E ﬃ ciency Improvement of MAPbI 3 Perovskite Solar Cells Based on a CsPbBr 3 Quantum Dot / Au Nanoparticle Composite Plasmonic Light-Harvesting Layer

: We demonstrate a method to enhance the power conversion efficiency (PCE) of MAPbI 3 perovskite solar cells through localized surface plasmon (LSP) coupling with gold nanoparticles:CsPbBr 3 hybrid perovskite quantum dots (AuNPs:QD-CsPbBr 3 ). The plasmonic AuNPs:QD-CsPbBr 3 possess the features of high light-harvesting capacity and fast charge transfer through the LSP resonance effect, thus improving the short-circuit current density and the fill factor. Compared to the original device without Au NPs, a 27.8% enhancement in PCE of plasmonic AuNPs:QD-CsPbBr 3 / MAPbI 3 perovskite solar cells was achieved upon 120 µ L Au NP solution doping. This improvement can be attributed to the formation of surface plasmon resonance and light scattering effects in Au NPs embedded in QD-CsPbBr 3 , resulting in improved light absorption due to plasmonic nanoparticles.


Introduction
As one of the most competitive materials in the photovoltaic field, organic-inorganic hybrid halide lead perovskites (ABX 3 , where A is a monovalent cation, B is a divalent cation and X is a halide ion) have attracted great attention in recent years [1][2][3][4]. Through structural engineering [5,6], surface interface engineering [7,8], solvent engineering [9,10] and incident light management engineering [11,12], current organic-inorganic hybrid perovskite-based photovoltaic devices have demonstrated power conversion efficiencies (PCEs) of over 25.2% [13]. They show broad application prospects in the new generation of the green energy industry. Organic-inorganic perovskite solar cells (CH 3 NH 3 PbX 3 = MAPbX 3 , X = Cl, Br and I) have excellent charge transport properties, long carrier diffusion length, full spectrum and strong light absorption. The properties of the perovskite light absorbing layer, electron transport layer and hole transport layer have an important influence on the efficiency of the perovskite solar cell, and most importantly, the perovskite film directly affects the photovoltaic performance of the device. Therefore, a high-quality, defect-free perovskite film is a necessary condition for the perovskite solar device to have high photovoltaic performance.
The literature indicates that the grain size of the roughened perovskite film can affect charge transport, and the addition of Br − or Cl − ions to MAPbX 3 can improve the crystallinity and stability of the material [14][15][16][17][18][19]. Although perovskite materials have a high absorption coefficient (10 4 cm −1 ), the thickness of the film is still an important factor limiting the light-harvesting efficiency. In past literature, the thickness of the perovskite layer was controlled below 500 nm. However, theoretical calculations show that sufficient light absorption can be achieved only when the thickness of the film exceeds 1 µm to promote further efficiency [20,21]. This increase in thickness adversely affects

Synthesis of Plasmonic AuNPs:QD-CsPbBr 3 Perovskite Solution
The perovskite QD-CsPbBr 3 was synthesized by using hot-injection and ice-water bath methods. First, a Cs-oleate precursor solution was prepared via the mixing of 81.4 mg cesium carbonate (Cs 2 CO 3 , 99.995%) in 3 mL of 1-octadecene (ODE, 90%) and 0.25 mL of oleic acid (OA, 90%); this mixture was then heat-treated at 200 • C for 30 min until it dissolved to a clear solution, indicating that the Cs 2 CO 3 reacted with OA. Subsequently, a mixture of 0.2 g lead (II) bromide (PbBr 2 , 99.999%), 10 mL ODE, 1 mL OA and 1 mL oleylamine (OAM, 70%) was stirred and heated at 140 • C for 20 min. Afterwards, the temperature was increased to 200 • C and the mixture was stirred until the PbBr 2 dissolved completely. A volume of 0.8 mL of Cs-oleate precursor solution was subsequently injected rapidly into the PbBr 2 reaction solution, and the reactant was immediately bathed using an ice-water bath for 30 s after 5 s of injection. The synthesized QD-CsPbBr 3 solution was added to 40, 80, or 120 µL of the citrate-containing Au NP solution and shaken for 10 s. Afterwards, 8 mL of ethyl acetate (EA, 99.8%) was added, and the mixture was centrifuged at 5000 rpm for 15 min to collect the mixed precipitate, which was then dispersed in 1 mL of toluene (99.99%) and oscillated by sonication for 3 min to complete the plasmonic AuNPs:QD-CsPbBr 3 perovskite solution.

Fabrication of Plasmonic AuNPs:QD-CsPbBr 3 /MAPbI 3 Perovskite Solar Cells
The procedure for fabricating plasmonic AuNPs:QD-CsPbBr 3 /MAPbI 3 perovskite solar cells can be found in Figure 1. Patterned indium tin oxide (ITO) glass was sequentially washed by sonication in acetone, alcohol and isopropanol solution for 30 min, and then treated with UV-ozone for 10 min. A 25-nm-thick plasmonic AuNPs:QDs-CsPbBr 3 light-harvesting layer was deposited by spin-coating a 40 mL green solution of AuNPs:QDs-CsPbBr 3 at 4500 rpm for 45 s, and then annealed at 120 • C for 5 min. A 70-nm-thick PEDOT:PSS hole transport layer was deposited by spin-coating at 5000 rpm for 30 s from a solution containing PEDOT:PSS (UR-AI4083, Uni-Onward Corp., New Taipei City, Taiwan) at a ratio of 1:6, and then annealed at 120 • C for 10 min. A 300-nm-thick perovskite MAPbI 3 absorber layer was deposited by two-step continuous spin-coating at 1000 rpm and 5000 rpm for 10 s and 20 s, respectively. When the second stage was spin-coated with a countdown of 3 s, the absorber layer was completed by dropping 100 µL of toluene and was then heated to 100 • C for 10 min. The perovskite MAPbI 3 solution was synthesized by a solvent-engineering method from a precursor prepared by mixing 159 mg methylammonium iodide (MAI) and 461 mg lead (II) iodide (PbI 2 , 99.9985%) dissolved in cosolvents of 0.5 mL dimethyl sulfoxide (DMSO) and 0.5 mL dimethylformamide (DMF), which was heated at 70 • C for 60 min and then left to stand for 24 h. A 20-nm-thick C 60 electron transport layer was evaporated at a rate of 0.05 nm/s at a vacuum pressure of 1 × 10 −6 torr. Ag electrodes were deposited at a rate of 2.5 nm/s to 3.0 nm/s with a vacuum pressure of 4.8 × 10 −5 torr for a total thickness of 100 nm. The device working area was about 2 mm × 5 mm in this research. After fabrication, the devices were encapsulated in a N 2 -filled glove box by sealing the active area using another piece of glass. The complete QD-CsPbBr 3 /MAPbI 3 and plasmonic AuNPs:QD-CsPbBr 3 /MAPbI 3 perovskite solar cell configurations are shown in Figure 2. The inset of Figure 2b shows a transmission electron microscopy (TEM) image of the AuNPs:QD-CsPbBr 3 . It can be seen that Au NPs were embedded in the QD-CsPbBr 3 . The diameters of the Au NPs and QD-CsPbBr 3 were about 5 nm (red circle) and 10 nm, respectively. (99.99%) and oscillated by sonication for 3 min to complete the plasmonic AuNPs:QD-CsPbBr3 perovskite solution.

Fabrication of Plasmonic AuNPs:QD-CsPbBr3/MAPbI3 Perovskite Solar Cells
The procedure for fabricating plasmonic AuNPs:QD-CsPbBr3/MAPbI3 perovskite solar cells can be found in Figure 1. Patterned indium tin oxide (ITO) glass was sequentially washed by sonication in acetone, alcohol and isopropanol solution for 30 min, and then treated with UV-ozone for 10 min. A 25-nm-thick plasmonic AuNPs:QDs-CsPbBr3 light-harvesting layer was deposited by spin-coating a 40 ml green solution of AuNPs:QDs-CsPbBr3 at 4500 rpm for 45 s, and then annealed at 120 °C for 5 min. A 70-nm-thick PEDOT:PSS hole transport layer was deposited by spin-coating at 5000 rpm for 30 s from a solution containing PEDOT:PSS (UR-AI4083, Uni-Onward Corp., New Taipei City, Taiwan) at a ratio of 1:6, and then annealed at 120 °C for 10 min. A 300-nm-thick perovskite MAPbI3 absorber layer was deposited by two-step continuous spin-coating at 1000 rpm and 5000 rpm for 10 s and 20 s, respectively. When the second stage was spin-coated with a countdown of 3 s, the absorber layer was completed by dropping 100 μL of toluene and was then heated to 100 °C for 10 min. The perovskite MAPbI3 solution was synthesized by a solvent-engineering method from a precursor prepared by mixing 159 mg methylammonium iodide (MAI) and 461 mg lead (II) iodide (PbI2, 99.9985%) dissolved in cosolvents of 0.5 mL dimethyl sulfoxide (DMSO) and 0.5 mL dimethylformamide (DMF), which was heated at 70 °C for 60 min and then left to stand for 24 h. A 20-nm-thick C60 electron transport layer was evaporated at a rate of 0.05 nm/s at a vacuum pressure of 1 × 10 −6 torr. Ag electrodes were deposited at a rate of 2.5 nm/s to 3.0 nm/s with a vacuum pressure of 4.8 × 10 −5 torr for a total thickness of 100 nm. The device working area was about 2 mm × 5 mm in this research. After fabrication, the devices were encapsulated in a N2-filled glove box by sealing the active area using another piece of glass. The complete QD-CsPbBr3/MAPbI3 and plasmonic AuNPs:QD-CsPbBr3/MAPbI3 perovskite solar cell configurations are shown in Figure 2. The inset of Figure 2b shows a transmission electron microscopy (TEM) image of the AuNPs:QD-CsPbBr3. It can be seen that Au NPs were embedded in the QD-CsPbBr3. The diameters of the Au NPs and QD-CsPbBr3 were about 5 nm (red circle) and 10 nm, respectively.   The QD-CsPbBr3, 40AuNPs:QD-CsPbBr3, 80AuNPs:QD-CsPbBr3, and 120AuNPs:QD-CsPbBr3 films were all annealed at 120 °C for 5 min before measurement. The surface morphologies were observed using a field emission scanning electron microscope (FESEM, ZEISS Sigma, ZEISS, Munich, Germany). The morphology and structure were analyzed using transmission electron microscopy (TEM) (JEM-2100F, JEOL, Tokyo, Japan). X-ray diffraction patterns (XRD) were recorded using an X-ray diffractometer with Cu Kα (λ = 1.5418 Å) radiation operated at 40 kV and 25 mA (X'Pert PRO MRD, PANalytical, Almelo, the Netherlands). The photoluminescence (PL) spectra were recorded using a fluorescence spectrophotometer (F-7000, Hitachi, Tokyo, Japan) under UV light illumination (λext = 365 nm). The transmittance and absorbance spectra were measured using a UV/VIS/NIR spectrophotometer (UH-4150, Hitachi, Tokyo, Japan). The current density-voltage (J-V) curves of the devices were tested using a Keithley 2420 source meter under AM 1.5G irradiation. External quantum efficiencies (EQE) were measured utilizing a spectral response measurement system (QE-R3015, Enli Technololy Co. Ltd., Taiwan).

Results and Discussion
The surface morphologies of QD-CsPbBr3, 40AuNPs:QD-CsPbBr3, 80AuNPs:QD-CsPbBr3 and 120AuNPs:QD-CsPbBr3 annealed at 120 °C for 5 min are shown in Figure 3, and the difference in grain size and agglomeration distribution can be clearly observed. The annealed QD-CsPbBr3 films can be seen to be agglomerated and incompletely dispersed, which is detrimental to QD dispersion due to the lower concentration of dispersant than of QDs. The annealed 40AuNPs:QD-CsPbBr3 films showed that the addition of Au NPs improved the QD dispersion and crystallinity, but there was still some agglomeration after heat treatment to form pore defects. When the Au NP solution volume was increased to 80 μL, the annealed 80AuNPs:QD-CsPbBr3 films were found to have reduced large pore defects and greater grain size uniformity, but the QDs still showed some agglomeration. In addition, when increased to 120 μL, the annealed 120AuNPs:QD-CsPbBr3 films, compared with the QD-CsPbBr3, 40AuNPs:QD-CsPbBr3 and 80AuNPs:QD-CsPbBr3 films, showed better QD grain size and crystallinity and significantly improved QD agglomeration. This indicates that large Au NP contents cause the enhancement of QD agglomeration, and thus the crystallinity of films becomes better. This may be due to the addition of large amounts of Au NPs, causing the Au-Br complex would play a key role in the assembly of AuNPs:QD-CsPbBr3 films. The QD-CsPbBr 3 , 40AuNPs:QD-CsPbBr 3 , 80AuNPs:QD-CsPbBr 3 , and 120AuNPs:QD-CsPbBr 3 films were all annealed at 120 • C for 5 min before measurement. The surface morphologies were observed using a field emission scanning electron microscope (FESEM, ZEISS Sigma, ZEISS, Munich, Germany). The morphology and structure were analyzed using transmission electron microscopy (TEM) (JEM-2100F, JEOL, Tokyo, Japan). X-ray diffraction patterns (XRD) were recorded using an X-ray diffractometer with Cu Kα (λ = 1.5418 Å) radiation operated at 40 kV and 25 mA (X'Pert PRO MRD, PANalytical, Almelo, the Netherlands). The photoluminescence (PL) spectra were recorded using a fluorescence spectrophotometer (F-7000, Hitachi, Tokyo, Japan) under UV light illumination (λ ext = 365 nm). The transmittance and absorbance spectra were measured using a UV/VIS/NIR spectrophotometer (UH-4150, Hitachi, Tokyo, Japan). The current density-voltage (J-V) curves of the devices were tested using a Keithley 2420 source meter under AM 1.5G irradiation. External quantum efficiencies (EQE) were measured utilizing a spectral response measurement system (QE-R3015, Enli Technololy Co. Ltd., Taiwan).

Results and Discussion
The surface morphologies of QD-CsPbBr 3 , 40AuNPs:QD-CsPbBr 3 , 80AuNPs:QD-CsPbBr 3 and 120AuNPs:QD-CsPbBr 3 annealed at 120 • C for 5 min are shown in Figure 3, and the difference in grain size and agglomeration distribution can be clearly observed. The annealed QD-CsPbBr 3 films can be seen to be agglomerated and incompletely dispersed, which is detrimental to QD dispersion due to the lower concentration of dispersant than of QDs. The annealed 40AuNPs:QD-CsPbBr 3 films showed that the addition of Au NPs improved the QD dispersion and crystallinity, but there was still some agglomeration after heat treatment to form pore defects. When the Au NP solution volume was increased to 80 µL, the annealed 80AuNPs:QD-CsPbBr 3 films were found to have reduced large pore defects and greater grain size uniformity, but the QDs still showed some agglomeration. In addition, when increased to 120 µL, the annealed 120AuNPs:QD-CsPbBr 3 films, compared with the QD-CsPbBr 3 , 40AuNPs:QD-CsPbBr 3 and 80AuNPs:QD-CsPbBr 3 films, showed better QD grain size and crystallinity and significantly improved QD agglomeration. This indicates that large Au NP contents cause the enhancement of QD agglomeration, and thus the crystallinity of films becomes better. This may be due to the addition of large amounts of Au NPs, causing the Au-Br complex would play a key role in the assembly of AuNPs:QD-CsPbBr 3 films. Perovskite QD-CsPbBr3 with and without the Au NPs were characterized by XRD to confirm the crystallinity influence of QD-CsPbBr3 films before and after the addition of the Au NPs. The XRD patterns of annealed QD-CsPbBr3, 40AuNPs:QD-CsPbBr3, 80AuNPs:QD-CsPbBr3 and 120AuNPs:QD-CsPbBr3 films are displayed in Figure 4. The 80AuNPs:QD-CsPbBr3 and 120AuNPs:QD-CsPbBr3 XRD patterns and the Rietveld profile refinement, using the Match program analysis software, as shown in Figures S1 and S2. According to an earlier report [41], CsPbBr3 has two phase transitions at different temperatures-one is an orthorhombic phase (lower than 88 °C) and the other is a cubic phase (higher than 130 °C for the bulk and 117 °C for the nanocrystals)-which were distinguished by the PbBr6 octahedral tilting in the orthorhombic phase. It can be seen that the annealed QD-CsPbBr3 films without Au NPs showed only two diffraction peaks, corresponding to the (100) and (200) planes. In addition, with increasing Au NP content, the crystal quality of annealed AuNPs:QD-CsPbBr3 films was better than that of the annealed QD-CsPbBr3 films without Au NPs; the (110) plane of AuNPs:QD-CsPbBr3 films gradually becomes apparent. On the other hand, it was also found that a small amount of Au NPs combined with Br to form Au-Br complex (Au2Br6) and other diffraction peaks appeared, which not only reduced the QD agglomeration and improved the crystallinity quality of the films, but also did not affect the main crystal structure of the QDs. This indicates that the Au-Br complex has a self-assembling effect on QD-CsPbBr3 to improve the film quality [42]. This is in good agreement with the SEM analysis. All the annealed AuNPs:QD-CsPbBr3 films showed three main peaks at 2θ = 15.55°, 21.55° and 30.65°, which were assigned to the (100), (110) and (200) planes, respectively. In particular, the annealed 80AuNPs:QD-CsPbBr3 and 120AuNPs:QD-CsPbBr3 films that appeared in crystal planes of (100), (110)  Perovskite QD-CsPbBr 3 with and without the Au NPs were characterized by XRD to confirm the crystallinity influence of QD-CsPbBr 3 films before and after the addition of the Au NPs. The XRD patterns of annealed QD-CsPbBr 3 , 40AuNPs:QD-CsPbBr 3 , 80AuNPs:QD-CsPbBr 3 and 120AuNPs:QD-CsPbBr 3 films are displayed in Figure 4. The 80AuNPs:QD-CsPbBr 3 and 120AuNPs:QD-CsPbBr 3 XRD patterns and the Rietveld profile refinement, using the Match program analysis software, as shown in Figures S1 and S2. According to an earlier report [41], CsPbBr 3 has two phase transitions at different temperatures-one is an orthorhombic phase (lower than 88 • C) and the other is a cubic phase (higher than 130 • C for the bulk and 117 • C for the nanocrystals)-which were distinguished by the PbBr 6 octahedral tilting in the orthorhombic phase. It can be seen that the annealed QD-CsPbBr 3 films without Au NPs showed only two diffraction peaks, corresponding to the (100) and (200) planes. In addition, with increasing Au NP content, the crystal quality of annealed AuNPs:QD-CsPbBr 3 films was better than that of the annealed QD-CsPbBr 3 films without Au NPs; the (110) plane of AuNPs:QD-CsPbBr 3 films gradually becomes apparent. On the other hand, it was also found that a small amount of Au NPs combined with Br to form Au-Br complex (Au 2 Br 6 ) and other diffraction peaks appeared, which not only reduced the QD agglomeration and improved the crystallinity quality of the films, but also did not affect the main crystal structure of the QDs. This indicates that the Au-Br complex has a self-assembling effect on QD-CsPbBr 3 to improve the film quality [42]. This is in good agreement with the SEM analysis. All the annealed AuNPs:QD-CsPbBr 3 films showed three main peaks at 2θ = 15.55 • , 21.55 • and 30.65 • , which were assigned to the (100), (110) and (200) planes, respectively. In particular, the annealed 80AuNPs:QD-CsPbBr 3 and 120AuNPs:QD-CsPbBr 3 films that appeared in crystal planes of (100), (110), means that Au NPs help the formation of Cs 4 PbBr 6 crystals. The Cs 4 PbBr 6 crystals exhibit a superior quantum yield and device performance than that of CsPbBr 3 crystals. [43,44] Energies 2019, 12, x FOR PEER REVIEW 6 of 13 crystals exhibit a superior quantum yield and device performance than that of CsPbBr3 crystals. [43,44] Figure 5a shows the photoluminescence (PL) spectra of annealed QD-CsPbBr3, 40AuNPs:QD-CsPbBr3, 80AuNPs:QD-CsPbBr3 and 120AuNPs:QD-CsPbBr3 films. It was observed that the PL intensity of all the AuPs:QD-CsPbBr3 films was higher than that of the QD-CsPbBr3 film without Au NPs, and the PL intensity did not increase as the Au NP amount increased. NPs can increase or decrease the overall fluorescence of QDs by plasmon resonance energy transfer, fluorescence resonance energy transfer and charge transfer. These mechanisms will determine whether fluorescence enhancement or quenching is observed in the QDs. In the case of 40AuNPs:QD-CsPbBr3, light can be coupled to the LSP modes and increase the optical energy density near the QDs, thereby enhancing absorption. Therefore, the plasmon resonance energy transfer from excited NPs to nearby QDs is dominant, which increases the luminescence of the QDs. On the other hand, in the case of adding larger amounts of Au NPs, the main reason for the fluorescence reduction and quenching of QDs is speculated to be fluorescence resonance energy transfer. This result clearly shows that the fluorescence emission of QD-CsPbBr3 was quenched by Au NPs. Since the PL spectrum of QD-CsPbBr3 and the localized plasmon absorbance spectrum (525 nm) of Au NPs overlap significantly, the PL quenching can be considered mainly as energy/charge transfer, which should enhance localized plasmon excitation more effectively [45]. This is anticipated to enhance the generation of a photocarrier in the active layer when the plasmonic AuNPs:QD-CsPbBr3 hybrid is used in perovskite solar cells. When the PL spectra were normalized (see Figure 5b), the annealed QD-CsPbBr3, 40AuNPs:QD-CsPbBr3, 80AuNPs:QD-CsPbBr3 and 120AuNPs:QD-CsPbBr3 films exhibited a green emission with a PL peak at 515-517 nm and full widths at half-maximum (FWHMs) of 21.4, 21.6, 22.4 and 24.2 nm, where the FWHMs became larger and were expected to increase light absorption. It can be observed that AuNPs:QD-CsPbBr3 generated an induced electric field under light irradiation, and QD-CsPbBr3 and Au NPs experienced plasmon coupling with each other, causing a red shift. The degree of red shift became significant with increasing amount of embedded Au NPs.  Figure 5a shows the photoluminescence (PL) spectra of annealed QD-CsPbBr 3 , 40AuNPs:QD-CsPbBr 3 , 80AuNPs:QD-CsPbBr 3 and 120AuNPs:QD-CsPbBr 3 films. It was observed that the PL intensity of all the AuPs:QD-CsPbBr 3 films was higher than that of the QD-CsPbBr 3 film without Au NPs, and the PL intensity did not increase as the Au NP amount increased. NPs can increase or decrease the overall fluorescence of QDs by plasmon resonance energy transfer, fluorescence resonance energy transfer and charge transfer. These mechanisms will determine whether fluorescence enhancement or quenching is observed in the QDs. In the case of 40AuNPs:QD-CsPbBr 3 , light can be coupled to the LSP modes and increase the optical energy density near the QDs, thereby enhancing absorption. Therefore, the plasmon resonance energy transfer from excited NPs to nearby QDs is dominant, which increases the luminescence of the QDs. On the other hand, in the case of adding larger amounts of Au NPs, the main reason for the fluorescence reduction and quenching of QDs is speculated to be fluorescence resonance energy transfer. This result clearly shows that the fluorescence emission of QD-CsPbBr 3 was quenched by Au NPs. Since the PL spectrum of QD-CsPbBr 3 and the localized plasmon absorbance spectrum (525 nm) of Au NPs overlap significantly, the PL quenching can be considered mainly as energy/charge transfer, which should enhance localized plasmon excitation more effectively [45]. This is anticipated to enhance the generation of a photocarrier in the active layer when the plasmonic AuNPs:QD-CsPbBr 3 hybrid is used in perovskite solar cells. When the PL spectra were normalized (see Figure 5b), the annealed QD-CsPbBr 3 , 40AuNPs:QD-CsPbBr 3 , 80AuNPs:QD-CsPbBr 3 and 120AuNPs:QD-CsPbBr 3 films exhibited a green emission with a PL peak at 515-517 nm and full widths at half-maximum (FWHMs) of 21.4, 21.6, 22.4 and 24.2 nm, where the FWHMs became larger and were expected to increase light absorption. It can be observed that AuNPs:QD-CsPbBr 3 generated an induced electric field under light irradiation, and QD-CsPbBr 3 and Au NPs experienced plasmon coupling with each other, causing a red shift. The degree of red shift became significant with increasing amount of embedded Au NPs.  Figure 6a presents the absorbance spectra of annealed QD-CsPbBr3, 40AuNPs:QD-CsPbBr3, 80AuNPs:QD-CsPbBr3 and 120AuNPs:QD-CsPbBr3 films. The transmittance spectra of all the films were greater than 85% in the wavelength range between 520 nm and 800 nm ( Figure S3). However, the light absorbance of the QD-CsPbBr3 perovskite decreased abruptly at wavelengths shorter than 520 nm due to the addition of Au NPs. In our study, the absorbance peak at about 525 nm was the unique LSPR band in the absorbance spectrum of pure Au NPs. In the short wavelength range from 400 nm to 520 nm, it was observed that the light absorbance of AuNPs:QD-CsPbBr3 films was significantly enhanced compared to that of QD-CsPbBr3 films without AuNPs. In addition, the absorbance of the samples improved with increasing amount of embedded Au NPs. Due to the addition of large amounts of Au NPs to generate fluorescence resonance energy transfer and quench the emission of QDs, which effectively enhances local plasmon excitation, the absorbance of 120AuNPs:QD-CsPbBr3 significantly increased in the whole wavelength range of 400-800 nm. The absorbance spectra of the AuNPs:QD-CsPbBr3/PEDOT:PSS/MAPbI3 perovskite films with and without Au NPs are shown in Figure 6b. The transmittance spectra of the AuNPs:QD-CsPbBr3/PEDOT:PSS/MAPbI3 perovskite films with and without Au NPs showed film transmittances (see Figure S4) that were below 15% between 400 nm and 700 nm, but above 20% at greater than 750 nm. In particular, 120AuNPs:QD-CsPbBr3/PEDOT:PSS/MAPbI3 perovskite films showed the lowest transmittance (less than 20%) in the whole wavelength range of 400-800 nm. It can be seen that the light absorbance spectra of the QD-CsPbBr3/PEDOT:PSS/MAPbI3 perovskite films were lower than those of AuNPs:QD-CsPbBr3/PEDOT:PSS/MAPbI3 perovskite films when Au NPs were not embedded. On the other hand, the 120AuNPs:QD-CsPbBr3/PEDOT:PSS/MAPbI3 perovskite films had the highest observed light absorbance. This means that the LSPR effect of AuNPs:QD-CsPbBr3 significantly enhanced the light-harvesting and photocarrier transmission of the perovskite solar cells in a wavelength range of 400-800 nm, which will help improve the PCE.  Figure 6a presents the absorbance spectra of annealed QD-CsPbBr 3 , 40AuNPs:QD-CsPbBr 3 , 80AuNPs:QD-CsPbBr 3 and 120AuNPs:QD-CsPbBr 3 films. The transmittance spectra of all the films were greater than 85% in the wavelength range between 520 nm and 800 nm ( Figure S3). However, the light absorbance of the QD-CsPbBr 3 perovskite decreased abruptly at wavelengths shorter than 520 nm due to the addition of Au NPs. In our study, the absorbance peak at about 525 nm was the unique LSPR band in the absorbance spectrum of pure Au NPs. In the short wavelength range from 400 nm to 520 nm, it was observed that the light absorbance of AuNPs:QD-CsPbBr 3 films was significantly enhanced compared to that of QD-CsPbBr 3 films without AuNPs. In addition, the absorbance of the samples improved with increasing amount of embedded Au NPs. Due to the addition of large amounts of Au NPs to generate fluorescence resonance energy transfer and quench the emission of QDs, which effectively enhances local plasmon excitation, the absorbance of 120AuNPs:QD-CsPbBr 3 significantly increased in the whole wavelength range of 400-800 nm. The absorbance spectra of the AuNPs:QD-CsPbBr 3 /PEDOT:PSS/MAPbI 3 perovskite films with and without Au NPs are shown in Figure 6b. The transmittance spectra of the AuNPs:QD-CsPbBr 3 /PEDOT:PSS/MAPbI 3 perovskite films with and without Au NPs showed film transmittances (see Figure S4) that were below 15% between 400 nm and 700 nm, but above 20% at greater than 750 nm. In particular, 120AuNPs:QD-CsPbBr 3 /PEDOT:PSS/MAPbI 3 perovskite films showed the lowest transmittance (less than 20%) in the whole wavelength range of 400-800 nm. It can be seen that the light absorbance spectra of the QD-CsPbBr 3 /PEDOT:PSS/MAPbI 3 perovskite films were lower than those of AuNPs:QD-CsPbBr 3 /PEDOT:PSS/MAPbI 3 perovskite films when Au NPs were not embedded. On the other hand, the 120AuNPs:QD-CsPbBr 3 /PEDOT:PSS/MAPbI 3 perovskite films had the highest observed light absorbance. This means that the LSPR effect of AuNPs:QD-CsPbBr 3 significantly enhanced the light-harvesting and photocarrier transmission of the perovskite solar cells in a wavelength range of 400-800 nm, which will help improve the PCE.
In order to estimate the influence of Au NP addition on the performance of perovskite solar cells, three QD-CsPbBr 3 precursors with different amounts of Au NP solution (µL) ranging from 40 µL to 120 µL were prepared to form AuNPs:QD-CsPbBr 3 films. The original QD-CsPbBr 3 film without the Au NPs was prepared under the same conditions. Figure 7a shows the J-V curves of the original QD-CsPbBr 3 /MAPbI 3 and three types of plasmonic AuNPs:QD-CsPbBr 3 /MAPbI 3 perovskite solar cells. The related parameters of perovskite solar cells with different amounts of Au NPs are listed in Table 1. Obviously, the addition of the Au NPs improved the performance of the perovskite solar cells. The short-circuit current density (Jsc) of the perovskite solar cells was apparently improved when Au NPs were embedded in the QD-CsPbBr 3 . The PCE of the devices increased from 8.53% to 10.9% with increasing Au NP solution addition from 0 µL to 120 µL. Further, the Jsc increased from 20.6 mA cm −2 to 22.5 mA cm −2 . The enhanced PCE was attributed to the Jsc increase of the device due to the incorporation of Au NPs with intense LSPR band into the device. This suggests that Au NPs in QD-CsPbBr 3 can effectively improve the carrier gathering ability. In addition, the photocurrent of the 120AuNPs:QD-CsPbBr 3 /MAPbI 3 perovskite solar cells was larger than that of the 40AuNPs:QD-CsPbBr 3 /MAPbI 3 perovskite solar cell, which may be attributed to the 120AuNPs:QD-CsPbBr 3 with fluorescence resonance energy transfer and charge transfer, which helps to improve the photocurrent and light-harvesting; 40AuNPs:QD-CsPbBr 3 experienced a plasmon resonance energy transfer, which only helps to increase the light-harvesting of the device. On the other hand, as the amount of added Au NPs increased, the fill factor (FF) of the devices also increased. It was also observed that the addition of Au NPs had an almost insignificant influence on the open-circuit voltage (Voc). We found that the series resistance value of plasmonic 120AuNPs:QD-CsPbBr 3 device (157.2 Ω) was smaller than that of the QD-CsPbBr 3 device (212.8 Ω), which indicates that the interfacial contact resistance of the fabricated devices was decreased by adding larger amounts of Au NPs, which directly benefited the improvement of Jsc and FF. This improvement factor can be attributed to 120AuNPs:QD-CsPbBr 3 has better film quality from SEM and XRD analysis. The optimal plasmonic 120AuNPs:QD-CsPbBr 3 /MAPbI 3 perovskite solar cells in our study showed excellent performance, with a Jsc of 22.5 mA cm −2 , a Voc of 0.9 V, a FF of 54%, and a PCE of 10.9%, which indicates that the addition of Au NPs in QD-CsPbBr 3 can efficiently improve the performance of perovskite solar cells. Because of the wavelength overlap between the fluorescence of QD-CsPbBr 3 and the localized plasmon peak of Au NPs, energy transfer that enhances LSP excitation can be obtained in the AuNPs:QD-CsPbBr 3 . In order to estimate the influence of Au NP addition on the performance of perovskite solar cells, three QD-CsPbBr3 precursors with different amounts of Au NP solution (μL) ranging from 40 μL to 120 μL were prepared to form AuNPs:QD-CsPbBr3 films. The original QD-CsPbBr3 film without the Au NPs was prepared under the same conditions. Figure 7a shows the J-V curves of the original QD-CsPbBr3/MAPbI3 and three types of plasmonic AuNPs:QD-CsPbBr3/MAPbI3 perovskite solar cells. The related parameters of perovskite solar cells with different amounts of Au NPs are listed in Table 1. Obviously, the addition of the Au NPs improved the performance of the perovskite solar cells. The short-circuit current density (Jsc) of the perovskite solar cells was apparently improved when Au NPs were embedded in the QD-CsPbBr3. The PCE of the devices increased from 8.53% to 10.9% with increasing Au NP solution addition from 0 μL to 120 μL. Further, the Jsc increased from 20.6 mA cm -2 to 22.5 mA cm -2 . The enhanced PCE was attributed to the Jsc increase of the device due to the incorporation of Au NPs with intense LSPR band into the device. This suggests that Au NPs in QD-CsPbBr3 can effectively improve the carrier gathering ability. In addition, the photocurrent of the 120AuNPs:QD-CsPbBr3/MAPbI3 perovskite solar cells was larger than that of the 40AuNPs:QD-CsPbBr3/MAPbI3 perovskite solar cell, which may be attributed to the 120AuNPs:QD-CsPbBr3 with fluorescence resonance energy transfer and charge transfer, which helps to improve the photocurrent and light-harvesting; 40AuNPs:QD-CsPbBr3 experienced a plasmon resonance energy transfer, which only helps to increase the light-harvesting of the device. On the other hand, as the amount of added Au NPs increased, the fill factor (FF) of the devices also increased. It was also observed that the addition of Au NPs had an almost insignificant influence on the open-circuit voltage (Voc). We found that the series resistance value of plasmonic 120AuNPs:QD-CsPbBr3 device (157.2 Ω) was smaller than that of the QD-CsPbBr3 device (212.8 Ω), To investigate the photocurrent response enhancement effect of AuNPs:QD-CsPbBr 3 on the perovskite solar cell properties, we conducted external quantum efficiency (EQE) measurements in the wavelength range from 350 nm to 800 nm. In Figure 7b, the EQE spectra of the original QD-CsPbBr 3 /MAPbI 3 and three types of plasmonic AuNPs:QD-CsPbBr 3 /MAPbI 3 perovskite solar cells are presented. Compared with the original perovskite solar cells, it was found that the EQEs of the three plasmonic perovskite solar cells with Au NPs were enhanced over the whole wavelength range of 350-800 nm. When the volume of added Au NP solution was 120 µL, the highest EQE values of the plasmonic perovskite solar cells were achieved, which indicated that the EQE measurement results were consistent with the PCE results. In addition, the EQE enhancement factors were obtained by dividing the EQE values of the plasmonic perovskite solar cell by the EQE values of the original perovskite solar cell, as presented in Figure 7c. This clearly shows an enhancement of the EQE values for all plasmonic perovskite solar cells based on a AuNPs:QD-CsPbBr 3 layer. When the amount of Au NP solution was more than 40 µL, the EQE enhancement factor was more apparent, which was mainly caused by the light absorption of AuNPs:QD-CsPbBr 3 in the devices. In particular, the plasmonic AuNPs:QD-CsPbBr 3 /MAPbI 3 perovskite solar cells exhibited strong light absorption and fluorescence emission, which could be due to light absorption by the active layer and LSPR excitation. This helped Energies 2020, 13, 1471 9 of 12 to significantly increase light absorption and photocarrier generation in the devices. It was worth noting that EQE enhancement factors of 120AuNPs:QD-CsPbBr 3 /MAPbI 3 perovskite solar cell shows the value larger than 1.2 over a wide wavelength range of 400-750 nm. However, the Jsc enhancement ratio of 120AuNPs: QD-CsPbBr 3 /MAPbI 3 perovskite solar cell (Jsc = 22.5 mA cm −2 ) was only 1.092 from the J-V curves compared to QD-CsPbBr 3 /MAPbI 3 perovskite solar cell (Jsc = 20.6 mA cm −2 ). The measurement variation may be attributed to different instruments. Figure 7d depicts an energy level diagram, according to which PEDOT:PSS and C 60 performed roles as hole transport and electron transport layers, respectively, while QD-CsPbBr 3 or AuNPs:QD-CsPbBr 3 was used as a light collection layer. It can be seen from the energy level diagram that AuNPs:QD-CsPbBr 3 is supposed to block holes. However, a sufficiently thin layer can allow for tunneling of a hole as indicated by the blue horizontal arrow in the diagram. Figure 7e depicts a schematic diagram of the enhanced PCE mechanism for the plasmonic perovskite solar cells. The Au NPs in the QD-CsPbBr 3 improved their performance due to the LSPR effect, which raised the optical path length by light scattering and increasing the light absorption of the whole perovskite solar cell, thereby efficiently improving the PCE.

Conclusions
Efficient plasmonic AuNPs:QD-CsPbBr 3 /PEDOT:PSS/MAPbI 3 perovskite solar cells were fabricated through embedding Au NPs in perovskite QD-CsPbBr 3 , which manage light-harvesting to enhance perovskite solar cells. Plasmonic AuNPs:QD-CsPbBr 3 films with increasing amounts of embedded Au NPs exhibited good crystal quality and strong light absorption properties. In addition, enhanced PL quenching could improve exciton ionization and charge separation due to LSPR excitation by resonance interaction, which could help to enhance the PCE. In comparison to the performance of the original perovskite solar cells, superior Jsc, FF, PCE and EQE properties of the plasmonic perovskite solar cells were confirmed. The PCE of the plasmonic perovskite solar cells was enhanced by 27.8% compared with that of the original perovskite solar cells. This is attributed to an improvement in the PCE and EQE of plasmonic perovskite solar cells caused by the LSPR effect in AuNPs:QD-CsPbBr 3 enhancing the light-harvesting and light scattering effect.