Highly stable photoluminescence in vacuum-processed halide perovskite core-shell 1D nanostructures

RESEARCH ARTICLE www.afm-journal.de Highly Stable Photoluminescence in Vacuum-Processed Halide Perovskite Core–Shell 1D Nanostructures Javier Castillo-Seoane, Lidia Contreras-Bernal,* T. Cristina Rojas, Juan P. Espinós, Andrés-Felipe Castro-Méndez, Juan-P. Correa-Baena, Angel Barranco, Juan R. Sanchez-Valencia,* and Ana Borras Hybrid organometal halide perovskites (HP) present exceptional optoelectronic properties, but their poor long-term stability is a major bottleneck for their commercialization. Herein, a solvent-free approach to growing single-crystal organic nanowires (ONW) is presented, along with nanoporous metal oxide scaffolds and HP, to form a core@multishell architecture. The synthesis is carried out under mild vacuum conditions employing thermal evaporation for the metal-free phthalocyanine (H2Pc) nanowires, which are the core, plasma-enhanced chemical vapor deposition (PECVD) for the TiO2shell, and co-evaporation of lead iodide (PbI2)and methylammonium iodide (CH3NH3I/MAI) for the CH3NH3PbI3 (MAPbI3/MAPI) perovskite shell. A detailed characterization of the nanostructures by electron microscopy, (S)-TEM, and X-ray diffraction, XRD, is presented, revealing a different crystallization of the HP depending on the template: while the growth on H2Pc nanowires induces the typical MAPI tetragonal structure, a low-dimensional phase (LDP) is observed on the 1D-TiO2nanotubes. Such a combination yields an unprecedentedly stable photoluminescence emission over 20 h and over 300 h after encapsulation in polymethyl methacrylate (PMMA) under different atmospheres including N2, air, and high moisture levels. Moreover, the unique 1D morphology of the system, together with the high refractive index, allows for a strong waveguiding effect along the HP nanowire length. J. Castillo-Seoane, L. Contreras-Bernal, J. P. Espinós, A. Barranco, J. R. Sanchez-Valencia, A. Borras Nanotechnology on Surfaces and Plasma Lab Institute of Materials Science of Seville (CSIC-US) Americo Vespucio 49, Seville 41092, Spain E-mail: [email protected];[email protected] The ORCID identification number(s) for the author(s) of this article can be found under https://doi.org/10.1002/adfm.202403763 © 2024 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH. This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. DOI: 10.1002/adfm.202403763 1. Introduction The last 15 years have witnessed an increasing interest in the development of halide perovskites (HP), such as the pure metal and hybrid organometal perovskites as photo-absorbers in third-generation solar cells with efficiencies reaching 26.1% in single-junction and 33.9% in tandem perovskite-silicon cells.[1]Even though issues such as long-term stability and reproducibility of the cell devices are yet to be solved, the photophysical properties of these materials make them impressively attractive for the development of optoelectronic systems.[2,3]Morphology, size, and structure of the HPs are critical in all these applications as well as it is essential to synthesize them in a highly reproducible way. In this context, low-dimensional perovskites such as 0D nanocrystals and nanoparticles, 1D nanowires, nanorods, nanotubes, and 2D nanosheets have been recently reported.[4,5]Most of these examples showcase a 3D-networked perovskite crystalline structure, different from the 2D Ruddlesden–Popper perovskites.[2]1D HP T. C. Rojas Institute of Materials Science of Seville (CSIC-US) Americo Vespucio 49, Seville 41092, Spain A.-F. Castro-Méndez, J.-P. Correa-Baena School of Materials Science and Engineering Georgia Institute of Technology Atlanta, GA 30332, USA Adv. Funct. Mater. 2024,34, 2403763 2403763 (1 of 13) © 2024 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH www.advancedsciencenews.com www.afm-journal.de nanostructures have already been explored as lasers, photo absorbers, and passivation fillers in solar cells, light-emitting diodes (LEDs), and photodetectors[2,3]due to their high crystallinity, superior charge carrier mobility, low defects, enhanced mechanical properties, and surface area, tunable bandgaps, anisotropic electrical/optical properties, and high photoluminescence quantum yield.[2,3]Direct on-substrate formation of these nanomaterials is also decisive to improve the level of integration in devices. Moreover, 1D perovskites possess a strong potential as a model system for a full elucidation of shape/size-dependent relationships and properties such as quantum confinement and charge transport.[6]Some examples in the literature have also demonstrated more robust stability than the counterparts of polycrystalline thin films, as the reduced lateral dimension would block the diffusion of water or oxygen.[6]Different approaches have been proposed to fabricate low-dimensional nanomaterials, including solution-based, vapor transport, vapor phase synthesis, hot injection methods, templating, and combined solutions.[2,3]One of the major advantages of the use of template-assisted methods is the high definition and reproducibility of the shape of the nanostructures, particularly in the case of nanowires and nanorods. The most extended approach is the use of hard templates based on anodized aluminum oxide (AAO) and porous alumina membranes that have already been exploited for the fabrication of MAPbI3, MAPbBr3,MASnI3,andCs 2SnI6.[2,7–11]However, an important drawback of this approach is the deleterious effect on the HP during the etching processes required for the removal of the template. In this article, we present an alternative vacuum and plasmabased soft-template method for the preparation of core@shell nanowires/nanotubes formed by methylammonium lead iodide (MAPI) perovskite on 1D supported nanostructures made of single crystalline metal-free phthalocyanine (H2Pc) as a softtemplate and nanoporous TiO2. We have recently demonstrated the advantages of using this soft template approach for the formation of core@shell and core@multishell nanowires (NW) and nanotubes (NT) with different organic, metallic, and metal oxide compositions.[12–16]The process is carried out under mild vacuum and temperature conditions (below 350 °C). It is fully compatible with processable substrates such as silicon, fused silica, glass, polymers (PET, PDMS), metal foils and meshes, and even cellulose.[16]To the best of our knowledge, this is the first time that this type of approach has been applied to the fabrication of 1D perovskite nanostructures. We will show below the fundamentals for the formation of H2Pc@MAPI nanowires and TiO2@MAPI nanotubes, the crystalline structure of the HP and their optical properties, focusing on the effect of the nanoscale arrangement on the photoluminescence and light-guiding properties. TiO2@MAPI nanotubes exhibit outstanding photoluminescence stability even in a rich-moisture environment. We thoroughly compare the performance of the 1D nanostructures with polycrystalline thin films prepared under the same conditions also, including encapsulation with PMMA. The long-lasting stability is discussed under the premise of nanostructure and crystalline characteristics and the formation of a low dimensional perovskite (LDP) crystalline phase. Although some LDP halide perovskite crystalline phases have been reported with a higher stability, such as the 2D Ruddlesden–Popper perovskites, the MAI-rich MAPI LDP phase reported here has not been yet related to a stability improvement.[17,18] 2. Results and Discussion Scheme 1shows the multistep vacuum procedure applied to synthesize H2Pc@MAPI NW and TiO2@MAPI NT. Step 0, in the schematic corresponds to the growth of the seed layer (in this case, an amorphous, mesoporous, and nanocolumnar TiO2thin film) required for the crystalline growth of the 1D soft templates. The first case for the synthesis of H2Pc@MAPI NW (top side way of the schematic) consists mainly of two steps (see Experimental Section for further details): 1-a) Growth of single-crystalline organic nanowires (ONW) through self-assembly by pi-stacking of planar extended aromatic H2Pc molecules by thermal evaporation. The approach for the growth of the ONW provides a reproducible and feasible route for the controlled formation of smallmolecule NW (regarding density and mean length) supported on an ample variety of substrates.[16]The H2Pc nanowires are highly flexible and grow along random directions with respect to the substrate, presenting a square or rectangular footprint and an extremely flat and pristine surface (see Figure S1, Supporting Information).[19–21]2) Formation of the perovskite shell by co-evaporation of the PbI2and CH3NH3I on top of the as-grown ONW. The procedure synthesizing HP crystals, in this case, MAPI by co-evaporation of organic and inorganic counterparts, was previously reported by other authors in references.[17,22,23] In this study, we have adapted the deposition methodology to our system, allowing the formation of high-quality MAPI crystals with precise control in the shell (and equivalent thin film) thickness. The corresponding SEM images of the H2Pc@MAPI NW at different magnifications are shown in Figure 1a,c,e in planar view, and a cross-section in panel (d) (in blue color). The images show how the perovskite is deposited along the ONW, producing, at first sight, a core@shell H2Pc@MAPI 1D nanostructure without the presence of other MAPI nanostructures such as platelets or conglomerates. The mean length and diameter are ≈4μmand 150 nm, respectively (see the statistical analysis of the diameters in Figure S2, Supporting Information). From Figure 1c,itcanbe noted that the length dispersion of the H2Pc NW is broad, and the longest ones reach up to 10 μm and are bent, forming arcs randomly oriented. The thickness of the HP shell is homogeneous along the NW’s length, with a slight decrease at the bottom near the interface with the substrate. This difference in thickness from the top to bottom apexes is the consequence of a self-shadowing effect intrinsic to the directionality of the vacuum deposition process. The higher magnification views (Figure 1e) show the distribution of the HP as a smooth and continuous shell. Surprisingly, a close inspection of the cross-sectional views of the NW reveals that the microstructure does not correspond to the expected coaxial architecture. The perovskite layer grows preferentially along one side of the ONW, as shown in Figure S3 (Supporting Information). This is likely due to the high directionality of the vapor evaporation technique, the low surface mobility, and the high sticking coefficient of the perovskite on the organic surface, resulting in a shell with an asymmetric section. The physical vapor deposition method applied in step 1-a (see Scheme 1) for the formation of the ONW is extendable to other Adv. Funct. Mater. 2024,34, 2403763 2403763 (2 of 13) © 2024 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH 16163028, 2024, 40, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adfm.202403763 by Readcube (Labtiva Inc.), Wiley Online Library on [17/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.afm-journal.de Scheme 1. Step-by-step procedure for the fabrication of H2Pc@MAPI nanowires and TiO2@MAPI nanotubes. Step 0) Thin layer deposition of TiO2 by Plasma Enhanced Chemical Vapor Deposition (PECVD), which serves as seeds for the next step. Step 1-a) Growth of H2Pc ONW by Physical Vapor Deposition (PVD). Step 1-b) TiO2thin film by PECVD deposited on top of the H2Pc ONW, followed by an air annealing at 350 °C to remove the H2Pc, thus remaining TiO2Nanotubes. The last step 2) consists of the co-evaporation of CH3NH3IandPbI 2to fabricate MAPI in a core@shell structure: If step 2) is performed on the nanostructures obtained at step 1-a) or step 1-b), the H2Pc@MAPI NW or TiO2@MAPI NT are obtained, respectively. Figure 1. Formation and microstructure of H2Pc@MAPI nanowires and TiO2@MAPI nanotubes. a,c–e) SEM micrographs of the H2Pc@MAPI NW (blue color): top-view a,c,e) and cross-section d). b,f–i) SEM micrographs of the TiO2@MAPI NW (red color): top-view b,f,h), cross-section g,i). The panel i) also includes a scheme of the TiO2@MAPI NTs. Adv. Funct. Mater. 2024,34, 2403763 2403763 (3 of 13) © 2024 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH 16163028, 2024, 40, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adfm.202403763 by Readcube (Labtiva Inc.), Wiley Online Library on [17/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.afm-journal.de Figure 2. Crystallinity and microstructure of core@shell H2Pc@MAPI NW and TiO2@MAPI NT. a) XRD patterns comparing the growth of 1D perovskitesupported nanostructures with the thin film counterpart. b,c) HAADF-STEM images showing an individual H2Pc@MAPI NW b) and a TiO2@MAPI NT c). Both figures (b,c) also include the scheme of the core@shell architecture. Panel f) shows the elements (Pb, I, Ti, and O) distributions along the arrow highlighted in c) obtained by energy-dispersive X-ray spectroscopy for the TiO2@MAPI NT. d,e) High-resolution TEM images of selected regions depicting lattice fringes compatible with the tetragonal and LDP crystalline phase for ONW@MAPI NW d) and DDP (inset)) and TiO2@MAPI NT e) and DDP (inset)). molecules such as porphyrins and perylenes,[19,20]all of them deploying at least one absorption band in the visible, which can interfere with the optoelectronic applications of the MAPI. To avoid interference of H2Pc with MAPI absorption bands, we included a second case of study using a fully transparent 1D template. Thus, in step 1-b (see Scheme 1), we generate a thin shell of nanoporous TiO2(nominal thickness ca. ≈5nm,see Figure S1, Supporting Information) formed at room temperature by plasma-enhanced chemical vapor deposition (PECVD) on the ONW and yielding a full conformal coverage. As previously reported, the plasma sheath electric field provides a higher verticality to 1D nanostructures, as shown in Figure 1g and Figure S1f (Supporting Information).[12–14,24]The post-annealing of the samples in air at 350 °C ensures the complete removal of the H2Pc organic core, forming transparent TiO2NT. Figure S4 (Supporting Information) gathers the evolution of the UV–vis transmittance spectra and the comparison between the blueish photograph characteristic of the H2Pc NW sample and the fully transparent view of the TiO2NT sample. At first sight, the formation of the MAPI shell on the TiO2nanotubes (Figure 1b,f–i)presents highsimilaritywiththeH2Pc@MAPI NW, but in high magnification (Figure 1h,i), the appearance of a conglomerate of particles can be observed, in some cases with angular shapes. It is important to mention that the thickness of the multi-shell structure can be easily tuned by controlling the deposition time for both TiO2and MAPI. Thus, it is possible to fabricate extremely high aspect ratio nanowires as those presented in the micrographs in Figure 1g,i. We have also analyzed the crystalline characteristics and microstructure of the H2Pc@MAPI NW and TiO2@MAPI NT (hereafter NW and NT) by XRD and TEM (Figure 2). XRD patterns in Figure 2a show the comparison between the NW, NT, and the equivalent thin film (TF), i.e., the layer grown on a bare fused silica substrate and Si wafer with an equivalent MAPI thickness (≈100 nm). It should be noted that both the 1D nanostructures (NW and NT) and the thin film were deposited simultaneously, and none of them show traces of the presence of PbI2 (which maximum peak should appear at 12.7°). The comparison between both systems is not straightforward. On the one hand, NW and thin film share the characteristic peaks at 14.1°, 28.4°, Adv. Funct. Mater. 2024,34, 2403763 2403763 (4 of 13) © 2024 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH 16163028, 2024, 40, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adfm.202403763 by Readcube (Labtiva Inc.), Wiley Online Library on [17/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.afm-journal.de and 43.2°, assigned to (110), (220), and (330) of the MAPI tetragonal crystalline phase. The thin film shows a higher surface texturization accounted for by the predominance of the (110) and (220) planes in contrast to the NW, which presents the expected diagram of a randomly distributed polycrystalline sample, with the presence of peaks related to a broader variety of planes. The highresolution TEM image of H2Pc@MAPI NW in Figure 2d displays the interplanar distances of 2.25, 2.79, and 3.15 Å compatible with (224), (310), and (220) of the tetragonal phase of MAPI perovskite. The Digital Diffraction Pattern (DDP) of the TEM image is shown in the inset of Figure 2d, where the presence of several crystals with different orientations can be easily noted. On the contrary, the XRD pattern corresponding to TiO2@MAPI NT presents a principal double peak in the range between 11°and 12.2°, which has been related by other authors with the formation of low-dimensional perovskites (LDP) in the crystalline meaning of the term.[17,18]The origin of such features at low diffraction angles is related to a higher interplanar distance (see d-spacing on the upper axis of Figure 2a) due to intercalated MA+cations. References [17]and [18]discussed the presence of the LDP as a consequence of an excess of MAI during the deposition when the sublimation temperature settled for the MAI is too high (>155 °C). However, in our case, the deposition conditions are the same for all the samples studied (please note that they have been deposited simultaneously). The High-Resolution TEM image of Figure 2e demonstrates a much higher interplanar distance of 7.6 Å, in agreement with the XRD measurements. The DDP pattern of the image also shows this increased planar distance (note that the DDP scale is the same in Figure 2d,e), with diffraction maxima close to the center. In addition to the XRD peaks in the range of 11°–12.2°, the low dimensional phase observed in TiO2@MAPI NTs also shows many additional minor peaks in the range of 15°–32°. In the reference,[25]Song et al. simulated the XRD of several LDP structures, that is, 0D, 1D, and 2D phases, all of them presented a complex X-ray diffractogram, and compared them with the XRD pattern of the TiO2@MAPI NTs. However, apart from the peaks associated with the 3D MAPI perovskite (with the only distinguishable peak at 14°), the diffractogram adjusts to the simplest low-dimensional structure, the 0D phase (0D-LDP), in which the (PbI6)4−units are as isolated quantum dots in an MAI matrix. Figure S5 (Supporting Information) shows the crystal lattice of this 0D-LDP from the reference,[25]as well as a detailed comparison of the X-ray diffractogram of the TiO2@MAPI NTs and the simulation of the 0D-LDP (including the assignment of every diffraction plane). The simulated diffractogram presents a highly similar XRD pattern to the one observed in the TiO2@MAPI NTs, with coincident peaks (in position) in the 11°–12.2°range and many minor peaks in the range of 15°–21°and 22°–29°. This can be easily observed in the zoom shown in Figure S5c (Supporting Information). Furthermore, the appearance of LDP is usually accompanied by an optical absorption band at 370 nm related to the groundstate exciton from the 0D MAPI perovskite structure.[25]This feature is also present in the TiO2@MAPI NT, as shown in Figure 3a (left, red curve highlighted with an arrow), corroborating the existence of LDP. It is worth mentioning that the appearance of this band in the H2Pc@MAPI NW cannot be easily observed since the B-band of the H2Pc appears at a similar wavelength (see Figure S4, Supporting Information).[26]Other authors have analyzed the growth of MAPI on TiO2(110) single crystals and also related the enrichment of MAI to the hydroxyl groups on the oxide surface.[27,28]However, in our case, the appearance of these peaks in the XRD for the case of the NT is linked to the formation of the LDP phase with a much higher abundance (as deduced from the higher relative intensity with respect to the main tetragonal (110) perovskite peak), indicating that the LDP phase cannot be restricted only to the surface (the equivalent thickness is 100 nm). An additional study on MAPI growth on different thin films and 1D surfaces is shown in Figure S6 (Supporting Information). First, the growth of MAPI on other surfaces, such as H2Pc TF and H2Pc@MAPI NW, does not produce the appearance of this LDP, which diffractogram depicts a very similar pattern to the growth of MAPI on glass or TiO2TF, but with a small amount of PbI2. Second, the growth on other TiO2surfaces, such as thin films or the H2Pc@TiO2 core@shell structure (the latter without annealing at 350 °Cto remove the organic phthalocyanine core), does not generate LDP, and only a minor peak related to this phase can be observed in the core@shell H2Pc@TiO2@MAPI structure. Note that both TiO2 TF and H2Pc@TiO2core@shell nanostructures (and hence the TiO2NTs) were deposited in the same batch by using flat substrates (such as Si or glass) and H2Pc NWs, respectively, so we do not expect any difference from the point of view of surface chemistry. Thus, the appearance of the LDP phase is linked to the hollow nanostructure, although a certain effect of the TiO2 surface chemistry cannot be excluded for several reasons: i) the preferential growth of the LDP is only fostered by empty TiO2NT, with the near disappearance of the tetragonal MAPI peak at ca. 14°. Our findings confirm that empty TiO2NTs promote higher adsorption of MAI molecules when they are deposited without PbI2. This is apparent from the samples’ whitish color, which we hypothesize serves as a reservoir providing MAI during perovskite deposition; ii) the growth of the LDP MAPI phase, albeit at a much lower intensity, in the H2Pc@TiO2core@shell structure; iii) the exclusive growth of the MAPI tetragonal phase on SiO2NTs (see Figure S6, Supporting Information). Thus, the formation of LDP requires both TiO2surface chemistry and nano structuration in the form of nanoshells. The crystallite sizes extracted from the XRD patterns show a smaller size for the 1D nanostructures, with mean crystallite sizes of 80, 35, and 30 nm for MAPI thin film, NT, and NW, respectively. Note that the crystalline size was calculated from the crystalline plane of tetragonal perovskite (110) for the three examples (MAPI thin film, NT, and NW). However, in the case of TiO2@MAPI NT, the tetragonal is not the majority phase. Thus, the accuracy of the crystallite is much lower because of the very weak intensity of the (110) peak. Unfortunately, the crystallite size cannot be extracted from the LDP peak since this phase comprises different components located at very similar diffraction angles (See Figure S5, Supporting Information). In any case, the calculated crystallite sizes are consistent with the SEM images of Figure 1, where grains with angular shapes and sizes below 50 nm can be observed for both 1D nanostructures. The TEM characterization provides additional detailed information on the perovskite shell microstructure. In good agreement with the SEM results, Figure 2b,c reveal a homogeneous distribution of perovskite crystals along the length and thickness of Adv. Funct. Mater. 2024,34, 2403763 2403763 (5 of 13) © 2024 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH 16163028, 2024, 40, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adfm.202403763 by Readcube (Labtiva Inc.), Wiley Online Library on [17/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.afm-journal.de Figure 3. Photoluminescence and waveguiding effect for core@shell MAPI nanowires and nanotubes. a) UV–vis absorptance and photoluminescence (𝜆exc =490 nm) spectra comparing the systems of MAPI thin films, H2Pc@MAPI NW and TiO2@MAPI NT for equivalent thickness of 100 nm; the insets show characteristic PL photographs of the three surfaces illuminated with a 𝜆=365 nm, and collecting the images using a longpass filter with a cut-on wavelength of 695 nm. Optical b,e) and PL c,f) (𝜆EX =460–490 nm, 𝜆EM>550 nm) images were acquired for H2Pc@MAPI NW b,c) and TiO2@MAPI NT e,f), respectively. The sequences in (e,f) are the corresponding microscope images at different sample-to-objective distances. d) Low magnification SEM top view image of the TiO2@MAPI NT. the 1D templates, which is also corroborated by the line-scan energy-dispersive X-ray spectroscopy analysis shown in Figure 2 f for the TiO2@MAPI NT (line scan performed at the location highlighted in c). These results demonstrate the homogeneous distribution of Pb and I of the MAPI in the outer shell of the nanostructure, while Ti and O are located in an inner shell (Figure S7, Supporting Information shows the linescan profiles ofbothH2Pc@MAPI NW and TiO2@MAPI NT, for comparison). Such homogeneity in composition contrasts with the results reported in the literature for the synthesis of 1D HP nanostructures by catalytic vapor-solid (VS) and vapor–liquid–solid (VLS) approaches, usually presenting an accumulation of the metal particles at the top or the bottom of the nanowires.[29] A significant advantage of our soft-template method is the direct fabrication of supported nanowires and nanotubes on processable substrates such as FTO and ITO/Glass[14,15]or flexible ITO/PET.[16]Here, we synthesized the NW and NT on a highly transparent fused silica substrate with a thin layer of TiO2working as a seed interface, allowing for the on-support characterization of their optical properties. Thus, the optical (UV–vis spectra) and photoluminescence (PL) emission of the core@shell MAPI nanostructures are shown in Figure 3(see also Figure S8, Supporting Information for the complete optical study) and have been analyzed in comparison with the thin-film counterpart. Figure 3a) displays the optical absorptance in the UV–vis for the ranges 350–600 nm (left) and 675–850 nm (right), this latter range also includes the normalized PL. The absorptance spectra of both 1D nanostructures (H2Pc@MAPI NW and TiO2@MAPI NT) and thin film show the typical bandgap onset at ∼780 nm, proper of the MAPI perovskite.[30]Moreover, the Tauc-plots shown in Figure S9 (Supporting Information) for the three types of samples show a similar band gap (1.62, 1.62, and 1.61 eV for thin film, H2Pc@MAPI NW and TiO2@MAPI NT, respectively) as expected for the bulk MAPI perovskite.[30]The fact that the bandgap is slightly lower for the TiO2@MAPI NT could also be indicative of the presence of the LDP MAI-rich crystalline phase. One important aspect extracted from the optical study is the enhanced absorptance of the 1D nanostructures. In the case of H2Pc@MAPI NW, this effect can be related to the presence of the Adv. Funct. Mater. 2024,34, 2403763 2403763 (6 of 13) © 2024 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH 16163028, 2024, 40, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adfm.202403763 by Readcube (Labtiva Inc.), Wiley Online Library on [17/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.afm-journal.de H2Pc, which have strong and wide absorption bands, the B- and Q- bands at 370 and 590 nm, respectively, also shown in Figure S4 (Supporting Information). However, in the TiO2@MAPI NT case, it can only be understood due to the presence of 1D nanostructures. This significantly enhanced absorptance with respect to the MAPI thin film counterpart in the visible range (≈five times higher in the 550–780 nm range), has been previously ascribed to light-trapping effects due to the nanostructuration of MAPI.[31] The enhanced absorption of light also has an impact on the PL properties. The different PL intensities can be observed in the insets of Figure 3a) (also in the non-normalized PL spectra in Figure S10, Supporting Information), which correspond to the IR photographs of the samples illuminated with a 𝜆= 365 nm and collecting the images using a long-pass filter with a cut-on wavelength of 695 nm. An enhanced PL emission of the TiO2@MAPI NT is obtained with respect to the thin film for the same illumination intensity and camera acquisition parameters. We have measured a very similar PL quantum yield (PLQY) for the three samples of 0.012, 0.007, and 0.018% for the thin film, NW and NT, respectively. This small difference between the PLQY indicates that the enhanced PL intensity is mainly due to the increased absorptance in the TiO2@MAPI NT system. In the case of the H2Pc@MAPI NW, the PLQY was even lower, which also suggests that a possible quenching effect can be occurring at the H2Pc/MAPI interface (also supported by the lower PL emission intensity shown in Figure S10, Supporting Information). However, the PLQY values are very low and thus are not fully reliable. To shed light on the PL process, we have characterized the nanostructures employing time-resolved PL. The lifetime curves and the fitting to a biexponential decay of the core@shell nanostructures are shown in Figure S11 (Supporting Information). In halide perovskites, a bi-exponential fit is usually performed, in which the two lifetimes, one short and the other long, are associated with trapmediated recombination (𝜏1) and radiative recombination (𝜏2), respectively.[32]However, comparing the perovskite isolated layer (MAPI thin film) with the one in contact with a charge extraction layer (TiO2or H2Pc) is not straightforward since the charge accumulation produced by the electron or hole injection affects the initial carrier concentration, as reported by other authors.[32] In addition, the radiative recombination rate (𝜏2) can also be affected by other factors, such as stoichiometry and crystalline characteristics (especially relevant for the TiO2@MAPI NTs), which have been proven in the manuscript. For these reasons, only a study of the average lifetime will be done here. The reference MAPI thin film deposited on glass presents an associated average lifetime of 〈𝜏〉=19.8 ns. This short lifetime compared to other MAPI systems is due to the small size of the crystal grains, which increases the number of trap states at the grain boundaries.[33]In the case of the core@shell nanostructures, it can be noted a drastic reduction of the average lifetime with respect to the reference (〈𝜏〉TF =19.8 ns)to〈𝜏〉TiO2@MAPI =10.1 ns and 〈𝜏〉H2Pc@MAPI =7.3 ns. The much lower average lifetime of the H2Pc@MAPI NWs can be understood as a more efficient charge injection to the phthalocyanine core[34]with respect to the TiO2@MAPI NTs, for which a very efficient electron extraction is not expected due to that TiO2tube is very thin and amorphous. In addition, the PL spectrum is slightly blue-shifted, from 769 nm for the thin film to 765 and 762 nm for the H2Pc@MAPI NW and TiO2@MAPI NT, respectively. Similar blue shifts have been reported for ultrathin vacuum-deposited MAPI layers due to quantum confinement.[35]On the one hand, the MAPI thin film appears to be blue-shifted from the expected PL emission wavelength for MAPI bulk (ca. 785 nm), which has been reported for thin layers of MAPI perovskites due to quantum confinement.[35] On the other hand, the PL blue shift observed for the NT and NW is also in good agreement with their smaller crystallite size. In the reference,[35]Laura Hertz and co-workers observed a blue shift in the MAPI PL to 770 nm for thin layers of 42 nm thickness. We have found a similar shift to762 and 765 nm for the TiO2@MAPI NT and H2Pc@MAPI NW, respectively (with a mean crystallite size of 30 and 35 nm, respectively). The minor inconsistency in the TiO2@MAPI NT maximum PL wavelength (which is slightly blue-shifted with respect to the H2Pc@MAPI NW) concerning the crystallite diameter calculated from XRD can be due to several factors: i) higher reabsorption for H2Pc@MAPI NWs samples due to the H2Pc core (see the higher absorptance close to the PL Emission wavelength in Figure 3a), which is reported to induce a red-shift in the PL for emitters with a low Stokes shift (small difference between emission and absorption maxima)[36] and ii) a low accurate calculation of the XRD signal width due to the low intensity of the (110) tetragonal MAPI peak. When the samples were exposed to a 490 nm light source, the emission was higher for the TiO2@MAPI NT, followed by the MAPI thin films, and lastly, by the H2Pc@MAPI NW. It is wellestablished that the presence of a scattering layer increases PL emission by enhancing light absorption, light trapping, waveguiding, and reducing reflection. In the case of TiO2@MAPI NT, not only an enhanced PL emission is observed at perpendicular, but also at oblique angles. This effect is shown in Figure S12 (Supporting Information), where while the PL of a MAPI thin film at 25°drops to ca. 50% (with respect to the perpendicular PL), the TiO2@MAPI NT maintains almost the same PL intensity, with a value above 99%. At more grazing angles (65º), this effect is even more pronounced, producing a drastic decrease of the PL of the thin film to 3%, while TiO2@MAPI NT is above 40%. In the case of the H2Pc@MAPI NW, the samples present a strong reduction of the photoluminescence emission. Two main factors can account for such a feature; on one side, the absorption of light by the H2Pc competes with the MAPI. On the other side, the quenching of the PL can also be related to efficient charge injection at the interface between the H2Pc single crystalline nanowire and the MAPI. In this regard, numerous authors have previously reported on the benefits of using porphyrins and phthalocyanines as HTL and ETL.[37,38,34,39]Specifically, when dealing with evaporated HP, organic and polymeric interfaces have been demonstrated to outperform the use of inorganic ETLs such as TiO2.[27,40] Long carrier lifetimes (fostered by reduced charge carrier trapping), long diffusion lengths, and low non-radiative recombination rates have made halide perovskite nanowires ideal materials for waveguiding and lasing.[41]Most of the examples of waveguiding effects in the literature develop around the exploitation of single crystalline HP nanowires, in which each NW serves as a waveguide, and the two end facets form a Fabry–Perot cavity for optical amplification.[41–44]In Figure 3b–f, we explore the Adv. Funct. Mater. 2024,34, 2403763 2403763 (7 of 13) © 2024 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH 16163028, 2024, 40, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adfm.202403763 by Readcube (Labtiva Inc.), Wiley Online Library on [17/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.afm-journal.de effect of the 1D arrangement of the polycrystalline MAPI on such a nanophotonic property. Figure 3b,c,e,f compares the optical (b,e) and PL (c,f) images for H2Pc@MAPI NW and TiO2@MAPI NT, correspondently. The PL images show that the light is predominantly emitted from the tips of the NW and NT. Indeed, in the case of the TiO2@MAPI NT, the vertical alignment of the nanostructures (see Figure 3d and 1g) allows us to observe a plane-by-plane enhanced emission from the tips. From Figure 1g, it can be noted that although most of the tubes have a length in the 2–3 um range, there is a significant dispersion, finding vertical tubes as long as 8 μm. Thus, Figure 3e,f showcases the light emitted by the NT of different lengths, in which every image corresponds to a z-step of ca. 2 μm (note that the 100x objective used with NA =0.9, has a depth of field of ≈1μm), with the top one corresponding to the highest separation (longer tubes). In a previous article, we reported on the waveguiding effect on perylene@TiO2 and perylene@ZnO nanowires enhanced by the high contrast between the refractive index of the metal oxide shells and the luminescence core.[24]In Figure 3, the waveguiding effect is induced by the high refractive index of the MAPI perovskite (n =2.55 @ 770 nm)[45]with respect to air. The TiO2@MAPI NT presents three optical interfaces: air-MAPI, MAPI-TiO2,andTiO 2-air. The critical angle at the air-MAPI interface is 23°(see Supporting Information Section S1). The incident light with an angle of 23° reaching the MAPI-TiO2interface will change the angle in the TiO2media but will also produce the total internal reflection at the TiO2-air interface. For the H2Pc@MAPI NW, the critical angles for light propagation by total internal reflection for the two interfaces, air-MAPI, and MAPI-H2Pc, are 23°and 46°,respectively. The incident light at the MAPI-H2Pc interface with angles in the range of 23°–46°, will go through the phthalocyanine core, refracted again to the MAPI layer, and will also be confined in the waveguide. Thus, the PL of the MAPI emitted at angles below 23°will escape from the waveguide, whereas the PL emitted at angles above 46°, will be confined at the MAPI shell. However, while the PL emitted in the range of 23°–46°will also be confined inside the nanostructure, it will pass many times through the H2Pc core. As the H2Pc at 770 nm is not transparent, the phthalocyanine core can lead to a decrease in the overall PL of the nanostructures, as was observed in Figure 3. However, as detailed above, the lower average lifetime of the H2Pc@MAPI NWs is ascribed to a more efficient charge injection to the phthalocyanine core,[34]which also decreases the PL intensity. Thus, it can be concluded that both effects should occur in the H2Pc@MAPI NWs, leading to a drastic decrease in the PL. Therefore, a more efficient waveguide is expected for TiO2@MAPI NT, since the losses at the TiO2are negligible. In addition, TiO2will act as internal cladding, minimizing the losses, and all the PL emitted at angles above 23°will be confined inside the nanostructure. In addition, the low value of the waveguide parameters for both nanostructures indicates that they behave as single-mode waveguides (see Supporting Information Section S1). This low V parameter value makes these waveguides sensitive to bend-associated losses as well as absorption losses at the cladding, which in our case is only at the internal core. Thus, although the waveguide effect is visible in Figure 3b,c,e,f,wedo not expect the core@shell nanostructures with the current architecture to be very efficient for their use as waveguides. However, our methodology allows fine control of the internal and external cladding media, as well as the thickness of every layer (including the perovskite). Thus, our approach presents a very high potential for developing efficient waveguides by using nanowires and nanotubes as templates. Further details about the waveguiding effect are discussed in Supporting Information Section S1 (Supporting Information). As it is well known, a critical drawback of using organometal halide perovskites as photoactive materials is their poor stability under room conditions (exposition to air, moisture, oxygen, and light). Specifically, the reaction between oxygen molecules and methylammonium cations in the presence of humidity produces the hydrolyzation of the MAPI back to the original PbI2 precursor. Previous articles have reported on the higher stability of 1D perovskite nanostructures in comparison with their thin film counterparts.[46]Thus, for instance, Z. Fan et al.presented the template fabrication of CH3NH3SnI3NW on anodic aluminum oxide with 840 times slower PL decreasing than the planar thin film.[29]However, this significant improvement in stability was attributed to the effective blockage of the diffusion of water and oxygen molecules due to the inorganic template. Fu et al. showed that FAPbI3nanowires with diameters in the order of 10 nm present a significantly enhanced photostability in comparison to higher-diameter wires or thin films.[47]Many authors have related this stability to the enhanced crystallinity of smaller clusters and reduced presence of grain boundaries.[48]Other authors have also considered the higher water contact angle of microstructured and functionalized HP arrangements as a positive factor to enhance stability.[49]With these antecedents in mind, we analyzed the PL stability of NW, NT, and TF (Figure 4). Under room air conditions (T =25 °C, RH =50%–60%) and continuous illumination with monochromatic light at 𝜆=490 nm and with a power of 0.35 mW/cm2, the evolution of the PL emission follows a similar trend for the three samples but with very different time characteristic. Note that Figure 4a left and middle curves are normalized to the maximum and initial PL values, respectively. Thus, the emission intensity increases at the beginning of the illumination to then decay to values lower than 20% of the initial (middle curves in Figure 4a) after only 1.2 h for the thin film, 5.2 h for the NW, but reaching more than 20 h for the NT (note the logarithmic scale for the time axis). The initial increase in the PL intensity induced by illumination has been previously reported. However, the explanation for such an increase is not unique. Some authors assign it to a light-induced lattice expansion of the perovskite,[50]surface selfpassivation of halide vacancies,[51]or surface chemical modifications in the presence of oxygen or moisture.[52]In our case, the increase is always present for the three samples studied, and the most significant change with respect to the original value occurs for the thin film (220%) and is slightly lower for the 1D MAPI nanostructures (185% and 175% for NW and NT, respectively). Nevertheless, the time needed to reach this maximum is very different: in the thin film case, this time is extremely low and occurs after 15 min, while for the NW and NT, it is 2.3 and 6 h, respectively. These results clearly show an enhanced photo-stability for the 1D nanostructures, especially for the TiO2@MAPI NT. Moreover, considering that the absorptance at 490 nm is much higher for the NT than for the reference thin film (77% vs 27%), this enhanced photo-stability is even higher since the total energy absorbed by the NT is almost three times higher than for the TF. Adv. Funct. Mater. 2024,34, 2403763 2403763 (8 of 13) © 2024 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH 16163028, 2024, 40, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adfm.202403763 by Readcube (Labtiva Inc.), Wiley Online Library on [17/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.afm-journal.de Figure 4. Photo-stability of MAPI thin film, nanowires, and nanotubes. a) Left and middle. Stability test of 1D MAPI nanostructures and thin films in air atmosphere and illuminated with monochromatic light (𝜆EX =490 nm, with a power of 0.35 mW cm−2), monitoring the intensity at the maximum emission wavelength. The time evolution curves have been normalized to the maximum (left) and to the initial (middle) PL intensity value. Note that the time axis in the middle graph is on a logarithmic scale. Right. IR photographs of the samples before (as grown) and after the stability test comparing nanostructured samples with equivalent thin films. b) Hard stressing stability test of TiO2@MAPI NT sample encapsulated in PMMA illuminated with monochromatic light (𝜆EX =490 nm), at intervals of 40 h sequentially exposed to N2, air (room conditions), N2+100% RH, and air+100% RH atmospheres (24 h relaxing period in inert atmosphere between cycles). The inset shows the fast decay on PL of the reference MAPI thin film exposed to the lab environment compared to the nanostructured sample (both encapsulated in PMMA). The PL pictures (taken in the same way as in the inset of Figure 3a), are shown in Figure 4a, right before (As grown) and after 5 h (for NW) and 20 h (for NT) of illumination time (after the PL evolution characterization). First, the horizontal elongated darker areas visible in the NW and NT cases after 5 and 20 h, respectively, show a total degradation of the perovskite luminescence. Although this elongated darker area is more prominent for the reference thin film than for the NW and much bigger than the NT case, it is worth mentioning that the excitation monochromator slits have remained the same for all the PL evolution experiences. The reason for this bigger degraded area is also related to the enhanced stability of the 1D nanostructures: the illumination area presents a certain Gaussian profile in intensity, thus provoking that close to the edges, even if the illumination intensity is much lower, is sufficient to induce the full degradation of the reference MAPI thin film. By contrast, for the NT sample, the lower intensity at the edges of the illumination area has not been enough to degrade the nanostructures, but to enhance the PL. This is visible by an increased PL intensity at the borders of the illuminated area for the NT, and to a lesser extent in the NW samples. In fact, the XRD analysis of the TiO2@MAPI NT system after several weeks stored in the lab at a relative humidity of ca. 40–60% only suffers modifications in the relative intensity between the LDP and (110) tetragonal perovskite peaks, without the presence of the PbI2characteristic peaks (see Figure S14, Supporting Information). This is a very striking result since the aging only produces the transformation of the LDP to the MAPI tetragonal structure, without producing any degradation to the PbI2, as it happens, for example, in the thin film aged for a similar time at the same storage conditions (see Figure S14, Supporting Information). This feature also demonstrates the higher stability of the TiO2@MAPI NT concerning the thin film counterpart. The encapsulation of hybrid metal halide perovskite solar cells and other optoelectronic devices with different polymeric solutions has been previously demonstrated to improve the stability of these devices under room environmental conditions.[53]The encapsulation diminishes the interaction with the atmosphere and allows the separation of the intrinsic stability of the perovskite with respect to the chemical reactions (which are usually accelerated by the continuous illumination) of the MAPI with the moisture and/or oxygen. To study the enhanced intrinsic stability of the TiO2@MAPI NT, the sample was encapsulated in PMMA (see Experimental section). Figure 4b shows the PL evolution during intervals of 40 h of continuous illumination under different atmospheres, that is, i) pure N2, ii) air (room conditions), iii) moisture-saturated N2(using a water bubbler), and iv) moisture saturated air. In the first 40 h of the experiment under an N2atmosphere, a continuous increase of the PL with the illumination time can be noted, which reaches ≈200% of the initial value. The effect of an inert atmosphere on the PL under continuous illumination is reported to induce either a photobrightening or photodarkening effect.[54]The explanation for this variability was assessed in the work of Hong et al.,[55]in which the simultaneous presence of brightening and darkening traps was identified as responsible for these opposing behaviors, which were assigned (supported by theoretical calculations) to Ii+1and Pbi2+as responsible for PL enhancement and decline, respectively. However, this assignment is still under debate since many authors suggest that the overall PL modification is a complex interaction of ion migration, defect formation and annihilation, and other light-induced processes that govern the behavior of PL evolution under an inert atmosphere or vacuum. Nevertheless, all the reported PL evolution under nitrogen or vacuum environments shows much faster kinetics, with the enhancement of PL in the range from seconds to several minutes. The much slower PL enhancement in a nitrogen environment for our TiO2@MAPI NTs embedded in PMMA, which is continuously increasing during more than 40 h (see Figure S15 (Supporting Information) for additional PL measurements of the encapsulated TiO2@MAPI NT during 57 h under N2environment), indicates an intrinsic different behavior, which can be tentatively assigned to the simultaneous presence of the MAI-rich LDP and tetragonal MAPI crystals. Adv. Funct. Mater. 2024,34, 2403763 2403763 (9 of 13) © 2024 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH 16163028, 2024, 40, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adfm.202403763 by Readcube (Labtiva Inc.), Wiley Online Library on [17/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License