Boosting the efficiency of quantum dot–sensitized solar cells over 15% through light‐harvesting enhancement

How to improve the capacity of light‐harvesting is still an important point and essential strategy for the assembling of high‐efficiency quantum dot–sensitized solar cells (QDSCs). A believable approach is to implant new light absorption materials into QDSCs to stimulate the charge transfer. Herein, the few‐layer black phosphorus quantum dots (BPQDs) are synthesized by electrochemical intercalation technology using bulk BP as source. Then the obtained BPQDs are deposited onto the surface of Zn–Cu–In–S–Se (ZCISSe) QD‐sensitized TiO2 substrate to serve as another light‐harvesting material for the first time. The experimental results have shown that BPQDs can not only increase the absorption intensity by photoanode but also reduce unnecessary charge recombination processes at the interface of photoanode/electrolyte. Through optimizing the size and deposition process of BPQDs, the champion power conversion efficiency of ZCISSe QDSCs is increased to 15.66% (26.88 mA/cm2, Voc = 0.816 V, fill factor [FF] = 0.714) when compared with the original value of 14.11% (Jsc = 25.41 mA/cm2, Voc = 0.779 V, FF = 0.713).


INTRODUCTION
Quantum dot-sensitized solar cells (QDSCs) use QDs as light-harvesting materials and have attracted some research interests due to their excellent semiconductor properties, including high absorption coefficient, adjustable light absorption range, good humidity, light and thermal stability, and the possibility to generate multiple excitons. [1][2][3][4][5]  beyond 16% in recent report. 6 But the value lags behind those of emerging photovoltaic devices such as perovskite solar cells. 7,8 The primary reasons for the relatively low efficiency could be accounted for the inadequate efficiency of light collection and the influence of charge recombination in QDSCs. [9][10][11] Therefore, improving the use of sunlight is a proper way for the preparation of high-efficiency QDSCs.
Large loading amount of QDs on the TiO 2 film surface is a valid strategy for enhancing light-harvesting in QDSCs. Co-sensitization and secondary/multiple deposition of QD as well as surface engineering strategies have also been explored. 9,[12][13][14] For example, the pre-sensitized photoanode film by cetyltrimethyl-ammonium bromide treated can achieve secondary loading of QD due to strong electrostatic interactions between TiO 2 /QD substrates and QD. 15 The obtained PCE has been improved from 9.35% to 10.42%. In addition, a metal oxyhydroxide layer has been applied to modify the photoanode surface, providing new adsorption sites to realize the redeposition of QD. 16 The obtained PCE value for the modification QDSCs enhanced to 15.31% from initial 13.54%, whereas the proposed surface engineering strategies usually cause the unwanted multilayer deposition and increase the possibility of charge recombination from QD toward TiO 2 substrate. [17][18][19] On the other hand, Zhong et al. recently displayed a cosensitization technology, whereby the TiO 2 substrate could be the redeposition of different types of QD after saturation toward a certain kind of QDs. 9,12 For example, the co-sensitization strategy allows Zn-Cu-In-S or CdSe QDs to be loaded on Zn-Cu-In-Se pre-sensitized electrode, resulting in an improved light absorption and photovoltaic properties for the achieved QDSCs. However, the abovementioned strategies are limited to I-III-VI and II-VI QDs, and the introduction of toxic heavy metal Cd is not conducive to the progress of QDSC devices. Therefore, the research for new light-absorbing materials to realize the deposition on TiO 2 film is an important task that constantly attracts many research interests.
Black phosphorus (BP) QDs with tunable bandgap, good absorption capacity, and excellent electron mobility have become a kind of promising nanomaterials in photovoltaic application. [20][21][22][23][24] Black phosphorus quantum dots (BPQDs) as appropriate materials have been introduced to organic photovoltaics [25][26][27] and dye-sensitized solar cells 28,29 recently. The optical absorption of organic solar cell based on BPQDs enhanced owing to the high broadband light absorption of BPQDs, which gives rise to high efficiency of the obtained cells. Moreover, BPQDs can also be used as a class of transporting material in perovskite solar cells to enhance the charge transfer process. [30][31][32] It is reported that BPQDs were used as suitable seedlike sites to adjust the growth and nucleation of perovskite films, 20 resulting in the crystallization of high-quality film. Based on these ideas, it is anticipated that the BPQDs as another absorbing material will also be suitable to prepare the high light-harvesting photoanode in QDSCs and increase the property of resulted cells.
In this work, we present a convenient and general method for improving PCE in QDSCs via depositing BPQDs onto the TiO 2 /QD film surface, which are promoted to the light-harvesting of photoanodes and charge transmission of devices. The BPQDs with aver-age size of 3.8 nm were obtained by electrochemical intercalation from bulk BP crystal, illustrating the nearinfrared absorption range. Zn-Cu-In-S-Se (ZCISSe) QDsensitized TiO 2 substrates redeposited by BPQDs are employed to prepare the photoanode (TiO 2 /QD/BP) for the assembly of QDSCs. The results of the experiment suggest that the light collection capacity of TiO 2 /QD/BP film electrode is increased significantly. Accordingly, the champion PCE value of the obtained TiO 2 /QD/BP QDSCs is enhanced to 15.66% (J sc = 26.88 mA/cm 2 , V oc = 0.816 V, fill factor [FF] = 0.714) from the incipient 14.11% (J sc = 25.41 mA/cm 2 , V oc = 0.779 V, FF = 0.713) for TiO 2 /QD QDSCs.

Preparation and characterization of BPQDs
The ultrasmall BPQDs was synthesized by electrochemical intercalation method from BP bulk crystals, [33][34][35] and the schematic illustration of preparation processes is exhibited in Figure 1A. Briefly, the two electrode systems, BP crystal and platinum foil, were set in TBA⋅HSO 4 /PC (tetrabutylammonium hydrogen sulfate dissolved in propylene carbonate) solution with a constant potential to complete the exfoliation process. Subsequently, the mixture solution was sonicated under the condition of ice bath to obtain a brown suspension. The resulting suspension was centrifuged at high revolutions to get precipitation and washed with isopropanol (IPA). Afterward, the washed solution was further centrifugated at different revolutions per minute (10k, 8k, and 6k rpm) to collect the corresponding supernatants successively, and the BPQDs with different particle sizes were obtained for use.
It has been well recognized that the size of QDs is important to the properties of QDSCs. The QDs with small size on the one hand are contributed to the deposition of TiO 2 films, whereas on the other hand can lead to the narrow light absorption range and reduce the capacity of light harvesting. The sizes of BPQDs can be changed by tuning the centrifugation speeds. The optical properties of BPQDs at different centrifugation rates (BP-6K, BP-8K and BP-10K) were investigated by UV-visible spectrophotometry and fluorescence spectroscopy. The redshift methodically is observed for the corresponding absorption onset and photoluminescence peaks with the increase of the centrifugal rates ( Figure 1B and Figure S1). The BPQDs at 6k rpm display the high absorption intensity and wide light absorption range. It should be noted that the concentration of obtained BP/IPA solution is related to the absorption intensity. In order to verify the reason of the redshift, the absorption spectra of BPQD-8K based on different concentrations (1.0, 0.5, and 0.3 mg/mL) have been measured, and the results are shown in Figure S2A. For the convenience of comparison, the absorption intensity has been normalized. It can be seen that the similar absorption onset was observed for different concentrations of BPQDs. However, a slight redshift of BPQDs shown in Figure S2B under the same concentration (1.0 mg/mL) has been viewed at different centrifugation speeds. Therefore, the redshift methodically is ascribed to the difference in the size of the prepared BPQDs, which is consistent with the conclusion reported in the previous literatures. 26 With the increase of BPQDs size, the light absorption spectrum shifts to long wavelength due to the decrease of QDs average bandgap induced by quantum confinement. Nevertheless, the photograph of BPQDs shown in the inset of Figure 1B illustrates the poor dimensional uniformity of prepared BP/IPA solution at 6k rpm, which may be inappropriate for the deposition of BPQDs on the surface of the TiO 2 /QD photoanode.
Transmission electron microscopy (TEM) images of a class of well-distributed BPQDs with uniform sizes are shown in Figure S3. Figure 1C illustrates the correspondingly enlarged TEM images of synthesized BPQDs with a narrow size distribution, and the corresponding diameter is approximatively 3.8 nm. The size is similar to ZCISSe QDs, which is contributed to the deposition procedure. The high-resolution TEM picture in Figure 1D describes 0.21 nm d-spacing value, which can be assigned to the (002) planes of BPQD crystal, confirming the successful synthesis of BPQDs. 36 Raman spectra were also applied to test the structure of BPQDs. As demonstrated in Figure 1E, Raman scattering of bulk BP exhibits three characteristically vibrational peaks of the out-of-plane phonon modes at 361.08, 437.51, and 464.64 cm −1 , which correspond to the A 1 g , B 2g , and A 2 g modes, respectively. The A 1 g , B 2g , and A 2 g modes of BPQDs are redshift about 2.1, 3.3, and 3.2 cm −1 in comparison with bulk BP. The redshift phenomenon can be ascribed to decreases in the thickness and the lateral dimensions, which is similar to the situations of graphene and MoS 2 QDs. 37,38 In addition, these sharp patterns further confirm that BPQDs decompose from large crystalline bulk BP crystal but retain the crystalline features. Additionally, the size of BPQDs was further investigated by an atomic force microscope, and the results are exhibited in Figure 1F. Figure 1F displays the topography of BPQDs, and the measured corresponding heights shown in Figure 1G are 2.95, 2.48, and 1.93 nm, suggesting that the number of layers of the obtained BPQDs is approximately five layers. Moreover, the chemical composition of BPQDs has been further analyzed by X-ray photoelectron spectroscopy, and the corresponding full spectrum of BPQDs is illustrated in Figure S4. Figure 1H describes the high-resolution P2p spectrum; two peaks at 130.2 and 129.5 eV are corresponding to the 2p 1/2 and 2p 3/2 binding energies, respectively, which are the representative feature of crystalline BP. 39 Furthermore, the P=O peak at the high-energy region of 133.9 eV arises from the exposure of BPQD samples to the atmosphere, and the phenomenon is also observed in previous measurements. 29,40 The results confirm the successful synthesis of BPQDs through the electrochemical intercalation method.

Characterization of photoanode film
ZCISSe-based QDSCs were selected as reference devices to investigate the effect of BPQDs on the QDSCs as the highest efficiency of QDSCs was obtained from the kind of devices currently. 6,16 The absorption spectroscopy was utilized to investigate the loading of two kinds of QDs, ZCISSe and BPQDs. For this experiment, TiO 2 film electrodes were produced by printing TiO 2 paste onto an FTO glass to form a transparent TiO 2 layer about 6.0 μm thickness. Figure 2A illustrates that the TiO 2 film is initially sensitized by MPAmodified ZCISSe QDs using a capping ligand-induced self-assembly method. The synthesized ZCISSe QDs aqueous solution was pipetted onto the electrode and standing for a certain period until deposition reached saturation, creating a pre-sensitized film electrode referred to as TiO 2 /QD in subsequent discussions. Subsequently, the TiO 2 /QD electrodes were treated by BP/IPA solution to get another photoanode, denoted as TiO 2 /QD/BP electrode. The absorption spectra of TiO 2 /QD/BP electrodes with various BPQDs sizes were determined and are displayed in Figure 2B. It can be found that as the centrifugal revolutions of BPQDs decreased, the absorption intensity of TiO 2 /QD/BP electrode increased gradually. The result demonstrates that BPQDs can be further deposited on the TiO 2 /QD film effectively and enhanced the light capture ability of photoanode. The BPQD loading can also be confirmed by TEM characterization. The TEM pictures of TiO 2 /QD and TiO 2 /QD/BP film electrodes are presented in Figure 2C,D respectively. The higher loading density of QDs on TiO 2 substrate could be obtained from TiO 2 /QD/BP film in comparison with the reference TiO 2 /QD sample. The TEM picture of TiO 2 /QD/BP electrode and the corresponding elemental mapping images are shown in Figure 2E,F. It is noted that, except for the distribution of Ti, O, Zn, Cu, In, S, and Se elements, the signal of P element is observed obviously, suggesting that the TiO 2 /QD/BP photoanode film electrode is prepared successfully. In a control experiment, the elemental mapping of TiO 2 /QD film was also investigated. As displayed in Figure S5, no P signal is found. Hence, it can be concluded that the ZCISSe QDs and BPQDs are uniformly deposited on TiO 2 substrate and that there is no aggregation or stacking in both photoanodes. In addition, the elemental analysis of TiO 2 /QD/BP photoanode cross section was also used to investigate the distribution of BPQDs in TiO 2 film, and the results are exhibited in Figure S6. It shows a uniform distribution for Ti, O, Zn, Cu, In, S, Se, and P elements throughout the TiO 2 film thickness, illustrating that BPQDs can penetrate into the TiO 2 film substrate effectively.

Photovoltaic performance of QDSCs
For the construction of QDSCs, TiO 2 film electrodes with 25.0 μm composed of a 20.0 μm transparent layer and a 5.0 μm light scattering layer were prepared by printing the corresponding pastes on FTO glass substrates. Specifically, QDSCs were assembled employing dual QDs (ZCISSe QDs and BPQDs) deposition TiO 2 film substrate, Cu 2 S/brass counter electrode (CE), and polysulfide (S 2− /S n 2− ) electrolyte. The device architecture for QDSC is presented in Figure 3A. The corresponding energy band diagram and charge transfer process of photoanode are depicted in Figure 3B. It could be concluded that the conduction band position of BPQDs is higher than that of ZCISSe QDs. The results facilitate effectively the electron transfer and suppress recombination at the interfaces of photoanode/electrolyte for TiO 2 /QD/BP QDSCs.
It should be noted that the concentration of obtained BPQDs in the IPA solvent can be adjusted by tuning the centrifugation speeds. The photovoltaic performances of TiO 2 /QD/BP QDSCs based on different centrifugation rates (BP-6K, BP-8K, and BP-10K) of BPQDs were first investigated under the same concentration of 1.0 mg/mL, and the deposition amounts of BPQDs were kept at 90 μL. The assemble QDSCs were detected under simulated AM 1.5 G sunlight with the intensity of 100 mW/cm 2 . The current density-voltage (J-V) curves and corresponding photovoltaic parameters are shown in Figure 3C and Table S1, respectively. The pristine TiO 2 /QD achieves a PCE value of 9.98%, and the TiO 2 /QD/BP illustrates a highest PCE of 11.29% at the condition of BP-8K.
In addition, the deposition amounts of BPQDs also significantly affects the cell efficiency under the different centrifugation rates of BPQDs. The detailed J-V curves of TiO 2 /QD/BP QDSCs under different BPQDs solutions with various amounts of BPQDs were tested, and the results are illustrated in Figure 3D-F. The average photovoltaic properties, such as PCE, short-circuit current density (J sc ), open-circuit voltage (V oc ), and FF for each condition, are summarized in Table 1. The QDSCs based on the treatment of BPQDs solution exhibit superior performance compared to the reference cells. The J sc , V oc , and FF photovoltaic parameters increase systematically with the enhancement of BPQDs amounts in the TiO 2 /QD/BP photoanode. Consequently, the final PCE of the obtained QDSCs exhibits a similar dependent trend on the variation of BPQDs, and results are summarized in Figure S7. Specifically, the PCE values increased as the amount of BPQDs solution increases and achieves the highest efficiency with the BPQD amount of 90 μL for TiO 2 /QD/BP-6K and TiO 2 /QD/BP-8K cells. However, the TiO 2 /QD/BP-10K obtain the highest PCE at the BPQDs amount of 120 μL. Figure 3G  Taking the TiO 2 /QD/BP-8K QDSCs as examples, five parallel devices were constructed for the TiO 2 /QD/BP to ensure the reliability of reported experimental data, and the corresponding photovoltaic parameters were detected. The measured results and J-V curves for each device are illustrated in Table S2 and Figure S8, respectively. This superior performance for three TiO 2 /QD/BP QDSCs could mainly originate from the significantly enhanced J sc value (24.84 vs. 26.88 mA/cm 2 ), which is approximate 8% improvement for J sc compared to TiO 2 /QD cells. The enhancement of J sc is ascribed to the absorption ability of BPQDs, which promotes effectively the light-harvesting capacity of photoanode. Moreover, a slightly increased in V oc and FF values can be seen in TiO 2 /QD/BP QDSCs. The BPQDs were deposited on the surface of photoanodes via a simple evaporation process. In this case, BPQDs may cover the surface of ZCISSe QDs and TiO 2 nanoparticles. It can be concluded that the exposed area of TiO 2 nanoparticles can be decreased via the deposition of BPQDs on the surface of TiO 2 /QD film. Reducing the direct contact between TiO 2 and electrolyte decreases effectively the possibility of photogenerated electrons captured by electrolyte at the TiO 2 /electrolyte interface. In addition, it is notable that the TiO 2 /QD/BP QDSC has a cascaded band structure with the BPQD deposition, which is favorable for efficient hole transport and can prohibit carrier recombination at the interface of QD/electrolyte. When the amount of BPQD solution is beyond the critical value, a decreased PCE is observed may be attributed to the existence of QD multilayer in photoanode. The obtained BPQDs lead to the poor size uniformity and dispersion at the centrifugation rates of BP-6K, which may be aggregated in the deposition process. Under these conditions, the efficiency of charge collection can be reduced due to the formation of charge recombination caused by the multilayer deposition of QDs. On the contrary, the absorption TA B L E 1 The photovoltaic features (J sc , V oc , FF, and power conversion efficiency [PCE]) of TiO 2 /QD/BP quantum dot-sensitized solar cells (QDSCs) with different amounts of black phosphorus quantum dots (BPQDs) based on Cu 2 S/brass counter electrodes (CEs) under the illumination of one full sun intensity (AM 1.5G, 100 mW/cm 2 ). intensity of BPQDs obtained decreased evidently at 10k rpm, which limited the photocurrent enhancement of fabricated QDSCs.

BPQDs Amount of BPQDs
To verify the effect of BPQDs in generating photocurrent, the TiO 2 /BP devices based on different centrifugation rates of BPQDs were constructed. The amounts of BPQDs were kept to be identical. Figure S9A displays the J-V curves of TiO 2 /BP cells under one sun illumination, illustrating that the only BPQDs could effectively generate photocarriers to assemble solar cell. Moreover, it can be found that the TiO 2 /BP-8K cells exhibit the highest PCE, which is mainly related to the sizes of BPQDs. The number of BPQDs in TiO 2 film decreased with the increase of particle sizes. However, BPQDs with smaller sizes could lead to the weak light harvesting, resulting in the decrease of PCE. The detailed photovoltaic parameters are shown in Table S3. The J-V curves shown in Figure S9B reveal a relatively low PCE for TiO 2 /BP in comparison with the TiO 2 /QD and TiO 2 /QD/BP QDSCs, which can be attributed to the inadequate deposition amounts of BPQDs.
In order to better understand the derivation of increased photocurrent, external quantum efficiencies (EQEs) of the prepared cells were evaluated. As described in Figure 3H, the curves of TiO 2 /QD/BP QDSCs based on the different amounts of BPQDs illustrate a similar trend in the photoresponse range, and the EQE values show an obvious increase within the response range of 350-700 nm. From the results, the corresponding integral current values according to the EQE spectra improve from 24.93 to 26.69 mA/cm 2 for TiO 2 /QD and TiO 2 /QD/BP QDSCs, respectively, which are in accordance with the J sc values in the J-V detection (shown in Table S4). The higher EQE value for TiO 2 /QD/BP QDSC is attributed to the excellent light-harvesting capacity of BPQDs, which illustrates an important effect on generating more carriers.
The stabilized power output measurement based on both TiO 2 /QD and TiO 2 /QD/BP QDSCs were also investigated. Figure 3I exhibits the steady-state current density of TiO 2 /QD and TiO 2 /QD/BP QDSCs tested on the potential of the maximum power point. The obtained QDSCs exhibit steady photocurrent output, suggesting the favorable stability of QDSCs under the condition of continuous illumination. In addition, the long-term stability of device is important to real applications, and the performances are also investigated. After each test, the QDSCs were stored under Ar atmosphere. The corresponding results are shown in Figure S10, and the PCE values decreased by about 17.07% for TiO 2 /QD and 18.05% for TiO 2 /QD/BP after 7 days. It can be concluded that both TiO 2 /QD and TiO 2 /QD/BP QDSCs exhibit relatively acceptable stability.

Electrochemical characterization
Both electrochemical impedance spectroscopy (EIS) and open-circuit voltage decay (OCVD) assessments were used to explore the internal fundamental of BPQD effect on the properties of TiO 2 /QD/BP QDSCs. The corresponding Nyquist diagrams of different QDSCs are specifically explored at different applied bias voltages (V app ) from −0.35 to −0.65 V under dark condition, and the results are illustrated in Figure S11. Then the obtained data are fitted by a standard simulation circuit (inset of Figure 4) to get the electrochemical parameters. Figure 4A,B displays the dependence of the extractive chemical capacitance (C μ ) and recombination resistance (R rec ) values on the whole range of forward bias, respectively. The R rec value is known to reflect the recombination process at the interface of photoanode/electrolyte. Similar C μ values are observed in different conditions, as shown in Figure 4A, indicating that the level of conduction band edge and the density of states of TiO 2 are not affected by the deposition of BPQDs. 41,42 However, different values of R rec shown in Figure 4B are gained with the different loading amounts of BPQDs. Particularly, in comparison to other structures, the TiO 2 /QD/BP QDSCs with BPQDs amount of 90 μL describe an obviously improve in R rec throughout the entire bias range. The exposed area of bare TiO 2 substrate after BPQD treatment can contribute to diminish the direct contact with polysulfide electrolyte, resulting in the high R rec value. If we take into account that the possibility of charge recombination is inversely proportional to the value of R rec , the significantly high R rec suggests that the charge recombination procedure in TiO 2 /QD/BP QDSCs has been greatly inhibited. Moreover, the Nyquist diagrams of TiO 2 /QD/BP QDSCs with various deposition amounts of BPQDs are displayed in Figure 4C at −0.65 V bias, which is close to the open-circuit voltage. Fitting parameters have been obtained and are presented in Table 2.
The results demonstrate that calculated electron lifetimes (τ n = R rec × C μ ) of TiO 2 /QD/BP QDSCs with 90 μL treatment are longer than those of other conditions, enhancing the photovoltaic properties of QDSCs. TiO 2 /QD and TiO 2 /QD/BP QDSCs were further applied via OCVD detection to investigate charge recombination dynamics that can give other information on the recombination of the photogenerated electrons. In this test, the electron lifetime (τ n ) can be obtained from the following equation, and the dependence of τ n on the potential is shown in Figure 4E 6 : where k B is the Boltzmann constant, T is absolute temperature, and e is electronic charge. Figure 4D reveals the OCVD detection results for different QDSCs. The potential decay rates are different for the cells that different deposition amounts of BPQDs. From Figure 4D, TiO 2 /QD/BP TA B L E 2 Impedance parameters of quantum dot-sensitized solar cells (QDSCs) at −0.65 V bias.

F I G U R E 5 (A) J-V curves of TiO 2 /QD and TiO 2 /QD/BP quantum dot-sensitized solar cells (QDSCs) based on NMC/Ti counter electrodes (CEs); (B) a summary of QDSC performances with different photoanodes.
QDSCs illustrate a comparatively slower decay compared to TiO 2 /QD cells, and the τ n curves of TiO 2 /QD/BP QDSCs displayed in Figure 4E are conspicuously larger than that of TiO 2 /QD device at the same potential. It should be noted that higher τ n implies less charge recombination occurring in QDSCs. This result may be caused by the existence of BPQDs on TiO 2 /QD/BP photoanode film, suggesting that the photogenerated τ n is greatly enhanced by deposition BPQDs on the photoanode film. Moreover, the trend is in accordance with prior EIS testing and further proves the effect of BPQDs on suppressing the charge recombination.

Photovoltaic property of QDSCs based on NMC/Ti CEs
Nitrogen-doped mesoporous carbon supported Ti mesh CEs (NMC/Ti) have been proved to exhibit better electrocatalytic properties compared to the typically used Cu 2 S/brass. [43][44][45][46] Moreover, the NMC/Ti electrode contributes to decrease the potential of S 2− /S n 2− redox couple, and accordingly the potential of resultant QDSCs is enhanced. Herein, the TiO 2 /QD and TiO 2 /QD/BP cells based on NMC/Ti CEs were prepared, and the obtained champion photovoltaic performance and J-V curves for both QDSCs are presented in Figure 5A. Specifically, five parallel J-V curves and the detailed cell data under each condition are illustrated in Figure S12 and Table S5, respectively. It is concluded that, as for NMC/Ti CEs, the TiO 2 /QD/BP QDSCs with 90 μL BPQD treatment can improve the champion PCE to 15.66% (J sc = 26.88 mA/cm 2 , V oc = 0.816 V, FF = 0.714) from the original 14.11% (J sc = 25.41 mA/cm 2 , V oc = 0.779 V, FF = 0.713) corresponding to TiO 2 /QD QDSCs. According to the abovementioned consequences, it can believe that the method for enhancing QDSCs property via depositing a certain amount of BPQDs on TiO 2 /QD film is effective for QDSC systems. Compared with the PCE values reported by other literatures for different photoanodes ( Figure 5B), it is obvious that we presented TiO 2 /QD/BP QDSCs features a relatively high PCE and excellent properties.

CONCLUSIONS
In this study, ultrasmall BPQDs were obtained by an electrochemical intercalation method. The as-prepared BPQDs feature a uniform size distribution and display the super capacity of light harvesting and charge transporting. By introducing a certain amount of BPQDs on the pre-sensitized TiO 2 /QD film substrate to achieve TiO 2 /QD/BP photoanode, a convenient and effective method for improving the PCE of QDSCs has been explored. We found that the presence of BPQDs promoted the ability of light harvesting and charge transport. In addition, the proportion of uncovered TiO 2 substrate surface is also decreased after BP modification. This effectively reduces the probability of needless charge recombination and enhances both photocurrent and photovoltage in the corresponding QDSCs. Finally, the QDSCs based on TiO 2 /QD/BP photoanode under optimal condition achieved a champion PCE of 15.66%. Our strategy proposes a convenient way to further develop high-efficiency QDSCs.

Preparation of the BPQDs
High-quality BP crystals were synthesized according to the previous report, 47 and stored in an Ar glovebox prior to use. The BPQDs were prepared from bulk BP crystals by a simple electrochemical intercalation method. Specifically, the intercalation process was performed using a twoelectrode system, in which BP crystal was used as cathode and a piece of platinum foil as anode, respectively. The electrodes were set in parallel with a constant distance of 2.0 cm. The bulk BP with dimension around 10 mm × 10 mm × 2 mm and weight around 50 mg were used. The electrolyte was prepared by dissolving TBA⋅HSO 4 in PC solution (0.1 M). It should be noted that the electrolyte needed to be filled with nitrogen to remove oxygen. Sub-sequently, a constant potential of -4.0 V was applied to start the exfoliation. After complete exfoliation, the exfoliated BP flakes along with electrolyte were sonicated with a power of 800 W for 30 min. The ultrasound probe was operated for 2 s and stopped for 2 s. The temperature in probe ultrasonic processing and ultrasound bath treatment was controlled by an ice bath. The acquired brown suspension was the PC solution of BPQDs. Afterward, the resulting suspensions were centrifuged for 20 min at centrifugal rate 15K per minute (rpm), and the precipitate was washed by IPA for two times. Finally, the resuspended solutions were centrifuged for 10 min at different rates (6K, 8K and 10K rpm) to obtain the upper layer solution for further use. Subsequently, the obtained BP/IPA solutions at different centrifugation speeds were further centrifuged at 15K rpm for 20 min to get the corresponding BPQDs precipitate, which were re-suspended in IPA solvent. The concentrations of different BPQDs in IPA were about 1.3 mg/ml for BP-6K, 1.0 mg/ml for BP-8K, and 0.6 mg/ml for BP-10K.

Preparation of ZCISSe QDs
The S precursor (1.0 M) was prepared by dissolving S powder in DPP, and the Se precursor (1.0 M) was prepared by dissolving Se powder in a mixture of DPP and OAm with a volume ratio of 1:1. For the synthesis of ZCISSe QDs, In(OAc) 3 (0.2 mmol), CuI (0.14 mmol), Zn(OAc) 2 (0.08 mmol), and OAm (8.0 mL) were loaded in a flask. The solution was degassed under vacuum at 100 • C for 5 min and then heated to 180 • C under N 2 environment. After that, 1.0 mL S/Se (4:6) precursor was quickly injected into the mixture, and then the system was heated to 220 • C and kept the reaction for 8 min. The obtained OAm-capped QDs were purified by centrifugation with the addition of ethanol. Phase transfer by a ligand exchange process with the use of MPA was carried out. After another precipitation and centrifugation cycle, the water-soluble QDs were dissolved in 1.0 ml of deionized water, and the pH of the solution was adjusted to 10 by addition of NaOH solution.

Preparation of photoanodes and assembly of QDSCs
For the construction of QDSCs, TiO 2 film electrode with 25.0 μm composed of a 20.0 μm transparent layer and a 5.0 μm light scattering layer were prepared by printing the corresponding pastes on FTO glass substrates. For the deposition of QDs onto TiO 2 substrate, 50 μL of the MPAcapped QDs aqueous solution was pipetted on the TiO 2 film and then stayed for 1 h at 60 • C before rinsing sequentially with water and ethanol, respectively. Then, a certain amount of BP/IPA solution was dropped onto the TiO 2 /QD electrode and the electrode placed in a furnace at a temperature of 50 • C to allow the complete evaporation of IPA solution and obtain the TiO 2 /QD/BP electrode. The electrodes were washed sequentially with water and ethanol, followed by drying with a flow of Ar. After coating with the ZnS blocking layer by the SILAR method, the photoanode was obtained. The CEs of Cu 2 S/brass and NMC/Ti were fabricated according to the previous literature. 48,49 Polysulfide aqueous solution (2.0 M Na 2 S, 2.0 M S, and 0.2 M KCl) was used as the electrolyte.

Characterization
TEM of BPQDs were performed on the JEM-F200 field emission transmission electron microscope at 200 kV. AFM was conducted on the drop-cast flakes on Si/SiO 2 substrates using the Bruker Dimension Icon atomic force microscope. Raman scattering was carried out on a Horiba LabRam HR Evolution confocal Raman microscope equipped with a 532 nm laser as the excitation source at room temperature. The UV-vis absorption spectra were measured by a HATACHI UH4150 spectrophotometer. The PL emission spectra were recorded on a HATACHI F-7100 spectrophotometer. The J−V curves were tested using Keithley 2400 source meter under AM 1.5G irradiation from a solar simulator (Enlitech model SS-F5-3A). Before each test, the power of the simulated solar light was calibrated to 100 mW/cm 2 by an NREL standard Si solar cell. EQE curves were measured on a Keithley 2000 multimeter under the illumination of a 300 W xenon lamp. EIS tests were carried out with an electrochemical workstation (Zahner, Zennium) under the dark condition, applying a 20 mV AC sinusoidal signal with the frequency ranging from 1 MHz to 0.1 Hz. The samples for OCVD measurements were illuminated by a white light-emitting diode (LED) with intensity of 100 mW/cm 2 , and the decay of photovoltage with time were recorded after switching off the light source.

A C K N O W L E D G M E N T S
This work is financially supported by the Postdoctoral Innovative Talents Support Program (No. BX2021349) and Basic and Applied Basic Research Foundation of Guangdong Province (No. 2022A1515110462).

C O N F L I C T O F I N T E R E S T S TAT E M E N T
The authors declare no conflict of interest.