Effect of TiO 2 addition on the microstructures,mechanical and dielectric properties of porous Si 3 N 4 -based ceramics

Abstract

Porous Si₃N₄-based ceramics with different TiO₂ contents were prepared by gas pressure sintering method. The effects of TiO₂ addition ranging from 0 to 25 wt-% on the phase compositions, microstructures, mechanical performance and dielectric properties were investigated. The addition of TiO₂ significantly promoted the density which increased from 1.64 to about 2.3 g cm⁻³. The mechanical properties of porous Si₃N₄-based ceramics with TiO₂ addition decreased first and then increased with the increase of TiO₂ content, and the flexural strength and elastic modulus are more than 167.4 MPa and 72.8 GPa, respectively, which were higher than that of the Si₃N₄ ceramic without TiO₂ addition. With the increase of TiO₂ content, both the dielectric constant and dielectric loss increased, and the dielectric constant enhanced obviously. These results suggested that the TiO₂ was beneficial for the improvement of mechanical properties and dielectric constant of porous Si₃N₄-based ceramics.

Keywords

microstructures dielectric properties

Introduction

Silicon nitride ceramics have been recognised as one of the most important materials for high temperature applications. High temperature mechanical strength [1 -3], good resistance to corrosion, wears and chemicals, low dielectric constant and loss, and super hardness with low density of silicon nitride [4 10] indicate a very optimistic future for high temperature radome application [11]. However, the dielectric constant of silicon nitride cannot meet the requirement of radar/infrared dual mode composite radome with higher dielectric constant.

It is often that radome materials are prepared to meet complicated conditions. In addition, in order to satisfy the needs of high dielectric constant of some radome materials, we will choose materials with high dielectric constant as additive to improve dielectric properties of porous silicon nitride ceramics. Although researchers have proved that the introduction of a second phase into the Si₃N₄ matrix like CaHPO₄ [12] can effectively increase the dielectric constant of porous silicon nitride ceramics, which cannot meet the requirement of the radome materials. TiO₂ mainly exists in three forms: rutile, anatase and brookite. The rutile displays better dielectric properties than the anatase and brookite. TiO₂ has potential for use as a microwave dielectric ceramic due to its excellent dielectric properties, i.e. its high dielectric constant and low dielectric loss [13]. Zhang et al. [14] reported the dielectric behaviour of nano-TiO₂ bulks, and found that the rutile bulk exhibits a high dielectric constant of ∼10⁴ at 10⁻² kHz. In addition, researchers have found that CuO ceramic [15], NiO ceramic co-doped with Li and Ti [16] and CaCu₃Ti₄O₁₂ single crystal [17] all exhibit a high dielectric constant of the order of magnitude of 10⁴. The titanium oxide possesses high dielectric constant, so we can improve the dielectric constant by adding titanium oxide to porous silicon nitride ceramics.

In this paper, porous Si₃N₄-based ceramics with TiO₂ addition have been prepared via the fabrication of mixed powders, moulding and gas pressure sintering process employing Si₃N₄ and TiO₂ as starting materials, and a small amount of Y₂O₃ was added as the sintering additive. The influences of TiO₂ addition on the phase compositions, microstructures, mechanical performance and dielectric properties have been studied.

Experimental procedure

Materials processing

Commercial P grade Si₃N₄ powder (D₅₀= 1.30 μm, α phase content: 95%) and TiO₂ (D₅₀= 2.30 μm, 96% R-TiO₂) were used as the starting powders. Y₂O₃ (D₅₀= 0.70 μm, 99.99% purity) was added as the sintering additive. The amount of adding Y₂O₃ particles was 6 wt-%. And the mass fraction of TiO₂ varied from 0 to 25%. Figure 1 showed the SEM image of Si₃N₄, TiO₂ and powder mixture. From Figure 1, it can be seen that they were homogeneous dispersion basically in spite of parts of TiO₂ particles were agglomerated. There was mainly R-TiO₂ (rutile) phase in TiO₂ powder, as can be seen in Figure 2.

The above raw powders were mixed and wet milled in anhydrous ethanol for 24 h with speed of 220 rev min⁻¹ in high-energy ball miller. The slurry was dried, and then passed through 50 mesh sieve. The mixed powders were formed into compacts of about Ø40 mm × 120 mm in size under an isostatic pressing of 80 MPa. And the green compacts were dried at 110°C for 3 h in order to reduce the humidity in the ventilation drying. And they were marked and weighed. Finally, the green bodies were embedded in the Si₃N₄ powders in a graphite crucible and sintered in a graphite furnace at 1700°C for 2 h in 0.3 MPa N₂ atmosphere.

Materials characterisation

The density and apparent porosity of the specimens were determined by Archimedes’ principle, which means that the samples were placed in distilled water under vacuum environment. Flexural strength and elastic modulus were measured on test specimens of 3 mm × 4 mm × 36 mm using a three-point bend method with a span of 30 mm and a cross-head speed of 0.5 mm min⁻¹. X-ray diffraction (XRD, D/max 2500, Tokyo, Japan) was carried out using Cu Kα (λ = 0.15405 nm) radiation to determine the phase composition. The microstructure of fracture surface of samples was observed by scanning electron microscopy (SEM, FEI Quanta 200, Brno, Czech) after being sputtered with gold film. Dielectric properties of the specimens with a size of 6.96 mm × 3.48 mm × 1.00 mm was recorded in the frequency of 35 GHz at room temperature by short circuited waveguide technique with the help of an Agilent Network Analyzer (model N5230A, Agilent Technologies Inc, Palo Alto, CA, USA).

Results and discussion

Density and porosity of porous Si₃N₄-based ceramics with TiO₂ addition

Variations in the density and apparent porosity of the samples are shown in Figure 3. The density and apparent porosity of porous Si₃N₄ ceramic without TiO₂ addition were 1.64 g cm⁻³ and 37.9%, respectively, detected by Archimedes’ method. The apparent porosity of porous Si₃N₄-based ceramics increased initially with the addition of TiO₂ up to 15 wt-%, while beyond that content the apparent porosity began to decline. The porosity was a combining result from density; increasing density decreased porosity. Compared with the density of porous Si₃N₄ ceramic without TiO₂ addition, it significantly increased for the porous Si₃N₄-based ceramics with TiO₂ addition. The main reason was that the formation of a small amount of intergranular glass phase and TiN phase. Rare-earth oxide additives, Y₂O₃ in this system, might form complex nitride or oxides intergranular phase by reaction with Si₃N₄ or SiO₂ [18]. The pore of porous Si₃N₄-based ceramics was stuffed, which improved the density. The density decreased first and then increased with the increase of TiO₂ contents ranging from 5 to 25 wt-%, on the contrary, the apparent porosity increased first and then decreased. The density of the material mainly depends on the main crystal phase of silicon nitride, as the amount of adding TiO₂ content increases from 5 to 15 wt-%. However, when TiO₂ particles were added between 15 and 25 wt-%, the existence of eutectic liquids between crystalline phase might accelerate the migration and rearrangement of the Si₃N₄ main grains and enhance the densification by filling the clearance between solid grains.

Phase compositions and microstructures of porous Si₃N₄-based ceramics with TiO₂ addition

Figure 4 shows the XRD patterns of the obtained porous Si₃N₄-based ceramics with different TiO₂ contents. After sintering, α-Si₃N₄ in all samples was transformed into β-Si₃N₄ completely regardless of the different formulation, as can be seen in Figure 4. Except a small amount of YSiO₂N phase formed by the reaction among Si₃N₄, Y₂O₃ and SiO₂ in the pure Si₃N₄ ceramic, where a small amount SiO₂ was present on the surface of Si₃N₄ particles because of oxidation of silicon nitride, no other intercrystalline phases were detected. There have been small amount of Si₂N₂O and Y₂Si₃O₃N₄ phase in other samples. Obviously, TiN phase appeared in all specimens with TiO₂ powders after sintering. The generation of these phases was obtained by the interactions of materials. It has been reported that the reaction between TiO₂ and Si₃N₄ can occur at temperature as low as 1000°C according to reaction (1), where SiO₂ is present in the amorphous state [19] and TiO₂ on the surface of the TiN nano-particles can promote the formation of liquid phase at early sintering stage via the reaction (1) and TiN grains grow through solution-reprecipitation process [20,21]. In addition, Si₂N₂O, YSiO₂N and Y₂Si₃O₃N₄ phases were obtained via reactions (2) and (3). (1) (2) (3)

Figure 5 displays the fracture morphology of porous Si₃N₄-based ceramics with different contents of TiO₂ ranging from 0 to 25 wt-%. As we all know, the sintering of Si₃N₄-based ceramics was a solution-reprecipitation process [22,23]. During the phase transformation, α-Si₃N₄ particles dissolve first into the liquid phase formed by sintering additives, and then β-Si₃N₄ grains reprecipiated out and tended to grow in the preferred orientation, leading to the formation of elongated, rod-shaped grains [24]. From Figure 5(a), it can be seen that there were many well-developed prismatic β-Si₃N₄ grains stacking together to form spatial network, as well as less short β-Si₃N₄ grains in the regular morphology growing around or in the pores. When the amount of adding TiO₂ particles increased from 5 to 15 wt-%, there were no apparent rob-like β-Si₃N₄ grains in Figure 5(a,b). When the content of TiO₂ reached 20 and 25 wt-%, there were some apparent rob-like β-Si₃N₄ grains, specifically, for the porous Si₃N₄ based with 25 wt-% TiO₂ particles. During the microstructure evolution stage of the sintering process, β-Si₃N₄ nucleus was first developed then, the growth of β-Si₃N₄ grains would be inhibited along its longitudinal direction when the glass phase occurred during the grain growth stage.

Mechanical properties of porous Si₃N₄-based ceramics with TiO₂ addition

It is well known that mechanical properties of the material are closely related to porosity, phase compositions and microstructures. Figure 6 shows the mechanical properties of the ceramics with the different TiO₂ contents. With the increase of TiO₂ content ranging from 5 to 25 wt-%, the flexural strength and elastic modulus of specimens decreased first and then increased, which had the same trends of density. For the porous Si₃N₄ ceramic without TiO₂ content, the flexural strength and elastic modulus were 150 MPa and 45 GPa, separately, which was lower than that of all porous Si₃N₄-based ceramics with TiO₂ addition. Furthermore, when the 15 wt-% TiO₂ particles were added, the flexural strength and elastic modulus can still reach 167.4 MPa and 72.8 GPa, respectively, while porosity of the specimen reached the maximum value of 33.2%, and the value of mechanical properties can meet the required value of wave transparent materials. In addition, the mechanical properties of the specimens are greatly related to their porosity. With the same phase components, mechanical properties of the material showed an opposite tendency to its open porosity. Ryshkewitch proposed an empirical equation of porous ceramics as follows [25]: (4) where σ₀ is flexural strength of the material without pores; β is structural parameter with value of 4–7; p is the porosity of testing samples. The equation revealed that the flexural strength of porous materials decreases exponentially with the increasing of porosity. One of reasons was that pores were regarded as a weak boundary phase and more pores caused less solid cross-section and larger actual stress, and the other was that pores induced the local stress concentration and weaken the bending strength. Besides, the influence of porosity on the mechanical properties of the porous Si₃N₄-based ceramics with TiO₂ addition can also be explained by the theories that less energy was required for crack propagation with the increase of pores when samples were bent. Apart from the porosity, phase compositions and microstructure have an effect on the mechanical properties of silicon nitride-based ceramic materials. From XRD patterns, it can be seen that the diffraction peaks of Si₃N₄ and Si₂N₂O when the content of TiO₂ was 15 wt-% (Figure 4(d)), being lower than those of other samples with TiO₂ addition. TiN phase can improve the flexural strength of the Si₃N₄-based ceramic materials, but glassy phase weaken the strength. Combined with microstructures, comprehensive factors affected the mechanical properties of the samples.

Dielectric properties of porous Si₃N₄-based ceramics with TiO₂ addition

Figure 7 shows the effects of the different contents of TiO₂ particles on the dielectric constant and dielectric loss tangent of porous Si₃N₄-based ceramics with TiO₂ addition in the frequency of 35 GHz. Obviously, the increasing amount of TiO₂ significantly enhanced the dielectric constant and dielectric loss tangent, which manifested as an expected enhancement in dielectric constant as TiO₂ content increased. The dielectric properties of porous Si₃N₄-based ceramics with TiO₂ addition strongly depend on their phase compositions and porosity. The effects of these elements on dielectric constant (ε) could be characterised by mixture law [26] using the following expression: (5) where ν_i is the volume of i phase, ε_i is the dielectric constant of i phase.

The dielectric constant and dielectric loss are two important microwave interaction properties of a dielectric material [27,28]. The dielectric loss predicts the ability of one material absorbing electromagnetic wave, and this material can be used as an excellent electromagnetic wave absorber if it possesses high dielectric loss. The dielectric loss of the Si₃N₄ ceramic is so small that the dielectric loss can only be measured exactly by using resonant cavity method.

Research results showed that the dielectric constant and dielectric loss of porous Si₃N₄ ceramic without TiO₂ addition were about 3.0 and 0.0032 in this experiment, respectively. The amount of TiN phase, ε and tanδ of porous Si₃N₄ matrix ceramics with TiO₂ addition all increased evidently with the increase of TiO₂ content. In addition, the values of ε and tanδ of pores can be regarded as 1 and 0, separately. From Figure 7, it can be seen that the dielectric constant of the silicon nitride-based ceramics with 5 wt-% TiO₂ was significantly higher than that of the silicon nitride ceramics without TiO₂. What is more, the dielectric constant of the specimen with 25 wt-% was 7.66, which was increased by 35.3% compared to the sample with 5 wt-% TiO₂, and we know the dielectric constant of Si₂N₂O is 6–7. Therefore, compared to Si₃N₄, we can know that the TiN phase has higher dielectric constant (ε). What is more, it was indicated that the dielectric loss increased with the increase of TiO₂ content, the reasons may be that the TiN phase increased and the crystal defects were formed by crystalline transformation from α-Si₃N₄ to β-Si₃N₄. Therefore, these results suggested that the increasing content of TiO₂ would lead to the increase in both the ε and tanδ value of porous Si₃N₄-based ceramics with TiO₂ addition.

Conclusions

Porous Si₃N₄-based ceramics with different amounts of TiO₂ particles were successfully fabricated by gas pressure sintering method. With the increase of TiO₂ content, the apparent porosity of the porous Si₃N₄-based ceramics with TiO₂ addition increased first and then decreased between 21.7 and 33.2%. The decrease of porosity and increase of density of porous Si₃N₄-based ceramics were able to promote the mechanical properties when TiO₂ particles were added. The flexural strength and elastic modulus first decreased and then increased with the increase of TiO₂ contents ranging from 5 to 25 wt-%. Although the flexural strength and elastic modulus appeared the minimum value for the Si₃N₄-based ceramics with TiO₂ particles, they were higher than that of the porous Si₃N₄ ceramics without TiO₂ addition. Furthermore, the dielectric constant and dielectric loss increased with the increase of TiO₂ content. The comprehensive performance of the obtained porous Si₃N₄-based ceramic with TiO₂ addition is fit for the some wave transparent materials.

Footnotes

Acknowledgement

Authors thank Shandong Industrial Ceramics Research & Design Institute Co., Ltd for the experimental conditions offered.

Disclosure statement

No potential conflict of interest was reported by the authors.

References

Ault

, Yeckley

Silicon nitride. Am Ceram Soc Bull. 1995;74(6):153–155.

Inagaki

, Kondo

, Ohji

High performance porous silicon nitrides. J Eur Ceram Soc. 2002;22:2489–2494. doi: 10.1016/S0955-2219(02)00107-3

Ohji

, Shigegaki

, Miyajima

, Fracture resistance behavior of multilayered silicon nitride. J Am Ceram Soc. 1997;80(2):991–994.

She

, Yang

, Daniel

, Thermal shock behavior of isotropic and anisotropic porous silicon nitride. J Am Ceram Soc. 2003;86(4):738–740. doi: 10.1111/j.1151-2916.2003.tb03367.x

Lee

, Moretti

, Readey

, Thermal shock resistance of silicon nitrides using an indentation-quench test. J Am Ceram Soc. 2002;85(1):279–281. doi: 10.1111/j.1151-2916.2002.tb00083.x

Toshimori

Shock synthesis of cubic silicon nitride. J Am Ceram Soc. 2002;85(1):113–116.

Dennis

, Elizabeth

, Quynhgiao

, Paralinear oxidation of silicon nitride in a water-vapor/oxygen environment. J Am Ceram Soc. 2003;86(8):1256–1261. doi: 10.1111/j.1151-2916.2003.tb03461.x

Backhaus-ricoult

, Guerin

, Huntz

, High-temperature oxidation behavior of high-purity α, β, and mixed silicon nitride ceramics. J Am Ceram Soc. 2002;85(2):385–392. doi: 10.1111/j.1151-2916.2002.tb00101.x

Zhang

, Cheng

, Srinivasarao

, Erosion response of highly anisotropic silicon nitride. J Am Ceram Soc. 2005;88(1):114–120. doi: 10.1111/j.1551-2916.2004.00014.x

10.

Andreas

, Markus

, Marcus

, Elastic modulus and hardness of cubic silicon nitride. J Am Ceram Soc. 2002;85(1):86–90.

11.

Barta

, Manela

, Fischer

Si₃N₄ and Si₂N₂O for high performance radomes. Mater Sci Eng. 1985;71:265–272. doi: 10.1016/0025-5416(85)90236-8

12.

Duan

, Jia

, Deng

, Mechanical and dielectric properties of gelcasted Si₃N₄ porous ceramic using CaHPO₄ as an additive. Ceram Int. 2012;38:4363–4367. doi: 10.1016/j.ceramint.2012.01.006

13.

Templeton

, Wang

, Penn

, Microwave dielectric loss of titanium oxide. J Am Ceram Soc. 2000;83:95–100. doi: 10.1111/j.1151-2916.2000.tb01154.x

14.

Zhang

, Zhang

, Wang

, Dielectric behaviour of nano-TiO₂ bulks. Phys Stat Sol(a). 1996;157:483–491. doi: 10.1002/pssa.2211570232

15.

Sarkar

, Jana

, Chaudhuri

, Copper (II) oxide as a giant dielectric material. Appl Phys Lett. 2006;89:212905. doi: 10.1063/1.2393001

16.

, Nan

, Lin

, Giant dielectric permittivity observed in Li and Ti doped NiO. Phys Rev Lett. 2002;89:217601. doi: 10.1103/PhysRevLett.89.217601

17.

Homes

, Vogt

, Shapiro

, Optical response of high-dielectric-constant Perovskite-related oxide. Science. 2001;293:673–676. doi: 10.1126/science.1061655

18.

Tanaka

, Itatani

, Uchida

, The effect of rare-earth oxide addition on the hot-pressing of magnesium silicon nitride. J Eur Ceram Soc. 2002;22(5):777–783. doi: 10.1016/S0955-2219(01)00380-6

19.

, Yu

, Hiragushi

, Phase composition of TiO₂-coated ZrO₂/Si₃N₄ composite. J Eur Ceram Soc. 1999;19:299–303. doi: 10.1016/S0955-2219(98)00204-0

20.

Fricke

, Nonninger

, Schmidt

Production of Si₃N₄/TiN nano composites. Adv Eng Mater. 2000;2:647–652. doi: 10.1002/1527-2648(200010)2:10<647::AID-ADEM647>3.0.CO;2-E

21.

Schmidt

HK.

The role of TiN in the intergranular phase-forming process in TiN-dispersed Si3N4 nanocomposites. J Am Ceram Soc. 2005;88:404–410. doi: 10.1111/j.1551-2916.2005.00054.x

22.

Greskovich

Preparation of high-density Si₃N₄ by a gas-pressure sintering process. J Am Ceram Soc. 1981;64(12):725–730. doi: 10.1111/j.1151-2916.1981.tb15895.x

23.

Mitomo

, Yang

, Kishi

, Influence of powder characteristics on gas pressure sintering of Si3N4. J Mater Sci. 1988;23:3413–3419. doi: 10.1007/BF00551328

24.

Wang

, Jiang

, Ge

, Interface structure related to growth of β-phase in hot pressed silicon nitride and its growth mechanism. J Chin Ceram Soc. 1993;21(5):393–397.

25.

Ryshkewitch

Compression strength of porous sintered alumina and zirconia. J Am Ceram Soc. 1953;36(2):65–68. doi: 10.1111/j.1151-2916.1953.tb12837.x

26.

Kingery

, Bowen

, Uhlmann DR.

Introduction to ceramics. New York: John Wiley & Sons; 1976.

27.

, Siores

, Ball

JAR.

Proc. CIPR Intern. Symp; 1997; Hong Kong. p. 612–619.

28.