Supporting Information Spatially Resolved Multi-Color CsPbX 3 Nanowire Heterojunctions via Anion Exchange Letian Dou, Minliang Lai, Christopher S. Kley, Yiming Yang, Connor G. Bischak, Dandan Zhang, Samuel W. Eaton, Naomi S. Ginsberg, Peidong Yang Dou et al.: 10.1073/pnas. 201703860 Additional information for Materials and Methods Synthesis of the CsPbBr 3 nanowires and plates. Reagents: Unless otherwise stated, all of the chemicals were purchased from Sigma- Aldrich Chemical and used as received. Synthesis of CsPbBr 3 nanowires (1): To grow CsPbBr 3 nanowires and nanoplates, 460 mg PbI 2 (99.999%) was dissolved in 1 ml anhydrous dimethylformide (DMF) and stirred at 70 C overnight before further use. The PbI 2 solution was spun onto a PEDOT:PSScoated glass substrates at 1000 rpm for 120 s, then annealed at 100 C for 15 min. The PbI 2 film was carefully submerged into a glass vial with 8 mg/ml CsBr (99.999%) solution in methanol (anhydrous, 99.8%), with the PbI 2 side facing up. The capped reaction vial was heated at 50 C for 12 h, and the substrate was removed and washed twice in anhydrous isopropanol (each time for 30 s). The sample was then dried by heating to 50 C for 5 min. Preparation of oleylammonium halide: Oleylammonium chloride and iodide were prepared via a reported approach developed by Nedelcu (2). Briefly, 100 ml ethanol and 0.038mol oleylamine were added into a 3-neck flask with vigorous stir. The reaction
mixture was cooled by an ice-water batch, and certain amount of HX (X = Cl, Br, I) (0.076 mol, HCl >37%; HBr 48%; HI 57%) was added. The reaction was kept under N 2 flow overnight. Then the solvent was evaporated under vacuum. The product was rinsed with diethylether for 3 times. The product was then dried in a vacuum oven overnight at room temperature. Anion exchange of CsPbBr 3 nanowires: To convert the CsPbBr 3 to CsPbCl 3, 10 mg of oleylammonium chloride was dissolved in 10 mg of 1-Octadecene (ODE) to make the conversion solution. Individual CsPbBr 3 nanowires were transferred onto a clean Si/SiO 2 substrate using a nano manipulator. The chip with nanowires was immersed into the conversion solution at room temperature for 16 hours for complete conversion. Then, the chip was taken out from the solution and washed with chlorobenzene twice and hexanes once to remove the extra salts left on the chip. The reaction dynamics can be tracked in situ by monitoring the photoluminescence emission of an individual nanowire (see characterization section and Figure S2 below). To convert the CsPbBr 3 to CsPbI 3, 10 mg of oleylammonium iodide was dissolved in 10 mg of ODE to make the conversion solution and the reaction was carried out at room temperature for 4 hours. Fabrication of the nanowire heterojunctions. To fabricate heterojunction devices, the as-grown nanowires were first transferred onto a 300 nm SiO 2 coated Si substrate by a micromanipulator. The substrate was then spincoated with PMMA, and baked at 130 C for 6 minutes. Electron beam lithography (EBL) was performed in a Crestec CABL-9510CC High Resolution Electron Beam Lithography System with acceleration voltage of 50 kev and beam current of 500 pa. After EBL, the substrate was dipped into developer (MIBK:IPA = 1:3) for 60 s followed by washing in IPA for another 20 s. The developer and IPA were dried with molecular sieves (Sigma- Aldrich) to remove water molecules. Subsequently, the NWs with partially coated PMMA were immersed into the conversion solution. The conversion reaction was carried out using the same recipe as described above. After conversion, the PMMA mask was removed by dipping the substrate into chlorobenzene and hexanes, respectively.
Note: although different ways of carrying out the anion exchange reaction have been reported (3-6), none of them are applicable here. For example, more polar solvents such as toluene, chlorobenzene, or isopropanol can either dissolve the PMMA mask and/or damage the nanowire. Even with ODE, higher temperature will also damage the nanowire morphology. For vapor phase conversion, heating at over 100 C is necessary to evaporate the organic halide precursors, which may also damage the PMMA layer. Therefore, a non-polar solvent, such as ODE, is critical. To dissolve the halide precursor in such non-polar solvent, an ammonium cation with long carbon chain is necessary. Characterization of the perovskite heterojunctions. Optical microscopy (OM) measurements: OM images of the thin sheets were taken using a Zeiss Axio Scope.A1 in bright field mode. Steady state photoluminescence (PL) measurements: Photoluminescence measurements were performed using a 325 nm He:Cd laser. The full intensity of the beam is 5 mw. Using an iris diaphragm and a focusing lens, a Gaussian beam spot with a waist of approximately 20 µm was obtained and used to excite individual 2D sheets. The output power of the excitation source was adjusted by a series of neutral density filters (normally 1 ~ 100 µw) and monitored with an external energy meter. Emission from nanowires was collected with a bright-field microscope objective (Nikon 50, N.A. 0.55, in a Nikon ME600 optical microscope) and routed via a bundled optical fiber to a UV-vis spectroscopy spectrometer (Princeton Instruments/Acton) equipped with a 1200 groove/mm grating blazed at 300 nm and a liquid N 2 -cooled charge-coupled device. During the measurement, there was no obvious sample damage by laser during the measurement. Atomic force microscopy (AFM) and scanning Kelvin probe force microscopy (KPFM) measurements: AFM and scanning KPFM measurements were performed on an AFM
system (MFP-3D Asylum Research, Oxford Instruments) equipped with an acoustic isolation chamber (AEK 2002). The samples were transferred from the preparation glovebox into the AFM chamber operated in nitrogen atmosphere. The AFM/KPFM measurements were performed at room temperature employing conductive platinum coated silicon cantilevers (Olympus AC240TM-R3). KPFM measurements were performed in a two-pass mode: the first scan for topographical imaging in AC mode followed by the interleave mode in which the conductive tip was lifted with constant separation relative to the specimen surface while acquiring the contact potential difference V CPD (7). To achieve highest lateral resolution, the influence of the lift height on the measured surface potential (8) was carefully checked by surface potential vs z spectroscopy, while only a minor influence of the lift height on the measured CPD signals was observed in the delta lift height region of -10 to 30 nm. Accordingly, for all surface potential measurements the delta lift height was set to result in a tip apex - surface distance of approximately 20 nm. An AC bias of 3 V amplitude at the first contact resonance frequency and a DC bias of 1 V was applied to the conductive probe. All measurements were performed at low scan rate of 0.2 Hz at a resolution of 256-by-256 pixel. During measurements, ambient light is minimized by the AFM isolation chamber, while a low-intensity ~ 5 mw infrared (IR) diode (860 nm) is used to detect probe deflection. In order to quantitatively determine the work function value Φ sample of the investigated sample (Fig. S8A), reference measurements on freshly cleaved highly ordered pyrolytic graphite (HOPG) substrates (Fig. S8B) with known workfunction of Φ HOPG = - 4.5 ev in air (HOPG, grade ZYA, SPI supplies) (9) were performed for each tip prior to the measurement. Fig. S8C shows a representative KPFM surface potential map obtained on HOPG with the corresponding surface potential distribution shown in Fig. S8D. The freshly cleaved HOPG crystal showed nearly identical surface potential values over its whole surface area. For tip calibration, HOPG surface potential values V CPD HOPG were obtained by averaging four potential values determined by Gaussian fitting measured at CPD different HOPG surface spots. With Φ tip - Φ HOPG = eδv HOPG, and exemplarily taking the measured value of V CPD HOPG = 600 mv, the tip workfunction can be assigned to Φ tip =
5.1 ev. With the sample signal V sample CPD and the reference signal V HOPG CPD, the absolute workfunction values can be calculated by (9) (Fig S8A) Φ sample = Φ HOPG + e(v sample CPD - V HOPG CPD ). We performed KPFM measurements on various pristine as-synthesized CsPbBr 3, chemically converted CsPbCl 3, and as-synthesized CsPbCl 3 nanowires. A representative set of topographic AFM images are shown in Fig. S9A, Fig. S9C, and FigS9E, respectively, with the corresponding KPFM surface potential maps shown in Fig. S9B, Fig. S9D, and Fig. S9F; respectively. The surface potential distributions, each measured on the corresponding nanowire, are shown in Fig. S9F for the as-synthesized CsPbBr 3 (bottom), converted CsPbCl 3 (middle), and as-synthesized CsPbCl 3 (top) nanowires. The corresponding workfunction values are determined by averaging the Gaussian fitted mean surface potential values of five different nanowires for each sample type (Fig. S9G). The as-synthesized CsPbBr 3 and converted CsPbCl 3 nanowires feature clearly different workfunction values of 4.771 ± 0.034 ev and 4.928 ± 0.029 ev, respectively. In addition, the workfunction of the CsPbCl 3 nanowires obtained via anion exchange closely resembles the workfunction value of 4.993 ± 0.033 ev obtained for the as-synthesized CsPbCl 3. Scanning electron microscopy measurements (SEM): SEM images were obtained using a JEOL JSM-6340F field emission SEM. To acquire SEM images of nanowires on Si/SiO 2 substrates, the substrates were sputter coated with 3-nm gold using a Denton Vacuum Desk IV sputtering system. SEM energy-dispersive X-ray spectroscopy (EDS): EDS spectra and images were obtained using a Zeiss Ultra-55 field emission SEM equipped with an EDAX EDS detector using a beam energy of 10 kv. Confocal PL mapping: Confocal PL mapping was performed using an Olympus IX83 laser scanning confocal microscope with a 40x 0.95 NA objective and a 405 nm laser excitation source. All images were 512 x 512 pixels collected at 10 ms/line. CsPbCl 3,
CsPbBr 3, and CsPbI 3 emission were collected over a spectra range of 410-450 nm, 500-550 nm, and 580-640 nm, respectively, using an emission grating. Multichannel images were collected in series or simultaneously using either a 473 nm or 560 nm dichroic. Lambda scans were performed by collecting a series of images while scanning the emission grating in 10 nm spectral windows to vary the collected wavelength range. References: 1. Eaton SW, et al. (2016) Lasing in robust cesium lead halide perovskite nanowires. Proc. Natl. Acad. Sci.,, 113, 1993-1998. 2. Nedelcu G, et al. (2015) Fast anion-exchange in highly luminescent nanocrystals of cesium lead halide perovskites (CsPbX3, X= Cl, Br, I). Nano Lett., 15, 5635-5640. 3. Akkerman QA, et al. (2015) Tuning the optical properties of cesium lead halide perovskite nanocrystals by anion exchange reactions. J. Am. Chem. Soc., 137, 10276-10281. 4. Nedelcu G, et al. (2015) Fast anion-exchange in highly luminescent nanocrystals of cesium lead halide perovskites (CsPbX3, X= Cl, Br, I). Nano Lett., 15, 5635-5640. 5. Wong AB, et al. (2015) Growth and anion exchange conversion of CH3NH3PbX3 nanorod arrays for light-emitting diodes. Nano Lett., 15, 5519-5524. 6. Pellet N, Teuscher JL, Maier J, Grätzel M, (2015) Transforming Hybrid Organic Inorganic Perovskites by Rapid Halide Exchange. Chem. Mater., 27, 2181-2188. 7. Jacobs H, Knapp H, Stemmer A, (1999) Practical aspects of Kelvin probe force microscopy. Rev. Sci. Instrum., 70, 1756-1760. 8. Liscio A, Palermo V, Müllen K, Samorì P, (2008) Tip sample interactions in Kelvin probe force microscopy: quantitative measurement of the local surface potential. J. Phys. Chem. C, 112, 17368-17377. 9. Hansen WN, Hansen GJ, (2001) Standard reference surfaces for work function measurements in air. Surf. Sci., 481, 172-184.
Fig. S1. SEM image of the as-synthesized CsPbBr 3 nanowires and nanoplates.
Fig. S2. Anion exchange dynamics from CsPbBr 3 to CsPbCl 3 on a whole single nanowire without any coating. (A) PL emission spectrum evolution with time. (B) Emission energy evolution with time. It is noted that the PL intensity of the wire drops gradually and then increase to the initial state as the reaction time increases (chloride content increases), indicating the PL quantum efficiency is lower for the mixed halides, compared to the purer bromide or chloride perovskites. This is consistent with recent observations made by Nedelcu et al. in during the anion exchange in quantum dots (4). The peaks are also broad in the middle region, suggesting that there is certain inhomogeneity within the entire wire.
Fig. S3. A schematic illustration showing the nanowire heterojunction fabrication process.
Fig. S4. SEM EDS spectrum of a CsPbBr 3 -CsPbCl 3 nanowire heterojunction. Because Pb M shell and Cl K shell have similar energy, the elemental mapping for the Cl may contain some contribution from Pb. As a result, the CsPbBr 3 part of the heterojunction nanowire shows slight Cl signal (Fig. 2). We believe this is not due to the diffusion of Cl atoms into the CsPbBr 3 part because we do not observe a continuous concentration gradient along the nanowire.
Fig. S5. SEM image of the CsPbBr 3 -CsPbCl 3 nanowire heterojunction. (A) The entire nanowire. (B) The end of the converted section showing that the nanowire after conversion maintains smooth surface.
Fig. S6. AFM topographical imaging of CsPbBr 3 -CsPbCl 3 heterojunction nanowires. (A) 2D view (top) and corresponding height profile (bottom) of the single-junction nanowire shown in Fig. 3. The scale bar is 2 µm. (B) 2D view (top) and corresponding height profile (bottom) of the multi-junction nanowire featuring four heterojunctions as shown in Fig. 3. The scale bar is 2 µm.
Fig. S7. Additional CsPbBr 3 -CsPbCl 3 heterojunction nanowire on SiO 2 /Si. (A) AFM topography image and (B) corresponding KPFM surface potential map. The CsPbBr 3 part shows, consistently over many heterojunction nanowires, higher surface potential values (red) compared to the CsPbCl 3 part (green). The scale bars are 2 µm (A,B).
Fig. S8. Tip calibration for quantitative KPFM measurements. (A) Schematic representation of the relevant energy levels for the performed KPFM measurements. (B) 2D AFM topographic image of a freshly cleaved HOPG surface employed for tip calibration prior to quantitative KPFM measurements. The scale bar is 20 µm. (C) Corresponding KPFM surface potential map. The scale bar is 20 µm. (D) Statistical distribution of the surface potential map (C) with overlayed Gaussian fit.
Fig. S9. Workfunction determination for pristine CsPbBr 3 and CsPbCl 3 nanowires. AFM topography images of pristine (A) as-synthesized CsPbBr 3, (C) converted CsPbCl 3, and (E) as-synthesized CsPbCl 3 nanowires. Corresponding KPFM surface potential maps of (B) as-synthesized CsPbBr 3, (D) converted CsPbCl 3, and (F) as-synthesized CsPbCl 3 nanowires. (G) Surface potential distributions measured on the as-synthesized CsPbBr 3 (bottom), converted CsPbCl 3 (middle) and as-synthesized CsPbCl 3 (top) nanowire. (H) Absolute workfunction values determined by averaging the Gaussian fitted mean surface potential values of five different nanowires for each sample type. The scale bar are 2 µm (A-D) and 3 µm (E,F), respectively.
Fig. S10. Confocal PL mapping on a single nanowire heterojunction. (A) The nanowire before conversion. The exposed part is 10.8 µm long and the covered part is 8.7 µm long. (B) The spatially resolved PL mapping of the nanowire after conversion. Blue color represents emission from 410 to 450 nm. Green color represents emission from 500 to 550 nm. After conversion, the blue part is 10.8 µm long and the green part is 8.6 µm long. This result indicates that the covered area is well protected and the anion exchange reaction is highly localized to the exposed area.
Fig. S11. Confocal PL mapping of a CsPbBr 3 -CsPbCl 3 nanowire heterojunction at different wavelength range.
Fig. S12. The periodic hetero structure with 1 µm pixel size characterized using conventional optical microscope under laser excitation. The small periodic feature cannot be fully resolved due to interference of the light.
Fig. S13. The hetero structure with 200 nm pixel size. (A) Optical image of the nanowire before conversion. The nanowire is coated with a layer of PMMA. The vertical lines are the openings created by e-beam lithography. The width of the openings is 200 nm. The spacing between the openings is 200 nm in the left part and is 500 nm for the right part. (B) The confocal PL mapping of the nanowire. The alternative feature is not very well resolved, probably due to the limitation of the instrument or the inter-diffusion of different anions within such small area.
Fig. S14. Confocal PL mapping of a CsPbCl 3 -CsPbBr 3 -CsPbBr x I 3-x nanowire heterojunction at different wavelength range.
Fig. S15. Stability of the nanowire heterojunctions. (A, D) Two samples before conversion. (B, E) The corresponding nanowires after conversion. (C, F) The two nanowire heterojunctions after 1 week of storage in nitrogen-filled glove box.