Efficient Exciton Dislocation and Ultrafast Charge Extraction in CsPbI3 Perovskite Quantum Dots by Using Fullerene Derivative as Semiconductor Ligand

CsPbI3 quantum dots (QDs) are of great interest in new-generation photovoltaics (PVs) due to their excellent optoelectronic properties. The long and insulative ligands protect their phase stability and enable superior photoluminescence quantum yield, however, limiting charge transportation and extraction in PV devices. In this work, we use a fullerene derivative with the carboxylic anchor group ([SAM]C60) as the semiconductor ligand and build the type II heterojunction system of CsPbI3 QDs and [SAM]C60 molecules. We find their combination enables obvious exciton dislocation and highly efficient photogenerated charge extraction. After the introduction of [SAM]C60, the exciton-binding energy of CsPbI3 decreases from 30 meV to 7 meV and the fluorescence emission mechanism also exhibits obvious changes. Transient absorption spectroscopy visualizes a ~5 ps electron extraction rate in this system. The findings gained here may guide the development of perovskite QD devices.

Nevertheless, there is still a huge gap compared with bulk CsPbI 3 solar cells [12]. Although the size effects and long alkyl chain ligands endow perovskite materials with higher stability and higher defect tolerance, they result in more bound excitons, poorer charge transportation, and more difficult charge extraction [13][14][15]. Using weakly polar anti-solvents, such as methyl acetate, can remove the long alkyl chain ligands during layer-by-layer preparation of QD film, facilitating charge transportation and extraction [16]. In this method, precise control of atmospheric moisture can aid the hydrolysis of methyl acetate to generate acetic acid that replaces native oleate ligands, while high atmospheric moisture results in the yellow phase transition [16]. It is also suggested that both processed time and numbers have a big influence on the QD film quality [17]. Although this method has gained outstanding success, the complicated processes and rigorous conditions make the method difficult to repeat. Another method is directly using short-chain ligands to replace long-chain ligands during synthesis or ligand exchange processes [18]. However, the colloidal solution of perovskite QDs with short ligands usually suffers from aggregation because of the weak steric effect of short ligands [19]. Additionally, the polar of short-chain ligands usually has negative effects on the chemical stability of CsPbI 3 QD. Currently, finding approximate ligands that can not only facilitate exciton dislocation and extraction, but also have no negative effects on the chemical structure of CsPbI 3 QD is urgently required for more efficient CsPbI 3 QD photovoltaics. In recent years, fullerene derivatives, as excellent electron acceptors, have been broadly used in organic solar cells and bulk perovskite solar cells [20,21]. Their combination with perovskite QDs was also revealed to be beneficial for photogenerated carrier separation [22,23]. However, their potential as semiconductor ligands for perovskite QDs and their impacts on the exciton properties have not been fully discussed.
In this work, we achieve efficient exciton dislocation and extraction in CsPbI 3 QDs by using a fullerene derivative, 4-(1 ,5 -Dihydro-1 -methyl-2 H- [5,6]fullereno-C60-Ih- [1,9c]pyrrol-2 -yl)benzoic acid (hereafter called [SAM]C60), chosen for its surface-anchoring carboxylic group and its ability to form type II heterojunction system with CsPbI 3 QDs (Scheme 1). We find the introduction of [SAM]C60 has no negative effects on the chemical structure of CsPbI 3 QD, but it obviously decreases the exciton-binding energy of CsPbI 3 QD and changes the fluorescence emission mechanism. Concurrently, ultrafast electron extraction (5 ps) from CsPbI 3 QDs to [SAM]C60 is directly visualized by transient absorption spectroscopy. The results gained here provide an alluring strategy for changing the transportation properties and achieving efficient carrier extraction in CsPbI 3 QDs.
both processed time and numbers have a big influence on the QD film quality [1 hough this method has gained outstanding success, the complicated processes and ous conditions make the method difficult to repeat. Another method is directly short-chain ligands to replace long-chain ligands during synthesis or ligand exchan cesses [18]. However, the colloidal solution of perovskite QDs with short ligands suffers from aggregation because of the weak steric effect of short ligands [19]. Ad ally, the polar of short-chain ligands usually has negative effects on the chemical s of CsPbI3 QD. Currently, finding approximate ligands that can not only facilitate dislocation and extraction, but also have no negative effects on the chemical struc CsPbI3 QD is urgently required for more efficient CsPbI3 QD photovoltaics. In years, fullerene derivatives, as excellent electron acceptors, have been broadly use ganic solar cells and bulk perovskite solar cells [20,21]. Their combination with per QDs was also revealed to be beneficial for photogenerated carrier separation [22,23 ever, their potential as semiconductor ligands for perovskite QDs and their impacts exciton properties have not been fully discussed.
In this work, we achieve efficient exciton dislocation and extraction in CsPbI3 using a fullerene derivative, 4-(1′,5′-Dihydro-1′-methyl-2′H- [5,6]fullereno-C60c]pyrrol-2′-yl)benzoic acid (hereafter called [SAM]C60), chosen for its surface-anc carboxylic group and its ability to form type II heterojunction system with CsPb (Scheme 1). We find the introduction of [SAM]C60 has no negative effects on the ch structure of CsPbI3 QD, but it obviously decreases the exciton-binding energy of QD and changes the fluorescence emission mechanism. Concurrently, ultrafast e extraction (5 ps) from CsPbI3 QDs to [SAM]C60 is directly visualized by transient tion spectroscopy. The results gained here provide an alluring strategy for chang transportation properties and achieving efficient carrier extraction in CsPbI3 QDs.  [24,25] and following absorption spectroscopy.

Results and Discussion
CsPbI3 QDs were synthesized using the previously reported hot injection m with slight modification [1,7]. The as-prepared CsPbI3 QDs exhibit an average diam 12.4 nm and high crystalline quality, as shown in Figure 1a [24,25] and following absorption spectroscopy.

Results and Discussion
CsPbI 3 QDs were synthesized using the previously reported hot injection method, with slight modification [1,7]. The as-prepared CsPbI 3 QDs exhibit an average diameter of 12.4 nm and high crystalline quality, as shown in Figure 1a [SAM]C60 QDs remain very well dispersed, which is attributed to the big steric effect of [SAM]C60 molecules. Moreover, the obvious fluorescence quenching in the complexes under ultraviolet lamp excitation is ascribed by photogenerated electron transfer from CsPbI3 QDs to [SAM]C60 molecules. According to the results of XRD in Figure 1c, both CsPbI3 QDs and CsPbI3-[SAM]C60 crystallized in the orthorhombic phase indicating that [SAM]C60 molecules have no negative effect on the chemical structure of CsPbI3 QDs.         To further reveal the influence of [SAM]C60 molecules on the exciton state of CsPbI3 QDs, we fit the band-edge absorption with the Elliott formula [26,27]: where A is constant, hω is the photon energy, Eb is the binding energy of the exciton, Eg is the optical bandgap, and δ and θ are Delta function and Heaviside functions, separately. The broadening of absorption in a nonideal lattice is simulated by convolving with Gauss functions (broadening width is T). The exciton and continuous band contributions of the band-edge absorption spectrum are separated by fitting the experimental data (see Figure  3a). The fitting parameters are tabulated in Table 1 Figure 3b shows that an excitation power-dependent PL spectrum fitted with a power law IPL~Lexc α , where IPL is the PL intensity, Lexc is the power of the exciting laser (wavelength 470 nm, much higher than the bandgap of QDs), and α is the coefficient related to the exciton emission mechanism, yielded α~0.98 for CsP-bI3 QDs and a~1.42 for CsPbI3-[SAM]C60 QDs. According to the well-established rule that the coefficient α = 1 is in the case of bound exciton emission and α = 2 is in the case of free exciton emission, the carriers in CsPbI3 QDs could be treated as excitons, although they are usually considered to have weak quantum confinement, and the increased α in CsPbI3-[SAM]C60 QDs substantiates the fact that there is exciton delocalization [28]. The type II energy-level alignment can result in local electric fields. The electric field can lower the potential barrier at the QD boundaries and further trigger the exciton delocalization. This kind of fullerene-derivative-induced exciton delocalization has also been reported in bulk perovskite film [29]. In addition, theoretical calculations demonstrated that delocalized states can be formed while QD and ligand experience each other's electric field [30,31]. In brief, the carriers in CsPbI3-[SAM]C60 QDs become more dislocated, which is of great benefit to carrier transportation in photovoltaic devices. To further reveal the influence of [SAM]C60 molecules on the exciton state of CsPbI 3 QDs, we fit the band-edge absorption with the Elliott formula [26,27]: where A is constant, hω is the photon energy, E b is the binding energy of the exciton, E g is the optical bandgap, and δ and θ are Delta function and Heaviside functions, separately. The broadening of absorption in a nonideal lattice is simulated by convolving with Gauss functions (broadening width is T). The exciton and continuous band contributions of the band-edge absorption spectrum are separated by fitting the experimental data (see Figure 3a). The fitting parameters are tabulated in Table 1 Figure 3b shows that an excitation power-dependent PL spectrum fitted with a power law I PL~L exc α , where I PL is the PL intensity, L exc is the power of the exciting laser (wavelength 470 nm, much higher than the bandgap of QDs), and α is the coefficient related to the exciton emission mechanism, yielded α~0.98 for CsP-bI3 QDs and a~1.42 for CsPbI 3 -[SAM]C60 QDs. According to the well-established rule that the coefficient α = 1 is in the case of bound exciton emission and α = 2 is in the case of free exciton emission, the carriers in CsPbI 3 QDs could be treated as excitons, although they are usually considered to have weak quantum confinement, and the increased α in CsPbI 3 -[SAM]C60 QDs substantiates the fact that there is exciton delocalization [28]. The type II energy-level alignment can result in local electric fields. The electric field can lower the potential barrier at the QD boundaries and further trigger the exciton delocalization. This kind of fullerene-derivative-induced exciton delocalization has also been reported in bulk perovskite film [29]. In addition, theoretical calculations demonstrated that delocalized states can be formed while QD and ligand experience each other's electric field [30,31]. In brief, the carriers in CsPbI 3 -[SAM]C60 QDs become more dislocated, which is of great benefit to carrier transportation in photovoltaic devices.
To directly visualize the electron extraction from C S PbI 3 QD to [SAM]C60, we measured the transient absorption (TA) spectra. Figure 4a shows the pseudo color TA maps of CsPbI 3 QD and CsPbI 3 -[SAM]C60 QD, respectively. Here, we use a 600 nm pump as the excitation wavelength. The pump intensity is very low to avoid the generation of hot carrier and Auger recombination. The red color regions are the bleaching signals that reflect the transition of electrons at valence band maximum to exciton energy level. As with the absorption spectra, we also observe an obvious redshift in the bleaching peak (from 1.  To directly visualize the electron extraction from CSPbI3 QD to [SAM]C60, we measured the transient absorption (TA) spectra. Figure 4a shows the pseudo color TA maps of CsPbI3 QD and CsPbI3-[SAM]C60 QD, respectively. Here, we use a 600 nm pump as the excitation wavelength. The pump intensity is very low to avoid the generation of hot carrier and Auger recombination. The red color regions are the bleaching signals that reflect the transition of electrons at valence band maximum to exciton energy level. As with the absorption spectra, we also observe an obvious redshift in the bleaching peak (from 1.     To directly visualize the electron extraction from CSPbI3 QD to [SAM]C60, we measured the transient absorption (TA) spectra. Figure 4a shows the pseudo color TA maps of CsPbI3 QD and CsPbI3-[SAM]C60 QD, respectively. Here, we use a 600 nm pump as the excitation wavelength. The pump intensity is very low to avoid the generation of hot carrier and Auger recombination. The red color regions are the bleaching signals that reflect the transition of electrons at valence band maximum to exciton energy level. As with the absorption spectra, we also observe an obvious redshift in the bleaching peak (from 1.

Conclusions
To summarize, we use [SAM]C60 as the semiconductor ligands of CsPbI 3 QDs. Compared with other short-chain ligands, the big steric hindrance of the [SAM]C60 molecule keeps the colloidal dispersions stable in solution. Meanwhile, its alignment of the energy levels with CsPbI 3 QD can form a type II heterojunction, accelerating exciton dislocation and transfer. We observe ultrafast carrier transfer (~5 ps) and highly efficient extraction efficiency (~99.1%) from CsPbI 3 QDs to [SAM]C60. Our research suggests the great potential of a semiconductor fullerene derivative as the ligand of perovskite QDs and paves the way for efficient CsPbI 3 QD photovoltaics.

Colloidal Synthesis of CsPbI 3 QDs
First, 2.1324 g OA, 0.6 g of Cs 2 CO 3 and 20 mL ODE were mixed in a 100 mL three-neck flask and degassed at 120 • C for about 40 min. Concurrently, 0.4 g of PbI 2 and 20 mL ODE were mixed in a 50 mL three-neck flask at 120 • C for 30 min under vacuum. A mixture of 2 mL OA and 2 mL OAm was injected into the PbI 2 precursor and kept at 120 • C for about 30 min. The mixture was heated to 160 • C under N 2 , then 3.2 mL of the Cs-oleate precursor was immediately injected into the reaction flask, reacted for 5~8 s, and then quickly quenched to room temperature in an ice bath. The colloidal solution was separated into four tubes and 45 mL MeOAc was added, then it was centrifuged at 8500 rpm for 4 min. The precipitated QDs were dispersed in 4 mL of hexane and stored in a refrigerator at 4 • C for 24 h, then centrifuged at 4000 rpm for 2 min. Before being used, the QDs solution was centrifuged at 7000 rpm for 5 min. The supernatant was dried with N 2 and then dispersed into 50 mg/mL CsPbI 3 QDs n-hexane solution.

Characterization
The phase identification was carried out using a powder X-ray diffraction (XRD, TTR-III, Rigaku Corp., Akishima-shi, Tokyo, Japan). The morphology and crystal structure of the prepared samples were characterized using a transmission electron microscopy (JEM-2100F, JEOL Ltd., Akishima, Tokyo, Japan). UV-vis absorption spectra were measured with a spectrophotometer (HITACHI, U-3900H, Minato-ku, Tokyo" Japan). Photoluminescence Quantum Yield (PLQY) and steady Photoluminescence (PL) was measured with the Absolute PLQY Measurement System (from C11347 Hamamatsu, Hamamatsu City, Japan) at an excitation power of 0.1 mW. Transient PL. PL are all recorded with the PL system from TOKYO INSTRUMENT, INC. A 473 nm pulsed diode laser (pulse width 90 ps, repetition up to 100 MHz, peak power 4 mW) was used as the excitation source. An adjustable neutral density filter was adopted to adjust the excitation intensity. The PL detection was used with a PMT together with a TCSPC module. Transient absorption (TA) measurements were performed using an fs TA setup. The laser source was a titanium/sapphire laser (CPA-2010, Clark-MXR Inc., Huron River, MI, USA) with a wavelength of 775 nm, a repetition rate of 1 KHz, and a pulse width of 120 fs. The light was separated into two parts. One part was used to stimulate a sapphire plate to generate white light for the probe beam. The other part was used to pump an optical parametric amplifier (OPA) (a TOAPS from Quantronix) to generate light pulses with a wavelength tunable from 290 nm to 3 µm. It was used as a pump light to excite the sample. In this study, a pump light with a wavelength of 600 nm was used to excite the QDs.