Study of TiO2 on the Voltage Holdoff Capacity of Cr/Mn-Doped Al2O3 Ceramic in Vacuum

With the development of vacuum electronic devices toward high power, high frequency, and miniaturization, the voltage holdoff capacity of the insulation materials in devices has also been raised to a higher demand. Cr/Mn/Ti-doped Al2O3 ceramics were prepared, and the bulk density, micromorphology, phase composition, resistivity, secondary electron emission coefficient, and surface flashover threshold in the vacuum of the Al2O3 were characterized. The results show that the addition of TiO2 to the Al2O3 ceramic can promote the sintering of the ceramic. The Cr/Mn/Ti-doped Al2O3 ceramic with a homogeneous microstructure can be obtained by an appropriate amount of TiO2 addition. In the process of the heat treatment, the TiO2 in the ceramics was reduced to a certain degree, which had an impact on the microstructure of the Al2O3 ceramic. Adding a small amount of TiO2 can improve the voltage holdoff performance in the vacuum. The value of the surface flashover threshold in the vacuum of the Cr/Mn/Ti-doped Al2O3 ceramic containing 1 wt.% TiO2 reached a value of 33 kV, which is 32% higher than that of the basic Al2O3 ceramic. The preparation of Al2O3 ceramics with a high voltage holdoff capacity in a vacuum provides fundamental technical support for the development of vacuum electronic devices.


Introduction
Al 2 O 3 ceramics play a significant role in vacuum electronic devices, such as high-voltage insulation, vacuum sealing, power transmission, and support fixation [1,2]. Studies have shown that as a solid insulator, the voltage holdoff capacity of Al 2 O 3 ceramics in a vacuum is often lower than vacuum gaps of the same size. The surface of the insulator is the weakest point in vacuum-insulation systems because its bulk voltage holdoff capacity is generally greater than the vacuum gap of the same size [3,4]. Surface flashover occurs in many vacuum electronic devices when the applied voltage exceeds a certain value, resulting in the failure of or damage to the device [5][6][7][8][9][10]. With the development of vacuum electronic devices toward high power, high frequency, and miniaturization, such as high-power klystron and high-power pseudo spark switch, the operating voltage of the device increases exponentially, making surface flashover of Al 2 O 3 ceramics become an important factor affecting their reliability and restricting the development of vacuum electronic devices [11][12][13][14][15]. Therefore, improving the surface voltage holdoff capacity of Al 2 O 3 ceramics has become one of the urgent problems to be solved in the field of vacuum electronics.
Surface flashover is related to many factors, such as the materials' characteristics, shape structure, and surface roughness of the ceramic. According to secondary electron emission avalanche theory (SEEA) [16][17][18], the initiation of a surface flashover is usually started by the emission of electrons (generally by field emission or thermal field emission) from the cathode triple junction (CTJ-the interface where the insulator, cathode, and vacuum are in close proximity). Some of the electrons impact the surface of the insulator, producing

Samples Preparation
The 95% Al2O3 ceramics were chosen as fundamental Al2O3 ceramic materials. SiO2 and CaO were introduced in the form of silica and calcium carbonate as additives to promote the liquid phase sintering of Al2O3 ceramics. Cr2O3, MnO2, and TiO2 were added to the ceramic as additives. The Al2O3 ceramics were prepared from high-purity raw materials (≥99.9%). The ceramic powders were prepared by spray granulation method, then were formed by cold isostatic pressing. After that, the green ceramic bodies were sintered under air atmosphere using a high-temperature muffle furnace. Figure 1 shows the preparation process of the Cr/Mn/Ti-doped Al2O3 ceramics. The parameters of the sintering process are shown in Table 1.  The prepared ceramics were processed into samples with dimensions of Φ 26 mm × 5 mm, Φ 26 mm × 1 mm, and Φ 40 mm × 2 mm via grinding for testing the surface flashover voltage in a vacuum, secondary electron emission characteristics, and resistivity, respectively. The samples were numbered according to the amount of TiO2 added, as shown in Table 2, and photos of the samples are shown in Figure 2.   The prepared ceramics were processed into samples with dimensions of Φ 26 mm × 5 mm, Φ 26 mm × 1 mm, and Φ 40 mm × 2 mm via grinding for testing the surface flashover voltage in a vacuum, secondary electron emission characteristics, and resistivity, respectively. The samples were numbered according to the amount of TiO 2 added, as shown in Table 2, and photos of the samples are shown in Figure 2. A portion of samples were loaded into a high-temperature hydrogen furnace and heat treatment was performed at 1450 °C under wet hydrogen conditions to study the effect of the heat treatment process on the properties of the ceramics, considering the requirements of metallization and welding when applied in vacuum electronic devices. The parameters of the heat treatment process are shown in Table 3. Table 3. Parameters of heat treatment process.  A portion of samples were loaded into a high-temperature hydrogen furnace and heat treatment was performed at 1450 • C under wet hydrogen conditions to study the effect of the heat treatment process on the properties of the ceramics, considering the requirements of metallization and welding when applied in vacuum electronic devices. The parameters of the heat treatment process are shown in Table 3. The bulk density of the samples was measured by Archimedes' drainage method. High resistance meter (SM7120, HITACHI, Tokyo, Japan) was used to measure the surface resistivity and volume resistivity of the samples. Scanning electron microscope (SU3800, HITACHI, Tokyo, Japan) was applied to observe the microstructure of the samples. The microscopic composition of the samples was analyzed using X-ray diffractometer (D8 Advance, BRUKER, Karlsruhe, Germany) and X-ray photoelectron spectrometer (SM7120, HITACHI, Tokyo, Japan).

Test of Secondary Electron Emission Characteristics
There will be charges accumulated on the surface when a ceramic is bombarded by electrons. In order to eliminate the influence of surface charge accumulation, the three-gun method was applied to measure the SEY of the ceramic materials. A diagrammatic sketch of the test system's structure is shown in Figure 3. The system works as follows: the first charge neutralizer gun, which is a low-energy electron gun, is activated to remove the positive surface charges when they appear on the ceramic surface. The second charge neutralizer gun is activated to remove the negative surface charges when they appear on the ceramic surface. The surface potential of the ceramics is stabilized to ground potential by the cooperation of two neutralizer guns to eliminate the charge accumulated on the ceramic surface, in order to ensure the accuracy of secondary electron emission testing [25]. During the secondary electron emission test, the incident electron beam is perpendicular to the sample; the primary electron energy range is 0~3500 eV; the incident current is 0.1 µA with a pulse width of 5 µs; the temperature during the test is kept as room temperature; and the vacuum degree of the cavity is lower than 1 × 10 −5 Pa. is 0.1 μA with a pulse width of 5 μs; the temperature during the test is kept as room temperature; and the vacuum degree of the cavity is lower than 1 × 10 −5 Pa.

Test of Surface Flashover Threshold in Vacuum
A DC high-voltage vacuum test system was used to test the surface flashover threshold of Al2O3 ceramics in a vacuum, and the diagrammatic sketch of the testing device is shown in Figure 4a. The electrodes were disc-shaped flat shape made of stainless steel, with a diameter of 66 mm, and a distance between electrodes of the sample height, i.e., 5 mm. A photo of the flat electrode and test structure is shown in Figure 4b. The test started when the vacuum degree of the chamber reached 1 × 10 −4 Pa. During the test, a negative polarity DC voltage with the linear rise rate of 500 V/s was applied to the sample until the surface flashover occurred. Then, we gradually reduced the voltage until the surface flashover completely disappeared. If this voltage is repeatedly applied to the sample 3 times, and the sample can avoid flashover and remain stable, then the voltage was recorded as the surface flashover threshold in a vacuum. The voltage holdoff capacity of Al2O3 ceramics in a vacuum is represented by the surface flashover threshold, denoted by Uho. Five samples were tested per group, and the average value of Uho was calculated as the final result.

Test of Surface Flashover Threshold in Vacuum
A DC high-voltage vacuum test system was used to test the surface flashover threshold of Al 2 O 3 ceramics in a vacuum, and the diagrammatic sketch of the testing device is shown in Figure 4a. The electrodes were disc-shaped flat shape made of stainless steel, with a diameter of 66 mm, and a distance between electrodes of the sample height, i.e., 5 mm. A photo of the flat electrode and test structure is shown in Figure 4b. The test started when the vacuum degree of the chamber reached 1 × 10 −4 Pa. During the test, a negative polarity DC voltage with the linear rise rate of 500 V/s was applied to the sample until the surface flashover occurred. Then, we gradually reduced the voltage until the surface flashover completely disappeared. If this voltage is repeatedly applied to the sample 3 times, and the sample can avoid flashover and remain stable, then the voltage was recorded as the surface flashover threshold in a vacuum. The voltage holdoff capacity of Al 2 O 3 ceramics in a vacuum is represented by the surface flashover threshold, denoted by U ho . Five samples were tested per group, and the average value of U ho was calculated as the final result. is 0.1 μA with a pulse width of 5 μs; the temperature during the test is kept as room temperature; and the vacuum degree of the cavity is lower than 1 × 10 −5 Pa.

Test of Surface Flashover Threshold in Vacuum
A DC high-voltage vacuum test system was used to test the surface flashover threshold of Al2O3 ceramics in a vacuum, and the diagrammatic sketch of the testing device is shown in Figure 4a. The electrodes were disc-shaped flat shape made of stainless steel, with a diameter of 66 mm, and a distance between electrodes of the sample height, i.e., 5 mm. A photo of the flat electrode and test structure is shown in Figure 4b. The test started when the vacuum degree of the chamber reached 1 × 10 −4 Pa. During the test, a negative polarity DC voltage with the linear rise rate of 500 V/s was applied to the sample until the surface flashover occurred. Then, we gradually reduced the voltage until the surface flashover completely disappeared. If this voltage is repeatedly applied to the sample 3 times, and the sample can avoid flashover and remain stable, then the voltage was recorded as the surface flashover threshold in a vacuum. The voltage holdoff capacity of Al2O3 ceramics in a vacuum is represented by the surface flashover threshold, denoted by Uho. Five samples were tested per group, and the average value of Uho was calculated as the final result.

Bulk Density and Micromorphology
The influence of TiO 2 on the bulk density and micromorphology of the Al 2 O 3 ceramics was studied. It can be seen from Table 4 that we can increase the bulk density of the ceramics by adding a little amount of TiO 2 to Cr/Mn-doped Al 2 O 3 ceramics. The bulk density of the Al 2 O 3 ceramics with 1 wt.% TiO 2 is highest, reaching 3.810 g/cm 3 . After that, the bulk density of the Al 2 O 3 ceramics decreases with the increase in the content of TiO 2 . The addition of TiO 2 to the Al 2 O 3 ceramics has an effect on their sintering properties. TiO 2 can form the finite solid solution with Al 2 O 3 when it is added to ceramics. In the range of 1300 • C-1700 • C, the solid solubility of TiO 2 in the α-Al 2 O 3 ceramics is 0.27 wt.% [26]. The TiO 2 residual is located at the grain boundary, in which case it can react with MnO 2 and form a low-temperature eutectic mixture. The low eutectic presents a liquid phase during sintering, which reduces the sintering temperature of Al 2 O 3 ceramics [27]. Secondly, in a TiO 2 -Al 2 O 3 solid solution, Ti 4+ replaces Al 3+ , which produces lattice deformation and a large number of cation vacancies. The energy potential barrier required for lattice particle diffusion is reduced and the lattice is activated, which is conducive to the diffusion and transfer of substances and promotes the sintering of ceramics. 3.716 Figure 5 shows the fractured surface of the Cr/Mn/Ti-doped Al 2 O 3 ceramics. We can find from Figure 5 that there is little difference in the microstructures between CMT05 and CMT10, and there are mainly elongated grains, with an average particle size of about 8-10 µm. As the amount of TiO 2 continued to increase, the grain size of the CMT20 and CMT30 samples increased, and large grains appeared locally, as shown in Figure 5c,d. The addition of excessive TiO 2 to ceramics makes a local abnormal grain growth during the ceramic sintering process. The pores in the ceramic cannot be discharged promptly, which is likely to cause pores wrapped up (as circled in the Figure 5) in the interior and boundary of the grains, resulting in the reduction in the bulk density of the ceramics, and the uniformity of the microstructure will also become worse.
was studied. It can be seen from Table 4 that we can increase the bulk density of the ce ramics by adding a little amount of TiO2 to Cr/Mn-doped Al2O3 ceramics. The bulk density of the Al2O3 ceramics with 1 wt.% TiO2 is highest, reaching 3.810 g/cm 3 . After that, the bulk density of the Al2O3 ceramics decreases with the increase in the content of TiO2. The addi tion of TiO2 to the Al2O3 ceramics has an effect on their sintering properties. TiO2 can form the finite solid solution with Al2O3 when it is added to ceramics. In the range of 1300 °C 1700 °C, the solid solubility of TiO2 in the α-Al2O3 ceramics is 0.27 wt.% [26]. The TiO residual is located at the grain boundary, in which case it can react with MnO2 and form a low-temperature eutectic mixture. The low eutectic presents a liquid phase during sin tering, which reduces the sintering temperature of Al2O3 ceramics [27]. Secondly, in TiO2-Al2O3 solid solution, Ti 4+ replaces Al 3+ , which produces lattice deformation and large number of cation vacancies. The energy potential barrier required for lattice particl diffusion is reduced and the lattice is activated, which is conducive to the diffusion and transfer of substances and promotes the sintering of ceramics. 3.716 Figure 5 shows the fractured surface of the Cr/Mn/Ti-doped Al2O3 ceramics. We can find from Figure 5 that there is little difference in the microstructures between CMT05 and CMT10, and there are mainly elongated grains, with an average particle size of about 8 10 μm. As the amount of TiO2 continued to increase, the grain size of the CMT20 and CMT30 samples increased, and large grains appeared locally, as shown in Figure 5c,d. Th addition of excessive TiO2 to ceramics makes a local abnormal grain growth during th ceramic sintering process. The pores in the ceramic cannot be discharged promptly, which is likely to cause pores wrapped up (as circled in the Figure 5) in the interior and boundary of the grains, resulting in the reduction in the bulk density of the ceramics, and the uni formity of the microstructure will also become worse.

Microstructure and Phase Composition
XPS and XRD were applied to analyze the microstructure and phase composition of the Al 2 O 3 ceramics. Since Ti is a variable valence element, its elemental valence and phase composition in the ceramics may change during the sintering and heat treatment. Figure 6 shows the XPS pattern of CMT30. The XPS pattern of Ti 2p in Figure 6b shows that the binding energy of the Ti 2p 1/2 energy level of Ti 4+ in the untreated samples is 464.4 eV, and the binding energy of the Ti 2p 3/2 energy level is 459.2 eV [28,29]. After the heat treatment, the binding energy of the Ti 2penergy level of the samples moves toward the direction of lower energy, indicating that hydrogen has a certain degree of reduction on TiO 2 during the heat treatment.

Microstructure and Phase Composition
XPS and XRD were applied to analyze the microstructure and phase composition of the Al2O3 ceramics. Since Ti is a variable valence element, its elemental valence and phase composition in the ceramics may change during the sintering and heat treatment. Figure  6 shows the XPS pattern of CMT30. The XPS pattern of Ti 2p in Figure 6b shows that the binding energy of the Ti 2p1/2 energy level of Ti 4+ in the untreated samples is 464.4 eV, and the binding energy of the Ti 2p3/2 energy level is 459.2 eV [28,29]. After the heat treatment, the binding energy of the Ti 2penergy level of the samples moves toward the direction of lower energy, indicating that hydrogen has a certain degree of reduction on TiO2 during the heat treatment.  Figure 7b shows that the characteristic peak of the Al2O3 at 2 Theta = 35.15 shifts toward a small angle direction. This is due to the solid solution of TiO2 and Al2O3. Ti 4+ replaces Al 3+ , which changes the cell parameters of Al2O3. Combined with Figure 8, we can find that in addition to the main crystal phase of Al2O3, the CMT30 sample also contains a small amount of other crystal phases such as CaAl2Si2O8, Al2SiO5, and MnTiO3, which are generated by the reaction of Al2O, CaO, SiO2, MnO2, and TiO2 in the ceramics. Figure 8 is the particle comparison of the XRD data of the A-0 and CMT30 samples after normalization, displaying that the CMT30 sample before and after the heat treatment contained both TiO2 and TiO0.5. And the content of the TiO2 phase decreased and the content of the TiO0.5 phase increased in CMT30 after the heat treatment. In addition, after the heat treatment, the Ti8O15 phase appeared in CMT30. Both the XPS and XRD results show that TiO2 in Al2O3 ceramics was reduced during the heat treatment process, resulting in an increase in the content of the anoxic phase in the Cr/Mn/Ti-doped Al2O3 ceramics.   Figure 7b shows that the characteristic peak of the Al 2 O 3 at 2 Theta = 35.15 shifts toward a small angle direction. This is due to the solid solution of TiO 2 and Al 2 O 3 . Ti 4+ replaces Al 3+ , which changes the cell parameters of Al 2 O 3 . Combined with Figure 8, we can find that in addition to the main crystal phase of Al 2 O 3 , the CMT30 sample also contains a small amount of other crystal phases such as CaAl 2 Si 2 O 8 , Al 2 SiO 5, and MnTiO 3 , which are generated by the reaction of Al 2 O, CaO, SiO 2 , MnO 2, and TiO 2 in the ceramics. Figure 8 is the particle comparison of the XRD data of the A-0 and CMT30 samples after normalization, displaying that the CMT30 sample before and after the heat treatment contained both TiO 2 and TiO 0.5 . And the content of the TiO 2 phase decreased and the content of the TiO 0.5 phase increased in CMT30 after the heat treatment. In addition, after the heat treatment, the Ti 8 O 15 phase appeared in CMT30. Both the XPS and XRD results show that TiO 2 in Al 2 O 3 ceramics was reduced during the heat treatment process, resulting in an increase in the content of the anoxic phase in the Cr/Mn/Ti-doped Al 2 O 3 ceramics.

Secondary Electron Emission Characteristics
According to the SEEA theory [16][17][18], secondary electron emission of ceramics is a factor which influences the surface flashover voltage directly. We measured the SEY of the Al2O3 ceramics with different contents of TiO2. The secondary electron emission curves and maximum value of the SEY are shown in Figure 9 and Table 5, respectively.

Secondary Electron Emission Characteristics
According to the SEEA theory [16][17][18], secondary electron emission of ceramics is a factor which influences the surface flashover voltage directly. We measured the SEY of the Al2O3 ceramics with different contents of TiO2. The secondary electron emission curves and maximum value of the SEY are shown in Figure 9 and Table 5, respectively.

Secondary Electron Emission Characteristics
According to the SEEA theory [16][17][18], secondary electron emission of ceramics is a factor which influences the surface flashover voltage directly. We measured the SEY of the Al 2 O 3 ceramics with different contents of TiO 2 . The secondary electron emission curves and maximum value of the SEY are shown in Figure 9 and Table 5, respectively.

Secondary Electron Emission Characteristics
According to the SEEA theory [16][17][18], secondary electron emission of ceramics is a factor which influences the surface flashover voltage directly. We measured the SEY of the Al2O3 ceramics with different contents of TiO2. The secondary electron emission curves and maximum value of the SEY are shown in Figure 9 and Table 5, respectively.   The results show that the SEY of the Cr/Mn-or Cr/Mn/Ti-doped Al 2 O 3 ceramics is significantly lower than that of the basic Al 2 O 3 ceramics, because Cr 2 O 3 with a low secondary electron emission coefficient as an additive can effectively reduce the SEY of the Al 2 O 3 ceramics. While the SEY of the ceramics increases with the increase in the content of TiO 2 . Particularly, the SEY of the ceramics increases significantly when the content of TiO 2 is more than 2 wt.%. This is mainly attributed to the effect of TiO 2 on the microstructure of the Al 2 O 3 ceramics. Adding TiO 2 to the Al 2 O 3 ceramics can promote the process of sintering. The grain size of the ceramics increases significantly when the content of TiO 2 is more than 2 wt.%.
According to the SEEA theory, the primary electrons are accelerated by the electric field and gain energy, then they collide with the surface of the ceramic and produce the secondary electrons. The secondary electrons collide with the surface again under the effect of the electric field, after they escape from the surface. The whole process is repeated and the number of electrons multiplies rapidly, eventually leading to the occurrence of the electron avalanche. Electron avalanches cause the release of adsorbed gas on the ceramic surface, which is ionized by high-energy electrons, producing a large amount of plasma. Eventually, surface flashover occurs. Usually, the main factors affecting the collisional ionization process are the energy of the electron before the collision and the ionization energy of the collided lattice molecule or atom. The collisional ionization coefficient α is commonly used to describe the collisional ionization. The collisional ionization coefficient of an electron is the number of collisional ionizations produced by an electron traveling a unit distance under the action of the electric field. And the collisional ionization coefficient can generally be expressed as follows: where V i is the ionization energy of the collided lattice molecule or atom, λ is the mean free path of the electron, and E is the electric field strength. A, B are constants related to materials. It can be found from the expression of the collisional ionization coefficient that the mean free path of the electron influences the coefficient directly. A longer mean free path of the electron means that the electron is subjected to the electric field force between the two collisions for a longer time under the effect of the electric field, making ionization easy to occur. The grain size of the ceramics increases significantly when the content of TiO 2 is more than 2 wt.%. And the mean free path of the secondary electrons increases in the process of escape. The ionization coefficient of recollision ionization also becomes larger, which makes ionization easier to occur, resulting in the increase in the SEY.

Surface and Volume Resistivity
The surface resistivity of the ceramics is another important factor that affects the voltage holdoff capacity in a vacuum. The high surface resistivity of the ceramics is not conducive to charge leakage, resulting in an increase in the local field strength. And the surface flashover is aggravated. Thus, the surface resistivity of the ceramics must be reduced to a certain extent in order to improve the voltage holdoff capacity. It can be seen from the resistivity of the Al 2 O 3 ceramics in Table 6 that the resistivity of the Cr/Mn-doped Al 2 O 3 ceramics did not change a lot before and after the heat treatment, while the resistivity of the Al 2 O 3 ceramics with TiO 2 changed significantly. When TiO 2 was added to the Cr/Mn-doped Al 2 O 3 ceramics, the change in the resistivity of the ceramics was not significant before the heat treatment, and the surface resistance of the ceramics was one order of magnitude lower than that of the basic Al 2 O 3 ceramics. With the increase in the content of TiO 2 , the surface resistivity of the ceramics did not change much. The volume resistivity of the CMT20 and CMT30 was reduced by two orders of magnitude compared with the base Al 2 O 3 ceramics A-0. After the heat treatment, the resistivity of the Cr/Mn/Ti-doped Al 2 O 3 ceramics decreased significantly. The surface resistivity and volume resistivity of the CMT10 decreased by 4 and 5 orders of magnitude compared with the base Al 2 O 3 ceramics, reaching 7.6 × 10 11 Ω and 3.47 × 10 11 Ω·cm, respectively. The resistivity exceeded the measuring range of the high resistance meter (<10 8 Ω) when the content of TiO 2 was more than 3 wt.%. The color of the samples changed significantly after the heat treatment, as shown in Figure 10. ceramics did not change a lot before and after the heat treatment, while the resistivity of the Al2O3 ceramics with TiO2 changed significantly. When TiO2 was added to the Cr/Mn-doped Al2O3 ceramics, the change in the resistivity of the ceramics was not significant before the heat treatment, and the surface resistance of the ceramics was one order of magnitude lower than that of the basic Al2O3 ceramics. With the increase in the content of TiO2, the surface resistivity of the ceramics did not change much. The volume resistivity of the CMT20 and CMT30 was reduced by two orders of magnitude compared with the base Al2O3 ceramics A-0. After the heat treatment, the resistivity of the Cr/Mn/Ti-doped Al2O3 ceramics decreased significantly. The surface resistivity and volume resistivity of the CMT10 decreased by 4 and 5 orders of magnitude compared with the base Al2O3 ceramics, reaching 7.6 × 10 11 Ω and 3.47 × 10 11 Ω·cm, respectively. The resistivity exceeded the measuring range of the high resistance meter (˂10 8 Ω) when the content of TiO2 was more than 3 wt.%. The color of the samples changed significantly after the heat treatment, as shown in Figure 10. According to the XPS results in Figure 6, the binding energy of the Ti 2p level in the sample moved to the direction of low energy after the heat treatment. TiO2 was reduced by hydrogen to a certain extent, and the O:Ti ratio declined. The XRD analysis results show, meanwhile, that the content of the anoxic phase in the ceramics increased after the heat treatment. According to the relevant literature [30,31], the formation of anoxic phase TinO2n−1 will significantly reduce the resistivity of the ceramics. And the higher the content of the anoxic phase, the darker the ceramic color and the smaller the resistivity are. This is also consistent with the pattern of the samples' color in the experiment. According to the XPS results in Figure 6, the binding energy of the Ti 2p level in the sample moved to the direction of low energy after the heat treatment. TiO 2 was reduced by hydrogen to a certain extent, and the O:Ti ratio declined. The XRD analysis results show, meanwhile, that the content of the anoxic phase in the ceramics increased after the heat treatment. According to the relevant literature [30,31], the formation of anoxic phase Ti n O 2n−1 will significantly reduce the resistivity of the ceramics. And the higher the content of the anoxic phase, the darker the ceramic color and the smaller the resistivity are. This is also consistent with the pattern of the samples' color in the experiment.

Voltage Holdoff Capacity in Vacuum
We measured the surface flashover threshold of the Al 2 O 3 ceramics, and the size of the samples was Φ 26 mm × 5 mm. The surface flashover threshold of the samples before and after the heat treatment were tested by experiments, and the results are shown in Table 7. As shown in Table 7, the surface flashover threshold slightly increased when a small amount of TiO 2 was added to the ceramics. However, the surface flashover threshold of the ceramics changed significantly after the heat treatment. The effect of the heat treatment process on the Al 2 O 3 ceramics with the TiO 2 additive was different from the basic Al 2 O 3 ceramics A-0 and the Cr/Mn-doped Al 2 O 3 ceramics CM11. The U ho of the Cr/Mn/Tidoped Al 2 O 3 ceramics with a small amount of TiO 2 was significantly improved after the heat treatment. The U ho of CMT11 reached 33 kV, which was 32% higher than that of the basic Al 2 O 3 ceramics. After the heat treatment with wet hydrogen, the U ho of the basic Al 2 O 3 ceramics slightly decreased, while the U ho of the Cr/Mn/Ti-doped Al 2 O 3 ceramics increased significantly. The U ho of the Cr/Mn/Ti-doped Al 2 O 3 ceramics in a vacuum is higher than that of the base Al 2 O 3 ceramics, which is mainly attributed to the following three aspects. Firstly, the SEY of the Cr/Mn/Ti-doped Al 2 O 3 ceramics is obviously lower than that of the basic Al 2 O 3 ceramics, which makes the secondary electron avalanche hard to happen on the surface of the ceramics when the voltage is applied; thus, the surface flashover threshold is larger. Secondly, the surface resistivity of the sample with TiO 2 did not change a lot compared with the basic Al 2 O 3 ceramics before the heat treatment, but the surface resistivity and volume resistivity of the Cr/Mn/Ti-doped Al 2 O 3 ceramics were significantly lower than those of the basic Al 2 O 3 ceramics after the heat treatment. The low surface resistivity of the ceramics can make the surface charge be conducted timely and effectively, so the probability of the flashover occurrence along the surface can be reduced. Finally, the bulk density and microstructure uniformity of the ceramics can be improved with a small amount of TiO 2 addition, which can increase the stability of the ceramics in the process of the applied voltage.
According to the data of the surface flashover threshold in a vacuum from Table 7, when the content of TiO 2 was more than 1 wt.%, the U ho of CMT20 and CMT30 were very low after the heat treatment, which were 18 kV and 15 kV, respectively. This is because the resistivity of the ceramics after the heat treatment with wet hydrogen was too low. When the voltage was applied to the ceramics, there was a large leakage current, resulting in the action of power failure protection, and the power of the high-voltage equipment was cut. So, the test value of the surface flashover threshold was very low.

Conclusions
In this paper, Al 2 O 3 ceramics with a high surface flashover threshold in a vacuum were prepared by bulk doping Cr 2 O 3 , MnO 2, and TiO 2 , and the effect of the content of TiO 2 on the properties of the Al 2 O 3 ceramics was studied. The main conclusions are as follows: (1) The addition of TiO 2 to the Al 2 O 3 ceramics can promote the sintering of the ceramics, and ceramics with a high volume density and uniform microstructure can be prepared by an appropriate additive amount of TiO 2 .
(2) TiO 2 in the ceramics was reduced during the heat treatment with wet hydrogen condition. The resistivity of the Cr/Mn/Ti-doped Al 2 O 3 ceramics decreased significantly with the appearance of the anoxic phase. (3) The microstructure of the Cr/Mn/Ti-doped Al 2 O 3 ceramics with 1 wt.% TiO 2 is uniform and compact, and the surface resistivity of the sample is reduced to 7.6 × 10 11 Ω. The surface flashover threshold in a vacuum reaches 33 kV, which is 32% higher than that of the basic Al 2 O 3 ceramics. (4) The resistivity of the ceramics will be too low when the content of TiO 2 exceeds 2 wt.%, resulting in a large leakage flow and low surface flashover threshold during the experiment.