The fluidized bed pyrolysis of bitumen-impregnated sandstone from the tar sand deposits of Utah

The influence of process operating variables on the product distribution and yields for the pyrolysis of bitumen-impregnated sandstone in a fluidized-bed reactor have been investigated in both laboratory (11/2 inch diameter) and pilot-scale (4 1/2 inch diameter) reactors. The quality of the liquid products has also been determined and related to the process operating variables. The process variables investigated included pyrolysis reactor temperature, sand retention time in the pyrolysis zone of the reactor, the fluidizing gas velocity, and the average feed-sand particle size. The ranges of the variables were as follows: temperature 698-973 K (425-700°C), feed sand retention time 15-35 minutes, fluidizing gas velocity one to four times the minimum fluidization velocity, and feed sand particle sizes from 359 micron up to 1/2 inch. In the case of the bench scale experiments, the reactor pressure and feed sand particle size were constant; the reactor pressure was maintained at atmospheric pressure and the average feed sand particle size was 359 microns. In the pilot-scale experiments the feed sand particle size varied from 359 microns up to 1/2 inch. A variety of physical, chemical, and spectroscopic analyses was used to characterize the native bitumens and the bitumen-derived hydrocarbon liquids produced in the fluidized bed pyrolysis experiments. The Utah tar sand deposits investigated included Sunnyside, Whiterocks, PR Spring, Tar Sand Triangle, Circle Cliffs, and Asphalt Ridge. The sand retention time appeared to be the most significant variable affecting the product distribution and yield of the bitumen-derived hydrocarbon liquid, whereas the fluidizing gas velocity had little effect on the product distribution, yields, and liquid product quality for the range of values- studies. The liquid product yield increased with decreasing sand retention time; however, the yield of the carbonaceous residue on the sand was insensitive to changes in sand retention time. The gas yield increased with increasing temperature while the liquid yield decreased. The coke yield decreased as the reactor temperature increased up to 723 K and remained constant as the reactor temperature increased above 723 K. The liquid product quality and yield pattern indicated that the optimum operating conditions should be those approximating visbreaking rather than coking. The unusual nature of the Circle Cliffs tar sand will also be discussed.