1. Introduction
Improvement in the thermal efficiency of internal combustion engines plays a significant role in energy conservation and emission reduction, which is also a major strategy adopted by the energy and power industries of countries across the globe [
1,
2,
3]. The thermal protection coating technology is one of the critical methods used to increase the thermal efficiency of internal combustion engines. Yttria-stabilized zirconia (YSZ) ceramic coatings have been widely used as thermal insulation coatings in the internal combustion engine sector [
4,
5,
6], as they have superior properties of low heat conductivity and high fracture toughness, which can dramatically improve the piston top temperature as well as the combustion efficiency.
However, although YSZ ceramic coatings have low heat conductivity, their volumetric heat capacity is relatively high (>2500 kJ/(m
3·K)) [
7,
8,
9]. A high volumetric heat capacity often causes coatings to store a large amount of heat during insulation and heat the air intake, which can reduce the number of oxygen atoms used for combustion, resulting in incomplete combustion. As a solution to this, coatings with low heat capacity are considered as a feasible way to increase internal combustion efficiency, as their surface temperature can decrease rapidly during the exhaust stroke, which can prevent the coatings from heating the air intake during the intake stroke [
10].
At present, most of the research on thermal protection coatings mainly focuses on their low heat conductivity, while only a few research teams have studied how to realize the thermal protection function using low volumetric heat capacity. In 2017, the research team New ACE from Japan [
11] designed a new kind of coating with low heat capacity using the plasma spraying method, and its porosity was as high as 60%. However, high porosity will adversely affect the strength of materials, resulting in coating spallation at an early stage. In addition, Kogo et al. [
12], from Toyota Motor Corporation, tried to form a multi-porous anodic alumina coating on the surface of the piston chamber of a diesel engine, whose volumetric heat capacity was as low as 1300 kJ/(m
3·K) at 500 °C. However, they did not conduct detailed research on the effect of experimental parameters on the thermal property. The PEO technology, as an extension of the aluminum anodization technology, possesses the advantage of in situ growth of high adhesion. In addition, since the heat conductivity of static air is as low as 0.026 W/(m·k), an increase in coating porosity will contribute to an improvement in the thermal insulation performance of coatings. In addition, the spark breakdown characteristic of the PEO technology can give rise to a large number of pores in coatings [
13,
14,
15]. Therefore, the PEO technology has an inherent advantage when producing coatings with low heat capacity and low heat conductivity.
As PEO coatings are produced based on arc discharge in the liquid phase, their properties are closely related to discharge parameters [
16,
17,
18,
19]. For instance, research conducted by Clyne [
20], from Cambridge University, revealed that coatings with much looser microstructures could be obtained at high frequency. Similarly, Seong et al. [
19], from Mokpo National Maritime University, discovered that the microstructures and electrochemical properties of coatings could be dramatically changed by controlling discharge parameters. However, so far, limited studies have been reported on the effect of discharge parameters on heat capacity and heat conductivity of coatings prepared by PEO.
Given the above-mentioned situations, this study aims to explore the influence of different discharge frequencies on the thermal physical properties of coatings produced by PEO. In this study, PEO coatings are prepared on AlSi and Al substrates at different frequencies, and internal relations between coating mechanisms and thermal physical properties are investigated.
2. Materials and Methods
2.1. Preparation of Ceramic Coatings
In the present study, Al and AlSi (ASTM: A0336) with sizes of 15 mm × 15 mm × 5 mm and φ 38 mm × 10 mm were used as substrates, whose corners were perforated for electrical connection. Before experiments, specimen surfaces were polished with 1500 # emery paper. The polished samples were dried after ultrasonic cleaning in propanol and then in ethanol. A Na
2SiO
3 electrolyte solution was employed in experiments. In the coating process, samples were immersed in the electrolyte and used as an anode, while a stainless-steel material of grade 304 was used as a cathode. The positive terminal of the WHD-30 direct current pulse power supply (Harbin Institute of Technology, Harbin, China) was connected to the anode, and its positive terminal was connected to the cathode. The current density in the anode of all samples was 0.12 A/cm
2. Sample numbers and electrical parameters used in this experiment are exhibited in
Table 1.
2.2. Characterization of the Ceramic Coatings
The surface morphology and elemental composition of the coatings were examined by a SUPRA55 scanning electron microscopy (SEM, ZEISS, Oberkochen, German). An X-ray diffractometry (XRD, Rigaku D/Max 2500, Matsumoto, Japan) with CuKα radiation in grazing incidence mode (2°) was used. The maximum power of this XRD was 9 kW, and its working current and voltage were 45 kV and 200 mA, respectively. The coating thickness was measured with an eddy current coating thickness gauge (Huayang HCC-25A, Zhengzhou, China). The thickness was taken as the average value of the thickness measured at 10 different locations of the sample. Before the density testing, samples of PEO coatings were immersed in an alkaline solution with a high concentration to separate basement membranes to obtain laminated coatings. An electronic balance with a precision of 10−5 g (Sartorius, Goettingen, Germany)was employed to measure the quality, and the thickness and area of the stripped coatings were obtained with the microscope and the method of counting pixels, respectively, based on which the coating volume was calculated. Porosity is the ratio of the sum of the pore area in the cross section to its gross area. Before testing, SEM was used to scan the cross-section morphology of the coatings at different angles, which was followed by the calculation of the cross-section porosity of coatings based on the gray-level method (different gray levels between the hole and other regions). A multi-function surface tester (MFT, MFT-4000, Lanzhou, China) developed by the Lanzhou Institute of Chemical Physics of the Chinese Academy of Sciences was employed to measure the roughness of the coatings.
The specific heat capacity (c) was determined using a differential scanning calorimeter (DSC STA 404C, Netzsch, Selb, Germany). Before testing, coatings were obtained by putting samples in a NaOH solution of 12 g/L for 10 h and the test temperature and the heating rate were set at 300 °C and 10 °C/min, respectively. Ar was used as the purge gas, and the flow rate was 20 mL/min. The DSC curves of the empty crucible, α-Al
2O
3, and coatings were measured using the continuous heating-up scanning method, and the specific heat capacity was calculated based on their quality and the rate of heat flow. The thermal diffusivity (α) of PEO coatings was measured by a laser heat conductivity testing instrument (LAF467 HyperFlash, Selb, Netzsch, German). The data acquisition rate of LAF467 HyperFlash (version: NETZSCH Proteus Software 7.1.0) was up to 2 MHz (mainly reduces the semi-heating signal), and the sampling time could be as low as 1 ms. So it is more accurate for thin film samples; the minimum sample thickness for LAF467 can be as low as 0.01 mm. The heat conductivity (γ) was derived from
where α is the thermal diffusivity; ρ and c are the density and the specific heat capacity; respectively; and c
v is the volumetric heat capacity.
3. Results
3.1. Growth Rate of PEO Coatings
Figure 1a shows the relationship between the current/voltage and time of samples (anode) in the PEO process, during which the power supply was in bipolar constant voltage mode. At 0–80 s, the current/voltage rose quickly but no plasma discharge was detected. This stage is attributed to anodic oxidation, in which amorphous alumina coatings were formed on the surface [
21]. After 80 s, plasma discharge was observed on the surface of the samples, and after 140 s, when the current reached the set value, sustained, stable discharges occurred on the surface while the growth rate of the voltage tended to reduce. As
Figure 1b shows, τ
a in the plot is the anode pulse stage, where discharge breakdown occurs on the surface of samples. τ
c in the plot is the cathode pulse stage, where the voltage of the negative electrode is about 75 V.
During the electric discharge machining process, the discharge frequency determines the breakdown in unit time as well as the arcing time of each pulse, thereby affecting the microstructures and properties of coatings.
Figure 2 compares the growth rates of different coatings. It can be seen that the growth rates of Al-PEO coatings are related to their looseness (porosity). According to
Figure 2, sample A1# had the loosest structure, allowing coatings to have higher growth rates. Sample A4# possessed the highest density, and its growth rate was the lowest correspondingly. Nevertheless, for AlSi-PEO coatings, there was no direct relationship between their growth rates and porosity. For example, although sample AS5# was looser than AS4#, its growth rate tended to decrease slightly. This can be explained by the changes in discharge characteristics induced by Si. During the discharge process, a greater breakdown voltage and breakdown energy are required to spark through Si, which has higher electricity resistivity than Al. At low frequency (50–1000 Hz), the energy provided by a single pulse is sufficient enough to initiate the breakdown effect in ceramic coatings. However, as the frequency increases, the energy of a single pulse decreases so dramatically that it fails to break areas that have large-size bulk Si. It only discharges in weak areas, which reduces the integrity of the coatings produced by plasma electrolytic oxidation.
3.2. Morphology and Phase Composition of PEO Coatings
During the electric discharge machining process, the discharge frequency determines the breakdown in the unit time as well as the arcing time of each pulse, affecting the microstructures and properties of coatings.
Figure 3 and
Figure 4 illustrate surface morphologies of coatings.
Figure 5 shows the roughness of all coatings. It can be seen that with frequency increasing, melting materials on the surface of Al-PEO coatings reduced and the surface roughness of the coatings decreased. However, further increase in the frequency (1000–1400 Hz) resulted in a slight increase in the surface roughness (
Figure 5). Therefore, compared to Al-PEO coatings, frequency has a more significant effect on the microstructure of AlSi-PEO coatings. According to
Figure 4, as the frequency increased, a crater structure, characterized by large-size pores, was observed on the surfaces of sample AS4# and sample AS5#. This is because a further increase in the pulse frequency gives rise to a continuous reduction in single-pulse energy, which fails to penetrate areas with a high thickness in ceramic coatings to allow coating growth to proceed. Concentrated arc discharges take place in weak areas of coatings. As a result, large-size open pores occur on the coating surface due to intense partial discharges. In addition, the surface roughness of the coatings increases with the presence of large-size pores.
Figure 6 compares partially enlarged views of AlSi-PEO coatings at low and medium frequencies and at high frequency. At low and medium frequencies, small-size micropores (1–2 μm) were observed on the coating surfaces, on which there were large coverage areas formed by melting. However, at high frequency, the micropore size was relatively large (2–5 μm) and the pores had an open structure connected to outside air.
To further analyze the microstructure and porosity of coatings, the microscopic morphology of coating cross sections was examined (
Figure 7) and the changes in Al-PEO coating density and porosity at different frequencies were compared (
Figure 8). Hussain et al. [
22] categorized discharges during the PEO process into three types. A-type discharges occur on the surface of PEO coatings. Type B refers to strong discharges between PEO coatings and the substrate surface, and during this process, a number of molten oxides eject instantaneously from the discharge channels, which forms a relatively large-size “pancake” structure on the surface of the coatings [
23,
24,
25]. C-type discharges occur under PEO coatings. In
Figure 3 and
Figure 7, it can be seen that of all Al-PEO coating samples, A1# exhibited a typical pancake structure induced by B-type discharges. The reasons for this phenomenon will be discussed in
Section 4 of this paper.
From
Figure 8, it is evident that for Al-PEO coatings, sample A1# had an obvious porous structure and its porosity tended to decrease as the frequency increased, which is consistent with the pattern of the microscopic morphology observed on coating surfaces. As
Figure 8 shows, sample A4# had the lowest porosity, as well as the highest density.
In this section, the cross-section morphologies of AlSi-PEO coatings are compared (
Figure 9) and the relationship between coating porosity and density is computed and analyzed (
Figure 10). It is evident that the changing pulse frequency can significantly affect the pore structure of coatings, which is similar to the change pattern of surface morphology. At a frequency of 50–600 Hz, with increasing frequency, the porosity of coatings tended to reduce gradually. However, a further increase in the frequency (600–1400 Hz) led to a gradual increase in the coating porosity. In addition,
Figure 10 suggests that sample AS3# had the lowest porosity and the highest density, whereas an increase in the porosity of sample AS4# and sample AS5# led to a dramatic reduction in coating density. A comparison between
Figure 9 and
Figure 10 demonstrates that the change pattern of coating density is not linearly correlated with its porosity, as they are also affected by the compositions of coatings.
3.3. Phase Structure of PEO Coatings
Figure 11 shows the XRD patterns of Al-PEO coatings. It can be seen that in Na
2SiO
3 solution, Al coatings were mainly composed of α-Al
2O
3, γ-Al
2O
3, Al
2O
3·2SiO
2, and an amorphous aluminum silicon compound. Sample A1# had many amorphous structures, and the diffraction intensity peaks of the phases were weak. As the pulse frequency increased, the diffraction intensity peaks of α-Al
2O
3 and γ-Al
2O
3 showed a tendency to rise first and then fall. In addition, it is evident that when the pulse frequency was greater than 50 Hz, an obvious mullite phase could be observed in the coatings, and their diffraction peak intensities decreased gradually with a further increase in frequency. This change is consistent with the research reported by Dehnavi et al. [
26], who believe that as the pulse frequency rises, an increase in the number of discharges will dissociate more SiO
32− in the solution from the growing coating surfaces. For Al substrates, since Si in 3Al
2O
3·2SiO
2 mainly comes from SiO
32− in the solution, the mullite phase in the coatings will decrease gradually due to increasing frequency.
Figure 12 shows the XRD patterns of AlSi-PEO coatings. Similar to Al-PEO coatings, AlSi-PEO coatings were also mainly composed of α-Al
2O
3, γ-Al
2O
3, 3Al
2O
3·2SiO
2, and an amorphous aluminum silicon compound, but the diffraction peak intensities of the phases exhibited different change tendencies. In AlSi-PEO coatings, the Si in 3Al
2O
3·2SiO
2 mainly comes from the AlSi substrates as well as the solution. Unlike Al-PEO coatings, in AlSi-PEO coatings, with rising frequency, the diffraction peak intensity of the mullite phase tends to increase, which can be explained by our previous research. In the AlSi substrate, the edge effect at the interface of the Si phase and the Al phase results from resistivity difference between the two (Al: 2.83 × 10
−8 Ωm; Si: 2.3 × 10
3 Ωm), because of which the edge becomes the weak area, where discharge breakdown can easily occur [
27]. Therefore, a rising pulse frequency leads to an increasing number of breakdowns at the interface of the Si phase and the Al phase in unit time, which causes more Si to melt into coatings, resulting in a higher diffraction peak intensity of the mullite phase.
According to XRD change patterns, it is evident that at a pulse frequency of 50–1000 Hz, the diffraction peak intensity of γ-Al2O3 in the coatings gradually increased with rising frequency and decreased slightly at a pulse frequency of 1400 Hz. This is because at low frequency, a single pulse has excessive energy and the high temperature resulting from the partial discharge adversely affects the production of the γ-Al2O3 phase. However, the increasing frequency leads to a reduction in single-pulse energy as well as temperature in a small area where discharge occurs, so a γ-Al2O3 phase can be observed in the coatings. When the frequency is greater than 1400 Hz, however, coating growth is mullite phase oriented, resulting in a decrease in the diffraction peak intensity of γ-Al2O3.
3.4. Heat Capacity of PEO Coatings
From the perspective of coatings, composition and pore structure are two major determinants of their volumetric heat capacity. To better analyze the effect of changing pulse frequencies on the heat capacity of the coatings,
Figure 13 compares the relationship between the heat capacities, porosity, and diffraction peak intensities of α-Al
2O
3 and γ-Al
2O
3.
As observed in
Figure 13, the volumetric heat capacity of Al-PEO coatings showed a tendency to increase first and then decrease, which is consistent with the changing pattern of porosity. Before the coating process, the volumetric heat capacity of Al is 2422 kJ/m
3·K, which is significantly lower than that after the PEO process. The coatings had the lowest volumetric heat capacity at a pulse frequency of 50 Hz, which can be explained by two reasons. First, sample A1# had the least content of α-Al
2O
3 and γ-Al
2O
3 and was mainly composed of Al
2O
3, with a low heat capacity, which leads to a decrease in the heat capacity. This phenomenon is consistent with the research reported by Clyne [
28]. Second, sample A1# was characterized by high porosity and low density. The positive correlation between volumetric heat capacity and density contributes to the low heat capacity of sample A1#. At a pulse frequency of 1000 Hz, the coatings had the highest porosity and their number of micropores decreased dramatically with porosity as low as 3.2%. However, the characteristics of low porosity and high density are the main causes of the highest volumetric heat capacity of the coatings.
Figure 14 illustrates the relationship between the volumetric heat capacity, porosity, and phase structure of AlSi-PEO coatings. Similar to the Al substrate, the heat conductivity of AlSi-PEO coatings (2349 kJ/m
3·K) was dramatically lower than that of the AlSi substrate. It can be seen that sample AS4# had the lowest volumetric heat capacity, which mainly resulted from the compositions and porosity of the coatings. First, at 300 °C, since the heat capacity of the γ-Al
2O
3 phase is lower than that of the α-Al
2O
3 phase, the lower the volumetric heat capacity, the weaker the intensity of the γ-Al
2O
3 diffraction peak with a moderate level of crystallinity in coatings. As for AlSi-PEO coatings, sample AS4# had the highest peak intensity. Second, the high porosity and low density of sample AS4# also contribute to its low heat capacity. Similarly, the high volumetric heat capacity of AS1# is attributed to its low porosity as well as the lowest intensity of the γ-Al
2O
3 diffraction peak. To sum up, the changing pattern of the volumetric heat capacity of AlSi-PEO coatings is not exactly the same as that of their porosity; instead, it is affected by changes in composition.
3.5. Heat Conductivity Coefficient of PEO Coatings
Table 2 shows the thermal diffusivity α and heat conductivity γ of substrates and coating samples, from which it can be concluded that the PEO process dramatically reduced the thermal diffusivity and heat conductivity of materials. According to
Table 2, the AlSi alloy had lower thermal diffusivity α and heat conductivity γ than Al, which is attributed to the addition of the Si element to the alloy. This result is consistent with the research conducted by Zhang et al. [
29], who found that an increase in Si in the AlSi alloy leads to a decrease in thermal diffusivity α and heat conductivity. Sample A5# had the lowest heat conductivity, which is attributed to its low thermal diffusivity and relatively low volumetric heat capacity. From the perspective of microstructure, the low heat conductivity of sample A5# mainly results from a large number of closed pores in coatings. As
Figure 7 shows, there were a number of closed pores in coatings, which store trace gases during the formation, forming a large quantity of micron-size closed cells. It is well known that the heat conductivity of static air is as low as 0.023 W/(m·K), so such closed pores lead to a dramatic reduction in the thermal diffusivity and heat conductivity of coatings.
A comparison betwee sample A1# and sample A5# suggested that both coatings had low heat conductivity. Although sample A1# had higher porosity, its heat conductivity was higher than that of sample A5#. This is because large-size open pores in coatings are connected to outside air, which enhances the thermal convection. As a result, sample A5# had higher thermal diffusivity, resulting in the higher heat conductivity of sample A5#.
The thermal diffusivity α and heat conductivity γ of Al-Si PEO coatings showed a tendency to fall first and then rise. Sample AS2# had low thermal diffusivity and heat conductivity, which resulted from the closed pores in the coating, similar to Al-PEO coatings. Similarly, although sample AS4# and sample AS5# had high porosity and low densities, a large number of open, large-size pores inside led to a dramatic increase in the thermal diffusivity of the coatings.
4. Discussion
The above-mentioned data and results show that frequency changes can have effects on the microstructures, mechanisms, and thermal physical properties of PEO coatings. The following section will discuss changes in the property of coatings prepared on different substrates at high and low frequencies.
At different frequencies, the difference between the Al substrate and the AlSi substrate in discharge mode gives rise to a change in the pore structure. During the growth process of a PEO coating on an Al substrate, since the electric field is uniform due to the homogeneity of the substrate, the increasing frequency will result in a reduction in the single-pulse energy as well as an increase of penetrations. This kind of low-energy discharge can cause an increased number of melting zones but a reduction in size, while forming a large quantity of small-size closed pores (see
Figure 7). This is consistent with the research reported by Sobolev et al. [
23,
24], as at a low frequency (A1#), the long duration of each pulse severely heats up the coating, while B-type discharges become dominant, resulting in a large number of open-structured pores on the surface (see
Figure 3). However, with frequency increasing, the pulse energy tends to reduce gradually, which allows A-type discharges and C-type discharges to be dominant, leading to a decrease in the porosity. Therefore, with a rise in the frequency, the roughness and porosity of Al–PEO coatings tend to decrease gradually.
Unlike an Al substrate, an AlSi substrate demonstrates a different discharge mechanism. Experiments have demonstrated that at a frequency of 50–600 Hz, A-type discharges and C-type discharges tend to be dominant with the rising frequency, which is attributed to the reduction in single-pulse energy. Nevertheless, a further rise in frequency results in a sudden increase in porosity, which is affected by the Si element in the alloy. First, the AlSi substrate contains a large number of bulk Si phases of different sizes. Due to the difference in the electricity resistivity between Al and Si, the edge effect and the electric field concentration can be easily induced at the interface of the Al phase and the Si phase. Discharge breakdown most likely occurs when the electric field strength is higher than the breakdown voltage of the coating thickness (this phenomenon has been studied in detail in our previous research [
27]). During the growth of AlSi-PEO coatings, with the increase in coating thickness, it requires higher breakdown energy to allow the micro-arc oxidation to proceed. However, as the increasing frequency causes a reduction in single-pulse energy, the electric arc concentrated discharges often take place in weak areas such as the interface of the Al phase and the Si phase. The continuous concentrated discharges produce large-size open pores in coatings at a high frequency. Second, discharge breakdown can produce an instantaneous high temperature that is higher than the melting point of Si (~1685 K), SiO
x (<2000 K), Al (~930 K), and Al
2O
3 (~2300 K). This enables materials around the electric melting pool to mix together in the state for melting, forming the Al
2O
3 phase, the SiO
x phase, and the Al-Si-O phase. Since the melting point of the SiO
x phase and the Al-Si-O phase is lower than that of Al
2O
3 phase, the surrounding areas of bulk Si phase are easier to penetrate and concentrated discharge will be induced. This can also explain why the size of pores on AlSi-PEO coatings increases with the rising frequency (increased penetrations).
The analysis reveals that the pore structure characterized by the coatings can reduce their heat capacity and heat conductivity and different pore morphologies have different impacts on both properties. These air-filled closed pores can reduce the heat conductivity of the coatings, thereby decreasing their heat capacity, which is consistent with the research reported by Professor Clyne, from the University of Cambridge. In his research, Clyne [
20] investigated the effect of loose and porous layers of PEO coatings on their heat conductivity and found that more air could occupy the pores during the heat transfer process, thus decreasing the heat conductivity. However, open pores have a different influence mechanism. Since the pores are connected to the outside air, the thermal convection in the coatings can be enhanced, which leads to increased heat conductivity and heat capacity. However, according Equation (1), the heat conductivity of the coatings with an open-pore structure can be indirectly decreased through a reduction in coating density.