Silicon carbide-based composites are candidate materials for high-temperature structural applications in heat engines and heat recovery systems [1, 2]. For these applications, the ceramic components are often needed to be manufactured by processes with near-net-shape capability to minimize the requirements to perform very expensive and difficult post-machining operations [3]. Among the various fabrication processes, the melt infiltration process, which has the advantages of near-netshape capability and short processing time, is proposed to be one of the potential processes for such applications [4]. However, SiC-based composites made by conventional melt infiltration processes showed poor high-temperature mechanical properties, which are primarily limited by the low melting-point residual phases as exhibited by the reaction-bonded SiC [5–7]. Much effort has thus been made to reduce or eliminate the low-melting-point residual phases in the composites. One way to reduce the residual phases is through alloy infiltration [8–10]. Better approaches have been demonstrated by Lim et al. [11] and Shobu et al. [12], where high-melting-point MoSi2 and Mo(Si,Al)2 were directly infiltrated into the SiC performs. Both the infiltrated SiC-MoSi2 and SiC-Mo(Si,Al)2 composites exhibited significant increases (more than a 40% increase at 1400 ◦C) in fracture strength between 1200 and 1400 ◦C. In a previous report [13], a novel SiCMo5Si3C composite was fabricated by directly infiltrating high-melting-point Mo5Si3C into the SiC preforms. The composite also showed significant increases in fracture strength between 1200 and 1500 ◦C, which was attributed to the toughening effects provided by the infiltrated Mo5Si3C phases at elevated temperatures. To clarify this point, the fracture toughness of the composite is studied in the present report. The fracture toughness for ceramics can be characterized by various techniques, such as the double torsion (DT) technique [14], the double cantilever beam (DCB) technique [15], the indentation fracture (IF) technique [16], the indentation strength (IS) technique [17], the single edge precracked beam (SEPB) technique [18], the chevron notched beam (CNB) technique [19], etc. The IS method was adopted in the present study because it has proved to be efficient and economical [20]. The preparation procedure for the composite as well as the infiltration characteristics have been reported elsewhere [13, 21], so they are only briefly introduced here. Namely, a SiC preform of dimensions 40× 16× 6 (mm) was prepared from the raw SiC powder (α, circa 3.2 μm, purity 99%, Showa Denko, Japan) through a cold isostatic pressing (CIP) process. An infiltrant of the Mo45Si30C5 composition was prepared from the raw powders of MoSi2 (circa 2.93 μm, Japan New Metals), SiC (the same as above) and Mo (circa 1.3 μm, purity 99.94%, Japan New Metals). The infiltration was performed at 2100 ◦C in an induction furnace in 1 atm argon. Typically, the temperature was ramped to 2100 ◦C in 5 min from 1000 ◦C, held for 20 min, followed by a 5-min cool down to 1000 ◦C, where the power was cut off. The powder mixture was found to melt above 2000 ◦C, and spontaneous infiltration could be achieved above 2050 ◦C. Although 20 min for infiltration was employed in the present study, several minutes were found to be sufficient for a full infiltration. The infiltrated composites are relatively dense, but the relative density could only reach 95% of the theoretical value at best. The remaining porosity is attributed to the uninfiltrated areas up to 20μm and closed pores in the SiC matrix that could not be reached during the infiltration process, as exhibited by the typical microstructure of the composite in Fig. 1. Also, the volumetric shrinkage during solidification would contribute to the porosity. Further improvement in product density is very difficult. The XRD and SEM/EDS investigations confirmed that the composite is only composed of SiC and Mo5Si3C. Beams for the IS testing were cut from the infiltrated samples using a diamond saw, and ground to the nominal dimensions of 3× 1.5× 12 (mm). The tensile surface was polished with 1 μm diamond paste to remove residual stress due to machining and to produce a finish for optical microscopic observations. The edges of the prospective tensile face were slightly chamfered.