function of excess carrier concentration Δ n in the silicon sample. A low saturation current density J 0 of only 119 fA cm −2 was extracted according to a linear fit of Eq. (2) to the measured data. It is known that V oc is determined by the reverse saturation current density (J 0 ) and J sc 38,39 , as Equation (3)

Conclusion
In conclusion, we implanted the intrinsic and n-type a-Si:H layers to the rear surface of normal PEDOT:PSS/c-Si HSC, which can reduce the carrier recombination rate and increase the PCE from 8.3% to 12.1%. The PCE is further improved to 16.1% using the inverted structure, in which the parasitic absorption of PEDOT:PSS has been eliminated, resulting in an improvement of 33% over the normal-type device. The improvement can be attributed to well-matched energy level, improved light utilization efficiency of inverted structure, and effective suppression of carrier recombination. Furthermore, the device stability has been prominently enhanced. It is expected that PCE of the inverted PEDOT:PSS/c-Si HSC can be further increased to 20% once the passivation is optimized. Noticeably, the HSC with PEDOT:PSS on the front as a window layer is defined as normal PEDOT:PSS/c-Si HSC. The schematic of normal PEDOT:PSS/c-Si HSC is shown in Fig. 1a. The finger grid of titanium (10 nm) and silver (200 nm) is thermally deposited as a top electrode (Fig. 1a) by a bus-finger mask. A 600 nm thick aluminum (Al) film was deposited on the a-Si:H layer along the whole back surface of the Si substrate as the back electrode. Oppositely, the inverted PEDOT:PSS/c-Si HSC with indium tin oxide (ITO) and a-Si:H as window layer is fabricated with PEDOT:PSS deposited on the rear c-Si wafer surface. The ITO is sputtered with CVD on the front and metalized with Ag finger grids. The schematic illustration of inverted PEDOT:PSS/c-Si HSC is shown in Fig. 1b. The structures of as-prepared HSCs are listed in Table 1.

Methods
Measurements. The current density-voltage (J-V) characteristic was measured using a KEITHLEY 2400 source-measure unit both under dark and AM1.5G illumination at 100 mW cm −2 . The J sc was confirmed by the integrated product of the external quantum efficiency (EQE). The EQE was measured using a quantum efficiency measurement instrument (CROWNTECH, QTEST STATION 500TI). The UV-Vis-NIR diffuse reflection spectrum was measured using the SHIMADZU UV-3600 spectrophotometer. The scanning electron microscope (SEM) was performed on a JEOL JSM-6700F. WCT-120 lifetime tester from Sinton Instruments was applied for minority carrier lifetime test.