Introduction in-situ combustion with oxygen instead of air has high potential for recovering heavy and medium oils. Fig. 1 potential for recovering heavy and medium oils. Fig. 1 illustrates the oxygen combustion process. Among the many advantages is a higher concentration of CO2 ( is greater than 80 %) in the flue gas resulting from the absence of nitrogen. This allows more CO2 dissolution in the oil that helps swell the oil and reduce its viscosity, thus enhancing oil recovery. Although oxygen combustion has been investigated in the laboratory before, the beneficial effect of CO2 has never been demonstrated explicitly. In an attempt to simulate the nitrogen-free environment of oxygen combustion and to study the CO2 effect, we conducted a novel combustion tube run in which a CO2/O2 gas mixture was used in place of air. For comparison, a parallel test was performed that involved use of air at the same operating conditions. Experiment The laboratory apparatus and the experimental procedure have been described previously. A 13.4API [0.9765-g/cm3] California heavy oil and sands from the same reservoir were used. The oil has a viscosity of 6,600 cp [6.66 Pa s] at75F [23.9C] and an atomic H/C ratio of 1.52. The pack properties and operating conditions are listed in Table 1. Both runs were made in the dry combustion mode with about 21 % oxygen in the injection gas. Operations of both tests were smooth and ignition with the CO2/O2 mixture was successful. Results Stabilized combustion was achieved between 2 and 7 hours after ignition for both runs. The burning characteristics averaged over this time period are summarized in Table 1. The reaction temperature, the combustion front velocity, and the product gas composition for both tests were similar, except that the air test had a lower oxygen utilization. The fuel consumptions were practically the same for both runs. This was expected practically the same for both runs. This was expected since the temperature profiles in both runs were nearly identical and the fuel deposition process was largely a function of temperature. The oxygen/fuel ratios were also the same, indicating that the elemental compositions of the fuel were identical. It therefore appeared that substituting N2 in air with CO2 did not change the combustion characteristics for this oil. The oil displacement vs. volume burned is shown in Fig. 2. The fill-in time for the CO2/O2 run was a little longer because of a somewhat higher gas saturation. Once in production, the displacement rates in both runs were almost identical. The injection pressure and pressure drop across the tube are shown in Fig. 3. In pressure drop across the tube are shown in Fig. 3. In both tests, the injection pressure was set at 250 psi [1.7 MPa] by regulating the backpressure. The pressure drop started to increase shortly after ignition, indicating the formation of an oil bank ahead of the thermal front. It reached a maximum at about 2 to 2.5 hours when the oil bank started producing. As seen, the air test had a higher pressure drop that was associated with a bank of oil at the pressure drop that was associated with a bank of oil at the original viscosity, The oil bank in the CO2/O2 run, on the other hand, contained oil at a lower viscosity because of CO2 dissolution. As a result, less pressure drop was required to displace the oil at the same rate. Conclusions The system pressure of the tests was rather low, only 250 psi [1.7 MPa], simulating conditions in a typical psi [1.7 MPa], simulating conditions in atypical California heavy oil reservoir. P. 1137