I. Core Research Summary

• Strain-induced second-order ferroelectric phase transition significantly enhances the electrocaloric effect.

A 1% biaxial uniform strain in the film plane induces a second-order ferroelectric phase transition at 243 K; over an ultra-wide temperature range of 100–300 K, it exhibits a reversible electrocaloric response, with a peak isothermal entropy change of 1.8 kJ·K⁻¹·m⁻³—about one order of magnitude higher than that of bulk STO under the same electric field—and, under low‑field conditions, the enhancement can reach up to 339 times.

• Near-perfect reversible electrothermal properties under a unipolar electric field

Addressing the drawback of large hysteresis losses in first-order phase-change materials, this strain‑induced STO thin film exhibits no significant hysteresis in its polarization–electric field curves under both unipolar and bipolar cycling, shows no performance degradation upon repeated charge–discharge cycles, and demonstrates excellent long-term stability in refrigeration cycles.

• High-precision micrometer-scale interdigitated electrode fabrication for aerosols

The direct-printed silver electrodes have a line width of 25 μm, an inter-electrode gap of 50 μm, and a film thickness of 2 μm; after low‑temperature annealing at 200 °C, the conductivity is uniform. Electromagnetic field simulations confirm that the in‑plane field nonuniformity is only 3%, meeting the requirements for high‑precision dielectric and electrothermal quantitative measurements.

• High-crystallinity, defect-free epitaxial thin-film system

XRD and STEM analyses confirm the film’s perfect epitaxial growth, with a dislocation-free interface and high interlayer lattice matching. The dielectric loss is only 2%–4%, significantly lower than that of conventional relaxed STO films, thereby reducing electrothermal energy losses.

• Correcting the Landau theory to achieve precise agreement with experimental data.

By modifying the classical Landau model to account for a 1% biaxial strain, and aligning it with the in‑plane zero‑field polarization characteristics of the <100> orientation, this approach can fully reproduce the experimental trends across dielectric, polarization, and electrothermal dimensions, thereby providing a theoretical framework for the design of strained perovskite electrothermal materials.

II. Complete Preparation Procedure for the Interdigitated Electrode of Aerosol Jet Printing

(1) Complete Preparation Process

• Fine polishing of DyScO₃ single-crystal substrates, followed by plasma cleaning to remove surface impurities;
• Clariant silver nanoparticle ink is diluted with deionized water at a 1:1 ratio, and ultrasonic dispersion is employed to eliminate agglomeration.
• AJ200 ultrasonic atomization, sheath‑gas focusing of the beam, 150 μm nozzle, and uniform direct printing of a 25 μm‑wide periodic interdigitated electrode.
• Anneal in a 200°C constant-temperature oven for 3 hours to completely remove the surfactant and enhance the electrical conductivity of the silver film.
• Optical microscopy and impedance analysis were used to characterize the electrode morphology and electrical conductivity uniformity, while a low-temperature vacuum probe station was employed for electrothermal testing.

(II) Five Core Advantages of Aerosol-Based Processes (Comparison with Lithography and Electron-Beam Evaporation)

• Maskless digital direct printing eliminates the entire lithography process.

No need for adhesive coating, exposure, development, stripping, or metal deposition; CAD allows direct modification of finger‑line width, spacing, and array period, reducing sample‑iteration time by 80%.

• Non-destructive, non-contact deposition on single-crystal hard substrates

A constant printing gap of 3–5 mm, with no mechanical pressure, prevents scratching of brittle DSO and STO single-crystal thin films, thereby ensuring the integrity of the epitaxial layer’s crystalline quality.

• Micrometer-scale, high-precision electrode controllable shaping

A stable fabrication process yields 25-μm‑diameter fine wires and 50-μm‑wide narrow gaps in interdigitated electrodes, with well‑defined electrode edge profiles and uniform electric field distribution, thereby meeting the requirements for precise electrothermal quantitative testing.

• Low-temperature annealing process that does not disrupt the thin-film strain architecture.

Post‑annealing at only 200°C does not trigger stress relaxation in the STO thin film, thereby fully preserving the enhanced electrothermal performance arising from a 1% biaxial strain.

• Significant savings on precious metal consumables

Only the test area is subjected to localized silver‑line deposition; lithography of the entire wafer followed by metal evaporation results in substantial material waste, whereas conducting mechanism studies on small batches significantly reduces costs.

III. Five Key Technological Innovations

Innovation 1: A New Paradigm for Fabricating Micrometer-Scale Interdigitated Electrodes Compatible with Perovskite Single Crystals

Innovation 2: A Novel System Featuring a Wide-Temperature-Range, Highly Reversible External Electrothermal Effect Induced by Strain

Innovation 3: Quantitative elucidation of the dual electric‑thermal enhancement mechanism under high and low electric fields.

Innovation 4: Adaptation of the Double-Axis Strain STO-Modified Landau Physical Model

Innovative 5: Low-Loss Single-Crystal Thin Film + Aerosol Precision Testing Integrated Solution

IV. Current Industry Challenges in the Process and Corresponding Breakthrough Solutions

Challenge 1: High-temperature processing in lithography and evaporation deposition can lead to strain relaxation in STO thin films, resulting in degradation of their electrothermal performance.

Challenge 2: The conventional lithography process is cumbersome, with each batch of sample preparation taking several days, resulting in slow iterative refinement of the underlying mechanisms.

Challenge 3: Brittle single-crystal substrates are easily scratched during lithography and film‑bonding processes, leading to defects in the thin film.

Challenge 4: In conventional bulk SrTiO₃, the electrothermal entropy change is very small, rendering it of no practical cooling value.

Challenge 5: Conventional first-order phase-transition ferroelectric materials exhibit severe electric-field‑induced hysteresis and substantial long-term cooling losses.

V. Future Diversified Industrial Application Scenarios

1. Miniature fluorine-free solid-state cooling chip

2 Aerospace Lightweight Cryogenic Temperature-Control Components

3 High-precision low-temperature sensor substrate temperature control

4 Flexible Strain-Responsive Electrothermal Thin-Film Devices

5. Perovskite Functional Materials Research and Development Platform for Universities

Figure 1 | A strained STO (001) thin film deposited on a DSO (110) substrate. a, Cubic unit-cell structure of STO; b, Orthorhombic Pbnm unit-cell structure of DSO; c, Distribution of the strained STO film within a pseudo‑cubic DSO unit cell (with the upper half of the rotated ScO₆ octahedra distorted to maintain structural continuity); d, Four pseudo‑cubic unit cells of the interfacial STO spanning half a DSO unit cell (orange triangles denote oblique unit cells). The in‑plane lattice parameters of STO are expanded by approximately 1% through epitaxial growth, with a ≈ 3.943 Å and bpc ≈ 3.944 Å; e, Schematic illustration of an STO film about 60 nm thick and a DSO substrate 0.5 mm thick, not drawn to scale. Interdigitated silver electrodes enable the electric field to be applied along the STO [100] plane, parallel to the DSO [110] direction, across a 50 μm gap.

Figure 2 | XRD measurement results for STO//DSO. a, 2θ–ω scan (counts per second, c.p.s.). The STO 002 peak indicates an out-of-plane lattice parameter of cpc = 3.884 ± 0.001 Å; Laue fringes reveal a film thickness of 58.2 ± 0.5 nm. b, STO ϕ‑scan acquired at 2θ = 77.48°. c, ϕ‑scans of DSO at the following angles: 2θ = 76.35° (range –45° ≤ ϕ ≤ 45°, corresponding to the 332° diffraction peak), 75.08° (range 45° < ϕ ≤ 135°, corresponding to the 240° diffraction peak), 76.35° (range 135° < ϕ ≤ 225°, corresponding to the 332° diffraction peak), and 77.71° (range 225° < ϕ ≤ 315°, corresponding to the 420° diffraction peak). d and e, reciprocal space maps (intensity scale shown in e) indicate that most of the film is under full strain, with in-plane lattice parameters apc = 3.944 ± 0.001 Å (d) and bpc = 3.943 ± 0.001 Å (e) (Qx, Qy, and Qz correspond to reciprocal lattice spacings along the Cartesian axes, with the z‑axis oriented perpendicular to the plane). Black dots are calculated based on the following parameters: the pseudo‑cubic lattice parameters of the STO film (see inset in d), the substrate lattice parameters of DSO (ao = 5.440 Å, bo = 5.713 Å, co = 7.887 Å), and the bulk STO lattice parameters (ac = 3.905 Å). Dashed lines connect the black dots representing the STO film and bulk STO. Data for Sample 1 were collected prior to IDE deposition.

Figure 3 | Surface and cross-sectional images of STO//DSO. a–c: Atomic step structures with an average step height of d = 0.39 nm, observed via 1 μm × 1 μm atomic force microscopy (AFM) images (a), the corresponding profile along the white line in (a) (b), and the height distribution density function ρH (c); d–g: Visualized images of the ScO₂–strontium oxide interface (indicated by arrows), with color coding identical to that in (d); d and e: HAADF‑STEM (d) and ABF‑STEM (e) images acquired along the STO [100] crystal plane and the DSO [110] crystal axis, respectively; the inset in (d) shows an enlarged view with superimposed colored atomic labels; oxygen atoms are clearly visible in (e); f: EDXS spectrum at atomic resolution; g: EELS spectra at atomic resolution for the Sc L₂,₃ edge (orange) and the Ti L₂,₃ edge (pink). Large‑area AFM, HAADF‑STEM, and ABF‑STEM images are provided in Supplementary Note 6; AFM data for pristine sample 1 prior to IDE deposition and STEM data for sample 3 can be found in the respective sections.

Figure 4 | In-plane dielectric response of the stressed STO thin film. a: At 260 K, the dielectric hysteresis in the isothermal εfilm(f) response—obtained from 1,614 frequency sweeps under zero bias—is indicated by orange (black) for the positive‑field and negative‑field branches, respectively; εfilm(T,f). c, d: εfilm(E,T) curves constructed from all 4,375 temperature values T and 200 frequency values f, using bipolar εfilm(E) data; the transition temperature TC ≈ 243 K (indicated by arrows) was determined from the peak at the intermediate frequency, corresponding to the positive‑field (c) and negative‑field (d) scans. b–e: Plots derived from isothermal εfilm(T) curves measured under a positive bias field; these are analogous to the εfilm(E) curves obtained from 200 data points recorded during bias‑field sweeps at 106 Hz in isofield measurements (see Supplementary Figure 23e). All data were acquired at f ≈ 1 Hz; b: Eight of the 4,375 measured temperatures yield bipolar εfilm(E) curves that reveal a weak cooling effect; data are based on measurements of Sample 1 on DSO substrate 1.

Figure 5 | Electro‑polarization characteristics of the stressed STO thin film. a, At 85 K, the bipolar Pfilm(E) curves as a function of electric field strength E(t) (inset; t denotes time), with E increasing by 1 kV·cm⁻¹ in each 10‑ms cycle; b, Bipolar Pfilm(E) curves measured at 100 K for selected frequencies f; c and d, for seven temperature points out of 668, the bipolar Pfilm(E) curves (c) and the unipolar Pfilm(E) curves (d), each comprising 1,000 and 500 data points, respectively, obtained using the same field‑sweep rate (f = 102 Hz in c; effective frequency feff = 102 Hz in d; see Supplementary Note 12); e, |Pfilm(T, E)| values corresponding to the field‑shift branches removed from all unipolar Pfilm(E) plots; f, Cross‑sectional plots of isofield Pfilm(T) along the e curves under zero field and selected positive fields. The data in a and b are measurements taken after zero‑field cooling; those in c–e were acquired during heating. All data are based on experiments performed on Sample 1 and DSO substrate 1.

Figure 6 | Charge‑coupling effects in stressed STO thin films. a, |ΔSfilm(T, E)| = |∫E0(∂Pfilm/∂T)E′dE′| (adapted from Fig. 5e). b, at a selected positive electric field strength, four isofield lines of |ΔSfilm(T)| pass through point a; corresponding data are shown for the bulk STO sample (Supplementary Fig. 18b). c, an isothermal line of |ΔSfilm(E)| passes through point a at a chosen temperature. d, the charge‑coupling magnitude |ΔSfilm(T, E)/E|. e, in the non‑dashed region (dashed region), |ΔSfilm(T, E)/ΔSbulk(T, E)| is greater than (less than) 1; the interpolated data span a range of 1 K and 0.02 kV·cm⁻¹. f, the isofield line passing through point e at a positive electric field strength corresponds to the same ratio as the isofield line in Fig. b (the inset shows the trajectory at 0.2 kV·cm⁻¹). Data are based on measurements from Sample 1, DSO substrate 1, and a single bulk STO sample (Supplementary Note 10).