Thermal Characterization of Ceramic Composites for Optimized Surface Dielectric Barrier Discharge Plasma Actuators

Abstract

Ice accretion is a significant drawback in an aircraft’s and wind turbine’s aerodynamic performance in cold climate weather. Plasma actuators are an attractive technology for ice removal; however, dielectric barriers are typically restricted to borosilicate glass and various polymers, such as Teflon^® and Kapton^®. Nevertheless, new materials capable of withstanding prolonged exposure to charged particles are needed. In this work, Y₂O₃-ZrO₂, MgO-CaZrO₃, and MgO-Al₂O₃ ceramic samples were manufactured and their thermal properties as DBD plasma actuators were measured. As foreseen, the results showed that the higher the power consumed, the higher the temperature surface of the plasma actuators. The Y₂O₃-ZrO₂ dielectric showed the highest power consumption and ceiling temperatures (20.7 W and 155 °C at 10 kVpp, respectively), followed by MgO-CaZrO₃ (9.6 W and 62 °C at 10 kVpp, respectively) and by MgO-Al₂O₃ (5.6 W and 47 °C at 10 kVpp, respectively). It was concluded that MgO-Al₂O₃ presented stable magnitudes across the entire dielectric area, whilst Y₂O₃-ZrO₂ showed a more concentrated temperature field. Therefore, considering that about 65 to 95% of the total power supplied to the DBD plasma actuator is dissipated as heat, it becomes natural to propose ceramic-based DBD plasma actuators as de-/anti-icing means for aero-dynamic structures.

1. Introduction

Ice accumulation is reported to severely decrease the aerodynamic performance of an aircraft due to induced flow separation, drag, and weight increase, which, in turn, diminishes thrust and lift forces [1]. Ice formation, followed typically by ice accumulation phenomena, promotes, for example, wing and engine blade profile modifications on aircraft surfaces. Consequently, in the United States of America, the National Transportation Safety Board [2] has identified aircraft icing as one of the safety concerns that urgently needs to be addressed. Among several reports, in 1994, an Embraer EMB120 aircraft was involved in an accident resulting from ice accretion on the wing’s upper surface, which promoted airflow separation and uncoordinated flight [3]. From 2000 to 2011, carburetor icing was pointed out to be the cause of about two hundred and fifty accidents, of which two per year were fatal [4]. From 2016 to 2019, Europe and Canada conducted the PHOBIC2ICE project aiming to develop technologies and predictive simulation routines designed to reduce the ice accretion phenomena [5]. Similarly, wind turbine icing fosters aerodynamic performance degradation due to weight imbalances in cold climates. Moreover, ice formation may lead to the automatic shutdown of wind turbines, and, therefore, loss of power. Ice falls from the wind turbines may endanger nearby people and infrastructure. Gao and Hu (2021) [6] reported that 19%, 72%, and 94% of wind turbines in Asia, North America, and Europe, respectively, have experienced icing events. Consequently, anti-icing and de-icing technologies have vastly improved over the years, e.g., liquid-based, pneumatic, ultrasonic-based, thermal, and electro-thermal systems [7,8]. Nonetheless, these mechanisms have shown associated problems, such as environmental pollution, mechanical damage (foreign object damage hazard), unreliability, or inefficiency in flight conditions due to supercooled degradation and demanding power conditions [8]. Given the abovementioned, further development of innovative de-icing/anti-icing systems is required.

Dielectric barrier discharge plasma actuators have been subject to numerous studies as mechanisms for active aerodynamic flow control, and, more recently, as de-icing/anti-icing technology due to their thermal power loss [8,9,10,11,12,13]. Anti-icing processes prevent the formation of ice layers, whereas de-icing processes involve the removal of existing ice layers [7]. Jukes et al. (2007) [14] performed thermal imagery of the gas temperature around the electrode and Mylar dielectric surface temperature, noting that both experienced heating due to the plasma discharge. Dong et al. (2008) [15] performed a breakthrough study on the impact of frequency and applied voltages on the power dissipation of a DBD plasma actuator. Moreover, the authors quantified plasma temperatures by spectroscopy emission measurements. Stanfield et al. (2009) [16] used emission spectroscopy to obtain rotational and vibrational temperature profiles for N₂ for the discharge region of a single-layer surface DBD as a function of input voltage. The authors concluded that the rotational temperature profiles fluctuate in the spanwise direction of the discharge and are attenuated in the induced flow direction. The observed results led the authors to believe that a relationship exists between the area that microdischarges attach to the exposed electrode and temperature fluctuations. Joussot et al. (2010) [17] analyzed the thermal characterization of a Kapton^®/Mylar^®-based alternating current (AC) DBD plasma actuator via an infrared camera to assess the influence of electric parameters. Likhanskii et al. (2010) [18] proposed dielectric barrier discharge as a de-icing device. The authors remarked that due to both force and heating effects, temperature increase does not interfere with the flow characteristics, nor generates “hot spots”. Rodrigues et al. (2018) [19] constructed different plasma actuators to study the influence of the dielectric thickness and material for film cooling and de-icing/anti-icing. The results showed that Kapton^®-, poly-isobutylene-, polylactic acid-, and acetoxy silicone-based dielectrics had average heat generation efficiencies between 57% and 88%. More recently, Abdollahzadeh et al. (2022) [20] performed a parametric optimization for the better ice sensing and de-icing performances of surface DBD (SDBD) plasma actuators considering Kapton^®, PIB rubber, PMMA, and Teflon^® dielectric layers. It was also shown that the ice sensitivity varied considering the dielectric materials used. Over the years, although several studies have focused on the de-icing, anti-icing, and ice-sensing capabilities of AC SDBD plasma actuators, the dielectric barrier materials used in these studies have primarily been polymers [21]. Therefore, new materials capable of withstanding long periods of exposure to charged particles are needed [22,23]. Shvydyuk et al. (2023) [24] manufactured ceramic-based SDBDs and captured infrared images of them assembled. Three compositions commonly applied in metallic substrates for thermal barrier coatings were used as ceramic composites. As a follow-up, in this study, Y₂O₃-ZrO₂, MgO-CaZrO₃, and MgO-Al₂O₃ ceramic composite dielectric barriers were thermally analyzed to propose original solutions for improving AC SDBD robustness for ice accretion mitigation. Specifically, the produced ceramics’ thermal conductivity analysis and temperature field distributions were evaluated.

2. Materials and Methods

Sample Manufacturing

In the current study, three ceramic dielectric plates were fabricated from commercially available powders. To produce the Y₂O₃-ZrO₂ dielectric specimen 3 mol% yttria-stabilized zirconia (t-3YSZ) (Tosoh-Zirconia, Tokyo, Japan), 8 mol% yttria-stabilized zirconia (c-8YSZ) (Tosoh-Zirconia, Tokyo, Japan), and monoclinic zirconia (m-ZrO₂) (Acros Organics, Antwerp, Belgium) powders of high purity (wt%: 99.8, 99.7 and 98.5 for t-3YSZ, c-8YSZ, and m-ZrO₂, respectively) were used. For the MgO-CaZrO₃ ceramic composite, magnesium oxide (MgO) and calcium zirconate (CaZrO₃) (Alfa Aesar, Ward Hill, MA, USA) powders of high purity (wt%: 96, and 99.2, respectively) were used. Considering the MgO-Al₂O₃ composition, magnesium oxide (MgO) (Alfa Aesar, Ward Hill, MA, USA) and aluminum oxide (Al₂O₃) (Acros Organics, Antwerp, Belgium) powders of high purity (wt%: 96 and 99, respectively) were applied. The mentioned raw powders were milled (3 h for Y₂O₃-ZrO₂ and MgO-CaZrO₃; and 6 h for MgO-Al₂O₃) in grinding bowls with 250 mL capacity in a high-energy planetary mill (Fritsch, Pulverisette 6, Idar-Oberstein, Germany) at 500 rpm in cycles of 30 min. The milling process was performed with a powder/isopropyl alcohol/balls ratio of 1/1/2. Next, the mixtures were dried in a stove at 80 °C for 24 h (Carbolite, NR200-F, Hope Valley, UK), and sieved (Retsch, AS200, Haan, Germany) using a 63 μm mesh. The sieved mixtures were each compacted with a universal testing machine (Instron 8800, 100 kN capacity, Canton, MA, USA) using rectangular dies of high-strength steel measuring 60 × 60 mm. The sintering process was then conducted in an electrical furnace (Termolab, MLR, Águeda, Portugal) (2 h at 1450 °C for Y₂O₃-ZrO₂ and MgO-CaZrO₃; and 1600 °C for MgO-Al₂O₃) with an initial heating rate of 5 °C/min. Lastly, to uniformize the dielectric ceramic plates, polishing with silicon carbide paper was performed (Struers DAPV, Ballerup, Denmark). After sintering, the ceramic plates were tested for apparent porosity and relative density according to the international standard ASTM C20-00 [25]. The theoretical densities of the diffractogram cards were used to compute the relative densities of the ceramic composites. The theoretical diffractograms XRD cards no. 50-1089 (t-3YSZ), no. 49-1642 (c-8YSZ), no. 37-1484 (m-ZrO₂), no. 35-0790 for (CaZrO₃), no. 46-1212 (Al₂O₃), no. 71-1176 (MgO), and no. 77-1193 (MgAlO₄) were used. Moreover, an X-ray diffraction (XRD) study (Rigaku DMAX III/C, Tokyo, Japan) was conducted with data collected within 10° to 90° (2θ) to confirm the chemical composition of the final material obtained. Scanning electron microscopy (SEM) (Hitachi S-2700, Tokyo, Japan) was performed to obtain microscopic images of Y₂O₃-ZrO₂, MgO-CaZrO₃, and MgO-Al₂O₃ ceramic samples. The samples were previously thermally etched at 1305 °C for zirconia-based and 1440 °C for alumina-based ceramics and coated with an ultrathin gold layer (Emitech K550 Gold Sputter Coater, Quorum Technologies, East Sussex, UK).

After obtaining the three ceramic plates of Y₂O₃-ZrO₂, MgO-CaZrO₃, and MgO-Al₂O₃ these were assembled into SDBD plasma actuators. Two electrodes—one exposed and one covered with an insulation tape (10 mm and 20 mm width, respectively)—were asymmetrically mounted on each side of the ceramic dielectric plates with a 1 mm gap. This way, the total resulting plasma spanwise length was approximately 30 mm. Moreover, the electrodes were made of copper tape with a thickness of 80 μm. The SDBD plasma actuators were operated by a sinusoidal AC high-voltage source (PVM 500 model, Information Unlimited, Amherst, MA, USA)—up to 20 kV peak-to-peak (kVpp) and 20 kHz to 50 kHz frequencies. Both voltage and current waveforms were monitored with a digital oscilloscope (PicoScope 5443 A, Pico Technology, Cambridgeshire, UK) connected to a high-voltage probe (MI074 Secondary Ignition Pickup, Lurgan, UK). The waveforms were obtained with a sampling rate of 125 MS/s and a 14-bit vertical resolution, providing data with an uncertainty of 1%. Power consumptions were calculated using the electric current method [26], using a metal resistor (Robert Mauser, Lda., Lousa, Portugal). The metal resistor had an impedance of 100 Ω and a tolerance of 1%. The resistor was chosen based on its temperature coefficient of 50 ppm, which means that the impedance is approximately constant during the plasma discharge process and is therefore neglected in the power consumption analysis. Considering the uncertainty of the digital oscilloscope and the tolerance of the resistor, the error of the power consumption analysis will be less than 2%.

The temperature fields induced by heat dissipation through the plasma discharge process were analyzed using an infrared technique via a thermal imaging camera (FLIR E50, Teledyne FLIR, Wilsonville, OR, USA) with a resolution of 240 × 180 pixels and an uncertainty of 2%. Infrared is a non-invasive technique that provides the user with a thermal map of the analyzed sample. The infrared technique measures the electromagnetic radiation emitted by objects and converts this radiation—at wavelengths around 700 nm and 1 mm—into temperature [27,28]. All bodies with a temperature above absolute zero emit electromagnetic (including infrared) radiation from their surfaces that is proportional to their intrinsic temperature. In practice, the emission, $e$, of the body varies between $0<e<1$. By applying the Steffan–Boltzmann law, the temperature of the dielectric layer can be computed based on the radiation it emits [15]. Since the samples varied in size due to distinct densification processes, the temperature fields were captured at different heights (130 mm for Y₂O₃-ZrO₂ and MgO-CaZrO₃, and 210 mm for MgO-Al₂O₃). To ensure temperature stabilization, the infrared images were captured after the plasma discharge stoppage, which lasted for 300 s [17,29,30]. It should be emphasized that the plasma actuators were first painted with black matte ink (emissivity of 0.97). Figure 1 shows the experimental setup.

Additionally, thermal conductivity was determined according to the laser flash analysis standard test ASTM E1461-13 [31]. For that purpose, disc specimens (diameter of approximately 12.55 ± 0.5 mm) were tested on a LFA 457 MicroFlash (Netzsch, Selb, Germany) after applying a high emissivity coating (Kontakt-Chemie 76009-AG, Graphit 33 Spray, Zele, Belgium) for improved energy absorption. The ceramic samples’ thermal conductivity parameter was tested in a controlled inert atmosphere with a reference material (Pyroceramic, Order number 6.256.1–94.0.03, Netzsch, Germany) for calibration aims.

3. Results and Discussion

The following sections, divided into the topics of AC SDBD consumed power, thermographic analysis, and spatial thermal variations, describe the study performed and the analysis made.

3.1. Dielectric Ceramic Sample Analysis

After the sintering phase, the apparent porosity and relative density of the Y₂O₃-ZrO₂, MgO-CaZrO₃, and MgO-Al₂O₃ ceramic samples were determined to ascertain the manufacturing process quality.

Table 1. Apparent porosities and relative densities of manufactured ceramic composite plates and their respective standard deviations.

Ceramic Sample	Y2O3-ZrO2	MgO-CaZrO3	MgO-Al2O3
Apparent porosity [%]	0.06 ± 0.04	0.04 ± 0.02	6.7 ± 0.3
Relative density [%]	98.0 ± 0.1	98.9 ± 0.3	71.3 ± 0.3

summarizes the values obtained for both parameters. The results in Table 1 are the average values of five measurements.

From Table 1, on the one hand, Y₂O₃-ZrO₂, MgO-CaZrO₃, and MgO-Al₂O₃ had porosities of 0.06, 0.04, and 6.7, respectively. On the other hand, the samples showed relative densities of 98.0, 98.8, and 71.3 for Y₂O₃-ZrO₂, MgO-CaZrO₃, and MgO-Al₂O₃ ceramic plates, respectively. Considering the above, it was concluded that although the MgO-Al₂O₃ ceramic mixture was milled for 6 h instead of the 3 h of the Y₂O₃-ZrO₂ and MgO-CaZrO₃ compositions, it still presented some unwanted porosity content. Consequently, it shows that the sintering process was successful for zirconia-based compositions but needs future adjustments in the sintering temperature for the alumina-based ceramic. Since the use of ceramics as dielectric plates for DBD plasma actuators has not been extensively reported, despite the MgO-Al₂O₃ 6.7% porosity content, the ceramic plate was considered for further thermographic study.

The XRD results of Y₂O₃-ZrO₂, MgO-CaZrO₃, and MgO-Al₂O₃ are shown in Figure 2.

Figure 2 shows the diffractograms of Y₂O₃-ZrO₂ (top figure), MgO-CaZrO₃ (middle figure), and MgO-Al₂O₃ (bottom figure) ceramic composites. For Y₂O₃-ZrO₂, three of the raw powder phases were found in the sample, that is, m-ZrO₂, c-ZrO₂, and t-ZrO₂. However, due to higher intensity peaks, the c-ZrO₂ and t-ZrO₂ are the dominant phases. The MgO-CaZrO₃ ceramic had the most pronounced presence of calcium zirconate rather than magnesium oxide. Lastly, in addition to the MgO and Al₂O₃ phases, spinel (MgAl₂O₄) was also detected in the MgO and Al₂O₃ diffractograms. Considering the results of the XRD analysis, it was concluded that no contamination of the samples occurred during the manufacturing process of the dielectric plates.

For the three compositions, the scanning electron microscopy is illustrated in Figure 3; i.e., Figure 3a exhibits the Y₂O₃-ZrO₂, while Figure 3b is representative of MgO-CaZrO_3, and lastly Figure 3c is of the MgO-Al₂O₃ ceramic sample. A major difference in densification is observed between Figure 3a and 3b compared to Figure 3c. Micrographs (a) and (b) show that for Y₂O₃-ZrO₂ and MgO-CaZrO₃, respectively, the ceramic mixing process and green body sintering fostered good powder compaction. The grain boundary is well defined with homogeneous grain sizes and shapes.

Contrastingly, the alumina micrograph (c) shows grains of different sizes and shapes. This phenomenon is representative of an incompletely sintered ceramic material. The micrographs show the data collected and determined regarding apparent porosities and relative densities.

3.2. Thermal Conductivity of the Ceramic Samples

Thermal conductivity was computed after the experimental determination of thermal diffusivity and specific heat properties based on the flash method standard ASTM E1461-13 [31]. In thermal characterization, ceramic thermal conductivity is closely related to the design and development of new materials, manufacturing process control, and material quality assurance via analysis of the operating safe temperature ranges. Figure 4 shows the ceramic manufacturing plates’ thermal conductivity evolution with the temperature rising from 30 °C to 200 °C.

Analyzing the plot of Figure 4 leads to two main conclusions. Firstly, regardless of the ceramic sample, the thermal conductivity tends to diminish with temperature increase. Next, the Y₂O₃-ZrO₂ ceramic plate has the lowest temperature conductivity behavior, followed by MgO-Al₂O₃, which, in turn, is lower than MgO-Al₂O₃.

Yttria-stabilized zirconia exhibited good results based on the commercially available datasheets, i.e., values ranging between 2 and 3 W/(mK) [32,33]. Moreover, Zhao et al. (2006) [34] reported an interval of 2.2–2.9 W/(mK) for Y₂O₃-ZrO₂ ceramic composite, whilst Schlichting et al. (2001) [35] supported that the ceramic’s thermal conductivity is lower with the increase in temperature. For magnesium-doped calcium zirconate, thermal conductivities are reported in the literature to be between 3 and 7 W/(mK) for temperatures ranging from 20 °C to 500 °C [36,37]. Additionally, the authors confirmed with experimental and simulation work that thermal conductivity diminishes with temperature for MgO-CaZrO₃. Lastly, considering the values for magnesium-doped alumina, these were lower than expected. At room temperature, the alumina’s conductivity varies from 30 to 40 W/(mK) [38]. Consequently, a gap exists between the values measured and the ones reported in the literature. This reasoning consists of the 6.7% porosity content, shown in Table 1. Sun et al. (2014) [39] concluded that by increasing the porosity content of a ceramic material, the solid phase heat conduction decreases despite the temperature range considered.

3.3. AC SDBD Consumed Power

The average power consumed by the Y₂O₃-ZrO₂, MgO-CaZrO₃, and MgO-Al₂O₃ AC SDBD plasma actuators was computed through the electric current method.

Table 2. Ceramic-based DBD plasma actuator average power consumption [W].

Input Voltage [kVpp]	Y2O3-ZrO2	MgO-CaZrO3	MgO-Al2O3
6	3.0 ± 0.060	2.9 ± 0.058	1.7 ± 0.034
7	4.8 ± 0.096	4.0 ± 0.080	2.4 ± 0.048
8	8.3 ± 0.166	5.7 ± 0.114	3.1 ± 0.062
9	13.2 ± 0.264	7.4 ± 0.148	4.2 ± 0.084
10	20.7 ± 0.414	9.6 ± 0.192	5.6 ± 0.112

summarizes the results obtained as a function of the applied voltage levels at 24 kHz. The average electric power consumption has an associated error lower than 2%, considering the 1% tolerance of the resistor and the 1% high resolution of the digital oscilloscope. The input voltage varied between 6 kVpp and 10 kVpp since these were the minimum and maximum voltage levels at which the plasma discharge generation started (at 6 kVpp) and before it showed unstable behavior (10 kVpp), respectively.

From Table 2, despite the AC SDBD dielectric material being considered, it is possible to infer that by augmenting the input voltage, the average power consumption tends to increase. The Y₂O₃-ZrO₂ consumed the highest average power, followed by MgO-CaZrO₃, and then by MgO-Al₂O₃. At the maximum input voltage level of 10 kVpp, the MgO-Al₂O₃ (5.6 ± 0.112 W) and MgO-CaZrO₃ (9.6 ± 0.192 W) consumed, approximately, a third and a half of the average power consumed by the Y₂O₃-ZrO₂ (20.7 ± 0.414 W). These results evidence, from a practical point of view, that the MgO-Al₂O₃ is the most cost-effective actuator due to its low electric consumption. Additionally, since Y₂O₃-ZrO₂ presents larger power consumption, it is prone to dissipating more heat energy. Figure 5 exhibits the evolution of the power consumption behavior of the three dielectric-based DBD plasma actuators.

3.4. AC SDBD Thermographic Analysis

Infrared thermographic images were taken for the thermal behavior analysis of the Y₂O₃-ZrO₂, MgO-CaZrO₃, and MgO-Al₂O₃ AC SDBD plasma actuators. Heat generation during the plasma actuators’ operation must be thoughtfully comprehended to assess the ceramic-based SDBDs’ suitability as de-icing/anti-icing/ice-sensing tools. Therefore, Figure 6 presents the thermographic images in quiescent air after a 5 min temperature stabilization process at 10 kVpp input voltage. It should be noted that evaluating plasma temperature through infrared imaging is not possible due to the emissivity of the plasma resulting from its filamentary and nonuniform nature. Thus, the temperature of the dielectric surface and the exposed electrode must be considered. The grey rectangle placed in the middle of the images represents the exposed electrode position. Moreover, the asymmetries (on the sides of the electrode’s rectangle) in the temperature field distributions are justified by the circuit cables connected to the AC SDBD plasma actuators. Overall, different ceiling temperatures were recorded for the three ceramic AC SDBD plasma actuators: approximately, 155 °C, 62 °C, and 47 °C for Y₂O₃-ZrO₂, MgO-CaZrO₃, and MgO-Al₂O₃, respectively. Additionally, the temperature fields of the different ceramic dielectrics showed dissimilarities in the function of the dielectric material. These results are justified by the different thermal properties of the studied ceramic composites [37,40]. Additionally, it is strongly highlighted that the order of the maximum temperature achieved depicted in Figure 6 coincides with the power consumption analyzed in Table 2. The Y₂O₃-ZrO₂ had the highest average power consumption and ceiling temperature (20.7 W and 155 °C), followed by MgO-CaZrO₃ (9.6 W and 62 °C), and MgO-Al₂O₃ (5.6 W and 47 °C).

The results achieved evince the premise that the heat dissipation—due to plasma discharge phenomena—is responsible for the dielectric surface/its surrounding heating, promoting the anti- and de-icing features of SDBD plasma actuators. On top of that, the temperature achieved by Y₂O₃-ZrO₂ is much higher than the ones achieved by typical polymers, e.g., Kapton^®, used as dielectric barriers [29]. After being switched on for 180 s, Rodrigues et al. (2017) [29] reported that the Kapton actuators heated up to maximum temperatures of 48 °C, 40 °C, 30 °C, and 25 °C for 0.3 mm, 0.6 mm, 0.84 mm, and 1.02 mm dielectric layer thickness, respectively.

3.5. AC SDBD Spatial Temperature Variation for 10 kVpp

Additionally, for the Y₂O₃-ZrO₂, MgO-CaZrO₃, and MgO-Al₂O₃ AC SDBD plasma actuators an analysis of the spatial temperature variation for 10 kVpp and 24 kHz along the x– and y–axis follows below.

Figure 7 represents the temperature spatial variation along the x-axis for the three ceramic-based AC SDBD plasma actuators between x/l = −0.2 and x/l = 1.2 at y/w = 0.0. In this section, $l$ is the length of the exposed electrodes, whereas 0 < x/l < 1 is their frontal region. The extended interval of [−0.2, 1.2] was adopted to investigate the x-axis spatial temperature variation comprehensively. The spatial temperature variations along the x-axis were analyzed on the adjacent points of the exposed electrode edge, the region where the maximum temperature levels are located.

For the MgO-Al₂O₃ ceramic SDBD, and owing to Figure 7, it is possible to infer that the temperature distribution across the x-axis presents distinctly stable magnitudes. Additionally, MgO-Al₂O₃ depicted a temperature profile with slightly pronounced variation modification in temperature levels from distances of x/l in the ranges of [−0.2, 0.0] and [1.0, 1.2]. Although reaching higher temperatures, the same analysis—as MgO-Al₂O₃—can be made for the MgO-CaZrO₃ ceramic AC SDBD, i.e., stable temperature magnitudes with some variations around the [−0.2, 0.0] and [1.0, 1.2] interval.

When compared to both MgO-Al₂O₃ and MgO-CaZrO₃, a significant change, however, can be observed for the Y₂O₃-ZrO₂ ceramic. The AC SDBD plasma actuator made of Y₂O₃-ZrO₂ had profound dissimilarity between the temperature of the electrode’s frontal central region (0.2 < x/l < 0.8) and its edges (−0.2 < x/l < 1.0 and 1.0 < x/l < 1.2). Moreover, in terms of maximum achieved temperatures, MgO-Al₂O₃ does not achieve temperatures above 50 °C, MgO-CaZrO₃ ceramic slightly surpasses 60 °C, whereas Y₂O₃-ZrO₂ reaches over 140 °C.

Figure 7 depicts the spatial variation of the temperature across the y-axis (perpendicularly to the exposed electrode length, i.e., x/l = 0.5 outwards) for the Y₂O₃-ZrO₂, MgO-CaZrO₃, and MgO-Al₂O₃ AC SDBD plasma actuators between y/w = 0.0 and y/w = 2.5. In this section, $w$ is the width of the exposed electrodes.

Based on Figure 8 and along the y-axis, the temperature recorded was higher near the exposed electrode edge, i.e., y/w = 0.0, at the onset of plasma discharge formation. The main dissimilarities between MgO-Al₂O₃ and MgO-CaZrO₃ plasma actuators and the Y₂O₃-ZrO₂ is that the latest had a much more pronounced temperature initial increase—within the 0.0 < y/w < 0.2—and stepped decrease for the rest of the y-axis—0.2 < y/w < 2.5. This behavior represents that the temperature of the plasma is higher than the temperature of the exposed electrode.

Lastly, non-uniformities (i.e., steep spikes that indicate that the plasma is close to the limit of the dielectric material) caused by hot spots, were not identified in either of these spatial temperature profiles. These non-uniformities are commonly present in other dielectric materials, such as Kapton^®, PLA polymers, and PIB rubbers [24]. It is emphasized that the absence of non-uniformities shows the good stability of the plasma discharge created on the surface of the MgO-Al₂O₃, MgO-CaZrO₃, and Y₂O₃-ZrO₂ dielectrics.

3.6. AC SDBD Spatial Temperature Variation for Different Input Voltage Levels

Similarly to the information presented in Section 3.4, Figure 9 shows the spatial variation of the temperature along the x- and the y-axis for Y₂O₃-ZrO₂ (graphs a and d, respectively), MgO-CaZrO₃ (graphs b and e, respectively), and MgO-Al₂O₃ (graphs c and f, respectively) for different input voltage levels. For the zirconia-based ceramics, 8 kVpp, 9 kVpp, and 10 kVpp voltages were considered. Contrastingly, the higher levels of 10 kVpp, 12 kVpp, and 14 kVpp for the alumina-based ceramic were studied to demonstrate that alumina-based ceramic can also achieve considerably high temperatures. It is noted that it was not possible to apply the same vast voltage range to every ceramic composite sample due to their diverse limiting capacity in a plasma discharge.

In detail, some ceramics showed unstable filamentary behavior much sooner than others. Therefore, the alumina-based ceramic was tested up to 14 kVpp (graphs c and f), whilst the zirconia-based ones were tested up to 10 kVpp (graphs a, b, d, and e).

Figure 9a–c represent the spatial variation of the temperature along the x-axis for the sintered ceramics, i.e., Y₂O₃-ZrO₂, MgO-CaZrO₃, and MgO-Al₂O₃ between x/l = −0.2 and x/l = 1.2. The temperature variations along the x-axis were analyzed on the adjacent points of the exposed electrode edge, where the maximum temperature levels are located. Overall, the higher the input voltage, the higher the temperature next to the edge of the exposed electrode. Particularly, both zirconia-based ceramics, in Figure 9a,b, somehow exhibited a smooth temperature distribution profile between 0.0 < x/l and < 1.0. Conversely, MgO-Al₂O₃ evidenced scattered values at 14 kVpp for x/l of [−0.1; 0.1].

The phenomenon observed can be explained by the gradual deterioration of the insulating tape used to secure and electrically insulate the connecting wires of the circuitry mounted. Despite that, the rest of the profiles showed a homogeneous temperature distribution between 0.0 < x/l and < 1.0.

Figure 9d–f depict the spatial variation of the temperature along the y-axis for the manufactured plasma actuators between y/w = 0.0 and y/w = 2.5. The spatial variation of the temperature is taken across the y–axis, i.e., perpendicularly to the exposed electrode length (outwards). For all samples and from the experimental tests performed it is concluded that along the y-axis, the temperature recorded was higher near the exposed electrode edge, i.e., y/w = 0.0, in the onset of plasma discharge formation. Moreover, for MgO-Al₂O₃, little increase in temperature is noted for y/w marginally over the start of the axis reference, i.e., for 0.0 < y/w < 0.2 interval. Contrastingly, for Y₂O₃-ZrO₂ and MgO-CaZrO₃ a pronounced increase in temperature corresponds with a slightly increased interval distance outwards of the exposed electrode. The same tendency in temperature rise is noted for the x-axis variation; that is, the higher the input voltage, the higher the surface temperature.

In Figure 9a,d of the Y₂O₃-ZrO₂ composites, for 10 kVpp (approximately, 160 °C) the temperature ceiling almost doubles when compared to the temperature ceiling for 9 kVpp (approximately, 80 °C). In graphs (b) and (e) of MgO-CaZrO₃, for 10 kVpp the temperature (approximately 60 °C) is 30% higher than that registered for 9 kVpp (approximately, 45 °C). Lastly, in graphs, (c) and (f) of MgO-Al₂O₃, for 14 kVpp the temperature (approximately 78 °C) is 40% higher than that registered for 10 kVpp (approximately, 46 °C).

Lastly, in Figure 9 the plots do not show the incident fluctuations along the temperature variation curves that are typically resultant of the instability of the plasma discharge across the dielectric plate. These anomalies (nonuniformities) caused by hot spots, which are typically visible in the filamentary plasma discharge at higher voltages, appear as sharp spikes, signaling that the plasma is approaching the threshold of the dielectric material. To emphasize, these specific features were not observed in these temperature profiles but are typically seen in studies of other dielectric materials, such as Kapton^® (polyimide), PLA (polylactic acid), and PIB (poly-isobutylene) rubbers [22,41]. This demonstrates that the developed ceramic composites allow the inducing of large surface temperatures without reaching the limit of the dielectric material and, thus, are good alternatives to polymeric materials for SDBD anti-/de-icing applications.

4. Conclusions

In this study, successfully assembled Y₂O₃-ZrO₂, MgO-CaZrO₃, and MgO-Al₂O₃ ceramic composite dielectrics were thermally analyzed to propose original solutions for improving the robustness of SDBDs.

The thermal characterization evidenced that, regardless of the ceramic composite, the thermal conductivity tends to decrease with increasing temperature. Specifically, the Y₂O₃-ZrO₂ ceramic dielectric plate exhibited the lowest temperature conductivity, followed by MgO-Al₂O₃, which in turn was lower than MgO-Al₂O₃.

For 10 kVpp, the DBD plasma actuator analysis showed that Y₂O₃-ZrO₂ had the highest average power consumption and ceiling temperature (20.7 W and 155 °C), followed by MgO-CaZrO₃ (9.6 W and 62 °C), and MgO-Al₂O₃ (5.6 W and 47 °C).

Furthermore, the Y₂O₃-ZrO₂, MgO-CaZrO₃, and MgO-Al₂O₃ AC SDBD plasma actuators were studied for spatial temperature variation along the x- and y-axes during the discharge process. Two separate sets of studies were performed, one at 10 kVpp for the three ceramic-based SDBDs, and another for different input voltages (8, 9, and 10 kVpp for the zirconia-based compositions, and (10, 12, and 14 kVpp) for the alumina-based composition). For the MgO-Al₂O₃ and MgO-CaZrO₃ ceramic SDBDs at 10 kVpp, the temperature distribution across the x-axis presents distinctly stable magnitudes. Contrastingly, Y₂O₃-ZrO₂ showed a profound dissimilarity between the temperature of the frontal central region of the electrodes and its edges. For the y-axis at 10 kVpp, the major difference between the MgO-Al₂O₃ and MgO-CaZrO₃ and the Y₂O₃-ZrO₂ plasma actuators is that the latter had a much more pronounced initial temperature increase and stepped decrease for the rest of the streamwise direction.

Considering the temperatures achieved and the spatial temperature variations, the heating of the ceramic SDBD plasma actuators is believed to be a smart solution to mitigate ice accretion in cold climates.