Tuning the light emission of a Si micropillar quantum dot light-emitting device array with the strain coupling effect

The study targets the challenge of integrating strain/pressure sensing with optical readout in a Si-compatible platform. While optical signals offer advantages such as visualization and low crosstalk for tactile sensing, prior approaches rely largely on piezophototronic LED arrays based on wurtzite semiconductors (e.g., ZnO, CdS, GaN), which present defect-related emission and limited compatibility with mainstream Si technology. Si’s indirect band gap leads to poor light emission, complicating Si-based LEDs, and its centrosymmetric lattice precludes piezoelectric polarization. The authors explore strain-induced modulation in Si via flexoelectricity—polarization induced by strain gradients, not constrained by lattice symmetry—to control carrier transport and thereby modulate QLED electroluminescence. The research question is whether the flexoelectronic (flexoelectric polarization–assisted) effect in Si micropillar arrays can effectively tune carrier injection/transport and enhance the electroluminescence of Si-based QLED pixels for pressure mapping, while maintaining CMOS compatibility.

Previous tactile sensor arrays employing piezophototronics utilized wurtzite semiconductors (ZnO, CdS, GaN) and nanowire LED arrays to map pressure distributions, showing promise for e-skin, human–machine interfaces, and pressure sensing. However, defects in piezoelectric materials often introduce parasitic defect emissions at p–n junctions, complicating optical signal acquisition and processing. Quantum dots (QDs) offer narrow, tunable emission spectra and high stability, providing a way to mitigate defect-emission issues. Despite Si’s dominance in the semiconductor industry, coupling pressure sensing with Si electronics remains challenging because Si lacks piezoelectricity. Flexoelectricity—polarization generated by strain gradients—can exist even in centrosymmetric crystals like Si. Although the theoretical flexoelectric coefficient of Si is ~1 nC m−1, experimental reports have observed much larger values (~78 nC m−1), and recent studies indicate flexoelectronic effects can modulate Schottky barriers and carrier distributions, tuning charge transport. This suggests a pathway to Si-based, strain-coupled optoelectronics.

Device fabrication comprised three parts: (1) Si micropillar array formation, (2) QLED stack fabrication, and (3) characterization under electrical bias and applied pressure. 1) Si micropillars: p-type Si wafers were solvent-cleaned (acetone, ethanol), dipped in 1% HF for 2 min, and O2-plasma treated for 10 min. Photoresist (SUN 9i 50 cP) was spin-coated at 4000 rpm for 60 s, patterned by UV lithography, and served as the mask for ICP etching (SI 500). Etch gases: SF6 30 sccm, O2 6 sccm, Ar 10 sccm. Remaining resist was removed in acetone. Resulting micropillars were approximately square with ~20 µm side length and ~100 µm pitch. 2) QLED array fabrication: The wafer was blanket-coated with SU8 1040 and RIE-etched (2000CE) to re-expose micropillar tips, leaving SU8 to planarize and electrically isolate inter-pillar spaces. A NiO film (~200 nm) was deposited by magnetron sputtering (PVD75) at 100 W for 30 min as a hole transport/electron-blocking layer on Si micropillars. CdSe/ZnS QDs in n-hexane (15 mg mL−1) were spin-coated at 2000 rpm for 40 s and annealed at 60 °C for 5 min to form the emissive layer. TPBi (~800 nm) was thermally evaporated as the electron transport layer. Patterned ITO (~100 nm) was sputtered as the transparent top electrode; Ag (~50 nm) was sputtered on the Si backside as the bottom electrode. The SU8 ensured a flat surface; only regions atop micropillars formed hole injection paths, defining pixel areas. 3) Characterization: Morphology was examined by FE-SEM (Hitachi SU8020). Electrical biasing used a Maynuo M8812 source meter; EL spectra were recorded with an Ocean Optics QE65000 spectrometer. I–V and I–t curves were measured on a Keithley 1500 system. Optical emission images were captured with a Zeiss Observer Z1. A dynamometer mounted on three-dimensional stages applied normal forces to the micropillar array over a small contact area, generating high local pressures at the Si beneath pillars (e.g., 20 N → ~5.6 GPa). Pressure-induced flexoelectric polarization in Si was used to modulate carrier transport and EL.

- The Si-based QLED array, using p-Si micropillars as hole transport paths and pixel definitions, exhibits uniform, isolated pixel emission without crosstalk. - Emission wavelengths are set by the QDs: green EL peak at ~530 nm (FWHM ~24 nm) and red EL peak at ~633 nm (FWHM ~33 nm), matching the PL peaks of the respective CdSe QDs and remaining stable with increasing bias. - Pixel geometry: square Si micropillars ~20 µm per side with ~100 µm spacing. - EL intensity increases with applied bias; current–voltage characteristics are consistent with the designed energy level alignment (ITO/TPBi/QDs/NiO/p-Si/Ag). - Applying external pressure via a dynamometer induces a strain gradient in Si, generating a flexoelectric polarization field that modulates local carrier concentrations and transport. - Under forward bias, device current increases monotonically with pressure; under reverse bias, current shows an asymmetric rise to a maximum, indicating contributions beyond symmetric piezoresistive effects. - The electroluminescence density of the Si-based QLED array increases by approximately 600% under ~13.8 GPa pressure. An example calibration indicates 20 N normal force corresponds to ~5.6 GPa at the Si beneath pillars. - The observed asymmetric pressure dependence and large EL enhancement are attributed primarily to the flexoelectronic effect in Si, which acts effectively as a gate field modulating injection/transport, with additional contribution from the (symmetric) piezoresistive effect.

The work demonstrates that strain-gradient-induced polarization in centrosymmetric Si can be harnessed to actively tune carrier injection and transport in a QLED architecture, thereby modulating electroluminescence. By using QDs as the emissive layer, the platform circumvents Si’s indirect-bandgap limitation while preserving full compatibility with Si microfabrication. The Si micropillar geometry concentrates stress, producing large local pressures and strain gradients, which generate a flexoelectric polarization field. This field effectively shifts local band alignments and carrier distributions, increasing forward-bias current and EL output, while causing asymmetric I–V changes that distinguish flexoelectronic effects from purely piezoresistive responses. The ability to define pixels by micropillar placement yields uniform, isolated emission sites suitable for arrayed tactile sensing with optical readout. The results support the feasibility of Si-based, strain-coupled photonic devices for pressure mapping, human–machine interfaces, and electronic skin, offering straightforward multiwavelength capability via QD selection and compatibility with Si processing for potential large-scale integration.

This study reports a Si-compatible QLED pixel array that leverages Si micropillars and QD emitters to realize optically readable, strain-tunable emission. The key contribution is showing that the flexoelectronic effect in Si—induced by strain gradients under applied pressure—can substantially modulate carrier transport and enhance EL output (up to ~600% at ~13.8 GPa), with stable, narrow-spectrum emission defined by QDs. The micropillar-defined pixels exhibit uniformity and negligible crosstalk, and the fabrication flow is compatible with Si technology. Potential future research directions include: reducing the required operating pressure via optimized pillar geometry/material stacks; quantitative separation of flexoelectronic and piezoresistive contributions; exploring alternative transport/blocking layers and QD chemistries for lower-voltage operation and higher efficiency; scaling to high-resolution, large-area arrays and integrating on-chip readout; and evaluating long-term stability and cycling performance under repeated mechanical loading.

- The device requires very high local pressures (on the order of several GPa, with EL enhancement reported up to ~13.8 GPa), which may limit practical deployment and raises questions about mechanical reliability and integration in typical tactile sensing scenarios. - While asymmetric I–V behavior supports a flexoelectronic mechanism, the precise quantitative decomposition between flexoelectronic and piezoresistive effects is not fully resolved. - The study focuses on EL intensity and spectral characteristics; data on lifetime, environmental stability, and cycling durability under repeated loading are not provided in the excerpt. - Complete affiliation details and some methodological specifics (e.g., exact pressure calibration across the array, uniformity statistics over large areas) are not fully detailed in the provided text.