The development of a new, robust, portable life support system (PLSS) is currently a crucial need for NASA as it continues making advances in space exploration and extends its missions beyond low Earth orbit. The control of carbon dioxide (CO2) produced by astronauts during extravehicular activities (EVAs) is a critical function of the PLSS. Although the metal oxide and lithium hydroxide canisters, currently used to remove CO2, have worked well, they have a finite CO2 absorption capacity. Therefore, the only way to extend mission times with current technology is to make the unit larger or use excessive energy to regenerate, which is undesirable. Therefore, new CO2 control technologies must be developed for EVA applications. Although much recent work has centered on sorbents that can be regenerated during the EVA, this strategy increases the system size, weight, complexity, and power consumption. A much simpler approach is to use a membrane that selectively vents CO2 to space. A membrane is desirable because it is a continuous system with no theoretical capacity limit and it requires no consumables or hardware for switching beds between absorption and regeneration. Unfortunately, conventional gas separation membranes do not have adequate selectivity for use in the PLSS. However, the needed performance could be achieved with a supported liquid membrane (SLM) using a liquid reagent that selectively reacts with CO2 over oxygen (O2). Reaction Systems is developing new liquid compounds that absorb CO2 and have effectively zero vapor pressure, making these compounds ideal candidates for use in a SLM. In this paper, we describe results obtained with two sorbents. The sorbents were first characterized with nuclear magnetic resonance, and then the reversible CO2 uptakes of each were measured. Finally, the liquids were impregnated in porous membranes. Permeance measurements were carried out with both CO2 and O2. While the initial results show that higher permeance and selectivities are needed, they also suggest that SLMs have the potential to control CO2 in an EVA application and identify specific paths towards achieving the needed performance.