Equations estimate ASTM distillation loss in gasoline

April 1, 2013
In the ASTM D86 distillation test, pentanes minimize hydrocarbon loss by "holding back," or suppressing the volatility of, butanes.

William Ed Smith IV
Consultant
Newark, Del.

William E. Morris
Consultant
Wilmington, Del.

Gasoline blending and storage are among the final steps in a modern complex refinery, as shown here (upper right quadrant) at National Cooperative Refinery Association's 85,000-b/d refinery at McPherson, Kan. Photo from NCRA.

In the ASTM D86 distillation test, pentanes minimize hydrocarbon loss by "holding back," or suppressing the volatility of, butanes.

Studying blends of a depentanized gasoline gathered from a service station, the authors determined the effects of isobutane, normal butane, isopentane and normal pentane on percent hydrocarbon loss. This article presents equations to predict percent loss due to isobutane, normal butane, and total pentane content.

Winter compositions

With unusual winter gasoline compositions, the loss of low boiling hydrocarbons in the ASTM distillation test can make it difficult for a refiner to meet the 95% recovery specification. Normal butane with a boiling point of 31° F. can pass through the condenser tube as a vapor and be lost.

The temperature at the surface of the condenser tube is probably slightly greater than 32° F. because of the heat released when vapor is condensed. Lower boiling isobutane and propane are of course lost much more easily. Isopentane with a boiling point of 82° F. condenses and would be expected to decrease the loss of butane that condenses with the isopentane.

If normal butane is blended with a high-boiling gasoline component, the blend will give a high percent loss in the ASTM test in the absence of pentanes to "hold back" the normal butane.

Test results

The authors conducted a study to define the effects of individual butanes and pentanes on distillation percent loss. They prepared a base gasoline by depentanizing service station gasoline. The depentanized gasoline contained only 0.09 vol % normal pentane based on gas chromatography analysis.

They added 14 combinations of butanes and pentanes, which gave RVP values of 14-17 lb. Table 1 shows composition, RVP, and ASTM distillation data for these blends. Running three to five repeat tests improved the accuracy of the average distillation loss.

Blend C consisting of 39% base gasoline and 61% isopentane gave little or no loss. Blends of the base gasoline with 13% isobutane or 26% normal butane both gave about 10% loss. Blends containing about equal amounts of pentanes and butanes gave about 5% loss with isobutane and 2% loss with normal butane. As might be expected, it is more difficult to hold back the lower boiling isobutane than to hold back normal butane.

With no pentanes present, 73% of the isobutane was lost in Blend A, and 39% of the normal butane was lost in Blend B.

Loss-prediction equations

The authors found that the percentage of either isobutane or normal butane lost could be related to the ratio of total pentanes to total butane (accompanying diagram). Normal butane can be held back by a small amount of pentanes, while a much larger amount of pentanes is required to hold back isobutane.

The accompanying box shows the equations for the curves in the diagram. Table 2 shows loss with blends containing both normal butane and isobutane.

The authors did not include in the study the effect of propane on percent loss. It is reasonable to assume, however, that all propane is lost in the ASTM distillation. Typical winter gasoline contains only 0.2 vol % propane.

Effects of varying RVP

For a given RVP and with a given quantity of pentanes to hold back the butanes, how much loss can be expected from isobutane and from normal butane? For a winter gasoline with 12-lb RVP and 14% pentanes, if isobutane is used as the pressurizing agent, the estimated loss would be 3.3%.

The accompanying box gives calculations. Similar calculations were made for 15-lb gasoline and for half the typical amount of pentanes with the results shown in Table 3.

With a pressurizing agent composition of 20% isobutane and 80% normal butane, the percent loss from a 15-lb RVP gasoline would become excessive only if the pentane content were abnormally low. In practice, gasolines with a high loss probably have high RVP, below average pentanes, above average isobutane, and above average propane.

To meet a 95% recovery specification, the blending percent loss should be less than 4% because distillation residue is usually about 1%. There is seldom a problem, but in one case a refiner was having difficulty meeting the 95% recovery specification for its premium gasoline. When advised of the importance of pentanes for holding back butanes, the refiner found a means of fractionating pentanes to obtain a small amount of additional isopentane for its premium gasoline.

Percent loss will seldom be a limitation in gasoline blending, but when it is, the equations presented here will be useful in deciding how to meet the specification.

End point

End point is another specification that will seldom be a limitation. In case it is, results of an excellent end point study reported more than 50 years ago may be useful.1

The highest end point component essentially controls the end point of the blend. There are, however, significant differences in the shapes of the end point vs. percent composition curve, depending upon the difference between 90% and end point temperature of the high end point component. Surprisingly, in certain cases, the end point of a blend of two components can be higher than the end point of either one.

When a refinery blend has too high an end point, it is difficult to make a correction. We speculate that a small amount of a narrow boiling component perhaps 20° F. less than the target end point might work. The assumption is that such a component would carry over some of the highest boiling hydrocarbons. If so, this would decrease the distillation rate at the end of the distillation and thereby lower the end point, by decreasing the heat flow to the thermometer.

Reference

1. Stanley, M.E., and Pingrey, G.D., "Prediction of ASTM End Points Of Blended Light Petroleum Products," Industrial & Engineering Chemistry, October 1954.

The authors

William Ed Smith IV ([email protected]) is an independent consultant working in Newark, Del. He served as an officer at the Naval Radiological Defense Laboratory 1959-62 and worked for DuPont Co. 1962-93, when he retired to begin his consulting practice. Smith holds a bachelors of nuclear engineering and an MS in physics from North Carolina State University. He is a member of Sigma Xi, the Scientific Research Society.

William E. Morris ([email protected]) is a consultant in Wilmington, Del., who provides gasoline blending computer programs via www.gasolineblendingplus.com. He worked for ExxonMobil Corp.'s Bayway, NJ, refinery for 7 years. Then he joined the DuPont Co.'s petroleum laboratory where he provided technical service for refineries in support of petroleum chemical sales. Morris holds a BS in chemical engineering from the University of Missouri.