Stable CsPbBr 3 Nanocrystals—Decorated Nanoporous Gold for Optoelectronic Applications

: Halide perovskite colloidal nanocrystals have recently gained much attention thanks to their superior stability compared with their bulk counterpart and to their unique optical properties. In this paper, two systems combining nanocrystals and nanoporous gold are studied to create an optimal metal semiconductor heterojunction that can be used in photocatalysis and photovoltaic devices. The perovskite degradation phenomenon is observed when the nanoporous gold powder is mixed into the hexane suspension of nanocrystals, while the charge separation efﬁciency is increased by synthesizing the nanocrystals directly onto the gold porous structure. The analysis of the structural and optical properties evidences an energy transfer efﬁciency of 47%, along with the high structural stability of the hybrid system.


Introduction
Since the 1990s, halide perovskites nanocrystals (NCs) have received increasing attention in an optoelectronic context for their intrinsic properties, finding potential applications in many different sectors and devices ranging from efficient light harvesters to photovoltaics, photodetectors, solar fuels, lighting, and displays [1]. Despite their promising characteristics, the efficiency of perovskite-based devices is far from the theoretically predicted values, probably due to poor charge transfer at the interfaces [2,3]. Further, their poor stability hampers the commercial applications of the devices, being sensitive to heat, light, and environmental stresses [4][5][6][7]. In photocatalysis and photovoltaics, for example the performance of the devices depends on their charge separation ability and the implementation of good NCs-electrode interfaces. The development of nanocrystal-metal heterostructures can help to improve both factors and provide new functionalities to the nanostructures, altering the optical properties of the semiconductors, notably with photoluminescence quenching or enhancement [8][9][10]. Noble metals are widely used for their intrinsic capacity to accept electrons, quickly capturing the photogenerated electrons at the semiconductor surface and promoting the charge separation and reduction processes. The presence of metal is also known to induce plasmonic field effects in semiconductor nanostructures, influencing their photocatalytic and photovoltaic performances [11,12]. Previously, several studies have been focused on the optical properties of lead halide perovskite-Au nanocomposites and their application in photocatalysis and photovoltaics [13][14][15]. In particular,

Synthesis and Purification of CsPbBr 3 NCs
The synthesis and purification of CsPbBr 3 NCs were performed based on previous reports [27]. First, the cesium precursor was prepared by loading 1 mmol of Cs 2 CO 3 and 10 mL of OTAc into a 20 mL vial, then the mixture was stirred for 10 min at room temperature. The PbBr 2 precursor solution was prepared by dissolving 1 mmol of PbBr 2 and 2 mmol of TOAB in 10 mL of toluene. For the synthesis of CsPbBr 3 NCs, 1.0 mL of a cesium precursor solution was swiftly added to 9 mL of a PbBr 2 -toluene solution in the conical flask. The reaction was magnetically stirred for 5 min at room temperature in open air. Subsequently, 3 mL of DDAB (in toluene 10 mg/mL) solution was added. After 2 min, ethyl acetate was added to the crude solution at a volume ratio of 2:1; the precipitate was collected separately after centrifugation and dispersed in toluene. The extra ethyl acetate was added into the toluene dispersion and the precipitate was collected and re-dispersed in n-hexane.

Precursor Alloy Fabrication
Au and Ag powders were mixed at a ratio of 30:70. 2 g of the mixture were ball milled in a SPEX 8000M Mixer/Mill ball mill for 16 h in a hardened steel vial with two hardened steel spheres measuring 8 g each. The powders were homogenized every 30 min in the first 2 h and then every 5 h.

NP Au Fabrication
NP Au powders were prepared via chemical dealloying in HNO 3 70% for 24 h [28,29]. The powders were washed in Milli-Q water 5 times and then dried under vacuum.

Synthesis and Purification of CsPbBr 3 -Au Hybrid Structure
CsPbBr 3 -Au hybrid structures were prepared following two different procedures. In procedure 1, 70 mg of NP Au was added into the CsPbBr 3 NCs (2 mg/mL) dispersed in hexane (3 mL) under ambient conditions. After stirring for 4 h, the powder was filtered and washed several times with hexane under vacuum and then dried overnight.
In procedure 2, the synthesis of CsPbBr 3 NCs was repeated in the presence of NP Au. NP Au (140 mg) was added to the PbBr 2 precursor solution, then the synthesis proceeded identically. The hexane dispersion was filtered and washed several times with hexane under vacuum, to separate the nanocrystals not linked to gold, and then dried overnight.

Characterization
X-ray measurements were performed with a Bruker D8 Advance diffractometer operating at 30 kV and 20 mA, equipped with a Cu tube (λ = 1.5418 Å) and a Vantec-1 PSD detector. The powder patterns were recorded in the 10 • ≤ 2θ ≤ 70 • range.
Raman measurements were collected using an MS750 spectrograph (sol-instruments) equipped with 600 gr/mm grating. The laser beam (785 nm) was focalized through an Olympus objective (10×), with a laser power of about 7.5 mW. Measurements were performed at room temperature, with a spectral resolution of 1 cm −1 .
Time-resolved photoluminescence (TR-PL) measurements were recorded by exciting the samples with 200 fs pulses sourced from an optical parametric amplifier (Light Conversion TOPAS-C) pumped by a regenerative Ti-sapphire amplifier (Coherent Libra-HE). The repetition frequency was 1 kHz and the TR-PL signal was recorded by a streak camera (Hamamatsu C10910) equipped with a grating spectrometer (Princeton Instruments Acton Spectra Pro SP-2300). All measurements were collected in the front-face configuration to reduce inner filter effects. Proper emission filters were applied to remove the reflected contribution of the excitation light.
Scanning electron microscopy (SEM) of nanoporous gold was performed using an FEI Quanta Field Emission Electron Microscope, equipped with an Everhart-Thornley secondary electron detector. Imaging was performed at an operating voltage of 20 kV 10 mm from the examined samples.
SEM studies of heterostructures were carried out using a Carl ZEISS Auriga microscope equipped with an energy-dispersive X-ray spectroscopy (EDX) detector or an ESEM FEI Quanta 200 microscope operating at 25 kV.
TEM measurements on CsPbBr 3 NCs and on the heterostructure were performed using a TEM JeolJEM 1400 Plus transmission electron microscope (120 kV).

Results and Discussion
An SEM image of the as-prepared NP Au is shown in Figure 1a. The NP Au shows a fine nanoporous structure of interconnected ligaments and pores with a mean diameter of around 20 nm. The CsPbBr 3 nanocrystals were synthesized following a room-temperature procedure, as adopted in previous reports, where the effects of the three ligands contibute to their stability over several months [27]. TOAB is used to increase the solubility of PbBr 2 , while DDAB and OTAc act as capping agents, which promote the uniform dispersion of NCs in hexane and the stability of the compound. The nanocrystals show a uniform morphology with a crystal size of about 15 nm, as shown in Figure 1b.
SEM studies of heterostructures were carried out using a Carl ZEISS Auriga microscope equipped with an energy-dispersive X-ray spectroscopy (EDX) detector or an ESEM FEI Quanta 200 microscope operating at 25 kV.
TEM measurements on CsPbBr3 NCs and on the heterostructure were performed using a TEM JeolJEM 1400 Plus transmission electron microscope (120 kV).

Results and Discussion
An SEM image of the as-prepared NP Au is shown in Figure 1a. The NP Au shows a fine nanoporous structure of interconnected ligaments and pores with a mean diameter of around 20 nm. The CsPbBr3 nanocrystals were synthesized following a roomtemperature procedure, as adopted in previous reports, where the effects of the three ligands contibute to their stability over several months [27]. TOAB is used to increase the solubility of PbBr2, while DDAB and OTAc act as capping agents, which promote the uniform dispersion of NCs in hexane and the stability of the compound. The nanocrystals show a uniform morphology with a crystal size of about 15 nm, as shown in Figure 1b. X-ray diffraction measurements were performed on a nanocrystal precipitate. Figure  2 reports the Rietveld refinement of the pattern obtained using MAUD software [30]. The CsPbBr3 NCs are in the orthorhombic -phase (Pbnm, space group No. 62, COD ID 1533062), with lattice parameters a = 8.219 Å, b = 8.254 Å, and c = 11.753 Å. X-ray diffraction measurements were performed on a nanocrystal precipitate. Figure 2 reports the Rietveld refinement of the pattern obtained using MAUD software [30]. The CsPbBr 3 NCs are in the orthorhombic γ-phase (Pbnm, space group No. 62, COD ID 1533062), with lattice parameters a = 8.219 Å, b = 8.254 Å, and c = 11.753 Å.
SEM studies of heterostructures were carried out using a Carl ZEISS Auriga microscope equipped with an energy-dispersive X-ray spectroscopy (EDX) detector or an ESEM FEI Quanta 200 microscope operating at 25 kV.
TEM measurements on CsPbBr3 NCs and on the heterostructure were performed using a TEM JeolJEM 1400 Plus transmission electron microscope (120 kV).

Results and Discussion
An SEM image of the as-prepared NP Au is shown in Figure 1a. The NP Au shows a fine nanoporous structure of interconnected ligaments and pores with a mean diameter of around 20 nm. The CsPbBr3 nanocrystals were synthesized following a roomtemperature procedure, as adopted in previous reports, where the effects of the three ligands contibute to their stability over several months [27]. TOAB is used to increase the solubility of PbBr2, while DDAB and OTAc act as capping agents, which promote the uniform dispersion of NCs in hexane and the stability of the compound. The nanocrystals show a uniform morphology with a crystal size of about 15 nm, as shown in Figure 1b. X-ray diffraction measurements were performed on a nanocrystal precipitate. Figure  2 reports the Rietveld refinement of the pattern obtained using MAUD software [30]. The CsPbBr3 NCs are in the orthorhombic -phase (Pbnm, space group No. 62, COD ID 1533062), with lattice parameters a = 8.219 Å, b = 8.254 Å, and c = 11.753 Å.  In Figure 3, the Raman spectrum of the nanocrystal precipitate is reported; in the inset is an enlarged view of the region covering 200-400 cm -1 . The Raman features confirm that CsPbBr 3 NCs are in the orthorhombic γ-phase, with an intense peak in the region of 60-80 cm -1 , a low-intensity peak around 130 cm −1 , and a broad second-order peak around 310 cm −1 [31,32].
In Figure 3, the Raman spectrum of the nanocrystal precipitate is report inset is an enlarged view of the region covering 200-400 cm -1 . The Raman feature that CsPbBr3 NCs are in the orthorhombic γ-phase, with an intense peak in the 60-80 cm -1 , a low-intensity peak around 130 cm −1 , and a broad second-order pea 310 cm −1 [31,32].     310 cm −1 [31,32].    The synthesis of NCs@NP Au followed two different procedures (detailed conditions are reported in Section 2.3). In procedure 1, a known amount of NP Au (70 mg) was incorporated into the hexane suspension of CsPbBr 3 NCs under ambient conditions. The concentration of the CsPbBr 3 NCs was 2 mg/mL. Figure 5 shows the pictures of the solution under UV light (375 nm), before and after the addition of NP Au. incorporated into the hexane suspension of CsPbBr3 NCs under ambient conditions. The concentration of the CsPbBr3 NCs was 2 mg/mL. Figure 5 shows the pictures of the solution under UV light (375 nm), before and after the addition of NP Au. Above, the solution turns colorless with a drastic quenching of the fluorescence; the slight blue fluorescence is due to the glass vial. The Raman measurements ( Figure 6) show the disappearance of the features typical of γ-CsPbBr3 and the appearance of two peaks (135 and 176 cm −1 ), which cannot be assigned but are in the range of Cs-Pb-Br vibrations [31,33,34]. This can be attributed to the instability of NCs in the presence of gold. Balakrishnan et al. already reported the instability of CsPbBr3 NCs in the presence of gold, depending on the procedure followed for the synthesis of the heterostructure, evidencing the degradation of perovskites in 5 min after the addition of gold NPs in a solution of colloidal CsPbBr3 NCs [35]. The capping agent (namely DDAB) is formed by the alkyl chain terminating with the nitrogen and methyl groups. It is widely accepted that there is a strong affinity of the Au surface with ammonium surfactants, and the capping agent likely leaves the nanoparticles to deposit on the Au nanostructures [36][37][38]. Above, the solution turns colorless with a drastic quenching of the fluorescence; the slight blue fluorescence is due to the glass vial. The Raman measurements ( Figure 6) show the disappearance of the features typical of γ-CsPbBr 3 and the appearance of two peaks (135 and 176 cm −1 ), which cannot be assigned but are in the range of Cs-Pb-Br vibrations [31,33,34]. This can be attributed to the instability of NCs in the presence of gold. Balakrishnan et al. already reported the instability of CsPbBr 3 NCs in the presence of gold, depending on the procedure followed for the synthesis of the heterostructure, evidencing the degradation of perovskites in 5 min after the addition of gold NPs in a solution of colloidal CsPbBr 3 NCs [35]. incorporated into the hexane suspension of CsPbBr3 NCs under ambient conditions. The concentration of the CsPbBr3 NCs was 2 mg/mL. Figure 5 shows the pictures of the solution under UV light (375 nm), before and after the addition of NP Au. Above, the solution turns colorless with a drastic quenching of the fluorescence; the slight blue fluorescence is due to the glass vial. The Raman measurements ( Figure 6) show the disappearance of the features typical of γ-CsPbBr3 and the appearance of two peaks (135 and 176 cm −1 ), which cannot be assigned but are in the range of Cs-Pb-Br vibrations [31,33,34]. This can be attributed to the instability of NCs in the presence of gold. Balakrishnan et al. already reported the instability of CsPbBr3 NCs in the presence of gold, depending on the procedure followed for the synthesis of the heterostructure, evidencing the degradation of perovskites in 5 min after the addition of gold NPs in a solution of colloidal CsPbBr3 NCs [35]. The capping agent (namely DDAB) is formed by the alkyl chain terminating with the nitrogen and methyl groups. It is widely accepted that there is a strong affinity of the Au surface with ammonium surfactants, and the capping agent likely leaves the nanoparticles to deposit on the Au nanostructures [36][37][38]. The capping agent (namely DDAB) is formed by the alkyl chain terminating with the nitrogen and methyl groups. It is widely accepted that there is a strong affinity of the Au surface with ammonium surfactants, and the capping agent likely leaves the nanoparticles to deposit on the Au nanostructures [36][37][38].
We argue that NP Au possesses a good affinity with the capping agent of CsPbBr 3 NCs. Stripped of all protection, NCs are no longer stable in hexane and degrade in different stoichiometric phases, as observed in the Raman spectra. To further confirm the hypothesis, the Au nanostructures were covered with the same capping agent and successively added to the CsPbBr 3 NCs hexane solution. No interaction between CsPbBr 3 NCs and Au nanostructures was observed in this case; the perovskites maintained the structural stability and their characteristic optical properties. To further confirm the hypothesis of the degradation, we performed an elemental analysis on different points of the surface of the sample obtained with procedure 1, using SEM-EDX measurements (Table 1). Paying attention to Cs, Pb and Br percentages, we evidenced the presence of lead or cesium or both, with the disappearance of bromine. This difference in the stoichiometry from the starting CsPbBr 3 NCs confirms the absence of perovskites and the formation of amorphous Cs-Pb compounds. In procedure 2, nanoporous gold was added during the synthesis of CsPbBr 3 nanocrystals, then the synthesis proceeded as mentioned in Section 2.1. Figure 7 reports the XRD pattern of the NC-NP Au sample compared to pure NP Au. It is possible to notice the presence of additional peaks (at 21.5 • , 31.5 • , 49.4 • and 59.1 • ) pertaining to pure CsPbBr 3 NCs (Rietveld refinement in Figure 2). hypothesis, the Au nanostructures were covered with the same capping agent and successively added to the CsPbBr3 NCs hexane solution. No interaction between CsPbBr3 NCs and Au nanostructures was observed in this case; the perovskites maintained the structural stability and their characteristic optical properties. To further confirm the hypothesis of the degradation, we performed an elemental analysis on different points of the surface of the sample obtained with procedure 1, using SEM-EDX measurements (Table 1). Paying attention to Cs, Pb and Br percentages, we evidenced the presence of lead or cesium or both, with the disappearance of bromine. This difference in the stoichiometry from the starting CsPbBr3 NCs confirms the absence of perovskites and the formation of amorphous Cs-Pb compounds. In procedure 2, nanoporous gold was added during the synthesis of CsPbBr3 nanocrystals, then the synthesis proceeded as mentioned in Section 2.1. Figure 7 reports the XRD pattern of the NC-NP Au sample compared to pure NP Au. It is possible to notice the presence of additional peaks (at 21.5°, 31.5°, 49.4° and 59.1°) pertaining to pure CsPbBr3 NCs (Rietveld refinement in Figure 2).   In order to confirm the success of the procedure and exclude the presence of degradation products, we performed SEM-EDX measurements on 6 different areas. The composition was found to be homogeneous across the sample surface, and Table 2 reports the atomic percentages of the elements detected in one of these areas, evidencing the correct stoichiometry of CsPbBr 3 NCs.  Figure 8 reports the TEM images of the NP Au and the hybrid structure with perovskites. The resolution of the images is affected by the nature of the material itself; NP Au powders have µm-range mean sizes, and for this reason it was only possible to acquire images of the smallest Au particles (around 200 nm in size), which were partially transparent to the electron beam. The thickness of the particles was still high for TEM imaging, which caused the low resolution of the images. However, looking at a small grain of NP Au in Figure 8b, the presence of the NCs at the surface of the gold structure is clearly visible, evidencing the close contact between the Au and CsPbBr 3 NCs. composition was found to be homogeneous across the sample surface, and Table 2 reports the atomic percentages of the elements detected in one of these areas, evidencing the correct stoichiometry of CsPbBr3 NCs.  Figure 8 reports the TEM images of the NP Au and the hybrid structure with perovskites. The resolution of the images is affected by the nature of the material itself; NP Au powders have µm-range mean sizes, and for this reason it was only possible to acquire images of the smallest Au particles (around 200 nm in size), which were partially transparent to the electron beam. The thickness of the particles was still high for TEM imaging, which caused the low resolution of the images. However, looking at a small grain of NP Au in Figure 8b, the presence of the NCs at the surface of the gold structure is clearly visible, evidencing the close contact between the Au and CsPbBr3 NCs.    The curves were fitted with a pseudo-Voigt function, a combination of Gaussian and Lorentzian profiles: The curves were fitted with a pseudo-Voigt function, a combination of Gaussian and Lorentzian profiles: where µ represents the Lorentzian character of the Voigt profile, and w L and w G are the width parameters of the Lorentzian and Gaussian components, respectively. Comparing the PL behavior of the NCs@NP Au sample and pure CsPbBr 3 NCs, it is possible to observe a slight red shift of the main peak at 518 nm (2.39 eV) and the appearance of a new less-intense peak at lower energies. The derived parameters are reported in Table 3. The variation in the positions of the peak implies an influence of the nanoporous gold matrix on the excited-state properties of CsPbBr 3 NCs. The possible reason for the appearance of the secondary peak could be the formation of NCs with different dimensions. Effectively, the synthesis procedure promotes the formation of NCs with dimensions of around 15 nm, although we cannot exclude the formation of aggregation or bigger nanocrystals inside the gold pores.
The time-resolved photoluminescence measurements (reported in Figure 10) performed on the samples give more insight into the properties of the recombination mechanism. With regard to the NCs@NP Au sample, the two peaks show the same decay behavior. For this reason, only one curve is reported for this sample in Figure 10. The curves are fitted with a biexponential decay function: Here, I(t) is the time-dependent PL intensity, I0 is the initial PL intensity, Ai is the amplitude, t is time, t0 is the initial time, and τi is the characteristic lifetime. The average lifetime is calculated using the following relation [39]: Here, I(t) is the time-dependent PL intensity, I 0 is the initial PL intensity, A i is the amplitude, t is time, t 0 is the initial time, and τ i is the characteristic lifetime. The average lifetime is calculated using the following relation [39]: The error in parameter A i derives from the fit, whilst in the case of τ i it is an experimental error resulting from the rising time, which is higher than the error provided by the fit. The error of the average lifetime is calculated by the propagation of uncertainties. Table 4 shows the fitting parameters, showing that CsPbBr 3 NCs are characterized by the presence of two decay components. Previous studies assigned the fast time component to the trap-assisted recombination and the long one to the intrinsic electron-hole recombination [40,41]. When CsPbBr 3 NCs are grown in the presence of NP Au, the average lifetime decreases from 3.6 ns to 1.9 ns, suggesting an interaction between CsPbBr 3 NCs and the nanoporous gold matrix, as evidenced by the shortening of the average lifetime with respect to the CsPbBr 3 NCs. Indeed, a faster decay time can be considered an indication of the formation of the NC-NP Au hybrid structure with the presence of a non-radiative recombination path and an interfacial charge transfer [14,35,42,43]. Actually, the same results could be due to other effects, such as the formation of surface defects, although the formation of a defective structure does not explain the high stability and the absence of a new feature, while an energy transfer is more likely, as already observed in other CsPbBr 3 NCs-Au systems [35]. The transfer efficiency can be calculated by the following relation: where τ 0 is the average lifetime in pure CsPbBr 3 NCs and τ is the average lifetime in the presence of gold. The calculated charge transfer efficiency is around 47%.

Conclusions
In summary, the goal of this study was to achieve a CsPbBr 3 -Au heterostructure. In one approach, CsPbBr 3 NCs dispersed in hexane were mixed with nanoporous gold. The result was the disappearance of the luminescence, the cause of which was assumed to be the degradation of NCs and the subsequent formation of amorphous compounds. The second approach involved the synthesis of NCs directly on the NP Au surface. In this case, it was possible to notice a red shift in the emission spectrum and a faster decay rate. Such behavior can be attributed to the achievement of a nanostructure of CsPbBr 3 NCs and NP Au, with an efficient charge transfer rate. The results showed the high potential for applications where an efficient charge transfer is required. Further, the use of perovskite NCs as an efficient sensitizer strongly suggests the possibility to achieve tunable photocatalytic devices active across the visible range. In this regard, further specific experimental tests need to be performed.