Piezoelectricity in chalcogenide perovskites

Mechanical-to-electrical energy conversion via the piezoelectric effect enables generation of electrical charge under applied mechanical stress in materials with a net dipole moment. Many of the highest-performing piezoelectrics (e.g., PZT-based) contain lead, raising toxicity and regulatory concerns (WEEE, RoHS), which limits biomedical and sustainable applications. The study aims to discover and validate lead-free, environmentally benign piezoelectric materials with strong performance. Chalcogenide perovskites (ABX3 with A = Ba, Ca, Sr; B = Ti, Zr, Hf; X = S, Se) are direct-bandgap materials of interest in optoelectronics. Here, the authors hypothesize and demonstrate that chalcogenide perovskites, despite often being centrosymmetric, can exhibit pronounced piezoelectricity due to a loosely packed unit cell that allows symmetry reduction under deformation. They experimentally verify piezoelectricity in BaZrS3 and explore energy harvesting using BaZrS3–polymer composites.

The authors reference common piezoelectric materials and their coefficients (Supplementary Table S1), noting top performers often contain lead (PZT variants, PZN-PT, PZNT). Prior studies have reported ferroelectricity in halide perovskites and piezoelectric applications of metal-free perovskites, but no prior report of piezoelectricity or ferroelectricity in chalcogenide perovskites. Environmental regulations and stability issues with halide perovskites (PbI2, halides) motivate exploration of lead-free, stable chalcogenide perovskites. The work positions BaZrS3 and related compounds as new candidates within this context.

Computational (DFPT): First-principles calculations were performed in VASP using PAW pseudopotentials and PBE-GGA exchange-correlation. Plane-wave cutoff was 800 eV; Monkhorst-Pack k-point mesh with ~0.02 Å⁻1 spacing; SCF convergence criterion 1e-5 eV. To evaluate piezoelectric response in centrosymmetric compounds, the authors introduced random ionic displacements of 1–2% (without changing cell volume/shape) to generate symmetry-broken configurations. For each of 20 distinct random structures per material, DFPT was used to compute piezoelectric stress tensors and elastic tensors; piezoelectric strain coefficients dij were obtained via tensor relations and averaged across the ensemble. Phonon band structures were checked to confirm dynamical stability (absence of imaginary modes). Flexoelectric contributions were also evaluated ab initio (Supplementary Figs. S14, S15). Input and structure files are provided in a public GitHub repository.

Synthesis of BaZrS3 powder: BaZrO3 powder (<10 µm) was sulfurized in a three-zone tube furnace using CS2/N2 at ~1050 °C for ~4 h (5 sccm flow, ~2 Torr). After cooling under N2 purge, fully sulfurized BaZrS3 powder was collected. XRD (Cu Kα) and Rietveld refinement confirmed orthorhombic Pnma structure (a = 7.0639 Å, b = 9.9809 Å, c = 7.0191 Å) with no secondary phases; SEM showed micron-sized crystallites comprised of nanocrystals.

BaZrS3 thin films: BaZrO3 precursor films were prepared by spin coating a propionic acid solution containing barium acetate, zirconium(IV) acetylacetonate, and polyvinyl butyral onto quartz, followed by air anneals (700 °C for ~90 min, then ~870 °C for ~40 min). Films were sulfurized in CS2/N2 at ~1050 °C for ~4 h to obtain ~300 nm BaZrS3 films.

PFM characterization: Out-of-plane PFM was conducted on an Asylum MFP-3D AFM using Ti/Ir-coated conductive probes (tip radius ~25 nm, k ~2.8 N/m). To enhance SNR, measurements were performed near contact resonance (~340 kHz) with a stiff cantilever to minimize electrostatic artifacts. Films were Ar plasma cleaned and sputter-coated with ~5 nm Pt to improve field uniformity under the tip. A 5×5 µm² area was scanned at ~0.6 Hz, 90° scan angle; AC bias 0–6 V was applied. Calibration used periodically poled lithium niobate (PPLN), yielding d33,eff = −25.81 ± 1.87 pm/V (within vendor specs). For BaZrS3, PFM amplitude and phase images were collected as a function of bias to extract d33,eff from the amplitude–voltage slope.

Composite film preparation (BaZrS3–PCL): BaZrS3 powder was dispersed in acetone via ultrasonication; PCL (Mw ~80,000) was dissolved in acetone at 60 °C with stirring to form solutions with 0–30 wt% BaZrS3. Solutions were solvent-cast on petri dishes and oven-dried at ~45 °C for ~4 h to yield flexible films. Microstructure and dispersion were examined by SEM (BSD) and EDS elemental mapping; particle size analyzed via Feret diameters. XRD verified retention of PCL and BaZrS3 phases.

Device fabrication and electrical testing: Circular films (~5.5 cm diameter, 0.35 ± 0.03 mm thickness) were sandwiched with copper tape electrodes. Open-circuit voltage (Voc) measured with an oscilloscope; short-circuit current (Isc) with a current amplifier; cyclic loading applied using a pneumatic actuator with controlled pressure and frequency. A full-bridge rectifier and external resistors/capacitors were used for DC output and charging tests. Performance was evaluated versus BaZrS3 loading (2–30 wt%), pressure (~30 PSI typical), frequency (1–4+ Hz), film thickness (0.2–1 mm), and resistive load. PMMA matrices were also tested as controls and as PMMA/BaZrS3 composites.

• BaZrS3 exhibits pronounced piezoelectricity despite an orthorhombic centrosymmetric (Pnma) structure. PFM on ~300 nm films yielded d33,eff = −21.38 ± 1.51 pm/V, comparable to lithium niobate (calibration: −25.81 ± 1.87 pm/V).
• DFPT with displacement-induced polarization predicts for BaZrS3 d33 ≈ 17.65 pC/N (pm/V), while BaZrO3 shows much weaker response (d33 ≈ −0.015 pC/N; maximum dij (d22) ≈ 2.46 pC/N). Calculations attribute enhanced response to a larger, loosely packed 20-atom unit cell in BaZrS3 allowing extended ionic displacements and symmetry reduction under load.
• Extension to the family: DFPT indicates high coefficients in other chalcogenide perovskites (e.g., CaZrS3 maximum dij = 61.74 pC/N for d14; BaHfS3 maximum dij = 102.58 pC/N for d21).
• Composite energy harvesters: BaZrS3–PCL films show optimal performance at ~15 wt% loading, delivering ~4× higher Voc and ~5.3× higher Isc than pure PCL at ~30 PSI and ~4 Hz. Peak areal power density ≈ 103.5 µW cm⁻² (at ~30 PSI, ~4 Hz; optimal load ≈ 1 MΩ).
• Body-motion harvesting: With full-wave rectification and 1 MΩ load, maximum DC voltages were ~−21.2 V (walking), ~−40.4 V (jogging), ~−72 V (running), and ~−5.3 V (elbow bending), ~−10.4 V (clapping), ~−19.8 V (hand tapping at ~1 Hz). A 1.0 µF capacitor charged to ~18.8/23.2/29.2 V (walking/jogging/running) in 60 s and to ~3.82/4.52 V (elbow bending/clapping). Capacitors up to ~10 µF were charged by hand tapping at ~4 Hz; higher tapping frequency accelerated charging. The device illuminated 35 commercial LEDs (3.0–3.2 V each) by hand tapping at ~4 Hz.
• Comparative tests: BaZrS3 additives significantly outperformed BaZrO3 additives at equal loading (~15 wt%) in composites, consistent with DFPT predictions.
• Thickness dependence: Voc increased with film thickness (0.2–1 mm) at constant ~10 wt% loading but saturated due to charge collection limits and polarization saturation.
• Control matrix: PMMA alone showed negligible piezo-response; PMMA/BaZrS3 (10 wt%) produced substantial Voc (−62.4 V), confirming BaZrS3 as the dominant piezoelectric contributor.
• Non-piezo effects: Ab initio analysis indicates minimal flexoelectric contribution; device architecture rules out triboelectricity (no dissimilar tribo-layers or separations).

The study addresses why BaZrS3, a centrosymmetric orthorhombic perovskite, displays strong piezoelectricity while BaZrO3 does not. DFPT calculations and structural analysis reveal that the larger, loosely packed BaZrS3 unit cell contains significant free volume, enabling larger ionic displacements under applied stress that reduce effective symmetry locally and generate a sizeable dipole moment. This displacement-mediated polarization mechanism explains the robust measured d33,eff (~−21.4 pm/V) and aligns with DFPT predictions (~17.65 pC/N). The mechanism generalizes to other chalcogenide perovskites (e.g., CaZrS3, BaHfS3) where even higher dij values are predicted. Experimental PFM, composite device measurements, and body-motion energy harvesting validate practical significance: BaZrS3-based composites outperform PCL composites with other additives reported in literature and surpass BaZrO3 in direct comparisons. Flexoelectric and triboelectric effects are shown to be negligible for the tested configurations, reinforcing the piezoelectric origin of the signals. Collectively, these findings demonstrate a new lead-free, environmentally benign class of piezoelectrics with strong potential for wearable and flexible energy harvesting.

The work establishes that lead-free chalcogenide perovskites exhibit piezoelectricity, contrary to expectations for centrosymmetric structures. A displacement-enabled symmetry reduction mechanism in loosely packed unit cells explains the effect and is supported by DFPT and experimental PFM on BaZrS3. Practical devices using BaZrS3–polymer composites harvest vibrational energy from common human motions to charge capacitors and power LEDs, outperforming comparable PCL-based systems reported previously. The materials are lead-free, environmentally benign, and relatively earth-abundant, making them attractive for wearable and biomedical applications. Future research should explore defect engineering to intentionally break local symmetry and further amplify piezoelectric response, extend experimental validation across the broader chalcogenide perovskite family, and optimize composite architectures (particle dispersion, interfaces, electrode design) for improved power density and durability.

• PFM measurement limitations include highly localized fields under AFM tips; while mitigated by a thin Pt top coat, nanoscale non-uniformities can introduce uncertainties in extracted d33,eff.
• Composite performance decreases at high BaZrS3 loadings (>15 wt%) due to particle agglomeration, weak particle–matrix interfaces, debonding, and void formation, reducing effective electromechanical coupling.
• Thickness scaling of devices shows saturation behavior, likely due to limited charge collection across thicker films and polarization saturation effects.
• Experimental verification focused primarily on BaZrS3; other chalcogenide perovskites were assessed via DFPT but not comprehensively validated experimentally.
• DFPT approach relies on artificially induced small ionic displacements to probe symmetry-broken configurations; while informative, it does not represent the relaxed ground state under zero load.
• Tests were conducted under controlled lab conditions (e.g., pneumatic loading, specific resistive loads); real-world variability and long-term environmental stability in devices were not fully explored.