Thermal Insulation Properties of Hollow Y2O3:Eu Spheres Laminated with Er2O3

Currently, hollow sphere insulating materials are of importance for applications such as energy storage and savings and cryogenic engineering. The structures are formed by single hollow spheres, which can be joined, for example, by sintering. In this study, a 15 wt% Er-EDTA complex aqueous solution in which hollow Y 2 O 3 spheres were mixed was used as the deposition body, and pencil spraying and sintering (PSS) was used to synthesize an Er 2 O 3 hollow Y 2 O 3 sphere composite lm on a polished Si substrate. The structure of the composite lm was successfully controlled by adjusting the 15 wt% Er-EDTA solution/hollow Y 2 O 3 sphere mass ratio and the jet-to-substrate distance in the PSS process. In addition, the thermal insulation capability of the lms was evaluated by the thermal steady-state method. The results show that the Er 2 O 3 /hollow Y 2 O 3 :Eu sphere composite lms have a higher thermal insulation capability at a jet-to-substrate distance of 150 mm and a mass ratio (g) of 3.5:1. For the composite lms with thicknesses of 38–92 µm, cross-sectional hollow ratio of 0.8–8.7% and void ratio of 6.3–13.1%, the temperature drop due to the porous (including hollow spheres and voids) structure lms at 440°C is ΔT f =47°C. This is mainly associated with the lm having more complicated microstructures. Therefore, the Er 2 O 3 /Y 2 O 3 :Eu composite lm has good thermal insulation performance, and a simple preparation method for many kinds of hollow sphere lms with complex structures and high porosities by using complex solutions with different compositions is provided.


Introduction
In recent years, hollow sphere structure materials have been developed, and compared with ordinary materials, they have light weight, low density, and large speci c surface area and are characterized by high speci c stiffness, the ability to absorb high amounts of energy at relatively low stress levels, the potential for noise control, vibration damping and thermal insulation, and reduced consumption of resources and environmental issues [1][2][3][4]. A combination of these different properties provides a wide range of potential multifunctional applications, especially in the automotive and aerospace industries [5].
Typical functional applications of porous materials include in insulation for furnaces, re retardant systems, thermal-mechanical aerospace structures, etc. [6][7][8][9]. Moreover, porous materials can be classi ed as either open-or closed-cell based on their structural properties. Compared to open-cell structures, closed-cell structures have advantages in thermal insulation due to their load-bearing cell walls [10][11][12][13] and have received considerable attention in recent studies. For lms with high thermal insulation performance, in addition to the design of porous structure lms, low thermal conductivity of the material itself is important, such as for yttrium oxide (Y 2 O 3 ) materials [14,15].
A considerable fraction of this kind of research has been devoted to Y 2 O 3 [16][17][18][19][20][21][22][23]. Previous investigations have shown that pores are of consequence in decreasing the thermal transfer of Y 2 O 3 lms, in which the deposited splat forms a honeycomb microstructure lm that exhibits a good thermal insulation capability [16]. Furthermore, hollow Y 2 O 3 spheres may be a potential candidate for thermal insulation materials by combining the low thermal conductivity of Y 2 O 3 with hollow structures. In a previous study, hollow (Y,Eu)-EDTA complex powder was obtained by the spray-drying technique, which has a morphological and compositional design ability. The results indicated that the hollow sphere form of Y 2 O 3 :Eu with a diameter of 1-30 µm was obtained by thermal decomposition of the complex powder [24].
In a follow-up study, a spray dryer with a two-uid nozzle was used to synthesize hollow Y 2 O 3 :Eu spheres, in which the size of the hollow spheres could be controlled. When the sintering temperature was 600°C, a hollow sphere with uniform size, a dense shell, and high porosity was obtained [25]. By combining the advantages of the low thermal conductivity and thermal stability characteristics of the Y 2 O 3 material, thermal insulation lms with complicated macroscopic and microscopic pores formed by the pencil spraying and sintering (PSS) method and controllable pores derived from hollow Y 2 O 3 :Eu spheres can be synthesized.
In this paper, Er 2 O 3 /hollow Y 2 O 3 :Eu sphere composite lms were successfully synthesized using a 15 wt% Er-EDTA complex aqueous solution as a green body for joining the spheres to form interdependent structures by the PSS process. The composite lms are designed to be constructed by sintered compact hollow spheres, which not only results in a closed-cell structure to ensure mechanical strength but also leads to relatively high porosity to guarantee a good thermal insulation performance. The effect of spray parameters (jet-to-substrate distance and mass ratio of the solution) on the structure of the composite lms was studied, and the connection microstructure between the hollow spheres was discussed based on energy-dispersive X-ray spectroscopy (EDX  Fig. 1(a) show many microscopic pores, which will form pores in Y 2 O 3 :Eu ceramics and guarantee high porosity. As shown in Fig. 1(b), only the cubic Y 2 O 3 phase exists in the calcined Y 2 O 3 :Eu.
As shown in Fig. 1(c), the mean particle size of the calcined Y 2 O 3 :Eu is 10.26 µm, the distribution is uniform, and this material will be bene cial to the fabrication of green bodies due to the suitable particle size distribution. Moreover, the uorescent agent Eu was added to determine whether the hollow spheres were uniformly present in the subsequent lm preparation. To investigate the effect of hollow spheres on the thermal insulation properties of the ceramic samples and optimize the content of hollow spheres in the layers, different spray parameters were used to prepare the lms, and the details are listed in Table 1. Figure 2 shows the spray equipment (SB1060, DENKEN-HIGHDENTAL Co., Ltd., Japan) used for deposition of the value in the quasi-stationary state (between 1500 and 1800 s) of the temperature difference between the temperature T 2 on the hot plate side and temperature T 1 on the back surface of the sample was taken as the temperature gradient ΔT for that sample (ΔT = T 2 -T 1 ). The thermal insulation capability ΔT f was calculated as the difference between the ΔT of each sample and the measured temperature gradient ΔT si of the Si substrate (ΔT f = ΔT-ΔT al ). The test model is shown in Fig. 3. From the above, the steady-state method was applied to perform a measurement of the thermal insulation performance when the sample had reached thermal equilibrium[16, 27].

Characteristics
The phases of the various synthesized coatings were characterized by X-ray diffraction (XRD, M03XHF22, Mac Science, Kanagawa, Japan) with Cu-Kα radiation over 2θ values of 10 90°. The hollow sphere morphology, microstructure and composition of the lms (or deposition bodies) were observed using eld-emission scanning electron microscopy (FE-SEM, JSM-6700F, JEOL, Japan) equipped with EDX (JED-2201-F, Japan). Secondary electron and backscattered electron (compositional) images of the synthesized composite lms were acquired. All the synthesized lms for cross-sectional analysis were rst embedded in transparent epoxy resin and then polished with water-resistant abrasive paper. The lm thickness is the average thickness of 20 transverse sections, which was estimated by analyzing the whole SEM image with the commercial software SmileView, and measurements were repeated three times. Moreover, the number of hollow spheres, hollow ratio (hollow sphere internal area ratio) and void ratio of the lm were evaluated based on 2D image analysis via ImageJ software. Image analysis has been established as a reliable method for determining the pore (or void) size and morphology in synthesized lms, keeping in mind its resolution limits [28][29][30]. Therefore, the numbers of hollow spheres (selecting an inner diameter of 0.5 µm or larger) and voids of the lms were measured from the SEM images. In addition, the acquisition parameters were xed to an image resolution of approximately 0.2 µm/pixel, and based on these conditions, a series of 5 images were recorded to obtain an accurate value. deposited at different spray distances and mass ratios. The peaks in the XRD pro les were assigned using International Centre for Diffraction Data (ICDD) cards as references. In Fig. 5(1)-(6), the hollow sphere composite lms deposited on Si substrates only exhibit cubic crystalline phases. Furthermore, diffraction peak splitting was observed in the deposited lms containing the Er-EDTA complex and Y 2 O 3 ceramic (after sintering EDTA (Y,Eu) H), the details of which were explained in our previous report [32]. In addition, the peak intensity of the composite lm phase hardly changes with changing spray conditions.

Results And
To survey the surface morphology and structure of the as-synthesized Er 2 O 3 /hollow Y 2 O 3 :Eu sphere composite lm formed by a one-time PSS process, SEM and EDX were carried out. As depicted in Fig.  6(a), three structures were formed as follows: a brous ribbon structure after sintering (red dotted lines I), a atbed layer structure (red dotted lines II), and a thin layer on the sphere wall (red dotted lines III) covering (or bonded to) the hollow sphere particles on the substrate. The elemental composition of the Er 2 O 3 /hollow Y 2 O 3 :Eu sphere composite lm was determined using EDX. The EDX images in Fig. 6(b) con rm the existence of Er (red) and Y (green) in the composite lm. Combined with the SEM and EDX images, the above three structures can be con rmed to be composed of Er. This means that the connected structure (red dotted lines I, II, and III) of the hollow spherical particles is composed of the metal oxide Er 2 O 3 . Then, thick lms were deposited after increasing the PSS process to three times in the following.
The surface morphology and structure of the as-synthesized samples were characterized by SEM, as shown in Figs. 7 and 8. When the jet-to-substrate distance is 150 mm, hollow Y 2 O 3 spheres and smaller particles ( 1 µm) are uniformly deposited on the substrate (Fig. 7(1)-(3)). In particular, the surface morphology of sample (1) is more compact. The magni ed images in Fig. 7(1 a )-(3 c ) reveal that hollow Y 2 O 3 spheres with diameters of 0.1-2 µm have a remarkable cavity, and the hollow Y 2 O 3 ceramic has a good spherical shape; thus, the internal pores are retained. There are gaps between the Y 2 O 3 spheres, and the distance between them is large, so the interface is not obvious. The composite lms also obviously consist of smaller and larger spheres bonded between them. Furthermore, a brous ribbon structure was observed on the surface of the Y 2 O 3 spheres at mass ratios of 5.2:1 and 3.5:1 (as shown in Fig. 7(2 with the red dotted line). When the jet-to-substrate distance was 200 mm, hollow spheres with an 6 µm diameter were uniformly deposited on the substrate, and the degree of dispersion of the hollow spheres on the surface became larger, as shown in Fig. 8(4)-(6) (compared to the jet-to-substrate distance of 150 mm). Similar to the above, a brous structure is also observed in Fig. 8(6  special structure: (a) a porous (or top) layer, (b) a middle layer, and (c) a bonding layer, as shown in Fig.   9(2). The reason why the composite lms are three layers is that the lms were deposited on the substrate by the PSS process three times under the same conditions. In addition, with the decrease in the mass ratio of the solution, the uidity of the Er-EDTA complex aqueous solution increases, and many obvious voids form between the hollow Y 2 O 3 spheres. These voids become more obvious when the mass ratio is decreased to 3.5:1 ( Fig. 9(3)). The magni ed images in Fig. 9(1 a ) and (2 b ) reveal hollow spheres constructed by interconnected dense Er 2 O 3 building blocks that ll the surrounding hollow Y 2 O 3 sphere architectures. Additionally, the hollow spheres obviously have a thin wall with a thickness of approximately 0.3 µm. Furthermore, in these images, the hollow structure of Y 2 O 3 because of the removal of the gas during calcination is visible. However, the cross-sectional image of Fig. 9(3 c ) shows a loose structure, which is due to the gaps formed after the Er-EDTA complex aqueous solution was deposited and solidi ed during the sintering process. It also makes connection of the composite lm and substrate easier. With a jet-to-substrate distance of 200 mm, sample (5) has a looser structure in the top layer and the middle layer than sample (3), as shown in Fig. 10 (5). For further characterization of the bonding layer (c) of the synthesized composite lms, EDX mapping is useful, and homogeneous dispersion of Er (red) and Y (green) over the particles can be observed in Fig. 11(b). Additionally, Er is obviously mainly distributed in the bonding layer (c). Therefore, this can also prove that the higher density of the bonding layer is due to the lling of Er 2 O 3 (combined with Fig. 11(a) and (b)). The denser layer of the synthesized   (1) has an average thickness of 38 µm, 67 hollow spheres, a hollow ratio (selecting a hollow diameter of 0.5 µm or larger) of 0.8%, and a void ratio of 6.3% when the jet-to-substrate distance is 150 mm. With a distance of 200 mm, sample (4) has almost the same lm thickness (40 µm), hollow ratio (0.7%) and void ratio (6.4%); however, the number of hollow spheres is less than that of sample (1). The dispersion of hollow spheres can be considered to become higher when the spraying distance becomes larger. This is also obvious in the results for sample (3) and sample (6) with the same mass ratio of 3.5:1. These results indicate that the cooling of the substrate is important for depositing composite lms with the three kinds of layers (shown in Fig. 9(2)). Such a microstructure (including voids) could improve the thermal insulation capability because hollow spheres with a loose distribution could form more pores (or deposition voids) to lower the thermal conductivity, and a more horizontal layer gap might be bene cial for resisting thermal conduction [33]. Thus, this material can be expected to be a good thermal insulation material.

Thermal insulation performance of Er 2 O 3 /hollow Y 2 O 3 :Eu composite lms
The thermal insulation performance of Er 2 O 3 /hollow Y 2 O 3 :Eu composite lms mainly depends on the special hollow structure and the number of hollow spheres distributed in the lms. Hollow Y 2 O 3 spheres could greatly suppress thermal conduction due to the presence of a large fraction of air cavities within the micrometer range, making them an ideal system for creating superinsulating materials. In this study, the thermal insulation capability was evaluated by the temperature drop across the composite lms (ΔT = T 2 -T 1 ). Figure 13 shows the recorded heating temperature curves of the heater ( lm surface, T 0 ), substrate (reference) backside (T 1 ), and backside of specimens synthesized with different mass ratios: (T 2 ) 6.9:1 and (T 3 ) 3.5:1. Figure 13 shows that as the heater temperature (T 0 ) increases, T 1 , T 2 , and T 3 all increase. The results of the temperature data indicate that T 1 , T 2, and T 3 become stable when the holding time at a T 0 of 440°C is longer than 20 min. The backside temperatures of all samples are lower than that of the substrate. This means that the synthetized samples contribute to decreasing the temperature of the substrate. Moreover, when the jet-to-substrate distance is 150 mm, there is an obvious distinction between sample (1) and sample (3) with increasing temperature, as shown Fig. 13(a). For a distance of 200 mm, the temperature curves of the samples ( (4) and (6)) are not obviously different in Fig. 13(b). Based on the evaluation results of the lms in Table 2, the better thermal insulation capability of sample (3) than that of sample (1) can be explained as being due to the values of the number of hollow spheres, hollow ratio (%), and void ratio (%) being much higher. This is also the reason why the heating curves of sample (4) and sample (6) are not obviously different. An interesting nding is that the temperature curves of samples (3) and (6) have obvious uctuations, but those of samples (1) and (4) Table 3. Apparently, the thermal insulation materials play a more important role in a higher temperature environment. Moreover, the tables also show the ΔT f values for the lms and ΔT f per micron. ΔT f depends on the porosity (hollow spheres and voids) of the composite lms. ΔT f increases with increasing composite lm porosity (hollow sphere/void ratio) from 0.8/6.3% to 8.7/13.1% at a jet-tosubstrate distance of 150 mm (see Table 2). However, when the jet-to-substrate distance is 200 mm, sample (5) exhibits higher ΔT f than sample (6) at similar lm thicknesses. This can be because when the concentration of hollow sphere particles is lower at a mass ratio of 5.2:1, the capability of the surface of the hollow spheres to adhere to the solution is reduced, and more voids are formed (shown in Fig. 10 (5) and (6) (3) and (6)) and ΔT f values of 47°C and 21°C, respectively, with a mass ratio (g) of 3.5:1 (shown in Table 3). Moreover, when the mass ratio (g) is 5.2:1, samples (2) and (5) exhibit similar ΔT f (33 and 32°C) at different thicknesses of 51 and 69 µm. This means that the thickness has no obvious in uence on the thermal insulation capability. This also means that the spray conditions could affect the formation of the microstructure in the lms, which is bene cial for improving the thermal insulation performance. In addition, the ΔT f per micron of the thinnest sample (2) is higher than that of the other coatings. This result can occur for several reasons. First, it could be due to the larger undulation in the outer layer, in which the hollow spheres are uniformly distributed in the top, bottom and inner layers. Second, as mentioned above, sample (2) has higher hollow (6.2%) and void (9.1%) ratios, but the structure itself also has two uniform dense layers, as evidenced in the cross-sectional SEM images ( Fig.  9(2)). Finally, a number of large hollow spheres and more voids were formed and oriented along the axis parallel to the lm-substrate interface, which improves the thermal transfer time, resulting in lowing of the temperature, due to the lm surface being perpendicular to the heat ow in this test.    Photographs of the Er2O3/ hollow Y2O3:Eu sphere composite lms deposited on Si substrates.   Surface SEM images of Er2O3/ Y2O3:Eu composite lms((1)-(3)), and higher magni cation images of Er2O3/ hollow Y2O3:Eu sphere composite lms((1a)-(3c)) synthesized at the jet-to-substrate distance of 150mm.

Figure 12
Flow schematic diagram of the composite lm synthesis. Figure 13