Silicon nitride is being extensively evaluated for use in making structural components exposed to high temperatures. Strength degradation during use, which governs the durability of silicon nitride, is generally caused by subcritical crack growth from a pre-existing flaw. One method for making a direct evaluation of subcritical crack growth is to measure the crack growth velocity as a function of the applied stress intensity. A stress rupture test for examining delayed fracture under a constant stress is usually conducted to evaluate strength degradation due to subcritical crack growth from a natural flaw. At high temperatures, however, other strength degradation mechanisms such as oxidation and cavity formation can also affect delayed fracture. It is therefore important in stress rupture tests to identify the fracture origins. Several studies [1-5] have been reported on the fracture origins of silicon nitride in stress rupture tests in air at high temperatures. Quinn et al. [1] examined the delayed fracture surface of hot-pressed silicon nitride; they reported that a short time-to-failure test of 1 h). Oxidation pits found in delayed fracture origins of silicon nitride were also reported for stress rupture tests in air [3]. Thus, flaw generation during testing needs to be minimized in order to evaluate the effect of subcritical crack growth. In the present study, stress rupture tests of sintered silicon nitride at high temperature were conducted in a nitrogen atmosphere to minimize flaw generation due to passive oxidation, which forms an oxide surface layer, and active oxidation, which causes the surface decomposition. The material used in this study was sintered silicon nitride having a density of 3.08 g cm -3 . This material was fabricated by sintering a compact of powder mixtures of 88 wt % S i 3 N 4 , 8 Wt % Y 2 0 3 and 4 wt % A120 3 in a nitrogen atmosphere under a pressure of 0.1 MPa at 1700 °C for 1 h. Specimens prepared for stress rupture tests were ground to the dimensions 2 mm × 8 mm x 35 mm. Fig. 1 shows a schematic diagram of the apparatus used for three-point bending stress rupture tests. The load was applied on the test specimen using a deadweight assembly with a cantilever arm. The experimental set-up was equipped with a microswitch to cut off power to the furnace and the timer at the instant specimen failure occurred. After the specimen was installed in the three-point bending fixture made of silicon carbide, the temperature was raised using graphite heating elements. Nitrogen gas was introduced into the furnace through an inlet located in the side, near the bottom. The load was applied on the specimen after reaching the test temperature of 1000 °C at a heating rate of approximately 300 ° C h 1 . Details of the design, construction and operation of the furnaces as well as the fixtures were described in [6]. Fourteen flexural stress rupture specimens of sintered silicon nitride were tested. The applied stress was varied from 296 to 508 MPa. Complete results are given in Table I. Eight specimens were fractured during the tests and tests for six specimens were interrupted before fracture occurred. The time-to-failure ranged from 0.0007 to 24.7 h. Fracture of specimen no. 1 occurred just after loading,
Read full abstract