Effect of Calcination on Alumina Ball-Milled Powders toward Lead-Free KNN-Type Ceramics Synthesis

1. Introduction

The piezoelectric materials are widely applied in several fields including piezoelectric actuators, piezoelectric transformers, and piezoelectric energy harvesters [1, 2, 3]. The corresponding market has been dominated by lead zirconium titanate (commonly abbreviated with PZT) ceramics, but these contain over 60 wt% lead and so, they are ineluctably toxic. Therefore, lead-free alternatives are sought [1]. Lately, the (Na_0.5K_0.5)NbO₃-based system (KNN) has received much attention thanks to its outstanding properties, including considerable piezoelectric coefficient d₃₃ of 300–400 pC N⁻¹ and high Curie temperature beyond 400°C, besides its lead-free content [4, 5].

Given the high complexity in the composition of KNN ceramics, many challenges are encountered during the synthesis, which would result in poorly reproducible functional properties [6]. The conventional route to synthesize KNN ceramics is based on the solid-state reaction between sodium and potassium carbonates and Nb₂O₅ powders. Despite many advantages, an important drawback of the conventional solid-state reaction consists in the hygroscopic nature of the carbonate precursors that leads to water content and thus, light stoichiometric changes that are able to induce important structural changes or appearance of secondary phase [7]. The non-homogeneous distribution of Nb and varying diffusion rates of Na and K ions could also lead to local heterogeneities, incorrect stoichiometry, and the presence of intermediate and secondary phases.

It has been shown that the ferroelectric and piezoelectric properties of KNN materials can be tailored by proper control of alkaline content and the K/Na ratio and by adjusting the content of intermediate and secondary phases. The optimal piezoelectric output in KNN systems was indicated for a K/Na ratio critical value of 1 [8]. On the other hand, during the synthesis, particle coarsening seldom takes place, which in turn diminishes the powder sinterability and lowers the piezoelectric response. Thus, a densification control is also required to improve the piezoelectric properties of KNN ceramics.

Regarding the synthesis, mechanochemical activation has been proposed as a suitable pre-treatment approach. Mechanochemical reactions are promoted via energy transfer between milling bodies and the powder being milled. Thus, the reaction between the powder precursors can be activated during the ball milling process given that the corresponding energetic barrier is reached by the cumulative kinetic energy provided by the high-energy employed in the milling [19]. The mechanochemical activation has important advantages, namely, it decreases the calcination temperature and overall synthesis duration, and it allows the control over the volatilization of alkaline species by reducing it. Thus, it enables the synthesis of KNN ceramics with enhanced chemical homogeneity and improved crystal structure [9, 10]. The common ball milling conditions refer to ball milling of precursor mixture with zirconia bodies for as long as 24 h, on/off time as 2/8 minutes. Following the ball milling mechanochemical activation of the precursor mixture, calcination is applied. Common calcination temperatures refer to 900°C and dwelling for as long as 5 h. Most of the works report a synthesis approach based on ball milling of the mixture followed by calcination and sintering or a second milling step after the calcination and another calcination step or varying sintering steps [11].

During the sintering, densification takes place together with grain growth and coalescence. In a first stage, a liquid phase is formed due to excess Nb₂O₅ and it helps to initiate the sintering when necking appears together with particle downsizing, thus inducing a light densification [12, 13]. Further, the densification continues as the grains get interconnected and the pores reduce their section area. In a last stage, the pores diminish greatly up to their complete disappearance.

The sintering temperature in conventional method was reported to be between 1000 and 1150°C. While a higher sintering temperature and dwelling were indicated to be useful in diminishing the intermediate phase and the residual porosity, an excess temperature or dwelling would result in grain growth and evaporation of the alkaline elements due to their low fusion points (892 and 760°C for Na and K), thus, in composition variation [14, 15, 16, 17].

The maximum densification in KNN structure of about 97% was observed at 1100°C. By increasing the sintering dwelling, the improvement of densification required for the enhancement of electrical properties of KNN ceramics can be achieved [13, 18]. An optimal dwelling at maximum sintering temperature was reported as 2 h [19]. Nevertheless, the dwell time and heat rate must be adjusted to completely remove the intermediate phase and impede the evaporation of alkaline elements by a prolonged dwelling.

Despite many reports on similar processing conditions, the significantly different properties of the obtained KNN ceramics show that more studies are required to define proper fabrication protocols that allow enhanced reproducibility of the final KNN properties. The mechanical energy supply during grinding and thermal energy delivery during calcination and densification have been indicated as key parameters to focus on due to the direct relation between particle sizes and powders’ reactivity [20]. In this work, a novel fabrication approach is proposed for KNN systems to be applied in the conventional solid-state route, namely by mechanochemical activation of downsized KNN precursors powders, and adjusting the calcination conditions. Both the downsizing of precursors and the mechanochemical activation of their mixture have been achieved by ball milling with alumina bodies as the cheaper alternative. The obtained results indicate the proposed synthesis approach is a valuable tool in the fabrication of reliable KNN ceramics.

2. Materials and methods

3. Results and discussion

The homogenized precursors employed in the synthesis of KNN ceramics were analyzed first in terms of particle size. All precursors have been downsized with respect to the as-received precursors of about 95%. Figure 1 presents the differential size distribution for the three reactants upon homogenization. As it can be observed, all reactants exhibit a bimodal distribution, except for Na₂CO₃ that presents a third one, which is attributed to agglomerates due to its hygroscopic nature.

The evolution of percentile size values below 10, 50 and 90% of sample content were retrieved according to the cumulative particle size distribution. The results obtained for the individual precursors upon ball milling homogenization are shown in Table 1. As it can be observed, the carbonates show similar small size distribution, with submicron average size, while the Nb₂O₅ average size particles have about 3 microns.

The individual precursors were further weighed in the corresponding stoichiometric values and mixed by ball milling. The evolution of percentile size values for the mixtures obtained with varying milling duration is depicted in Table 2. As it can be observed, the mixtures show intermediate values with respect to the individual precursors. The median particle size of the mixture is lower after the ball milling mixing, which is indicative of further downsizing of the reactants in the mixing step. On the other hand, it is observed that prolonged mixing by ball milling results in increasing size distribution across the percentiles, which could be due to limited downsizing followed by agglomeration.

The next step in the synthesis of the KNN ceramics is represented by the calcination of precursor mixture. To observe the effect of the milling duration for the mixing of the reactants on the behavior to solid-state reaction in the calcination step, the percentile size values were obtained. For exemplification, the results for the calcination at 700°C for 2 h, with a rate of 3°Cmin⁻¹ (method CA) are depicted in Table 3.

It was observed that the particle size values increased with respect to the mixtures prior to calcination, which is attributed to the decomposition and coalescence of the reactant particles during the solid-state reaction in the calcination step. Moreover, the mixture milled for 3 h resulted in particle size values greater than the one milled for 2 h, which might be explained by the larger particle size of the mixture prior to calcination. The solid-state reaction was further monitored by employing a parameter described as particle size increase factor (PF) obtained by dividing the percentile values obtained upon calcination by those prior to calcination as shown in Table 3. The obtained values show similar PF for up to 50% of the sample, irrespective of the mixing duration. On the other hand, the PF for the particles up to 90% of the sample is three times higher for the mixture obtained for 2 h mixing, which can be attributed to improved solid-state reaction due to the smaller particle size of the mixture subjected to calcination.

To get an insight into the effect of the calcination conditions on the solid-state reaction, the percentile values were obtained and analyzed by the PF with respect to the particle size values corresponding to the mixture prior to calcination (as depicted in Table 2). For exemplification, the evolution of particle size values for the mixture obtained by 2 h mixing is shown in Table 4.

As it can be observed, the particle size values increase with calcination conditions in the order CA < CC < CB. The particle size increase factor (PF) for each percentile exhibits the same trend, that is PF CA < PF CC < PF CB. The obtained particle size and PF values indicate that for the same calcination temperature value, as low as 700°C, a longer dwelling induces improved reaction despite a higher calcination rate (calcination condition CB) with respect to a shorter dwelling and slower calcination rate (calcination condition CA). The calcination CB also appears to induce improved coalescence with respect to condition CC, despite the lower temperature (700 vs. 800°C), which is explained by the increased dwelling.

Following the calcination step, the mixtures were sieved, uniaxially compacted and subjected to sintering step in the same conditions, namely 1100°C for 2 h and 10°C min⁻¹ rate. The solid-state reaction was further monitored by recording the weight loss (%) upon individual synthesis steps, that is calcination step and sintering one, by reporting the final weight to the initial weight of the mixture subjected to each step, as well as total weight loss obtained by summing the individual ones. Table 5 depicts the evolution of weight losses with the calcination conditions and milling duration of the reactant mixing.

It is observed that the weight loss values are higher for the mixture obtained by 3 h mixing than 2 h mixing, indicating that the particles undergo more decomposition and evaporation and less formation reaction, due to less contact surface of the larger particles. After the calcination step, the weight loss is observed to be the lowest for the CA calcination condition, attributed to the lower calcination temperature. The weight loss is similar for the CB and CC calcination conditions, and higher than CA condition.

The weight loss after the sintering step has an opposite trend, that is it decreases from condition CA to CB=CC, which is attributed to a larger extent of decomposition due to the high temperature applied in the sintering step. As the CB and CC mixtures already initiated the formation reaction in the calcination step, they lose less weight upon sintering one, where the formation reaction is completed.

The total weight loss recorded after the two steps (calcination and sintering) decreased in the order CA > CB > CC, indicating a higher yield of KNN formation in calcination conditions CB and CC and that proper control of the three parameters such as calcination rate, temperature and dwelling time is needed to adjust the synthesis of KNN ceramics, besides the control of reactant particle size.

The solid-state reaction synthesis of KNN ceramics was further analyzed using thermogravimetric measurements (TGA). The weight changes of the individual precursors indicated that Nb₂O₅ is stable over the whole temperature range while Na₂CO₃ and K₂CO₃ dehydrate around 100°C and start to decompose from 625°C, with temperature peaks at 820°C Na₂CO₃ and 830°C for K₂CO₃, in agreement with the melting points [22]. The thermograms obtained for the mixtures indicate varying behavior as a function of the ball milling mixing duration, as shown in Figure 2. The weight losses recorded below 100°C are attributed to water evaporation. The water loss from 2 h mixture amounts 1.9% while the 3 h one reaches double percentage. Another weight loss is recorded at about 280°C and it is assigned to water loss from hydrated carbonates together with CO₂ release [7]. This weight loss increased from 2% for the 2 h mixture to 2.5% for the 3 h one. The formation of an intermediate phase (K,Na)₂Nb₄O₁₁ at the Nb₂O₅ surface upon decomposition of the carbonates together with the initiation of the formation of stoichiometric (K_0.50Na_0.50)NbO₃ at the surface of the intermediate phase [23] is indicated by the wide peak centered at 447°C. The associated weight loss is about 3.2 and 3.6% for the 2 h and 3 h mixture, respectively. Another peak at about 630°C is assigned to the completion of the formation of KNN and diffusion of alkali elements toward the unreacted Nb₂O₅ [23].

While all temperature peaks were located at similar values, irrespective of the mixing duration, the weight losses (%) were higher for the mixture obtained for 3 h than for 2 h, indicating higher particle experience evaporation due to their larger size. Considering the weight loss observed up to 700°C, the mixture obtained for 2 h reached a value of 11.46%, which is closer to the theoretical weight loss for the KNN precursors, that is, 11.34% [23], than the one recorded for the mixture obtained for 3 h, namely 12.84%. Moreover, the smaller weight change at the temperature assigned to the KNN formation completion, that is, 630°C, for the mixture obtained for 3 h is indicative of the lower contact surface between the particles which affects the balance between Nb₂O₅ and carbonate particles and thus KNN formation completion.

The morphology of the sintered KNN was investigated by FESEM measurements and the typical morphologies are depicted in Figure 3 as a function of precursor mixing duration and calcination conditions. It is observed that the grain shape changed toward typical cube-like one [24, 25] by varying the calcination conditions in the order CA < CB < CC for both cases of mixing durations.

In general, the size of the grains was larger for the mixture obtained by 3 h mixing. The low calcination rate in CA condition results in decomposition and evaporation of material and in combination with a low contact surface due to larger particle size as the 3 h mixture exhibits, some of the calcined grains appear smaller than in the case of 2 h calcined mixture. By a faster calcination rate as in CB condition, although the large dwell allows for the formation reaction between the smaller particles, the bigger ones do not have enough time to decompose with the same degree; therefore, the grains exhibited by 3 h calcined mixture appear larger than the 2 h one. Finally, the combination of the higher calcination rate and temperature in CC condition allows for a more homogeneous grain size distribution within the sample, while the 3 h mixture still presents larger grains. The grain size ranged between 1 and 2 μm, and it agrees with other reports [26].

The elemental analysis by energy dispersive spectroscopy (EDS) was performed to obtain further information regarding the solid-state reaction and formation of KNN ceramics. K ions have a lower diffusion rate; thus, it is highly affecting the final properties of KNN ceramics. For this reason, the total alkaline content, A/(A + Nb) and K/A (where A is the alkaline content as K + Na), were investigated, as they reflect the formation completion for KNN structure due to particle size induced reactivity. Table 6 presents the evolution of such indicators with the mixing duration and calcination conditions. The theoretical values for the KNN and the intermediate phase, K₂Na₂Nb₄O₁₁ for A/(A + Nb), are 0.50 and 0.33, respectively, while the value of K/A is 0.50 for both.

As it can be observed in Table 6, the values for the A/(A + Nb) and K/A are closer to the theoretical values corresponding to the intermediate phase (K,Na)₂Nb₄O₁₁. The A/(A + Nb) increases in the order of calcination conditions CA < CB = CC, while the K/A decreases in the same order, for both mixture cases, toward the theoretical values obtained for perovskite KNN phase. The evolution of these two parameters is indicative of the increasing formation of KNN structure with proper calcination conditions achieved in CB and CC for the 2 h mixture and in CC for the 3 h mixture. Nevertheless, the two indicators are closer to the theoretical KNN phase for the 2 h mixture than the 3 h one. The calcination conditions CB and CC appear to result in similar A and K contents for the 2 h mixture, indicating that the smaller particle size allow more reactivity and KNN formation due to a larger contact surface and it can be achieved also at lower calcination temperature with further adjustment of dwelling time. On the other hand, for the larger particles in the 3 h mixture, only calcination condition CC appears mostly favorable for KNN formation due to the higher calcination temperature.

The crystal structure of KNN ceramics was further analyzed using XRD measurements. The calcined mixtures were first investigated with respect to the individual precursors. Knowing that high calcination temperature induces good KNN formation reaction, to observe the effect of the precursor mixing duration in this aspect, the calcination conditions such as the lowest calcination temperature and dwelling time were considered as relevant. Thus, Figure 4 depicts the XRD spectra of the mixtures calcined in condition CA (700°C, 2 h dwelling, 3°C min⁻¹ rate) vs. precursors. The precursors show the typical peaks [26]. The temperature of 700°C was selected for the calcination in agreement with the temperature revealed for the synthesis by TGA results (above 650°C). Indeed, both spectra of the mixtures depict the typical perovskite peaks and low content of the intermediate phase (intermediate phase (K,Na)₂Nb₄O₁₁ prior to the formation of KNN). Both spectra present a sharp (110) plane at about 32° of maximum intensity, indicative of the perovskite formation. The coexistence of both orthorhombic and tetragonal phases with dominancy from orthorhombic one is shown by the separation in two of the peaks [27].

The peak at 45° corresponding to the orthorhombic and tetragonal coexistence [28] was found to increase in intensity and show a marked split upon sintering as shown in Figure 5a and b, for both mixtures, obtained by 2 and 3 h, respectively. The smaller full-width half maximum of the peak located at 32° is indicative of the reaction formation where the particles fuse and result in larger structure. The insets in Figure 5 depict a different behavior during the synthesis depending on the precursor mixing duration. On one hand, the sintered 2 h mixture shows a decrease in the angle f with the calcination condition or the 45° peak, namely in the order CC < CB < CA. This aspect indicates changes in the composition, and they are attributed to improved reactivity and diffusion of alkaline elements toward the formation of pure perovskite KNN structure, in agreement with the EDS results in Table 6 for 2 h mixture. On the other hand, the sintered 3 h mixture exhibits an opposite behavior, that is, the peak at 45° shifts toward increasing angle CA < CB < CC. The peak intensity increases for both mixtures from CA to CB and CC calcination conditions.

The peak at 45° is further depicted in a magnified image in Figure 5c, to show the difference between the mixtures at the highest calcination temperature. It is observed that the angle corresponding to the 2 h mixture is lower than the 3 h one, attributed to improved K⁺ diffusion (larger cation radius than Na⁺) [29] in the KNN structure thanks to the larger contact surface conferred by smaller particles in the 2 h mixture.

The density measurements indicated the densification of the KNN ceramics depended greatly on the calcination conditions as well as on the reactant mixing duration, as shown in Table 7. The relative density ranged between 92.10 and 96.31%, in agreement with other reports. The relative density values of the KNN ceramics were higher for the 2 h mixture, which can be explained by the smaller average particle size of reactant mixture subjected to solid-state reaction as induced by proper precursor mixing duration by ball milling. On the other hand, the relative density increased in the order CA < CB < CC of calcination conditions, indicating that a higher calcination temperature is most determining for the densification, most probably due to the reducing porosity between the grains.

Despite the interest raised by KNN ceramics, there are only a few reports on their mechanical properties [30, 31, 32]. Here, the mechanical properties of the KNN ceramics were evaluated by microhardness measurements. The HV0.1 evolution with the average particle size induced by reactant mixing by ball milling and with calcination conditions is presented in Table 8. The KNN ceramics obtained by 2 h mixing recorded higher hardness values than their counterparts obtained by 3 h mixing, in agreement with the density evolution in Table 7. On the other hand, the obtained values show a different evolution of microhardness with the calcination condition if the mixtures were obtained for varying mixing time, that is, for the case of smaller particles in 2 h mixture, the best conditions to improve the microhardness appear as CC, where higher calcination temperature is combined with a faster heating rate and 2 h dwell. In the case of larger particles in 3 h mixture, the similar values obtained for calcinations CB and CC indicate tailoring could be achieved by considering the proper combination of dwelling, calcination temperature and heating rate, where dwelling exhibits a marked influence.

4. Conclusion

A novel approach based on mechanochemical activation of the mixture of downsized precursor powders is proposed for the conventional solid-state reaction synthesis method of KNN ceramics. The mechanochemical activation and downsizing of the precursors have been performed by ball milling with alumina bodies. The approach was tested in varying conditions of calcination, including temperature, heating rate and dwelling. The potential of the proposed approach for the control of the powder reactivity toward the suitable alkaline content, K/Na ratio, presence of secondary phase and grain homogeneity in the final KNN ceramic was demonstrated by the decreased reaction temperature in agreement with TGA results and particle size, as well as XRD ones. The results indicate that the application of ball milling with alumina bodies for 2 h achieves small enough particles to enable proper reaction by calcination at a temperature as low as 800° where the heat rate is 10°min⁻¹ and the dwelling is 2 h and achieve the best K content. Such results could be exploited in the fabrication of KNN with suitable piezoelectric properties for a large range of applications.

Acknowledgments

This publication is part of the grant PID2021-128548OB-C21 and CNS2023-144190 “NextGenerationEU”/PRTR funded by MCIN/AEI/10.13039/501100011033, and CIGRIS/2022/077 project funded by Generalitat Valenciana.

Conflict of interest

The authors declare no conflict of interest.