Thickness and strain dependence of piezoelectric coefficient in BaTiO3 thin films
Summary¶
Piezoresponse force microscopy (PFM) with interferometric displacement sensing (IDS) is used to quantify converse piezoelectric coefficients (d₃₃) in epitaxial BaTiO₃ films on (001) SrTiO₃ as thickness varies from ~10 nm to ~80 nm. Experiments show a strong thickness-driven drop in d₃₃ for ultrathin films. ReaxFF-based MD with an ab initio–derived BaTiO₃ parameterization explains the trend via thickness- and strain-dependent evolution of depolarization-field screening regions under applied electric fields. The study connects device-relevant ultrathin ferroelectric metrics to atomistic screening physics that are difficult to sample with DFT alone at similar length scales.
Methods¶
Experiment (epitaxial BTO, PFM/IDS). BaTiO₃ (BTO) films are grown by pulsed laser deposition on (001) SrTiO₃ with a ~5 nm SrRuO₃ bottom electrode (thickness series 10–80 nm in the abstract / intro). AFM and X-ray reciprocal-space mapping characterize morphology and in-plane strain vs thickness. Piezoresponse is mapped with band-excitation PFM; quantitative d₃₃ use interferometric displacement sensing (IDS) to limit spurious electrostatic contamination of the mechanical response (see Phys. Rev. Materials article and Supplement).
1 — MD application (ReaxFF on BTO under E). The paper uses ReaxFF-based MD with a recent fully ab initio–derived BTO parameter set in the ReaxFF form to study thickness- and strain-dependent depolarization-field screening under applied electric fields (context in abstract and Sec. I of the PDF). N/A — this note does not transcribe LAMMPS-level input (ensemble, timestep, supercell thickness, number of unit cells, PBC, thermostat, barostat); take those from the main text and SI of pdf_path. Duration / production length in ps or ns for the ReaxFF runs is specified in the article/SI (not tabulated in this wiki). Target temperature in K for the MD stages is in the source (not copied here). Barostat / pressure for MD: N/A or stated in PDF for any NPT stages—confirm in source. External electric field in MD: yes in the sense of the manuscript (applied E to probe ferroelectric response / screening), not N/A. Shear / shock / enhanced sampling: N/A — not the stated focus in the abstract-level summary here.
2 — Force-field training. N/A — the BTO ReaxFF is adopted as a prior ab initio–derived set; this article is not a new ReaxFF fit (it cites that parameterization).
3 — Static QM / DFT-only. N/A — the contribution is experiment + ReaxFF dynamics for screening physics, not a new static DFT benchmark in this article as the lead result.
Findings¶
Outcomes, comparisons, and mechanisms. IDS-based d₃₃ is about 20.5 pm/V for the ~80 nm film and falls to under ~2 pm/V for the ~10 nm film in the same epitaxial series (values as stated in the Phys. Rev. Materials abstract and body). PFM and Strain/structure ( RSM ) track higher in-plane compressive strain in thinner BTO on this stack, co-plotted with piezoresponse. ReaxFF MD is reported to show a thicker screening region of the depolarization field in ultrathin BTO under compressive in-plane strain when an E-field is applied, i.e. interfacial / screening physics is tied to the d₃₃ trend (not a single intrinsic coefficient only).
Sensitivity, limitations, corpus honesty. d₃₃ at the ultrathin extreme is gated by how E-induced bound charge is screened ( thickness- and strain-dependent in the ReaxFF story). Classical ReaxFF does not recover full DFT--level electronic polarization; PFM/IDS uncertainties and tip-field effects stay in the measured-data discussion of the source.
Limitations¶
Classical reactive MD cannot capture full electronic polarization changes at the DFT level; quantitative agreement relies on the fitted ReaxFF surface/screening physics encoded in the parameterization. PFM quantification also depends on tip calibration and electrostatic artifacts; the article’s IDS protocol is intended to mitigate spurious piezoresponse, but cross-lab absolute d₃₃ comparisons still require careful metadata. Strain maps from X-ray reciprocal-space mapping should be co-registered with film thickness when interpreting MD screening thickness trends. Electrode geometry in PFM may couple to domain patterns not represented in periodic MD cells.
Relevance to group¶
Joint ORNL + PSU (van Duin) effort tying ferroelectric thin-film measurements to ReaxFF modeling of screening and strain—a template for perovskite functional property studies in the wiki.