PSI - Issue 66

Johannes L. Otto et al. / Procedia Structural Integrity 66 (2024) 256–264 Johannes L. Otto et al. / Structural Integrity Procedia 00 (2025) 000–000

260

5

3. Results and Discussion 3.1. Microstructure and corrosion analysis

As it can be seen in Figure 4a) on the cross section of the brazed joint by scanning electron microscope (SEM) images using the backscatter electron (BSE) detector, there was complete isothermal solidification as shown schematically in Figure 2. In the diffusion zone, the Cr-rich borides can be seen as dark phases on the grain boundaries. Large grains can be seen away from the brazed seam, while the Cr-rich borides near the solidification zone appear to have inhibited grain growth. Isolated pores can be seen inside the solidification zone. To provoke a strong corrosive attack, the samples were boiled at 106°C in an aqueous solution of 25 wt.-% sulfuric acid (H 2 SO 4 ), 50 g/l copper sulfate (CuSO 4 ) and 100 g/l of pure copper chips for seven hours. The topography was measured afterwards using confocal microscopy with focus variation, shown in Figure 4b). It was found that there was a loss of material, but it was limited to a depth of less than 500 nm inside the solidification zone. Also, some material loss was observed at grain boundaries near the Cr-rich borides which indicates a Cr-depletion that has weakened the passive layer locally, but overall, the test demonstrated good corrosion resistance for a brazed joint.

Fig. 4. (a) Brazed joint microstructure in scanning electron microscope (SEM) with backscattered electron (BSE) detector, (b) surface topography using confocal microscopy after seven hours of corrosion attack in a sulfuric acid and copper sulfate solution with copper chips boiled at 106°C.

3.2. Fatigue and corrosion fatigue performance In a first step, pure fatigue tests (with cyclic creep) were carried out in air with different stress amplitudes to have a comparison basis for the drop in fatigue performance under corrosive environments. In air specimens withstood even at maximum stress σ max = 500 MPa more than N f = 10 5 cycles to failure. As in the following test series, the specimen fractured in the brazed joint, unless otherwise stated. The fracture mechanism will be discussed detailed in chapter 3.3. In a second step, it was examined whether the fatigue life changed under the influence of demineralized water (H 2 O). Here, a slight decrease at higher stress amplitudes was discovered, which is partly within the range of the scatter. The causes could lie in possible smallest quantities of impurities in the water, the improved heat dissipation by the constant flow of the surrounding medium or the mechanical influence of a liquid medium inside cracks which has a much higher bulk modulus than air. A classic corrosive attack can be ruled out, since the SN-curve comes close to that of air with a longer test duration, which conflicts with corrosion mechanisms and their effect on fatigue life. The synthetic exhaust gas condensate K2.2 (short: EGC) was used as the first real corrosion medium. It contains small amounts of formic acid and acetic acid, as well as a small amount of NaCl. The exact composition is given in Table 2. With a pH-value of 3.5, it represents a moderate environment, but nevertheless it showed a significant effect on the SN-curves. Compared to the SN-curves in air and water, the lifetime N f at σ max = 350 MPa is only one third. No corrosion products could be visually detected after the tests, and the shift of the overall SN-curve again appears to indicate no classic corrosion attack with large material loss that could have accelerated crack growth. Instead, the point at which cracks initiate and crossing the threshold of stress intensity for crack propagation ∆ K th appears to have