Effect of High Heat Flux of Helium and Hydrogen Plasma Jet on the Material Properties of Piezoelectric PZT-Ceramics

Abstract

A set of experimental and measurement techniques to study the influence of a plasma jet on the main material parameters of piezoelectric ceramics has been presented. A series of plasma experiments has been carried out using a pulsed plasma jet system. It allows of a metered-dose exposure to plasma of different composition and fluence with a constant particle flux density of 10²¹/m², energy flux density of 0.1 MJ/m² and average particle energy of 100–200 eV in a pulse duration of 15 μs. The study of the effects that a repeated exposure to an extreme heat flux of helium and hydrogen plasmas has on the near-surface layer structure and basic material parameters of mass-produced piezoelectric ceramic samples has been presented. The main result of the research is an experimental confirmation of the surface micro-structuring starting after just a few cycles of plasma exposure while only a slight decrease of the main material parameters as well as the preservation of polarization has been observed for two types of different compositions of PZT-ceramics. A further increase in the number of exposure pulses leads to practically no change of main material parameters of both ceramics, even showing a tendency for recovery instead.

1. Introduction

Irradiation with a gas plasma flow is an effective method of applying a controlled dosed effect on the structure of various materials. It significantly expands experimental possibilities for studying their physical properties in terms of resistance to extreme external influences. In recent decades, the vast majority of research in the field of plasma interaction with materials has been aiming for the search of materials resistant to extreme thermal and radiation loads during long-term operation in fusion reactors. However, the studies related to interaction of plasmas of various compositions, temperatures and energies with solids are not limited solely to the study of the material resistance to plasma’s destructive effects. Scattering and implantation of incident ions, emission of particles from a material under the influence of ion bombardment, as well as changes in the structure and properties of a material near-surface layer due to plasma-induced defects are actively studied. The above-mentioned effects are already used in various technological processes for the formation of a predetermined structure of the surface and its analysis. The predominance of a particular effect depends on the ion’s energy, charge and angle of incidence, the structure and crystallographic orientation of a sample, as well as on the heat load on its surface. With an increase of plasma ion energy to 100–200 eV, the intensity of the sputtering process increases by 3–4 orders of magnitude, ions begin to penetrate into the crystal lattice of the material, and an equilibrium state is established. With a further increase in the kinetic energy of the plasma flow, more volumetric lattice disruptions, as a result of a deeper penetration of ions, become a dominant factor instead. This is also used in various technological processes for micro-structuring the material surface without significantly changing its bulk properties, for example, in plasma cleaning, activation, deposition and etching of the surface, in particular for thermal barrier coating applications [1]. The creation of such coatings is in demand in various industries, since it can potentially improve both the safety of operation and durability of large energy systems and vehicles. In particular, extreme thermal loads that occur during skidding or emergency braking of a railway rolling stock can lead to several kinds of thermo-mechanical defects (sliders, so-called white spots, etc.) appearing on the wheel treading surface, thus necessitating urgent repairs or otherwise risking catastrophic failures [2]. An impressive effect of the combined use of helium and hydrogen gas plasma for hardening metals [3], which opens up wide possibilities for the use of plasma technology in the railway industry, should also be noted.

High-energy mono-kinetic ion flows of varying intensity are also used to form localized defect layers at predetermined depths inside materials with an accuracy of a few nanometers, which is the basis of controlled cleavage (smart-cut) and ion irradiated Deleted oxide Cut (DeleCut) techniques in manufacturing of silicon-on-insulator dislocation-free structures. Such structures are widely used for developing microelectronic components with a high radiation and a thermal resistance [4].

At the same time, similar studies of piezoelectric materials are rare and mainly consider the destructive effects of neutron radiation on the properties of piezoelectric ceramics [5,6]. Such materials are widely used as active elements for ultrasonic transducers, various sensors, electromechanical actuators for micro-displacements and piezoelectric motors capable of operating effectively in harsh conditions, including pulsed plasma jet systems [7,8]. A first study of the hydrogen plasma exposure at different (named weak and strong) thermal and radiation loads (energy flux density of 0.05 MJ/m², particle fluence 10²¹/m² and energy flux density of 0.15 MJ/m², particle fluence 2·10²²/m², respectively) on samples of piezoelectric ceramics CTSNV-1 mass-produced in Russia, was presented in Ref. [9]. The experiments were carried out using the pulsed plasma setup at the Ioffe Institute, developed to simulate plasma flows with energy densities comparable to the Edge-Localized Modes (ELM) events that can deposit high-energy plasma loads onto tokamak-type-reactor walls. The main objective of [9] was to determine the degree of preservation of dielectric, piezoelectric and electroacoustic properties of piezoelectric ceramics after plasma exposure. The preservation of electrical polarization with some reduction in main material parameters was confirmed even under weak plasma exposure. At the same time, increasing the intensity of plasma exposure up to a strong one led to practically no further deterioration of material parameters, which even showed a tendency towards their restoration. It was suggested that a surface layer that increases the radiation and thermal resistance of ceramics is formed even at low doses of plasma exposure. The structural changes of the ceramic surface turned out to be very similar to those previously described in the studies of the surface morphology changes for various tungsten modifications under the influence of hydrogen plasma [10,11,12,13]. Currently, tungsten is the main functional material for the first wall and divertor armor for the International Thermonuclear Experimental Reactor (ITER) [14]. Similar studies of structural changes and related material parameters in piezo- and ferroelectric materials are important for a fundamental understanding of processes occurring in them under extreme external influences as well as for the development of functional elements with an increased radiation and thermal stability.

This paper considers experimental techniques for quantitatively assessing the influence of plasma action parameters on the properties of piezoelectric materials. The study of the effects of the dosed pulses of hydrogen (H₂) and helium (He) plasma jet (energy flux density of 0.1 MJ/m²) on main material parameters (such as dielectric, piezoelectric and elastic constants, as well as the electromechanical coupling coefficient) of piezoelectric ceramics, mass-produced in Russia is presented.

2. Experimental

2.1. Objects Under Study

Solid solutions based on lead zirconate-titanate (PZT) are the basis for piezoelectric ceramics. A number of compositions have been developed, determined by the set of functional characteristics for specific practical applications such as ultrasonic transducers and various sensors, including pressure sensors for plasma setups [15]. The objects under study were commercially available piezoceramic materials CTS-19 and CTSNV-1 (Aurora-ELMA, Ltd., Volgograd, Russia), characterized by high values of piezoelectric constants and electromechanical coupling coefficients, as well as moderate dielectric losses. The use of such compositions as active piezoelectric elements in diagnostic and protective equipment for ITER is discussed in [16]. This is an additional incentive to study the stability of the piezoelectric material parameters, primarily against the effects of H₂- and He- plasma of varying energy and particle density, as well as to analyze the possibilities of restoring the performance characteristics of devices operating on their basis.

The CTSNV-1 sample intended for the study was a poled disk of 67 mm diameter and 3 mm thickness with burnt Ag electrodes. For the measurements, it was cut to get square samples of approximately 10 × 10 mm² in size. One of them was kept as a reference (the virgin sample), while the others were exposed to various doses of H₂- and He-plasma radiation. The CTS-19 samples were commercially available poled disks (10 mm diameter and 1 mm thickness) with burnt Ag electrodes. Prior to irradiation, the electrodes were grinded off from one surface of all samples.

The exact composition of the ceramics is know-how and not reported by the manufacturer. The principal difference is the ratio of molar concentrations (Zr/Ti ratio), Zr/Ti = 58/42 for CTS-19 samples, and Zr/Ti = 53/47 for CTSNV-1 samples.

The poling of the samples was performed by the manufacturer in accordance with the technical conditions of the company. The manufacturer also pre-aged the samples.

To ensure identical experimental conditions a fine silver paste was applied, before and after plasma exposure, to the surface of the samples, previously cleaned from the original burned-in silver electrodes. At the initial stage of our experimental cycle, measurements of the key material parameters were carried out on virgin samples of poled CTSNV-1 and CTS-19 ceramics. In general, obtained estimates were close to the manufacturer data.

The experimental cycle included producing the samples with different values of the fluence at the same intensity of the heat load density per pulse for both H₂- and He- plasma jet. As a result, 16 samples subjected to different plasma exposure were obtained, which were designated as follows: plasma-forming gas—number of plasma pulses. For example, He-30 and H2-30 are ceramic samples subjected to 30 pulses of helium and hydrogen plasma, respectively.

2.2. Experimental Methods and Techniques

In this study we used a pulsed plasma jet setup (plasma gun) developed at the Ioffe Institute to study the interaction of plasma with the protective materials of the first wall of the ITER and DEMO research thermonuclear reactors [7]. A block diagram of the setup is presented in Figure 1.

The jet source is a modification of a coaxial plasma accelerator with an intensive gas injection that determines the isotopic composition of the plasma. The parameters of the plasma jet and methods of their control are considered in detail in Ref. [15]. The jet diameter at the accelerator outlet is about 45 mm, the speed of the ionization front is about 100 km/s, the particle energy is up to 300 eV, and the energy flux density can reach 0.5 MJ/m². The energy flux is limited by the existing requirements for the voltage, current and discharge duration, necessary for minimizing the impurity input from the accelerator electrodes. Such parameters closely correspond to the permissible thermal loads of the ELMs level and disruptions in ITER [13]. An important parameter of the plasma jet is a particle fluence, F = dN/dS, where dN is the number of ionized gas particles crossing an area dS, perpendicular to the beam, over a certain time interval. According to our estimates, the number of the plasma-forming gas particles in the jet during one pulse is about 10¹⁹, which, with a jet diameter at the accelerator outlet of 45 mm, gives a fluence value of 6.3·10²¹/m² per pulse. The sample fluence F_Σ is determined by the total number of plasma pulses in a series. In our experiments, series of 10, 30 and 70 pulses were used, which corresponds to fluence values of 6.3·10²²/m², 19·10²²/m² and 44·10²²/m², respectively. The time interval between pulses was about 3 min. During this time, a sample surface temperature returns to its initial value.

Hydrogen and helium were used as plasma-forming gases. The samples were installed at a fixed distance of 370 mm from the accelerator outlet. A voltage of 4 kV on a 400 μF capacitive storage at a discharge current of 90 kA provides a plasma pulse of 15 μs duration with an energy flux density of 0.1 MJ/m², which determines the magnitude of the heat load and the expected temperature increase on the sample surface. For pulsed plasma heat loads the most informative parameter is the heat flux factor F_HF, [MW/(m²s^0.5)] [11], since it considers the duration and, consequently, the penetration depth of the thermal wave in the material at the initial stage of diffusion thermal processes distinctive to the short-term plasma exposure. In our experiments, the heat flux factor is of 25 MW/(m²s^0.5). The plasma jet pressure at the sample location was recorded by piezoelectric and interferometric sensors and reached 0.15 MPa for the selected operating mode. The pressure in the vacuum chamber was 4∙10⁻³ Pa.

The impact of ionized particles of plasma-forming gas at sufficiently high energies can obviously lead to significant defect formation and changes in the properties of the functional material for reactors [3,10,11,17]. Ceramic materials are also no exception to such processes [9,16]. These studies involve experiments using a wide range of techniques: optical and electron microscopy, X-ray structural analysis, electrical and ultrasonic measurements, etc.

The frequency spectra of the sample impedance required for the subsequent estimation of dielectric parameters (dielectric constant ε₃₃ and loss tangent tan(δ)) and the electromechanical coupling coefficient were measured using a PSM 1735 Analyzer. The electromechanical coupling coefficient of the thickness mode k_t was estimated both via the frequency ratio of the series (f_s) and parallel (f_p) resonances [18], as well as the use of non-equidistant frequencies of higher odd overtones [19]. The longitudinal V_L and shear V_S sound velocities were measured by the conventional pulse-echo technique at a frequency of 10 MHz. Independent measurements of V_L along the polarization direction and dielectric constant provide the value of the piezoelectric coefficient e₃₃ [20].

To measure the piezoelectric coefficient d₃₃, both a static method of measuring the displacement of a sample under DC electric field and a dynamic acoustoelectric method were used. The experimental setup for the static method included an electronic measuring micrometer head with a resolution of 0.01 µm and a high-voltage source. The acoustoelectric method [21] was implemented with the use of a home-made ultrasonic pulse-echo setup including a longitudinal piezoelectric transducer with an ultrasonic delay line. The idea of the method is that a piezoelectric transducer with a delay line, when fed with a short radio frequency burst, forms a mechanical burst of the same duration applied to the piezoceramic sample. The radio burst generated by the direct piezoelectric effect in the sample is amplified by a highly sensitive receiver and is recorded using a digital oscilloscope. It should be noted that, given the expected change in piezoelectric properties, the acoustoelectric method is fast and convenient for comparative measurements of the piezoelectric coefficient d₃₃ of ceramic samples subjected to plasma exposure of varying intensity.

3. Results

3.1. Micro-Structural and Phase Purity Analysis (XRD, Optical and SEM Microscopy)

The X-ray diffraction (XRD) patterns of CTSNV-1 and CTS-19 virgin samples, as well as CTS-19 samples subjected to maximum fluence of both H₂- and He- plasma, are shown in Figure 2. The X-ray diffraction patterns for irradiated CTSNV-1 sample are presented in our previous paper [9].

The patterns were obtained using a Dron-3 X-ray diffractometer (CuKα radiation, λ = 0.15406 nm) at room temperatures and show the presence of the pseudo-cubic single phase for all the samples. The diffraction files from the JCPDS X-ray data for the compositions of Rh- and T- phases were used to index the diffraction peaks, due to the fact that the PZT compositions under study are near the ratio of molar concentrations Zr/Ti = 52/48 known as the morphotropic phase boundary (MPB) [22,23]. Thus, a coexistence of T- and Rh- phases with closely related crystal lattice parameters may be expected. It is known that within the MPB, the dielectric constant, the electromechanical coupling coefficient and piezoelectric constants reach maximum values. It is also known that under high electrical and mechanical loads a shift in the equilibrium between the T- and Rh-phases is possible, which can persist after the loads are removed and lead to significant changes of the piezo- and dielectric properties [22].

The cubic parameter calculated from the (111) diffraction pattern peak of CTS-19 virgin sample is a = (4.067 ± 0.005) Å, and is close to the calculated one for the T-phase. The results of the analysis of the CTS-19 diffraction patterns after the exposure to both H₂- and He- plasma flows, showed that the samples generally retain a single-phase structure and have no signs of amorphization after cyclic irradiation with a flow up to the fluence of 4.4·10²³/cm².

The cubic parameter a, calculated for the irradiated samples, remains constant with the accuracy inherent in the data obtained. The evolution of the (002) and (112) peaks (highlighted in blue in Figure 2) is presented. Although no clear splitting is detected in the same peak groups of virgin and irradiated CTS-19 samples, obviously asymmetric peaks may indicate possible morphotropic transitions between two ferroelectric phases. The diffraction patterns of the irradiated samples of CTS-19 (He-70 and H₂-70) contain diffraction pattern peaks that do not belong to the T- and Rh- phases (highlighted in red in Figure 2), similar to those previously found in CTSNV-1 ceramics after the exposure to a jet of hydrogen plasma [9]. The bar diagram of scrutinyite (α-PbO2) structure presented in Figure 2e corresponds also to X-ray diffraction patterns of nanocrystals of ZrTiO4 [24]. The presence of lead-free crystals of ZrTiO4 in the modified layer corresponds to the results of elemental analysis of the surface by 20 pulses of H₂-plasma jet CTSNV-1 [9], demonstrating the significantly non-uniform distribution of lead over the modified surface of the sample and the presence of areas with a complete absence of lead. Such crystals have a black color; thus, their presence is additionally confirmed by the observed change in coloration of irradiated samples regardless of the plasma composition.

Figure 3 shows a series of microphotographs of the CTS-19 samples surface, both virgin and exposed to a series of pulses of both He- and H₂ plasma. It should be noted that the heat load in each of the pulses in the series remained constant and equal to a heat flux factor of 25 MW/(m²s^0.5). As can be seen from Figure 3, the surfaces of irradiated samples undergo a significant modification, acquiring a pronounced relief with a clearly distinct scale of inhomogeneities depending on the plasma type. He-plasma forms a finer structure of inhomogeneities than hydrogen one, at the same levels of particle energy and heat load densities on the sample surface. For H₂- plasma, a meso-structure with a distinct scale of a few to tens micron is formed within the first 10 pulses, which then undergoes no significant qualitative changes with a further increase in the number of pulses. In the case of He- plasma, the mesa-structure is formed between 10 and 30 pulses, and also remains without qualitative changes with a further increase in the number of pulses. Thus, it may be assumed that a series of 30 consecutive plasma pulses forms a surface modification that can be considered as the steady state. Figure 4 shows a SEM image of the end face of the CTSNV-1 ceramic sample after 30 pulses of hydrogen plasma exposure. A clearly defined relief is visible in the form of a quasi-ordered columnar mesa-structure with a characteristic dimensions of about 10–20 μm.

In Refs. [10,11,25] the formation of similar columnar crystallites and a recrystallized surface layer of tungsten with a depth of 40–50 μm after only two pulses of exposure to He-plasma with a heat flux factor of 90 MW/(m²s^0.5) is noted. The similarity of the above-mentioned effects in terms of changes in the structure of piezoelectric ceramics and tungsten samples under the influence of plasma loads suggests that the observed modifications of the surface and adjacent layers have the same nature.

Unfortunately, at the moment we have no estimate of the critical heat loads magnitude that causes a surface temperature rise to levels above the recrystallization and melting temperatures for the ceramic materials under study. However, considering the significant difference in the densities and thermal conductivities of tungsten and ceramic materials (ρ_W = 19,350 kg/m³, χ_W = 162 W/(m·K) for tungsten and ρ_Cer ≈ 7400 kg/m³, χ_Cer = 1.32 W/(m·K) for PZT ceramics), it may be assumed that the critical thermal factor for the studied samples is significantly lower than those for tungsten. Thus, a plasma flow with an energy density of 0.1 MJ/m² can provide the critical thermal load causing a temperature rise to recrystallization and melting temperatures of the surface layer of PZT-type ceramic samples. It is obvious that the observed surface modification should affect the material parameters of the piezoelectric PZT ceramics under study.

3.2. Dielectric Properties

The compositions under study initially have high values of the dielectric constant and relatively low dielectric loss. According to the manufacturer, these values are 2250 for CTSNV-1 and 1725 for CTS-19 (deviations of values within ±20% are acceptable) and dielectric loss tangent is approximately 0.025 for both compositions. Our estimates for virgin samples show that the values of the dielectric constant ε₃₃⁽¹⁾ = 2590 and ε₃₃⁽²⁾ = 1802 obtained at 1 kHz remain practically unchanged up to the frequencies of the first electromechanical resonance in the range of 150–200 kHz, corresponding to oscillations along the sample length (or radius) due to the piezoelectric constant d₃₁. The dielectric loss tangent for virgin samples of both types is less than 0.02 for frequencies up to 100 kHz. Figure 5 shows the frequency dependences of the dielectric constant and the dielectric loss tangent for virgin and irradiated ceramic samples of both compositions.

It is evident that for the CTS-19 samples, the exposure to plasma leads to a decrease in the value of the dielectric constant, and that the effect of the He- plasma is significantly stronger than that of the H₂-plasma. The behavior of the dielectric constant for CTSNV-1 samples also reveals the difference in the effect between the flows of the H₂- and He- plasmas. The effect of the He-plasma leads to a noticeable increase in dielectric constant, while the H₂-plasma has practically no effect on the dielectric constant values in the frequency range up to 100 kHz. At the same time, loss tangent increases by approximately three times for the samples of both compositions. It should be noted that a significant change in dielectric constant values and an increase in losses are observed even at the low exposure (the fluence of 6·10²²/m²). They change slightly with an increase in the exposure to fluence values up to 4.4·10²³/m², the maximum fluence in this series of experiments, for both types of plasma and even show a tendency for recovery towards the initial values. This may indicate that the modified layer formed on the surface of the samples, as a result of an exposure to few plasma pulses (Section 2.1), can act as a sink for non-equilibrium defects [26] and thereby increase the phase stability of the ceramics to subsequent ion radiation, which promotes the formation of defects. This assumption is in good agreement with the results of X-ray analysis and the modification of the surface microstructure observed by optical and electron microscopy.

3.3. Piezoelectric Constant d₃₃

The piezoelectric constant d₃₃ measurements using the standard static method (inverse piezoelectric effect) result in the values of 500 pm/V and 450 pm/V for virgin samples of CTSNV-1 and CTS-19 ceramics respectively. These values are consistent with the manufacturer data. For samples exposed to a series of consecutive plasma pulses (30 pulses), i.e., when the surface modification and dielectric properties may be considered as steady-state (Section 2.2), the decrease in the d₃₃ value in the static mode at E ≥ 500 kV/m did not exceed 10% for both He- and H₂- plasma compositions and remained within the parameter spread for virgin samples. Similar results were obtained using the acoustoelectric method, the advantage of which over the static method is the lack of need to apply a strong DC electric field to the samples. Figure 6 shows the radio frequency responses to a short mechanical burst from a piezoelectric transducer applied to piezoceramic samples exposed to different radiation doses of He- and H₂- plasma. Considering the adjustment due to the capacitance of each sample, the response amplitude is proportional to the value of the piezoelectric constant d₃₃ [21]. For the purpose of data uniformity, all results were obtained with the same setup parameters: amplitude, frequency and duration of the probing mechanical burst, as well as the gain of the receiver and the scale of the oscilloscope amplifier.

Figure 6 shows that for samples of both compositions, the exposure to the H₂-plasma has a lesser effect on the piezoelectric constant d₃₃ reduction than the exposure to the He- plasma. In both cases, an unusual behavior of the piezoelectric response is observed: the largest decrease is observed at a sufficiently low fluence, i.e., after the first 10 plasma pulses, while showing a tendency towards recovery with a subsequent increase in the number of pulses, which is equivalent to an increase in the particle fluence. Moreover, the effects of weakening at low fluences and restoration with a further fluence increase are more pronounced for CTS-19 samples, which belong to semi-hard PZT-ceramics by their composition, in contrast to the CTSNV-1 samples, which belong to semi-soft PZT-ceramics [23]. For all measured samples, CTSNV-1 ceramics shows a small (not exceeding 10%) deviation of the d₃₃ from the initial value for the virgin sample. This is in close agreement with the estimates of d₃₃ using the static method. For CTS-19 samples, the deviation of d₃₃ is more pronounced. At a fluence corresponding to 30 pulses, the response is restored only to a level of 85% relative to the virgin sample in the case of the H₂-plasma and 70% for the He-plasma. This differs from the results of measuring piezoelectric constant d₃₃ by the static method, which do not show such a noticeable decrease. This may be due to the fact that CTS-19 is a more ferroelectrically hard ceramics, as well as to the specificity of the acoustoelectric method, which is characterized by a significantly higher measurement frequency [27]. However, for both types of piezoelectric ceramics, He- plasma has a stronger effect than H₂- plasma.

3.4. Electromechanical Coupling Coefficient k_t and Piezoelectric Constant e₃₃

When the CTSNV-1 and CTS-19 samples are considered as piezoelectric resonators, their thickness mode electromechanical coupling coefficient k_t can be calculated basing on measurements of characteristic frequencies of the electromechanical resonances in accordance with the following formula [18,28]:

$$k_t^2={\displaystyle \frac{\pi }{2}}{\displaystyle \frac{f_s}{f_p}}\tan \left[{\displaystyle \frac{\pi }{2}}\left(1-{\displaystyle \frac{f_s}{f_p}}\right)\right],$$

(1)

where f_s and f_p are the series and parallel resonance frequencies respectively. Alternatively, it can be obtained from the non-equidistance of odd harmonics method, using the frequency ratios of fundamental and overtone resonances, tabulated as a function of a coupling factor [19]. In the latter case the resonant frequencies correspond to the roots of the equation tan(X) = X/k_t², where X = ω∙d/2∙V_L, ω is the frequency, d is the resonator thickness, V_L is the sound velocity, and k_t is the electromechanical coupling coefficient of the thickness mode. The analysis of the electromechanical resonance’s spectra for the CTS-19 and CTSNV-1 samples, both virgin and exposed to the He-and H₂- plasma flows, shows that all samples exhibit resonance responses at odd overtones corresponding to the thickness mode of oscillations, indicating the preservation of volume polarization and allowing for both methods to be used for calculations. The electromechanical coupling coefficient k_t calculated in accordance with Equation (1) is practically the same as the one obtained using the non-equidistance of odd overtones method, and equals 0.28 and 0.24 for virgin samples of CTS-19 and CTSNV-1 respectively.

The irradiation with both types of plasma at the initial stage of exposure (10 pulses) leads to a noticeable decrease in k_t, especially for the He-plasma, by approximately 30% for CTSNV-1 samples and by more than 50% for CTS-19 samples. However, with a further increase in the number of pulses, a restoration of k_t is observed, practically to the initial values for virgin samples. This result indirectly confirms the formation of a certain modified layer at low doses of exposure that plays the role of a barrier for the subsequent thermal and defect-forming plasma effects. The calculated values of the coupling coefficient for all the studied samples are presented in Figure 7, together with a set of the main material parameters highlighting the effect of the H₂- and He- plasma flows on the CTS-19 and CTSNV-1 ceramics.

The velocities of longitudinal V_L and shear V_S ultrasonic waves along the polarization direction were measured in the samples using the ultrasonic pulse-echo technique at a frequency of 10 MHz. The values of both velocities differed slightly from sample to sample for both compositions within 4250–4500 m/s and 1800–2000 m/s ranges for V_L and V_S respectively. The calculated corresponding elastic constants are C₃₃^D = ρ·V_L² = 134–150 GPa and C₄₄^E = ρ·V_S² = 24–30 GPa with ρ = 7400 kg/m³ for CTS-19 and ρ = 7300 kg/m³ for CTSNV-1. Using already obtained values for k_t and dielectric constant, and known equations for elastic constants of the poled ceramics:

$$C_{33}^D=C_{33}^E+{\displaystyle \frac{e_{33}^2}{{\epsilon }_{33}}}, C_{33}^E=C_{33}^D(1-k_t^2),$$

(2)

the piezoelectric constant e₃₃ is calculated for all samples. Similarly to d₃₃, the piezoelectric constant e₃₃ shows a noticeable decrease at low doses of plasma exposure, when a distinct columnar structure is formed in the near-surface region of the samples under the effect of a pulsed plasma loads. However, at 30 pulses, corresponding to the fluence of 18·10²² /m², e₃₃ is subsequently restored to the near-initial values for both types of samples. For the CTS-19 samples, it remains practically unchanged even at 70 pulses of plasma exposure, which corresponds to the fluence of 4.4·10²³/m². We assume that CTSNV-1 ceramics is also resistant against the radiation with doses corresponding to a fluence greater than 2.0·10²³/m². However, additional experiments are needed to confirm this assumption.

4. Discussion

Our experimental results, summarized in Figure 7, demonstrate qualitatively similar fluence dependencies of dielectric and piezoelectric parameters for both H₂- and He- plasmas and two PZT ceramic compositions. Dielectric parameters tend to saturate with increasing fluence (total number of impact pulses), and piezoelectric parameters show anomalous behavior. We observed a noticeable deterioration in the piezoelectric parameters after the first few plasma pulses and subsequent recovery to the original values. However, there are some quantitative differences for the He- and H₂- plasma and different PZT ceramic compositions. For both types of piezoelectric ceramics, He- plasma has a stronger effect than H₂- plasma on dielectric (Figure 5) and piezoelectric properties (Figure 6) properties.

A distinctive feature of high heat flux plasma exposure is the limited depth of changes in the structure of the material depending on the energy of ionized particles.

The changes consist of a gradual (with an increase in the number of pulses) modification of the structure and elemental and phase composition of the surface layer. The data presented in Section 2.1 suggest partial formation of a new nanocrystalline structure of lead-free ZrTiO4 at temperatures above 600 °C. It is known that grain boundaries can serve as defect sinks for absorbing and annihilating radiation-induced defects [29]. Therefore, nanocrystalline structures, which contain a high density of grain boundaries, demonstrate enhanced radiation tolerance compared to the initial PZT structure. This may explain the observed weak fluence dependence of the properties of the studied samples at a fluence greater than 2·10²³/m².

The formation of a modified near-surface layer is associated with a gradual deterioration of the considered measured properties of the samples. In this case, there is a non-uniform violation of the polarization of this layer due to heating above the Curie point. When the layer is completely depolarized, it practically does not affect the piezoelectric properties of the entire sample, since its thickness does not exceed several percent of the total thickness. The result is the observed restoration of the values of the piezoelectric constant and electromechanical coupling coefficient in all cases.

The process of formation of a new structure apparently depends on the plasma type and the initial PZT structure. As was mentioned previously, the PZT compositions under study are located near the MPB, where a strong dependence of the dielectric properties on the ratio of molar concentrations Zr/Ti exists. Therefore, a possible shift in the Zr/Ti ratio due to the plasma exposure may be one of the reasons for different effects in CTSNV-1 and CTS-19 samples. The observable difference in the effect of H₂- and He- plasmas is possibly due to the difference in chemical activity of He+ and H+ ions. For example, H+ ions are more active in reducing pure metals from oxides by forming OH bonds and are more easily removed from the modified layer in the form of water molecules. On the other hand, He+ ions are known to promote the formation of a ‘fuzz’ nanostructure on the tungsten surface with radically different properties, but it has not yet been possible to model this structure [30].

The detailed mechanisms of plasma effect on piezoelectric ceramics is currently unclear. Despite the existence of theoretical works on the introduction of H+ ions in PZT ceramics, for example, [31,32,33], their results can provide only possible ideas of such mechanisms, because they obtained using standard methods of ions introduction.

The effect of plasma of different isotopic composition on the properties of metals have been carried out for a long time [34,35]. And even for metals, which are in some sense simpler objects compared to piezoelectric ceramics, various plasma-induced phenomena cannot be fully explained and described within the framework of existing theoretical models.

Akhmanov et al. [36] proposed a thermal model for describing physical phenomena arising in the surface layers of metals, semiconductors and dielectrics under the influence of powerful pulsed laser radiation [36]. The model assumes various physical mechanisms for the excitation of highly nonequilibrium states. The model shows that pulsed laser annealing is accompanied by rapid (several tens of nanoseconds) heating of the crystal lattice and subsequent rapid melting of the surface layer of the material, leading to the formation of ordered spatial-temporal structures due to spatially inhomogeneous heating of the surface. The quasi-periodic structure on the surface of irradiated samples under study (Figure 3 and Figure 4) are well described within the framework of the thermal model. The model is in good agreement with the thermal mechanism of nonuniform formation of the nanocrystalline structure of ZrTiO4 and depolarization of the modified layer.

We believe that Akhmanov’s theory can be used in the future to explain physical phenomena arising under the influence of a pulsed high-energy plasma flow and the mechanisms preventing the degradation of the bulk properties of piezoceramics.

5. Conclusions

For the first time, the experimental results of a comprehensive study of the dynamics of changes in the surface structure and a set of dielectric and piezoelectric parameters under the influence of a powerful heat flow of plasma jet are presented for helium and hydrogen plasmas and two PZT-ceramics compositions in the MBP region.

An important obtained result is the preservation of polarization of bulk samples of piezoelectric ceramics despite multiple short-term extremely-high-energy loads.

Anomalous fluence dependencies of dielectric and piezoelectric parameters were revealed for both compositions of the studied piezoelectric ceramics. A noticeable deterioration in the parameters was observed already after the first plasma pulses with a total fluence of about 5·10²²/m². A further increase in fluence leads to the recovery of the piezoelectric properties to values inherent in the non-irradiated samples and saturation of the dielectric properties both when exposed to helium and hydrogen plasma.

Specific columnar crystal structure with a modified composition [9] and characteristic dimensions of several tens micron was formed on the surface of the samples by the tenth pulse (a fluence of 6·10²²/m²) for hydrogen plasma and by the 30th pulse (a fluence of 2·10²³/m²) for helium plasma.

Possible mechanisms of the observed effect of restoration and further preservation of piezoelectric properties with an increase in the number of plasma pulses are proposed in terms of non-stationary high energy thermal processes leading to recrystallization of the surface layer with a change in the elemental composition.

The results obtained demonstrate the possibility of modifying the surface layer of PZT ceramics that prevents the destructive effect of shock extremely-high-energy loads on their piezoelectric properties.

The nature of this phenomenon is currently not fully understood. Additional experiments are planned for further establishment of the main mechanisms.