Nanoscale Doping and Its Impact on the Ferroelectric and Piezoelectric Properties of Hf0.5Zr0.5O2

Abstract

Ferroelectric hafnium oxide thin films—the most promising materials in microelectronics’ non-volatile memory—exhibit both unconventional ferroelectricity and unconventional piezoelectricity. Their exact origin remains controversial, and the relationship between ferroelectric and piezoelectric properties remains unclear. We introduce a new method to investigate this issue, which consists in a local controlled modification of the ferroelectric and piezoelectric properties within a single Hf0.5Zr0.5O2 capacitor device through local doping and a further comparative nanoscopic analysis of the modified regions. By comparing the ferroelectric properties of Ga-doped Hf0.5Zr0.5O2 thin films with the results of piezoresponse force microscopy and their simulation, as well as with the results of in situ synchrotron X-ray microdiffractometry, we demonstrate that, depending on the doping concentration, ferroelectric Hf0.5Zr0.5O2 has either a negative or a positive longitudinal piezoelectric coefficient, and its maximal value is −0.3 pm/V. This is several hundreds or thousands of times less than those of classical ferroelectrics. These changes in piezoelectric properties are accompanied by either improved or decreased remnant polarization, as well as partial or complete domain switching. We conclude that various ferroelectric and piezoelectric properties, and the relationships between them, can be designed for Hf0.5Zr0.5O2 via oxygen vacancies and mechanical-strain engineering, e.g., by doping ferroelectric films.

Synergetic contributions of chemical doping and epitaxial stress to polarization in ferroelectric HfO2 films

“…2 (d,f) show similar piezoresponse ≈20 pm/V in good agreement with ab-initio studies [32]. The fact that the images show similar piezoresponse despite of the different amount of orthorhombic phase, except some contrast resulting from charging effects, denotes that the presence of monoclinic grains does not contribute significantly to the overall signal due to its small size compared with the tip diameter [33]. This is not the case of the film grown on YAlO 3 , neither of similar HZO or polycrystalline samples most probably because of the larger monoclinic phase amount [11,17].…”

Quantification of the Electromechanical Measurements by Piezoresponse Force Microscopy

“…Hence, PFM‐based investigations of the sign of the piezoelectric coefficient require proper calibration of the PFM phase signal and careful selection of the measurement conditions. This is particularly relevant for investigation of the piezoelectric properties of the hafnia‐based ferroelectrics, which have been the subject of significant controversy: while theoretical modeling predicts and some experimental studies observe a negative d 33,eff coefficient, [ 76–78,84 ] majority of the experimental works report a positive d 33,eff in these materials. [ 85–88 ]…”

“…Hence, PFM-based investigations of the sign of the piezoelectric coefficient require proper calibration of the PFM phase signal and careful selection of the measurement conditions. This is particularly relevant for investigation of the piezoelectric properties of the hafnia-based ferroelectrics, which have been the subject of significant controversy: while theoretical modeling predicts and some experimental studies observe a negative d 33,eff coefficient, [76][77][78]84] majority of the experimental works report a positive d 33,eff in these materials. [85][86][87][88] To address these issues, we have adopted two approaches to correct the parasitic phase offset typically present in the cantilever displacement measurements by the OBD method: The first approach uses a reference sample with a known d 33,eff sign, and the second approach involves the detection of the cantilever-sample electrostatic interactions to determine the parasitic phase offset.…”

Giant Electromechanical Effect in Piezoelectric Nanomembranes Based on Hf0.5Zr0.5O2

“…As was previously revealed by synchrotron Xray diffraction, the HZO film predominantly consisted of a ferroelectric orthorhombic Pca2 1 structural phase with a negligibly small fraction of non-polar monoclinic P2 1 /c phase. 14 Every capacitor structure contains four areas with bridge thin-film membranes oriented perpendicular to each other (Figure 2a). Each area includes bridge membranes with a lateral size of the membranes in the range from 80 to 1200 nm.…”

“…Thus, the electromechanical response on the membranes improves by up to 10 times compared to HZO on a rigid substrate. The sign “–” is selected based on the previous results of the measurement of the piezoelectric coefficient of HZO by in situ synchrotron X-ray microdiffractometry on similar capacitors …”