Accessing the intrinsic polarization-switching time and switching kinetics of thin-film ferroelectric capacitors below circuit-limited timescales

A 2004 photoconductive-switch experiment (Li, Nagaraj, Liang, Cao, Lee and Ramesh, Appl. Phys. Lett. 84, 1174 (2004)) drove 200-nm Pb(Nb,Zr,Ti)O3 thin-film capacitors with jitter-free, step-like electrical pulses of 68 ps rise time and, from fits of the switching transients with nucleation-and-growth models, extracted a characteristic switching time of roughly 70-90 ps — yet found no saturation of this time constant down to the smallest achieved pulse rise times (~50 ps), leaving an ambiguous fitted switching dimensionality of 3 and unresolved multistage-kinetics questions. The open problem is to measure and physically interpret polarization switching in thin-film ferroelectric capacitors in the regime where the material, not the measurement circuit, sets the timescale: demonstrate (or refute) saturation of the intrinsic characteristic switching time once both the input-pulse rise time and the device RC time constant are pushed below it, settle the physical origin of the fitted switching dimensionality, and establish whether finite-size-corrected multistage kinetics occur. Two decades of follow-up work has substantially advanced the enabling hardware and reframed the circuit-convolution issue (circuit-limited versus material-limited switching, nucleation-limited switching models, circuit-aware kinetic extraction), but the specific saturation demonstration and the mechanistic interpretation on the original perovskite-film system remain unreported.

Background

A ferroelectric is a crystalline material with a spontaneously polarized unit cell whose polarization can be reversed between two stable states by an applied electric field. This reversal, called polarization switching, is the operating principle of ferroelectric random access memories (FeRAM), where a capacitor stores a bit as the polarization direction; the write speed of such a memory is set by how fast the polarization can actually be switched. Since Merz's classic work, the switching time of a ferroelectric is known to decrease exponentially with applied field (the Merz law), and the microscopic picture is nucleation of small reverse domains followed by growth of their boundaries (domain walls). The standard quantitative description is the Kolmogorov-Avrami-Ishibashi (KAI) model, adapted for ferroelectrics by Ishibashi: the switched fraction of the polarization follows $f(t)=1-\exp\{-[\exp(-\alpha/E)(t/t_{0})^{n}]\}$, where $E$ is the field, $\alpha$ an activation-field parameter, $t_{0}$ the characteristic (intrinsic) switching time, and $n$ the Avrami exponent, an effective domain-growth dimensionality that is larger by 1 when nucleation continues at a constant rate during growth ('continuous nucleation') than when all nuclei pre-exist at the start (in Ishibashi's notation these two scenarios are the $\alpha$ and $\beta$ models, so a given $n$ can mean 2D growth with continuous nucleation or 3D growth from pre-existing nuclei). Shur and co-workers extended such theories with finite-size corrections, in which growth stops when domains impinge after a characteristic time $t_{m}$, which can produce multistage (multi-slope) switching transients. For polycrystalline films, Tagantsev and co-workers and Jo and co-workers later showed that switching transients spread over many decades in time and are better described by nucleation-limited switching (NLS), in which a distribution of local waiting times for nucleation, grain by grain, replaces a single Avrami exponent.

By 2003, the fastest directly measured switching times in PZT-family (PZT = Pb(Zr,Ti)O3, the archetypal perovskite ferroelectric) thin-film capacitors were of order a few hundred picoseconds, whereas nucleation-and-growth estimates suggested intrinsic times of order tens of picoseconds. The gap was widely suspected to be an artifact of the measurement itself: standard pulse generators have nanosecond-scale rise times, transmission lines disperse and reflect, and every real capacitor has a parasitic $RC$ time constant (the product of its series resistance and capacitance) that limits how fast charge can be delivered, so the observed current transient convolves circuit response with the ferroelectric kinetics. Li et al. (2004) attacked this with a semiconductor photoconductive switch — a silicon device that becomes conducting when illuminated by a femtosecond laser, generating jitter-free, step-like voltage pulses with rise times tunable down to about 50 ps — delivered over a matched 50-ohm coplanar transmission line to 200-nm-thick Pt/LSCO/PNZT/LSCO/Pt capacitors of a few micrometers in lateral size (LSCO is lanthanum strontium cobaltite, a conductive oxide electrode, and PNZT is the Pb(Nb,Zr,Ti)O3 ferroelectric layer). They measured a switching time of about 220 ps under a 5 V, 68 ps-rise-time pulse, and fits to the Merz-Ishibashi and Merz-Shur model forms gave $t_{0}\approx70-90$ ps with an integer-constrained dimensionality $n=3$. Crucially, $t_{0}$ extracted this way kept increasing with input-pulse rise time, with no sign of saturating at the smallest rise times achieved, and the estimated device $RC\approx45$ ps was itself comparable to $t_{0}$ — so even this state-of-the-art experiment could not certify that the fitted $t_{0}$ is the true intrinsic material time, nor could the polycrystalline, parasitic-convoluted data distinguish the $\alpha$-model from the $\beta$-model interpretation of $n=3$, while the finite-size (Shur-type) multistage form did not fit the data well. The problem is to reach and interpret the regime in which the intrinsic ferroelectric kinetics, rather than the pulse rise time or the circuit $RC$, is the slowest element in the measurement — a question that later work has rephrased as distinguishing circuit-limited from material-limited switching (e.g. Parsonnet et al. 2020 on the multiferroic BiFeO3; Chiang et al. 2025 on nanoscale capacitors of HZO (Hf0.5Zr0.5O2, the CMOS-compatible hafnium zirconium oxide ferroelectric) and La-substituted BiFeO3; Liang, Kezer and Heron 2026 on circuit-distorted voltage waveforms and the reliability of the extracted Avrami exponent).

Problem Statement

Determine the intrinsic polarization-switching kinetics of thin-film ferroelectric capacitors in the regime where the material, not the measurement circuit, sets the timescale, as left open by Li et al. (Appl. Phys. Lett. 84, 1174-1176 (2004)). Specifically: (1) Measure the true intrinsic characteristic switching time $t_{0}$ of ferroelectric thin-film capacitors (fitted value in the source work: about 70-90 ps for 200-nm LSCO/PNZT/LSCO capacitors) and establish whether it saturates — reaches a lower bound that no longer decreases — when the input-pulse rise time is reduced below the smallest value achieved in the source work (about 50 ps), which requires pulses with 10-90% rise times and device $RC$ time constants both smaller than the intrinsic $t_{0}$. (2) Determine the physical origin of the fitted switching dimensionality $n=3$: whether it corresponds to the $\alpha$ model (continuous nucleation during growth, two-dimensional domain growth) or the $\beta$ model (pre-existing nuclei, three-dimensional growth). (3) Establish whether multistage switching kinetics, as predicted by finite-size-corrected nucleation-and-growth theories of the Shur type (with a domain-impingement time), occur in these films under suitable pulse and device conditions, given that the source data were not well fitted by the multistage form. (4) Realize the device geometries and pulse-delivery hardware (matched, low-dispersion transmission with rise time $\le t_{0}$ and $RC\le t_{0}$) that make such intrinsic-regime measurements possible, and apply the same approach across additional ferroelectric compositions. All four components serve one overarching objective: observe and physically interpret ferroelectric polarization switching in thin-film capacitors free of circuit convolution, so that the intrinsic speed limit and the nucleation-and-growth mechanism of the switching transient can be read unambiguously from the measurement.

References

1. J. Li, B. Nagaraj, H. Liang, W. Cao, Chi. H. Lee, R. Ramesh, 'Ultrafast polarization switching in thin-film ferroelectrics', Applied Physics Letters 84, 1174-1176 (2004). DOI: 10.1063/1.1644917, https://doi.org/10.1063/1.1644917
2. Y. Ishibashi, 'A Model of Polarization Reversal in Ferroelectrics', Journal of the Physical Society of Japan 59, 4148-4154 (1990). DOI: 10.1143/jpsj.59.4148, https://doi.org/10.1143/jpsj.59.4148
3. V. Sh. Shur, E. L. Rumyantsev, 'Kinetics of ferroelectric domain structure during switching: Theory and experiment', Ferroelectrics 151, 171-180 (1994). DOI: 10.1080/00150199408244739, https://doi.org/10.1080/00150199408244739
4. A. K. Tagantsev, I. Stolichnov, N. Setter, J. S. Cross, M. Tsukada, 'Non-Kolmogorov-Avrami switching kinetics in ferroelectric thin films', Physical Review B 66, 214109 (2002). DOI: 10.1103/PhysRevB.66.214109, https://doi.org/10.1103/PhysRevB.66.214109
5. J. Y. Jo, H. S. Han, J.-G. Yoon, T. K. Song, S.-H. Kim, T. W. Noh, 'Domain Switching Kinetics in Disordered Ferroelectric Thin Films', Physical Review Letters 99, 267602 (2007). DOI: 10.1103/PhysRevLett.99.267602, https://doi.org/10.1103/PhysRevLett.99.267602
6. A. Grigoriev, M. B. Azad, J. McCampbell, 'Ultrafast electrical measurements of polarization dynamics in ferroelectric thin-film capacitors', Review of Scientific Instruments 82, 124704 (2011). DOI: 10.1063/1.3665209, https://doi.org/10.1063/1.3665209
7. N. Gong, X. Sun, H. Jiang, K. S. Chang-Liao, Q. Xia, T. P. Ma, 'Nucleation limited switching (NLS) model for HfO2-based metal-ferroelectric-metal (MFM) capacitors: Switching kinetics and retention characteristics', Applied Physics Letters 112, 262903 (2018). DOI: 10.1063/1.5010207, https://doi.org/10.1063/1.5010207
8. M. Si, X. Lyu, P. R. Shrestha, X. Sun, H. Wang, K. P. Cheung, P. D. Ye, 'Ultrafast measurements of polarization switching dynamics on ferroelectric and anti-ferroelectric hafnium zirconium oxide', Applied Physics Letters 115, 072107 (2019). DOI: 10.1063/1.5098786, https://doi.org/10.1063/1.5098786
9. X. Lyu, M. Si, P. R. Shrestha, K. P. Cheung, P. D. Ye, 'First Direct Measurement of Sub-Nanosecond Polarization Switching in Ferroelectric Hafnium Zirconium Oxide', 2019 IEEE International Electron Devices Meeting (IEDM), 15.2.1-15.2.4 (2019). DOI: 10.1109/iedm19573.2019.8993509, https://doi.org/10.1109/iedm19573.2019.8993509
10. R. Mankowsky, A. von Hoegen, M. Först, A. Cavalleri, 'Ultrafast Reversal of the Ferroelectric Polarization', Physical Review Letters 118, 197601 (2017). DOI: 10.1103/PhysRevLett.118.197601, https://doi.org/10.1103/PhysRevLett.118.197601
11. E. Parsonnet, Y.-L. Huang, T. Gosavi, A. Qualls, D. Nikonov, C.-C. Lin, I. Young, J. Bokor, L. W. Martin, R. Ramesh, 'Toward Intrinsic Ferroelectric Switching in Multiferroic BiFeO3', Physical Review Letters 125, 067601 (2020). DOI: 10.1103/PhysRevLett.125.067601, https://doi.org/10.1103/PhysRevLett.125.067601
12. C. Wang, H. Qiao, Y. Kim, 'Perspective on the switching behavior of HfO2-based ferroelectrics', Journal of Applied Physics 129, 010902 (2021). DOI: 10.1063/5.0035652, https://doi.org/10.1063/5.0035652
13. E. Mishina, V. Bilyk, N. Sherstyuk, A. Sigov, 'Ferroelectric switching by (sub)-picosecond electromagnetic pulse', Ferroelectrics 577, 1-12 (2021). DOI: 10.1080/00150193.2021.1916345, https://doi.org/10.1080/00150193.2021.1916345
14. T. Yang, B. Wang, J.-M. Hu, L.-Q. Chen, 'Domain Dynamics under Ultrafast Electric-Field Pulses', Physical Review Letters 124, 107601 (2020). DOI: 10.1103/PhysRevLett.124.107601, https://doi.org/10.1103/PhysRevLett.124.107601
15. H. Tamim, R. Khadka, P. Keblinski, 'Limits of high-frequency polarization switching in BaTiO3', Applied Physics Letters 125, 102905 (2024). DOI: 10.1063/5.0218240, https://doi.org/10.1063/5.0218240
16. T. Chiang, J. J. Plombon, M. K. Lenox, I. Mercer, P. Debashis, Mahendra DC, S. Trolier-McKinstry, J.-P. Maria, J. F. Ihlefeld, I. A. Young, J. T. Heron, 'Material-Limited Switching in Nanoscale Ferroelectrics', arXiv:2507.12353 (2025). arXiv ID: 2507.12353, https://arxiv.org/abs/2507.12353
17. Y. Liang, P. Kezer, J. T. Heron, 'Historical Foundation and Practical Guideline for Ferroelectric Switching Kinetic Studies', arXiv:2604.05328 (2026). arXiv ID: 2604.05328, https://arxiv.org/abs/2604.05328
18. Y. Liang, S. Kim, T. Chiang, M. K. Lenox, I. Mercer, J. J. Plombon, J.-P. Maria, J. F. Ihlefeld, W. Sun, W. Lu, J. T. Heron, 'Ferroelectric dynamic-field-driven nucleation and growth model for predictive materials-to-circuit co-design', arXiv:2602.02957 (2026). arXiv ID: 2602.02957, https://arxiv.org/abs/2602.02957