Nanowizard 2

Square-shaped electrodes (500 µm × 500 µm each) were fabricated on glass slides (2 cm × 2 cm × 170 µm) as detailed in the Supplementary Information (Fig. S1a). Prior to use, all electrodes were cleaned by washing with acetone, ethanol, ultrapure water, 10% NaOH aqueous solution at 80 °C (for 15 min), and again ultrapure water. They were studied with AFM (using a NanoWizard II from JPK Instruments) as well as by cyclic voltammetry and EIS carried out using a PGSTAT128N potentiostat (from Metrohm Autolab).

The hydrodynamic diameter and surface charge of NPs were measured with a Zetasizer Nano ZS (Malvern Incorporated, Malvern, UK). NPs were dispersed in ultra-pure water, and 20 measurements were conducted per sample. The morphology and distribution of NPs was visualized by TEM (H-7600, Hitachi, Tokyo, Japan) and AFM (NanoWizard II, JPK Instruments, Berlin, Germany). Samples were coated onto carbon film on 200 mesh copper grids (Electron Microscopy Science, Hatfield, PA, USA), dried overnight at room temperature, and visualized by TEM (ARM200F, JEOL, Tokyo, Japan). The levels of AuNP, CPEI, and HP in sCGNPs and mCGNPs were determined by various tools based on the physical and chemical properties of each component. AuNPs were characterized using a UV-vis spectrometer (TECAN PRO 200, Männedorf, Switzerland) in the wavelength range from 450 to 700 nm. CPEI and HP were detected on a fluorescence plate reader (TECAN PRO 200) using a ChemiDoc imaging system (BR170-8265, Bio-Rad Laboratories, Hercules, CA, USA), after which the fluorescence intensity was calculated.

All AFM images of CPE45 nanocrystals and single crystals were recorded in tapping mode on a JPK Nano wizard II. For the analysis of nanocrystal thickness, AFM images of nanocrystal films were plane-flattened. A background surface was subtracted, derived from areas showing the bare substrate. A height histogram as shown in Figure 2D was calculated.
The lateral size of the nanocrystals was extracted by measuring cross sections of closely packing nanocrystals. Phase images as shown in Figure 2 were used. The approximate radius of curvature of the AFM-tip represents the experimental error Δx = 10 nm. A bin size of 5 nm for the cross section was used to determine the distribution of particle diameters. For the AFM images of the nanocrystals, AFM-tips with a nominal radius of curvature of 8 or 10 nm and a resonance frequency around 150 kHz or 320 kHz were used. Both had force constants of the cantilever of about 40 N/m. For the images of the functionalized/non-functionalized CPE45 crystals, sharp AFM tips with resonance frequencies of 40–80 kHz and force constants of 0.12 N/m or 0.24 N/m were used.

Surface topographies of perovskite thin films were inspected using Nano Wizard II (JPK Instruments). The scanning area was 10 μm × 10 μm.

To evaluate the Young's modulus of the PDMS, stress–strain experiments were carried out at room temperature with a tensile machine (Instron model 5542). PDMS rectangular arrays of 26 mm in perimeter were tested at an initial strain rate of 2 mm min^–1. The slope at the start of the deformation was used to calculate the Young's modulus E. We found that by curing PDMS at 65 °C for 6 h, the E was 3±0.3 MPa. After measuring the micropillar dimensions by SEM, we calculated micropillar spring constants according to:

where I is the moment of inertia and L the length of the micropillar61 . Alternatively, we used an AFM (Nanowizard II, JPK Instruments) to directly evaluate the spring constant of shorter micropillars with lengths of 8.2 μm and of longer micropillars with lengths of 13.7 μm (Supplementary Fig. 2). To do this, the top of a micropillar was brought into contact with an AFM cantilever. The AFM cantilever was used to apply a force to a micropillar and the micropillar deflection was measured. From this deflection, we determined the average spring constant of the shorter micropillars.