Tip-induced strain, bandgap, and radiative decay engineering of a single metal halide perovskite quantum dot
Hyeongwoo Lee, Ju Young Woo, Dae Young Park, Inho Jo, Jusun Park, Yeunhee Lee, Yeonjeong Koo, Jinseong Choi, Hyojung Kim, Yong-Hyun Kim, Mun Seok Jeong, Sohee Jeong, Kyoung-Duck Park
TTip-induced strain, bandgap, and radiative decay engineering of asingle metal halide perovskite quantum dot
Hyeongwoo Lee † , Ju Young Woo † , Dae Young Park, Inho Jo, JusunPark, Yeunhee Lee, Yeonjeong Koo, Jinseong Choi, Hyojung Kim, Yong-Hyun Kim, Mun Seok Jeong, Sohee Jeong ∗ , and Kyoung-Duck Park ∗ Department of Physics, Ulsan National Institute of Science and Technology (UNIST),Ulsan 44919, Republic of Korea Manufacturing Process Platform R & D Department,Korea Institute of Industrial Technology (KITECH),Ansan 15588, Republic of Korea Department of Energy Science, Sungkyunkwan University (SKKU),Suwon 16419, Republic of Korea Department of Physics, Korea Advanced Institute of Science and Technology (KAIST),Daejeon 34141, Republic of Korea (Dated: February 5, 2021)
Abstract
Strain engineering of perovskite quantum dots (pQDs) enables widely-tunable pho-tonic device applications. However, manipulation at the single-emitter level has neverbeen attempted. Here, we present a tip-induced control approach combined with tip-enhanced photoluminescence (TEPL) spectroscopy to engineer strain, bandgap, andemission quantum yield of a single pQD. Single CsPbBr x I − x pQDs are clearly resolvedthrough hyperspectral TEPL imaging with ∼
10 nm spatial resolution. The plasmonictip then directly applies pressure to a single pQD to facilitate a bandgap shift up to ∼ ∼ for the strain-inducedpQD. Furthermore, by systematically modulating the tip-induced compressive strainof a single pQD, we achieve dynamical bandgap engineering in a reversible manner.In addition, we facilitate the quantum dot coupling for a pQD ensemble with ∼ a r X i v : . [ phy s i c s . a pp - ph ] F e b etal halide perovskite quantum dots (pQDs) are promising candidates for various optoelec-tronic device applications. Specifically, their excellent optical and electrical properties, suchas high photoluminescence (PL) quantum yields (PLQY), widely tunable bandgap, narrowemission width, high absorption coefficient, and high defect tolerance, enable a quantumleap in light-emitting [1–5] and photovoltaic devices [6, 7]. To improve the performanceof perovskite-based optoelectronic devices, many strategies have been proposed, such ascomposition engineering, interface modifications, and device structure engineering. As an-other of these approaches, strain engineering is an attractive method for improving deviceperformance significantly by reversibly and widely tuning the optoelectronic properties ofperovskites.Although a range of strain engineering techniques, including hydrostatic pressurization[8, 9], electrostriction [10], annealing [11, 12], and mechanical bending [13, 14], have beenstudied, controlling strain at the nanoscale or single-crystal level has never been attempteddue to the difficulties of local strain control as well as in-situ nano-optical characterizations.In addition, to widely tune the bandgap or even surpass the threshold of the bond-breakingenergy in nanocrystals, a high pressure at the ∼ GPa scale is generally required [8, 15]. How-ever, applying hydrostatic pressure of ∼ GPa scale and maintaining it in device platformsis highly challenging, which restricts its practical applications. Hence, to understand thefundamental nano-optomechanical properties of single nanocrystals, control their proper-ties at the nanoscale, and further adopt these advantages in practical applications, strainengineering at the single-nanocrystal level is highly desirable.In this work, we present a tip force control approach combined with tip-enhanced pho-toluminescence (TEPL) spectroscopy to engineer the strain of single CsPbBr x I − x pQDswhile simultaneously investigating their modifying emission behaviors [16]. Using Purcellenhanced TEPL signal by a factor of ∼ , we probe single CsPbBr x I − x pQDs with aspatial resolution of ∼
10 nm. We then systematically press the single pQDs through atomicforce microscopy (AFM) control of an Au tip. The applied compressive strain to a singlepQD gives rise a gradual bandgap redshift up to ∼
62 meV as well as dramatic enhance-ment of PLQY by the Purcell effect in the plasmonic nano-cavity, which are quantitativelyexplained by theoretical density functional theory (DFT) calculations and finite-differencetime domain (FDTD) simulations [17–20]. Furthermore, the bandgap energy of a singlepQD is gradually shifted back to the original state when we retract the Au tip from it, i.e.,2
IG. 1.
Schematic of experimental setup and characterization of single pQDs. (a)Schematic illustration of TEPL spectroscopy to probe and control the radiative emission of singleCsPbBr x I − x pQDs. Individually distributed CsPbBr x I − x pQDs are encapsulated by an Al O layer (0.5 nm thickness). Three-dimensional tip-positioning and atomic force tip control allowsingle CsPbBr x I − x pQDs to be found and induce local pressure on them. (b) Absorption (blue)and PL spectra (green) of CsPbBr x I − x pQDs measured from the ensemble state. Inset: TEMimage of drop-cast CsPbBr x I − x pQDs. Scale bar is 40 nm. (c) AFM topography image of spin-coated CsPbBr x I − x pQDs on the Au substrate. Scale bar is 50 nm. (d) Time-dependent far-fieldPL intensity of a few CsPbBr x I − x pQDs exhibiting a typical blinking behavior. systematic bandgap engineering of a single pQD is achieved in a reversible manner. Thisdynamic control is achieved by exploiting the tip indentation region of ∼
10 nm , whichenables local pressure up to ∼ x I − x pQDensemble characterized by structural distortion and spectral redshift up to ∼ Pre-characterizations of single isolated CsPbBr x I − x pQDs CsPbBr x I − x pQDs are spin-coated onto a template-stripped Au substrate with a thin di-electric layer (0.5 nm thick Al O ). The single isolated pQDs are then covered again with anAl O capping layer to prevent degradation under ambient conditions [15]. We use TEPLspectroscopy with a side-illumination geometry for nano-optomechanical characterizations3f the single pQDs, as illustrated in Fig. 1a. For the nanoscale positioning of the tip ontothe single pQDs and vertical distance control between the tip and sample with a preci-sion of < ∼
15 nm) to obtain strong tip-enhanced PLresponses of CsPbBr x I − x pQDs at an excitation wavelength of 632.8 nm. In addition, theAu tip can dynamically induce pressure on single CsPbBr x I − x pQDs through atomic forcecontrol at nanoscale (see Methods for more details). Before we prepare the single isolatedCsPbBr x I − x pQDs, we conduct pre-characterizations of absorption, PL, and transmissionelectron microscopy (TEM) for the ensemble state to understand the fundamental opticaland structural properties, as shown in Fig. 1b. The broad spectral range of absorptioncovers the excitation wavelength (632.8 nm) of our TEPL spectroscopy experiment. ThePL spectrum exhibits a sharp emission peak at ∼
650 nm, and the TEM image verifies auniform size distribution of CsPbBr x I − x pQDs. We then dilute the solution and spin-coat iton the Au substrate to prepare the single isolated pQDs (see Methods for more details). Tocharacterize the distribution of pQDs, AFM measurement is performed, as shown in Fig. 1c.The single CsPbBr x I − x pQDs are distributed well on the Au substrate for dynamic strainengineering of them individually. Fig. 1d shows the change of far-field PL intensity of theCsPbBr x I − x pQDs over 350 s, exhibiting distinct blinking behaviors, which also indicatesthe small number of pQDs in the focused beam spot. TEPL measurements of the single isolated CsPbBr x I − x pQDs To verify the tip-enhanced effect of the TEPL response as the plasmonic tip approachesthe pQDs, we measure TEPL spectra as a function of distance, as shown in Fig. 2a. Theevolution of the TEPL spectra shows a dramatically increasing emission intensity with de-creasing distance. This result describes the near-field enhancement of the CsPbBr x I − x pQDs emission through plasmon − exciton coupling in the nano-cavity [16, 25]. Because weuse CsPbBr x I − x pQDs as a model system to induce resonance coupling between the plas-mon response and pQD exciton PL at ∼ x I − x pQDs, as shown in Fig. 2b.Without the pQDs, only the gap-plasmon response is observed (black) through optical cou-pling between the Au tip and flat Au substrate. The broad linewidth of ∼
235 meV is ingood agreement with the results of previous studies [16, 26]. By contrast, as the Au tip is4
IG. 2.
TEPL measurements of single pQDs. (a) Evolution of TEPL spectra of singleCsPbBr x I − x pQDs as a function of the distance between the Au tip and sample. (b) TEPL spectra(green) of single CsPbBr x I − x pQDs and the gap plasmon response (black) from the junctionbetween the Au tip and Au substrate. The spectra in (a) and (b) are fitted with the Voigt function.(c) Contour plot of TEPL spectra when the tip moves along the lateral direction of CsPbBr x I − x pQDs with a constant tip − sample distance. positioned on the single CsPbBr x I − x pQDs, the gap plasmon resonance disappears becausea single CsPbBr x I − x pQD with a height of ∼
10 nm is sandwiched between the Au tip andAu substrate. Instead, the large TEPL signal of the single CsPbBr x I − x pQDs is observed(green) with a narrow linewidth of ∼
170 meV.We then spatio-spectrally resolve the TEPL response of CsPbBr x I − x pQDs by laterallyscanning the Au tip, as shown in Fig. 2c. In the lateral scanning of ∼
130 nm, a few distinctspots exhibiting a strong TEPL response from CsPbBr x I − x pQDs are observed. Theyshow different emission intensities which are possibly attributable to the different dipoleorientations of pQDs (see Fig. S1 for more details). It should be noted that the PL response5 IG. 3.
Strain, bandgap, and PLQY engineering of a single pQD. (a) Spectra of thegap plasmon response when the Au tip approaches the Au substrate and retracts from it. (b)Reversible TEPL spectra of a single CsPbBr x I − x pQD with compressive stress changed by thetip. The spectra in (a) and (b) are fitted with the Voigt function. The orange curves in (a) and (b)show the spectra at the minimum height (h min ) and under maximum pressure, respectively. Therelative height between the Au tip and Au substrate during the modulation process is the same asthat for the gap plasmon, as in the illustrations at the top of (a) and (b). (c) Peak shift (red) andintensity change (contour plot) of the TEPL spectra derived from (b). of these single pQDs cannot be resolved by far-field PL measurement. Dynamic strain, bandgap, and quantum yield engineering of a single pQD
Beyond tip-enhanced spectroscopy, we then present tip-induced control of strain and opticalproperties of a single CsPbBr x I − x pQD. Before we perform the tip-induced strain engineer-ing, we confirm the optical response of the gap plasmon when the distance between the Autip and flat Au substrate is changed. We measure the evolution of the gap plasmon spectrathrough dynamic tip control in the vertical axis with respect to the Au surface. As shownin Fig. 3a, the plasmon intensity increases slightly as the gap decreases without a spectralshift. Therefore, we can exclude the effect of gap plasmon in the bandgap engineering exper-iment for single pQDs. By constrast, distinct spectral redshifts are observed when the Au6ip moves the same distance on a single CsPbBr x I − x pQD, as shown in Fig. 3b. From thisAFM control, the Au tip can directly apply local pressure and induce compressive strain ona single pQD. When the tip is retracted, the pressure and strain applied on the single pQDare released. This implies the capability of modulating the strain of a single pQD as well asthe mode volume (V) of the plasmonic nano-cavity, which modifies the spontaneous emissionrate of the pQD by the Purcell effect ( ∝ / V). Hence, when we increases the tip pressureon the pQD, a gradual redshift of the TEPL peak caused by bandgap reduction is observedwith increasing TEPL intensity. Interestingly, when we release the pressure by retractingthe tip, the bandgap energy and emission intensity revert to their original state. As can beseen in Fig. 3c, robust bandgap and PLQY modulations are achieved in a reversible mannerwith a maximum energy shift of ∼
62 meV.In the situation of applying pressure larger than a certain value, a phase transition or anirreversible crystalline change has been reported [8, 9, 15]. To experimentally characterizethis threshold pressure and prevent these irreversible changes to a CsPbBr x I − x pQD, weanalyze the TEPL intensity change with respect to the tip pressure, as shown in Fig. 3c.When the perovskite crystals experience a gradual increase in pressure, the PL intensityshould be decreased due to lattice distortion and bending of chemical bonds [14, 15]. In ourresults, because the field enhancement and Purcell effect are increased as the tip − samplegap decreases, the PL emission of the CsPbBr x I − x pQD is enhanced until it reaches thethreshold ( d = 8.5 nm). However, from this point, because the degree of quenching exceedsthe field enhancement, the intensity starts to decrease, as can be seen in Fig. 3c. Thisthreshold is in good agreement with previous studies on the pressure-dependent intensityevolution of halide perovskite nanocrystals [9, 15]. Deterministic generation of nanoscale heterogeneity in pQD ensemble
As an extension of the tip-induced control study, we press the CsPbBr x I − x pQD ensemblewith the Au tip to investigate the interactive effect between pQDs under applied local pres-sure (see Fig. S2 for more details). When we compress and decompress the CsPbBr x I − x pQD ensemble, it shows an apparent topographic change in the pressed region, as shown inFig. 4a (top). Simultaneously, we obtain a corresponding TEPL image (bottom of Fig. 4a)to investigate the optical characteristics of the modified pQD ensemble. The TEPL imageshows the increased emission intensity at the tip-indented region, and this intensity change is7 IG. 4.
Nanoscale heterogeneity in the pQD ensemble. (a) AFM image (top) of thestructurally modified CsPbBr x I − x pQD ensemble induced by tip pressure and corresponding TEPLpeak intensity image (bottom). (b) TEPL spectra of the pQD ensemble obtained before (red dashedline) and after (red line) pQD coupling. (c) Topographic line profile (black) derived from line L1in the AFM image (a, top) and peak energy shift (red) corresponding to line L2 in TEPL image(a, bottom). (d) Numerical simulation of the stress distribution when the Au tip presses theCsPbBr x I − x pQD ensemble. (e) Illustration of the facilitated pQD coupling region attributed tothe decreased inter-pQDs distance as the Au tip presses the ensemble. correlated with the topographic change, as shown in Fig. 4a. This correlation is understoodby the enhanced PLQY as the cavity mode volume, i.e., distance between the Au tip andAu surface, is decreased [27]. When we compare the TEPL spectra before and after the Autip presses the CsPbBr x I − x pQD ensemble, they exhibit a distinct spectral redshift with alinewidth broadening, as shown in Fig. 4b. To check the correlation between the pressureand spectral redshift, we derive the topographic line profile and peak energy shift profilefrom the corresponding locations of lines L1 and L2 in Fig. 4a. As shown in Fig. 4c, the sizeof the tip-indented hole is comparable to the apex size of the Au tip (see Fig. S3 for moredetails). The magnitude of the redshifted energy changes proportionally with respect to thedistortion of the ensemble surface, which is confirmed by the simulated stress distribution.As shown in Fig. 4d, the stress generated from the Au tip is distributed across the broad8egion of the pQD ensemble. Specifically, the maximum-stress region is located on the rightunderneath of the Au tip, and the induced stress is gradually decreased as the distance in-creases from the maximum region. This stress distribution shows that a significant decreasein the inter-pQD distance occurs at the center, and a relatively smaller distance decrease isexpected at the near regions, as illustrated in Fig. 4e. These observed nano-optomechanicalcharacteristics are in good agreement with previous studies on pQD coupling based on thesame mechanism that high pressure causes a significant reduction in the inter-pQD distance[21–23]. The reduced inter-pQD distance then leads to pQD coupling with characteristicsof reduced energy for exciton transition and spectral broadening by forming multiple statesin the coupled pQDs [22, 23]. Therefore, our tip-induced local pressure of > DFT calculations for strained CsPbBrI perovskite To understand the electronic and optical properties of mechanically strained pQDs,we perform first-principles DFT calculations on uniaxially-strained CsPbBrI perovskite(Fig. 5a), which is compositionally very close to the pQDs used in our experiments (seeFig. S4 in Supplementary Information for more details). We obtain the equilibrium latticeconstants a and b at a specific c by finding equilibrium volumes using the Murnaghan equa-tion of state [28] from DFT calculations, which give rise to a Poisson’s ratio of 0.143. Theapplied stress is obtained by differentiating the total energy curve as a function of unit cellvolume in Fig. 5b. We find that the bandgap decreases almost linearly as the compressivestrain increases along the c -axis. Fig. 5b shows the shift and applied stress depending on the c -axis strain. The reduction of ∼
60 meV corresponds to 1.34 % compressive strain. Thus,we derive a compressive stress of 0.78 GPa applied to the CsPbBrI structure. Fig. 5c showsthe overall density of states (DOS) of CsPbBrI perovskite. The valence band maximum(VBM) is derived from the p states of Br and I with contribution from the Pb s states.Whereas, the conduction band minimum (CBM) predominantly consists of the p orbitalof Pb. Under the compressive strain of 1.34 %, the s orbital of Pb and p orbital of thehalide are up-shifted in the valance band edge, while the p orbital of Pb in the conductionband edge is down-shifted, resulting in the decrease (Fig. 5d). Specifically, as the CsPbBrI crystal is compressed by 1.34 % along the c -axis, the in-plane Pb − I bond length increases9
IG. 5.
Electronic and optical properties of mechanically strained CsPbBrI per-ovskite. (a) Atomic structure of CsPbBrI perovskite. (b) Bandgap shift and applied pressuredepending on the c-axis strain. The free energy graph is independently plotted in the background.(c) Projected density of states (PDOS) for the Pb s orbital and Pb, Br, and I p orbitals of CsPbBrI equilibrium structure. The dotted line indicates the Fermi energy. (d) PDOS for the s and p statesof Pb and p states of Br and I near the Fermi energy under 0 % (dashed line) and 1.34 % (solidline) strains. The two energy scales are aligned with respect to their core level. by 0.006 ˚A, while the Pb − Br bond length decreases by 0.040 ˚A. Because the VBM andCBM are the antibonding states of the Pb s and halide p orbitals and the Pb p and halide p orbitals, respectively [29–32], the CBM moves downwards alongside the increased Pb − Ibond lengths, and the VBM moves upwards alongside the decreased Pb − Br bond lengths(see Fig. S5 in Supplementary Information for more details).
Conclusions
In summary, we have demonstrated the tip-induced dynamic control of strain, bandgap, andquantum yield of single CsPbBr x I − x pQDs by using a controllable plasmonic nano-cavitycombined with TEPL spectroscopy. This tip-induced nano-engineering approach with local10ressure of GPa order not only provides an in-depth understanding of the optical proper-ties of metal halide perovskite pQDs and their coupling nature but also offers a practicalway to tune the mechanical and electronic properties at the single-pQD level for potentialapplications in ultracompact nano-optoelectronic devices. Specifically, we envision thatthis dynamical single-dot manipulation will enable the tunable nano-LEDs [33], realizingultra-high definition display with high efficiency. In addition, extremely high tip inducedlocal pressure will allow pressure-induced recrystallization and phase transition at the few-nanometer-length scale to further improve the grain quality and optical properties of 2Dperovskite ensembles as a post-fabrication process. Furthermore, the extrinsically modifiedphysical properties of pQDs by the plasmonic cavity, e.g., reduced non-radiative recombi-nation, modulation of carrier dynamics and PL lifetime, and enhanced radiative decay rate,can significantly improve the efficiency of perovskite photovoltaics [34, 35]. MethodsSynthesis of CsPbBr x I − x pQDs. Cesium carbonate (Cs CO , 99.9 %, Sigma-Aldrich),lead iodide (PbI , 99.99 %, TCI), oleic acid (OA, technical grade, Sigma-Aldrich), oleylamine(OLA, technical grade, Sigma-Aldrich), zinc bromide (ZnBr , 99.999 %, Sigma-Aldrich), and1-octadecene (ODE, technical grade, Sigma-Aldrich) were used. Cesium oleate precursorsolution was prepared based on previously reported procedures [36, 37] with modifications.0.163 g of Cs CO , 0.5 mL of OA, and 8 mL of ODE were placed in a round-bottomed3-neck flask and heated to 110 ◦ C under vacuum. After degassing for 2 h, the solutionwas further heated to 150 ◦ C under an inert atmosphere. Also, 0.347 g of PbI , 0.085 g ofZnBr , 2.0 mL of OA, 2.0 mL of OLA, and 20 mL of ODE were placed in a round-bottomed3-neck flask and degassed at 110 ◦ C for 2 h. Then, the solution was heated to 180 ◦ C underan inert environment. Finally, 1.6 mL of preheated cesium oleate precursor solution wasquickly injected, and the reaction solution was rapidly quenched in an ice bath. The pQDswere collected from the crude reaction solution after multiple purification steps. Specifically,the crude reaction solution was centrifuged at 13000 rpm for 30 min, and the precipitateswere redispersed in 6 mL anhydrous toluene. The pQD toluene solution was centrifuged(6000 rpm for 5 min) with an additional 4 mL of methyl acetate, and the precipitates wereredispersed in 8 mL of anhydrous hexane. The QD hexane solution was again centrifugedat 6000 rpm for 5 min. Finally, the supernatant was collected and filtered using a Teflon11yringe filter (0.45 micron). Note that all procedures were conducted in an inert atmosphere.
TEPL spectroscopy setup.
The prepared pQDs spin-coated on the template strippedgold substrate were loaded on a piezo-electric transducer (PZT, P-611.3X, Physik Instru-mente) for XY scanning and application of compressive strain to single pQDs with < ∼ − Ne laser (632.8 nm, < ∼ µ m) and collimatedagain using an aspheric lens. The collimated beam was then passed through a half-waveplate to make the excitation polarization parallel with respect to the tip axis. Finally, thebeam was focused onto the Au tip using a microscope objective (NA = 0.8, LMPLFLN100X,Olympus) with a side illumination geometry. To ensure highly efficient laser coupling tothe Au tip, the tip position was controlled with ∼
30 nm precision by Picomotor actuators(9062-XYZ-PPP-M, Newport). TEPL responses were collected using the same microscopeobjective (backscattering geometry) and passed through an edge filter (LP02-633RE-25,Semrock) to cutoff the fundamental laser line. TEPL signals were then dispersed onto aspectrometer (f = 328 mm, Kymera 328i, Andor) and imaged with a thermoelectricallycooled charge-coupled device (CCD, iDus 420, Andor) to obtain the TEPL spectra. Beforethe experiment, the spectrometer was calibrated with an Argon Mercury lamp. A 150g / mmgrating blazed to 800 nm (spectral resolution of 0.62 nm) was used for PL measurements. DFT calculations.
For the first-principles DFT calculations, we employed the projector-augmented wave potentials [38] with ionic core potential and the Perdew-Burke-Ernzernhofexchange-correlation functional [39], as implemented in the Vienna
Ab-initio
SimulationPackage (VASP) [40]. We used a kinetic energy cutoff of 400 eV and an (8 × ×
8) K-pointsgrid for sampling the Γ-point. Gaussian smearing of 0.05 eV was employed. All atomicforces were relaxed to less than 0.01 eV / ˚A. 12 DTD simulations of Purcell factor and optical field distribution.
We used a com-mercial FDTD simulation software (Lumerical Solutions, Inc.) to characterize the Purcellenhancement factor and optical field distribution at the Au tip − Au substrate nano-gap.The Au tip with a 5 nm apex radius was placed in close proximity (10 nm gap) to theAu substrate. As a fundamental excitation source, a linearly polarized monochromatic 633nm light was projected with 45 ◦ . For estimation of the Purcell enhancement factor, a fluo-rescent dipole with emission wavelength 650 nm was positioned at the center of the nano-gap. Numerical analysis of the tip-induced stress distribution.
A three-dimensional me-chanical simulation was performed to analyze the distribution of the pQD ensemble andthe pressure applied to the contact region between the Au tip and pQDs. The pressureand force were calculated through the ANSYS Mechanical Enterprise module. The lowestposition of the Au tip was set as the position of the surface of the pQDs to model theexperimental conditions. The bottom plane of the pQDs was considered to be fixed at thesubstrate surface because the pQDs were rigidly covered by a thin Al O capping layersupressing their lateral movement. Then, the mechanical properties, including stress andpressure, were simulated under gradual increases in the force applied to the Au tip. Theapplied force was defined on the top surface of the Au tip in the perpendicular directionwith respect to the pQD surface. The mechanical properties of Young’s modulus, Poissonratio, and density of the pQDs were obtained from the literature [41]. Data availability.
The data that support the plots within this paper and other findings ofthis study are available from the corresponding author upon reasonable request. [1] Liu, X.-K. et al.
Metal halide perovskites for light-emitting diodes.
Nat. Mater. , 10–21(2021).[2] Stranks, S. D. & Snaith, H. J. Metal-halide perovskites for photovoltaic and light-emittingdevices. Nat. Nanotechnol. , 391–402 (2015).[3] Dong, Y. et al. Bipolar-shell resurfacing for blue LEDs based on strongly confined perovskite uantum dots. Nat. Nanotechnol. , 668–674 (2020).[4] Swarnkar, A. et al. Colloidal CsPbBr perovskite nanocrystals: luminescence beyond tradi-tional quantum dots. Angew. Chem. , 15644–15648 (2015).[5] Dirin, D. N. et al.
Harnessing defect-tolerance at the nanoscale: highly luminescent lead halideperovskite nanocrystals in mesoporous silica matrixes.
Nano Lett. , 5866–5874 (2016).[6] Swarnkar, A. et al. Quantum dot–induced phase stabilization of α -CsPbI perovskite forhigh-efficiency photovoltaics. Science , 92–95 (2016).[7] Han, T.-H. et al.
Interface and defect engineering for metal halide perovskite optoelectronicdevices.
Adv. Mater. , 1803515 (2019).[8] Liu, G. et al. Isothermal pressure-derived metastable states in 2D hybrid perovskites showingenduring bandgap narrowing.
PNAS , 8076–8081 (2018).[9] Ma, Z. et al.
Pressure-induced emission of cesium lead halide perovskite nanocrystals.
NatureCommun. , 4506 (2018).[10] Chen, B. et al. Large electrostrictive response in lead halide perovskites.
Nat. Mater. ,1020–1026 (2018).[11] Steele, J. A. et al. Thermal unequilibrium of strained black CsPbI thin films. Science ,679–684 (2019).[12] Xiao, Z. et al.
Solvent annealing of perovskite-induced crystal growth for photovoltaic-deviceefficiency enhancement.
Adv. Mater. , 6503–6509 (2014).[13] Zhu, C. et al. Strain engineering in perovskite solar cells and its impacts on carrier dynamics.
Nat. Commun. , 815 (2019).[14] Chen, Y. et al. Strain engineering and epitaxial stabilization of halide perovskites.
Nature , 209–215 (2020).[15] Liu, S. et al.
Manipulating efficient light emission in two-dimensional perovskite crystals bypressure-induced anisotropic deformation.
Sci. Adv. , eaav9445 (2019).[16] Park, K.-D. et al. Tip-enhanced strong coupling spectroscopy, imaging, and control of a singlequantum emitter.
Sci. Adv. , eaav5931 (2019).[17] Lee, H. et al. Tip-enhanced photoluminescence nano-spectroscopy and nano-imaging.
Nanophotonics , 3089–3110 (2020).[18] Purcell, E. M., Torrey, H. C. & Pound, R. V. Resonance absorption by nuclear magneticmoments in a solid. Phys. Rev. , 37 (1946).
19] Schuller, J. A. et al.
Plasmonics for extreme light concentration and manipulation.
Nat.Mater. , 193–204 (2010).[20] Russell, K. J., Liu, T.-L., Cui, S. & Hu, E. L. Large spontaneous emission enhancement inplasmonic nanocavities. Nat. Photon. , 459–462 (2012).[21] Williams, K. J. et al. Strong electronic coupling in two-dimensional assemblies of colloidalPbSe quantum dots.
ACS nano , 1532–1538 (2009).[22] Cui, J. et al. Colloidal quantum dot molecules manifesting quantum coupling at room tem-perature.
Nat. Commun. , 5401 (2019).[23] Li, B. et al. Pressure-dependent photoluminescence of CdSe/ZnS quantum dots: Critical pointof different pressure regimes.
Phys. Lett. A , 1483–1486 (2019).[24] Park, K.-D., Jiang, T., Clark, G., Xu, X. & Raschke, M. B. Radiative control of dark excitonsat room temperature by nano-optical antenna-tip Purcell effect.
Nat. Nanotechnol. , 59–64(2018).[25] Groß, H., Hamm, J. M., Tufarelli, T., Hess, O. & Hecht, B. Near-field strong coupling ofsingle quantum dots. Sci. Adv. , eaar4906 (2018).[26] Kravtsov, V., Berweger, S., Atkin, J. M. & Raschke, M. B. Control of plasmon emission anddynamics at the transition from classical to quantum coupling. Nano Lett. , 5270–5275(2014).[27] Park, K.-D. et al. Hybrid tip-enhanced nanospectroscopy and nanoimaging of monolayer WSe with local strain control. Nano Lett. , 2621–2627 (2016).[28] Murnaghan, F. The compressibility of media under extreme pressures. PNAS , 244 (1944).[29] Filippetti, A. & Mattoni, A. Hybrid perovskites for photovoltaics: Insights from first princi-ples. Phys. Rev. B , 125203 (2014).[30] Brandt, R. E., Stevanovi´c, V., Ginley, D. S. & Buonassisi, T. Identifying defect-tolerantsemiconductors with high minority-carrier lifetimes: beyond hybrid lead halide perovskites. MRS Commun. , 265–275 (2015).[31] Ravi, V. K., Markad, G. B. & Nag, A. Band edge energies and excitonic transition probabilitiesof colloidal CsPbX (X= Cl, Br, I) perovskite nanocrystals. ACS Energy Lett. , 665–671(2016).[32] Umebayashi, T., Asai, K., Kondo, T. & Nakao, A. Electronic structures of lead iodide basedlow-dimensional crystals. Phys. Rev. B , 155405 (2003).
33] Liu, Y. et al.
Theoretical study of piezo-phototronic nano-LEDs.
Adv. Mater. , 7209–7216(2014).[34] Kim, W. et al. Oriented grains with preferred low-angle grain boundaries in halide perovskitefilms by pressure-induced crystallization.
Adv. Energy Mater. , 1702369 (2018).[35] Wang, T. et al. Indirect to direct bandgap transition in methylammonium lead halide per-ovskite.
Energy & Environ. Sci. , 509–515 (2017).[36] Protesescu, L. et al. Nanocrystals of cesium lead halide perovskites (CsPbX , X= Cl, Br, andI): novel optoelectronic materials showing bright emission with wide color gamut. Nano Lett. , 3692–3696 (2015).[37] Woo, J. Y. et al. Highly stable cesium lead halide perovskite nanocrystals through in situ leadhalide inorganic passivation.
Chem. Mater. , 7088–7092 (2017).[38] Bl¨ochl, P. E. Projector augmented-wave method. Phys. Rev. B , 17953 (1994).[39] Perdew, J. P., Burke, K. & Ernzerhof, M. Generalized gradient approximation made simple. Phy. Rev. Lett. , 3865 (1996).[40] Kresse, G. & Joubert, D. From ultrasoft pseudopotentials to the projector augmented-wavemethod. Phys. Rev. B , 1758 (1999).[41] Rakita, Y., Cohen, S. R., Kedem, N. K., Hodes, G. & Cahen, D. Mechanical properties ofAPbX (A= Cs or CH NH ; X= I or Br) perovskite single crystals. MRS Commun. ,623–629 (2015). Acknowledgements
This work was supported by a National Research Foundation of Korea (NRF) grant fundedby the Korea government (MEST) (2019K2A9A1A06099937 and 2020R1C1C1011301).S.J. acknowledges a Creative Materials Discovery Program through the National Re-search Foundation (NRF) of Korea funded by the Ministry of Science and ICT (NRF-2019M3D1A1078299, 2019R1A2B5B03070407). Y.-H.K acknowledges 2019M3D1A1078302.
Author contributions
H. Lee and K.-D Park conceived the experiments. H. Lee performed the TEPL spectroscopyand control experiments. J.Y. Woo, J. Park, and S. Jeong designed and prepared thesamples. I. Jo, Y. Lee, and Y.-H Kim designed and performed the simulations. Y. Koo, D.Y.Park, and H. Kim performed pre-characterizations with the PL and AFM measurements.16. Choi designed and developed the hyperspectral imaing processes. H. Lee, J.Y. Woo, S.Jeong, and K.-D. Park analyzed the data, and all authors discussed the results. H. Lee andK.-D. Park wrote the manuscript with contributions from all authors. K.-D. Park supervisedthe project.
Additional information
Supplementary information is available for this paper.