LIQUEFACTION AND FRACTIONATION OF NATURAL GAS

The authors have made some laboratory experiments on the liquefaction and fractionation of natural gas. The apparatus (Pl. VIII, p. 75) used consisted of a Töpler mercury pump by means of which vessels could be exhausted of their contained gases, and the gases trapped, measured, and analyzed; Dewar vessels for maintaining liquids at low temperatures for hours; gas-analysis apparatus; specially designed glass gas containers; and liquid air and solid carbon dioxide for obtaining low temperatures.

NATURAL GAS USED AT PITTSBURGH

The first experiments were made with the natural gas used at Pittsburgh, Pa. This gas, as shown by combustion analyses at the time the fractionation experiments were recorded, contained 79.2 per cent of methane, 19.6 per cent of ethane, and 1.2 per cent of nitrogen. The percentage probably includes some of the other inert rare gases, such as helium.61

FRACTIONATION OF PARAFFINS

As stated before, the analyses by combustion show only the two predominating paraffins. The authors attempted to separate the paraffins in the gas mixture by liquefaction and fractional distillation, as follows: About 90 or 100 c. c. of the gas was introduced into a previously exhausted glass vessel of about 100-c. c. capacity, and the vessel was immersed in liquid air in a Dewar flask. After 10 or 15 minutes the glass vessel, still immersed in liquid air, was connected to a Töpler pump and as much gas as possible was pumped off. The container was then removed from the Dewar flask, allowed to come to room temperature, and the residual gas collected. The first fraction was again liquefied, and as much gas as possible pumped off at the temperature of liquid air. The container was then again removed and allowed to come to room temperature, and the residual gas was added to the residual gas from the first operation. The fractions were then analyzed by the slow-combustion method. All measurements were made over mercury.

By this method of fractionation, methane can be completely separated from the other homologues in a gas mixture, as the vapor pressures of the paraffins higher than methane are almost zero at the temperature of liquid air (-190° C.). To separate the methane completely, a second liquefaction and fractional distillation was employed, the process being outlined above.

61 Rogers, G. S., Helium-bearing natural gas; U. S. Geol. Survey, Prof. Paper 121, 1921, pp. 28-29. Shepherd, Martin, and Porter, Frank, An improved method for the separation of gas mixtures: Jour. Ind. Eng. Chem., vol. 15, 1923, pp. 1143-1146.

The boiling points of ethane (-93° C.), propane (-45° C.), and butane (1o C.) lie so close together that a clean separation of those constituents is not easily made. At any temperature at which all the ethane can be removed by means of the mercury pump the vapor pressure of the propane is so great that part of the propane is removed also. However, after the separation of the methane at the temperature of liquid air such a temperature can be realized that all of the ethane and part of the propane can be obtained in one fraction, and the remainder of the propane and all of the butane can be collected in the second fraction. According to Lebeau and Damiens,62 the optimum temperature at which all the ethane and part of the propane can be separated from the remainder of the propane and all the butane is -130° C. After the first fraction has been removed, the vessel containing the residual gas is removed from the Dewar flask containing the refrigerant, and the remainder of the propane and all of the butane in this second fraction are collected at normal temperatures by means of the mercury pump.

After the fractions are collected they can be analyzed by the slowcombustion method, as each contains only two constituents. The composition of the mixture will then be accurately known.

If gaseous paraffin hydrocarbons and the vapors of liquid hydrocarbons exist in a mixture, separation of the gaseous paraffins can be made from the vapors of the liquid paraffins by subjecting the mixture to a temperature of 100° C., when the gaseous constituents can be collected by means of the mercury pump (Pl. VIII), the vapors of the liquid paraffins being left behind. The condensed vapors can then be recovered by allowing them to evaporate at room temperatures. Such mixtures are found in casing-head gases and gases drawn from the earth under reduced pressure in connection with the manufacture of gasoline from natural gas.

CONSTITUENTS OF PITTSBURGH NATURAL GAS

By using the above method for determining the percentage of the various constituents in the natural gas used at Pittsburgh, the following results were obtained:

Comparison of data obtained in fractionation tests

62 Lebeau, P., and Damiens, A., Sur une méthode d' analyse des mélanges d' hydrogen et d'hydro-carbures saturés gazeux; mélanges complexes: Compt. rend., t. 156, No. 4, 1913, p. 325,

These analyses show fair agreement among themselves. The greatest discrepancy is in the ethane and propane content, principally because of the difficulty encountered in exactly analyzing a small sample of either of these two gases. By taking larger samples of the original gas, however, this difficulty may be overcome.

There still remains the possibility of the presence of butane in the natural gas. By the method outlined above methane alone can be separated from the other homologues, as its boiling point lies 67° C. from that of ethane, the next higher hydrocarbon.

The boiling point of ethane, however, differs from that of propane by only 48° C., and the boiling point of propane differs from that of normal butane by 44° C.; consequently, no sharp separation can be made by the above method for ethane, propane, and butane, but such a temperature can be utilized that after separation of the methane from the other paraffins at -190° C. all the ethane and part of the propane can be obtained in one fraction and the remainder of the propane and all the butane can be collected in the second fraction. This temperature is 130° C. The authors subjected residual fractions (paraffins higher than methane) from the natural gas of Pittsburgh to this temperature, and by means of the mercury pump removed as much of the gas as possible. The removed fraction was found to consist of ethane and propane and the residual fraction of propane only.

TEST FOR GASOLINE CONTENT OF NATURAL GAS

As a result of the condensing of certain paraffin hydrocarbons in a natural-gas mixture and the use of the resultant liquid as gasoline, a demand has arisen for a test that will show the amount of liquid obtained per unit volume of natural gas at definite pressures and temperatures. Since by present procedure it is impossible to determine the separate paraffins in a mixture, the demand has been met in part by determining the density of the gas and dissolving the gas in a solvent that has a greater solvent action on the higher than on the lower paraffin hydrocarbons. Claroline oil, olive oil, and paraffin oil have been used.63

The solubility test is not quantitative because all of the dissolved constituents represent the gasoline to be obtained in actual practice, and because certain paraffins are dissolved to the entire exclusion of the others. Even methane is soluble to some extent in the solvents named, but the higher paraffin hydrocarbons are much more soluble;

63 Burrell, G. A., and Jones, G. W., Methods of testing natural gas for gasoline content: Tech. Paper 87, Bureau of Mines, 1916, 26 pp.

consequently, it is assumed that under the same conditions of plant operation the natural gas that is most soluble will produce the most gasoline.

RELATIVE EFFECTS OF SOLVENTS

To determine the relative solvent effects, the authors have subjected some samples of natural gas to solution in different oils. The

following table shows the solubility in several varieties of oil of Pittsburgh natural gas, at a temperature of 20° C. The specific gravity of the gas as compared to air is 0.63 at 0° C. and 760 mm. pressure. It is not adapted to the commercial production of gasoline, but was used to show the solvent action mentioned before.

In making a test, 35 c. c. of the oil was placed over mercury in a Hempel pipette (fig. 27), and was shaken for about 3 minutes with 100 c. c. of the gas until there was no further reduction in volume of the gas.

Results of tests of solubility of the natural gas used at Pittsburgh in different

The above determination could be checked within 0.5 per cent. Considerable uniformity regarding the solubility of the natural gas in the different oils will be noticed. The claroline oil had the following characteristics, as determined by I. C. Allen, former petroleum chemist of the Bureau of Mines:

The solubility of pure methane in claroline is 11.0 per cent and in cottonseed oil 9.5 per cent. The solubility of pure ethane oil is 68.5 per cent.

It was found that results equivalent to those obtained from the use of claroline oil could be obtained by using 50 c. c. of ethyl alcohol in the same manner as claroline oil was used.

DETERMINATION OF DENSITY OF NATURAL GAS

WEIGHT METHOD

Determination of the density of natural gas consists in weighing the gas in a globe of known capacity under definite conditions of temperature and pressure. The globe used is a spherical glass bulb sealed to a capillary glass stopcock that has been worked very carefully. To eliminate errors due to changes in the atmospheric temperature and pressure and consequently to change in the buoyancy of the globe, a closed bulb having the same external volume as the density globe is used as a counterpoise during the weighing. The capacity of the globe is determined at about 16° C.