Conductive passivating contact for high fill factor monolithic perovskite/silicon tandem solar cells

Perovskite/silicon tandem solar cells (PK/Si TSCs) blaze the way in pushing power conversion efficiency (PCE) beyond the single‐junction Shockley–Queisser limit. Meanwhile, localized defects in perovskite subcells result in a lower fill factor (FF), which limits further improvement of PCE in PK/Si TSCs. Herein, we report a conductive passivation contact layer by posttreatment of bis(2‐hydroxyethyl)dimethylammonium chloride (BDAC) zwitterion molecule on the perovskite surface. It can passivate the positive and negative localized defects, inhibit the formation of Pb0, and spontaneously convert the perovskite surface to be a more n‐type conductive contact layer for charge separation. These combined enhancements enabled a PCE of 21.4% with an enhanced VOC of 80 mV and an FF of 82.84% for the inverted single‐junction device prepared by the two‐step method. Moreover, BDAC passivation achieved a PCE of 28.67% with an FF of 80.02% for PK/Si TSCs. In addition, the scaling‐up device with an active area of 11.879 cm2 delivers a PCE of 24.46%, and a minimodule with power conversion over 2 W is designed and fabricated.

In 1961, Shockley and Queisser proposed the theoretical limit of 29.4% (S-Q limit) of single-junction solar cells. [1]ntil now, the highest power conversion efficiency (PCE) of 26.81% has been achieved by single-junction silicon (Si) solar cells, [2] which is very close to the S-Q limit.Therefore, further enhancing the PCE of single-junction Si solar cells has become a big challenge.Perovskite/ silicon tandem solar cells (PK/Si TSCs) by integrating wide-bandgap perovskite and narrow-bandgap Si subcells have been widely investigated, which are considered competitive candidates for further boosting the PCE beyond the efficiency limit of the single-junction Si device. [3]Tremendous efforts have been tried to increase the efficiency of PK/Si TSCs, for example, by perovskite composition control, [4] interface passivation, processing method optimization, [5] recombination layer selection, [6][7][8] and so on, the open circuit voltage (V OC ) of PK/Si TSCs have reached up to 1.97 V which is close the limitation. [9]n addition, via device configuration design, [10,11] light management, [12][13][14][15] nano-optical designs, [16] and so on, the short-circuit current density (J SC ) of PK/Si TSCs has already been over 20 mA/cm 2 .Recently, PK/Si TSC has obtained a PCE of 33.7%, [9] confirming the significant potential of this technology.
Many previous reports have demonstrated that the quality of perovskite/hole transfer layer or perovskite/ electron transfer layer (ETL) interfaces has been correlated with device performance, including PCE, hysteresis, stability, and reliability.However, during the thermal annealing processing for the perovskite film, the top surface of perovskite/ETL in a p-i-n type perovskite solar cell (PSC) is prone to defect formation due to easier escape of volatile species such as organic molecules and iodine (I), leading to various defect states associated with undercoordinated Pb 2+ cations, I − /Br − anions, organic cation (MA + /FA + /Cs + ), and halide vacancies.Therefore, perovskite/ETL top surface engineering is significantly important to promote p-i-n type PSCs achieving a higher PCE.Currently, there are a lot of reports focusing on localized defects in single junction perovskite or PK/Si TSCs.[24] However, these passivation layers are dielectric films that can passivate the interface defects very well, yet their insulator nature impedes further carrier transport.On the other side, C 60 bisadduct (indene-C 60 bisadduct) [25] or evaporated cesium bromide thin layer [26] is introduced between the perovskite layer and the electron or hole transport layer to construct a conductive contact layer to promote the carrier transport, and to improve the fill factor (FF).However, those materials that possess many free electrons do not provide a passivation effect.This means that the interface treatment between perovskite and electron transport layer can simultaneously passivate defects and facilitate electron transport, which is particularly important for further improving device performance.
Currently, most defect passivation work focuses on devices prepared by the one-step antisolvent method, whereas sequential deposition, in which inorganic components and organic salts spin-coat sequentially to interact to form a perovskite film without antisolvent, deserves more attention.Narrow-bandgap devices with an efficiency of 25.6% prepared by the two-step method were reported by You's group, [27] demonstrating that this is an effective method to fabricate dense and uniform perovskite film and highly efficient devices.Recently, some efforts were made toward two-step sequential deposition method in PK/Si TSCs. [4,28]However, the PCE of PK/Si TSCs based on the two-step method is far below the devices prepared by the one-step antisolvent method.The reason mainly stems from the undesirable crystallization caused by the solid-solid phase transition during the two-step sequential deposition.In particular, when introducing Br − to form wide-bandgap PSCs, the incomplete solid-solid phase transition further leads to a rougher perovskite surface which will significantly hinder the PCE elevation.Therefore, the development of an efficient passivation method during two-step widebandgap process is of great importance for the advancement of PK/Si TSCs.
In this work, we prepared wide-bandgap (1.63 eV) perovskite devices by a sequential deposition method, a zwitterion molecule of bis(2-hydroxyethyl) dimethylammonium chloride (labeled as BDAC) was introduced in perovskite surface which acted as a conductive passivation contact (c-PC) layer for simultaneously suppressing both the positive and negative localized defects recombination as well as promoting charge-carrier selectivity to produce high-quality perovskite films for utilization in efficient singlejunction and tandem PSCs.The disappeared Pb 0 4f peak and the enhanced contact potential difference (CPD) confirmed the c-PC function of the BDAC.With this surface engineering, the PCE of the p-i-n singlejunction device prepared by a two-step method achieved 21.4% with a FF of 82.84% and PK/Si TSCs increased from 26.71% to 28.67% with FF from 76.51% to 80.02%, and a PCE of 24.46% is achieved in a large area of 11.879 cm 2 .More importantly, a minimodule with a power conversion of 2.116 W is designed and fabricated, which will promote the industrialization of PK/Si products.
The structure of the single-junction p-i-n PSCs in this study is glass/ITO/PTAA/perovskite/PCBM/BCP/Cu, as shown in Figure 1A.Here, 1.63 eV bandgap PSC with an active layer consisting of (FAPbI 3 ) 1−x (MAPbBr 3 ) x was formed via a two-step method of depositing organic salts (FAI, MABr, MACl) precursor onto prepared inorganic film (PbI 2 , PbBr 2 ), followed by an annealing process.The bandgap of 1.63 eV was confirmed by the absorption spectra (Figure S1).The zwitterion molecule of BDAC (0.5 mg/mL in isopropanol) was deposited via spin coating on the perovskite film and then annealed at 100°C for 5 min.The BDAC whose chemical structure is shown in Figure 1A  (D) J-V curves of the champion control and target (with BDAC surface treatment) p-i-n PSCs under 1-sun (100 mW/cm 2 ) illumination.(E) EQE and integrated J sc spectra of the target device.(F) J-V curves from reverse (V oc to J sc ) and forward (J sc to V oc ) scan of the target device.BDAC, bis(2-hydroxyethyl)dimethylammonium chloride; PSC, perovskite solar cell; XRD, X-ray diffraction.
ion.From Figure 1B, BDAC-treated perovskite shows suppression of the PbI 2 diffraction peak at 12.7°along with an enhancement of the perovskite (100) diffraction peak, especially the formation of a more (111)-dominated crystalline perovskite films, which have been proven to have extraordinary moisture-proof stability. [29]In addition, a new diffraction peak appeared at 9.7°, as shown in Figure 1C, which may belong to the two-dimensional (2D) perovskite derived from the in-situ reaction of PbI 2 and BDAC.The effect of BDAC on the morphology of perovskite films was investigated by scanning electron microscopic (Figure S2), the surface of perovskite films treated by BDAC is blurred, and grain boundaries become inconspicuous, indicating that BDAC treatment leads to surface reconstruction of perovskites, and further supporting the formation of 2D perovskite.
Figure 1D shows the current J-V curves of the control and target devices, the control device shows performance with a J SC of 21.47 mA/cm 2 , a V OC of 1.12 V, an FF of 79.72%, and a PCE of 19.20%.The BDAC-processed device shows an improved PCE of 21.41% with significantly increased V OC of 1.2 V and FF of 82.84%, and an equivalent J SC of 21.49 mA/cm 2 , which agrees well with the integrated J SC value of 24.1 mA/cm 2 obtained from the external quantum efficiency (EQE) spectrum shown in Figure 1E.Table S1 presents the survey of widebandgap (≥1.63 eV) PSCs fabricated by the sequential deposition method.The PCE and FF in this work are higher than most of the published results, and the V OC loss is relatively small.This improvement can be attributed to the healing of perovskite surfaces by BDAC, in which the organic functional groups with positive charge can passivate the negative charge localized defects (the FA/MA vacancies), and chloride ion with a negative charge passivates the positive charge localized defects (the iodine vacancies).Previous studies have reported that the interface-recombination-dependent FF increases with V OC according to the following equation: where n is the ideality factor, k is the Boltzmann constant, T is the absolute temperature, and q is the electron charge.Thus, BDAC one material simultaneously achieves two types of localized defect suppression, which will further enhance both of V OC and FF of the PSCs.The BDAC-passivated device shows negligible hysteresis under forward and reverse scanning, as shown in Figure 1F.We demonstrated that BDAC can improve device performance compared with common passivating compounds such as GABr and CC, as shown in Figure S3 and Table S2, GABr-or CC-treated devices had slightly higher V OC and FF and lower J SC , whereas BDACmodified devices achieved higher V OC and FF simultaneously without sacrificing J SC .It is further proved that the zwitterionic molecules of BDAC can passivate the local defects of positive and negative charge, which provides an effective way to improve the PCE of PSCs.
The statistical photovoltaic parameters of the singlejunction perovskite devices (15 cells) with GABr, CC, and BDAC treatments are summarized in Figure S4, further indicating that BDAC significantly improved the FF and V OC of the devices as well as good repeatability.
To investigate the potential interactions of the perovskite with BDAC, high resolution X-ray photoelectron spectroscopy (XPS) is carried out.From Figure S5, the characteristic peaks of the hydroxyl group at 531.5 eV enhanced in the O 1s spectra for the BDACpassivated sample, indicating the BDAC was incorporated into the perovskite film.Figure 2A,B is the XPS spectra of Pb 4f peaks of the control and BDAC films, respectively.The binding energies at 142.9 and 138.0 eV are associated with 4f 5/2 and 4f 7/2 of the divalent Pb 2+ , respectively.For the BDAC treatment perovskite film, those peaks move to higher binding energy by over 200 meV (143.1 and 138.2 eV).This may be due to the surface Pb bond formation with the R 2 NH of BDAC, which is conducive to the organic cation vacancy passivation.The two lower shoulder peaks at 141.2 and 136.3 eV at the binding energy are assigned to metallic Pb 0 (Figure 2A).For the BDAC treatment surface, the Pb 0 peaks are decreased significantly.In fact, the existence of Pb 0 is an indication of the formation of iodide and cation vacancies on the perovskite surface.The absence Pb 0 peak illustrated that the surface vacancies are inhibited by the BDAC treatment.In the I 3d XPS spectrum, the peak intensity of BDAC treatment film is higher than that of control without passivation, as shown in Figure 2C,D.The atomic ratio of the elements of perovskite film with or without BDAC is shown in Table S3.The ratio of (I + Br)/Pb is increased from 2.34 to 2.74, which also confirms that the halide vacancy is suppressed by BDAC treatment.Figure 2E,F presents the C 1s peaks of the perovskite film with or without BDAC, respectively.The 284.8, 286.5, and 288.4 eV of the banding energy are assigned to C-C, C═NH 2+ , and C═O, respectively.Oxidized carbon species of C═O is attributed to the surface decay which may originate from the perovskite film exposed to the ambient environment.We can observe that the peak area of C═O shows a visible decrease after BDAC treatment, illustrating the inhibition of the oxidation process and improvement of film stability.As reported, the C = NH 2+ in the XPS spectrum originates from the organic components (FA + /MA + ). [30]he enhancement peak area of the perovskite film by treating with BDAC also verifies the dual ion passivation function of the BDAC.
The characteristics of perovskite films with or without BDAC treatment are investigated.Figure 3A,B displays the AFM image of control and BDAC perovskite films, respectively.A smoother surface with the RMS of the film decrease from 19.7 to 14.4 nm is achieved by using BDAC treating with the perovskite.Then the conductive atomic force microscopy (C-AFM) of control and BDAC films are tested in Figure 3C,D, respectively.For the control film, there are obvious sites with high current values that act as leakage channels leading to lower V OC and FF.On the contrary, C-AFM image of BDAC-treated perovskite is homogeneous without leakage point, which is beneficial for the realization of high efficiency PSCs.
Furthermore, the Kelvin probe force microscope (KPFM) of control and BDAC films are also employed for detecting the CPD of the perovskite surface as shown in Figure 3E,F.CPD distance is shown in Figure 3G, we can find that the perovskite surface treated with BDAC is 100 mV higher than that of the control film without BDAC.The higher CPD illustrates that a more n-type surface is formed at the perovskite surface.To verify the accuracy of the experiment, the surface potential of perovskite before and after passivation was further investigated by using ultraviolet photoelectron spectroscopy (UPS), as shown in Figure 3H and Figure S6.According to the equation: φ = hν − (E cutoff − E fermi ), the work function of the control and BDAC-treated film are 5.20 and 4.26 eV, respectively.This result is consistent with the KPFM measurement, indicating an n-type layer is formed on the perovskite surface by treating with BDAC. Figure 3I shows the energy level diagram of control and BDAC perovskite films.The KPFM and UPS results illustrate that zwitterion molecule BDAC not only passivates the surface defects but also spontaneously forms an n/n + conductive contact between the perovskite and ETL interface, achieving a c-PC function in the PSCs.
Figure 4A,B is the transient absorption spectrum images of control and BDAC films, respectively, which further verify the c-PC function of the BDAC.The higher intensity at the wavelength of 750 nm associated with the bandgap is shown in Figure 4B for the BDAC-treated perovskite, which may originate from the higher lifetime of the carriers due to the passivation effect of BDAC.carrier lifetime.Moreover, TRPL spectra of control and BDAC films were also carried out to measure the carrier transport and recombination characteristics (Figure 4D).Fast and slow decay times were fitted by double exponential function from decay curves for devices with or without BDAC, as summarized in Table S4.The perovskite films with BDAC exhibit carrier lifetimes (τ) that are much longer (τ ave = 253.79ns) than those of pristine films (τ ave = 77.88ns), which indicates that the carrier nonradiative recombination was significantly suppressed upon the introduction of BDAC.
Space-chargelimited current measurements were performed on both control and BDAC-treated perovskites with configurations of electron-only and hole-only devices to characterize the positive and negative trap densities in the devices.The trap density (N t ) is determined by the trap-filled limit voltage based on the following equation: N t = (2εε 0 V TFL )/qL 2 , where q, L, ɛ, ɛ 0 , and V TFL is the elementary charge, the perovskite film thickness, the relative dielectric constant, the vacuum permittivity, and the trap-filling voltage, respectively.As shown in Figure S7a, the electron trap densities are estimated to be 1.81 × 10 16 and 1.28 × 10 16 cm −3 for control and BDAC-based perovskite films, respectively.Similarly, the corresponding hole trap densities are estimated to be 1.94 × 10 16 and 1.42 × 10 16 cm −3 (Figure S7b).The significant decrease in positive and negative trap densities of BDAC-based devices illustrated the much-enhanced defect passivation effects, leading to the reduced non-radiative recombination losses of the devices.In addition, transient photovoltage and transient photocurrent (TPC) curves of the control and BDAC PSCs are employed to validate the enhanced charge extraction and suppressed trap states in the perovskite films.Signals of the TPC and photovoltage of the devices are obtained under a simulated illumination.As can be seen in Figure 4E,F, the charge recombination time increased from 0.41 to 4.68 μs (Figure 4E), while the photocurrent decay time decreased from 71 to 56 ns (Figure 4F).The longer recombination time indicates the interaction of BDAC with defective perovskite film and then results in suppressed surface trap density, as well as the n-n + conductive contact promoting the extraction of charge carriers.All of the test results verify the c-PC function of the BDAC at the perovskite surface.
Then, we fabricated a PK/Si TSC by using a two-step method as reported in our previous work. [4]The concentration (0.3-0.9 mg/mL) of BDAC is optimized in PK/Si TSCs (Figure S8).As the concentration of BDAC increases from 0.3 to 0.5 mg/mL, the PCE increases.Yet when the concentration of BDAC increases to 0.9 mg/mL, the devices exhibited suppressed performance.The performance parameters are shown in Table S5, the decreased PCE is in accordance with FF which is mainly because of the increased series resistor.Figure 5B shows the J-V characteristics of tandem device, exhibiting a champion PCE of 28.67% with a V OC of 1.86 V and an FF of 80.02% under forward scan, which is among the highest value of two-step perovskite/silicon monolithic tandem solar cell reported so far, although there is a slight hysteresis, this can be addressed by component management and stress regulation in next work.The front surface of the silicon cell is polished to ensure high quality perovskite film, while the antireflection film is matched to reduce reflection loss, and the rear surface is textured to increase light harvesting, resulting in a J SC of 19.25 mA/cm 2 , which is in excellent agreement with that the J SC integrated from the EQE spectra (Figure 5C).In addition, as shown in Figure S9 and Table S6, BDAC exhibited superior optimization effects compared to typical passivators such as GABr, MP (MAI + PEAI), and CC in tandems.The tandems show excellent operational stability under accelerated tests (Figure 5D), a steady-state PCE of about 28.5 ± 0.2% was attained during constant illumination of simulated AM1.5 G light for 3600 s.Upon increasing the aperture area up to 11.879 cm 2 , the BDAC-based tandem cells exhibit a PCE of 24.46% with a V OC of 1.84 V. Possible origins of the low FF and J SC of our device could be attributed to the absence of an optimized front grid.We also employed the BDAC-treated tandem devices to fabricate minimodules including nine subcells with an active area of 11.879 cm 2 , as displayed in Figure 5E and Figure S10, which showed performance with an output power of 2.116 W.

| CONCLUSION
In summary, a zwitterion molecule of BDAC is used to heal the positive and negative defects at perovskite/ETL interfaces, and BDAC treatment promotes growth along with (111) facets with better moisture resistance.KPFM and UPS show that an n-n + conductive contact layer is formed on BDAC-treated perovskite surface for promoted charge-carrier selectivity.Taking these advantages, the 1.63 eV bandgap inverted device with BDAC passivation delivers a PCE of 21.4% with improved FF of 82.84% and 80 mV higher V OC than the control device.Moreover, the BDAC-based tandem cell exhibits a PCE of 28.67% and an FF of 80.02% with an active area of 0.5003 cm 2 , as well as a minimodule with an output power over 2 W is proposed by composing of nine large areas PK/Si TSCs.This work will pave the way for the wide use of c-PC concepts in high-efficiency PSCs and perovskite-based tandem solar cells or other metal halide optoelectronic devices soon.

| Fabrication of p-i-n singlejunction PSCs
The tin-doped indium oxide (ITO) glass was successively cleaned by sonication with deionized water, ethanol, acetone, and isopropanol as agents followed by UV-ozone for 20 min.PTAA (5 mg/mL in chlorobenzene) was spincoated on the ITO substrates at 4000 rpm for 30 s with postannealed at 100°C for 10 min.The perovskite layer was via a two-step spin-coating process.In the sequential process, 1.3 M PbI 2 /PbBr 2 (0.83/0.17, molar ratio) in a mixture of DMF and DMSO (9:1, vol/vol) was spin-coated on PTAA-ITO substrates at 2500 rpm for 30 s and then annealed at 70°C for 1 min.After cooling down, the lead halide layer was spin-coated by a solution of FAI/MABr/MACl (60:6:6 g/mL in IPA) at 2700 rpm for 30 s.The as-prepared perovskite films were then annealed on a hot plate at 150°C for 15 min in ambient air conditions (30-40% humidity).The passivating agents, including guanidine bromide (GABr), choline chloride (CC), and BDAC, were 0.5 mg/mL, dissolved in IPA, and were spun on the surface of perovskite at 4000 rpm for 30 s, and then annealed at 100°C for 5 min.PCBM was spun on the modified perovskite film at 2000 rpm for 30 s, and BCP solution (0.5 mg/mL in IPA) was spin-coated at a speed of 5000 rpm for 30 s. Finally, 80-nm-thick copper was deposited using a thermal evaporator to form a completed device.A 0.08875 cm 2 mask was used to define the accurate aperture area for the cells.

| Fabrication of PK/Si TSCs
PTAA solution (5 mg/mL in CB) is spin-coated on the substrate with ITO tunnel junction at 5000 rpm for 30 s, and anneals at 100°C for 5 min.The perovskite absorption layer is fabricated as mentioned above.20 nm of C 60 is thermally evaporated as the electron transport layer in another evaporation system at a working pressure of <2 × 10 −6 mbar and the evaporation rate was 0.7 Å/s (±1.5%), calibrated by the quartz crystal monitor.A 35 nm SnO 2 prepared by atomic layer deposition is used as the buffer layer.Subsequently, 100 nm sputtered indium zinc oxide is used as a transparent electrode, and the active area of 0.5003 or 11.879 cm 2 is calibrated by the thermally evaporated grid Ag/Al electrode and a shadow mask.Finally, a light management antireflective foil of polydimethylsiloxane polymer with a pyramid size of 3-5 µm is put on the top of the active area.

| Device characterizations
The current density-voltage (J-V) characteristics of the devices are measured by Keithley 2400 Source meter under AM 1.5G illumination with Xenon-lamp based solar simulator (Enli.Tec.).The intensity of the light is calibrated with the Si reference cells, which KG-5 and QTZ filters were used for the measurements of wide band gap PSCs and PK/Si TSCs, respectively.External quantum efficiency (EQE) spectra were conducted by Enli.Tec.measurement system for the single-junction PSCs from 300 to 900 nm with a scanning step of 10 nm.For the tandem devices, to measure the EQE of the perovskite subcell, the silicon subcell was saturated using a light-emitting diode (LED) with 800 nm peak emission and applying a bias voltage of 0.5 V.The silicon subcell was measured by saturating the perovskite subcell with blue light from an LED (600 nm) and applying a bias voltage of 1 V.The photoluminescence and time-resolved photoluminescence (TRPL) spectroscopies were measured with a laser wavelength of 475 nm and a power of 0.2 mW (Edinburgh FS5).The root mean square (RMS) roughness of different films was taken from atomic force microscopy (AFM) images measured with Bruker (Dimension Icon).
Han, and Qian Huang helped sputter indium-zinc oxide.Bingbing Chen, Jin Wang, and Ningyu Ren contributed equally to this work.All authors discussed the results and revised the manuscript.
can be formed by a group of bis(2-hydroxyethyl)dimethylammonium and a chloride F I G U R E 1 (A) Schematic diagram of device structure and interaction mechanism between BDAC and perovskite.(B) XRD patterns of control and BDAC-treated perovskite.(C) XRD patterns of the control and BDAC-treated perovskite films in the local range (8-15°).
Figure 4C displays that perovskite film with BDAC treatment presents a faster charge extraction and a longer F I G U R E 2 High-resolution X-ray photoelectron spectroscopy of the perovskite film without (Control) and with (BDAC) passivation.(A, B) Pb 4f of Control and BDAC films, respectively.(C, D) I 3d of Control and BDAC films, respectively.(E, F) C 1s of Control and BDAC films, respectively.BDAC, bis(2-hydroxyethyl)dimethylammonium chloride.

F I G U R E 3
Properties of Control and BDAC perovskite films.(A, B) Atomic force microscopy images of Control and BDAC perovskite films, respectively.(C, D) Conductive atomic force microscopy of Control and BDAC films, respectively.(E, F) Kelvin probe force microscope of Control and BDAC films, respectively.(G) Contact potential difference linear spectra of Control and BDAC films.(H) Ultraviolet photoelectron spectroscopy of Control and BDAC films.(I) Energy level diagram of Control and BDAC perovskite films.BDAC, bis(2-hydroxyethyl)dimethylammonium chloride.

F I G U R E 4
Properties of Control and BDAC perovskite films.(A, B) Transient absorption spectrum images of Control and BDAC films, respectively.(C) Exciton decay dynamics of Control and BDAC films.(D) Time-resolved photoluminescence (TRPL) spectra of Control and BDAC films.(E, F) Transient photovoltage and transient photocurrent curves of the Control and BDAC PSCs.BDAC, bis(2-hydroxyethyl)dimethylammonium chloride; PSC, perovskite solar cell.