Tungsten (W) has unique advantages for hightemperature applications such as having the highest melting point (3680 K) of all metals, a high modulus of elasticity (315 GPa at 2273 K) and a low selfdiffusion coef®cient. Its poor low-temperature ductility, however, has limited its applications. Extensive alloying studies have focused on tungsten to improve its low-temperature fabricability and high-temperature strength, and it was found that the addition of rhenium (Re) increased room-temperature ductility and the high-temperature creep strength of tungsten [1]. Strengthening tungsten with ®ne precipitates of second-phase particles is a very effective way of improving high-temperature tensile and creep properties at temperatures above 0.5 Tm (Tm is the melting temperature of tungsten in Kelvin). Among potential strengtheners, hafnium carbide (HfC) has the highest melting point, the highest thermodynamic stability and the lowest vapor pressure. Combining the low-temperature ductilizing effect of Re and the high-temperature strengthening effect of HfC makes W±Re±HfC alloy one of the candidate structural materials for use at service temperatures above 1500 K. Because creep is one of the major concerns for high-temperature materials, the creep behavior of W±Re±HfC alloy was studied in this research. The material used in this research was an arcmelted and swaged W±4 wt % Re±0.32 wt % HfC (W±4Re±0.32HfC), which was supplied by the National Aeronautics and Space Administration. As-received materials was in the form of 25.4 mm diameter rods. Plate-type specimens with a gage section of 12:7 mm 3 3:2 mm were prepared by electric discharge machining. The specimens were ®rst mechanically polished with abrasive papers and then lapped with alumina powders. After the mechanical polishing, the specimens were chemically polished in a 10% NaOH solution to ensure a smooth surface. Prior to testing, all the specimens were annealed at 2500 K for 1 h to produce recrystallized grains. The creep test in this research was conducted in a custom-built ultrahigh-temperature, ultrahigh-vacuum (UHV) creep test station. The specimens were tested under stresses from 40 to 70 MPa, with 10 MPa intervals, at 2200, 2300 and 2400 K. Specimens were heated by passing alternating current through them. A mechanical pump, a sorption pump and an ion pump ensured a vacuum level of less than 1:33 3 10y6 Pa (1 3 10y8 Torr) during the creep tests. The specimen temperature was measured by a micro-optical disappearing ®lament pyrometer that was calibrated with a standard tungsten ribbon ®lament lamp provided by the National Institute of Standards and Technology. Blackbody temperature of the specimen was veri®ed by high-temperature emissivity data from a sample with the same alloy composition. The maximum uncertainty of the measured temperature was 5 K for the entire temperature range used in this research. Load was applied to the specimen through stainless-steel bellows that were attached to the UHV chamber, and corrections were made for changes in the spring force of the bellows caused by extension of the load column. Specimen elongation was measured by a linear variable differential transformer interfaced with a data acquisition system, which made it possible to measure the creep strains with 0.03% accuracy. The specimens were examined before and after the creep test with a transmission electron microscope (TEM) to study the microstructures that developed during the high-temperature deformation. Samples for the TEM study were cut from the gage section of the post-test specimens, and the thicknesses of the samples were reduced to approximately 120 im. From these 120 im samples, disc-shaped blanks with a diameter of 3 mm were produced using an ultrasonic cutter, and both sides of the blanks were ground to a thickness of 70 im. The blanks were dimpled to a thickness of 30 im at the center using a 6 im diamond paste for 2 h and then were polished with a 1 im diamond paste for 5 min. After the mechanical thinning procedures were complete, the specimen blanks were ion milled at an angle of 15±208 for 10±15 h. The samples were then examined with a high-resolution TEM at magni®cations of 5000, 15 000, 30 000, 60 000 and 100 000 times at 200 kV. All specimens fractured at the center of the gage section. There were no changes to the composition after the tests. The results of the strain-time creep curve [2] showed three regions of a creep curve: