As regards a natural gas of low specific gravity and low absorption percentage (known as a "lean" gas), the safest recourse is to test by means of a portable outfit consisting of a gas meter, small gas engine, compressor, cooling coils, and receiver. Such an outfit can be hauled from place to place on a wagon. This method is in all cases to be recommended as having distinct advantages over laboratory tests. However, it is true that tests made with the portable outfit may be misleading unless in charge of a careful and experienced person.

The authors have also used a small stationary outfit consisting of a meter with a capacity of 15,000 cubic feet per 24 hours, a small compressor, driven by a steam engine, 100-foot cooling coils made of 1-inch pipe, immersed in a tank of water, and a storage tank 5 feet high made of a 6-inch piece of pipe. To the latter was attached a relief valve which could be set to operate at the desired pressure. A trap was installed between the compressor and the cooling coils to catch oil that was sometimes brought from the wells with the gas. A glass gage was connected to the storage tank to indicate the volume of condensate produced.

In conducting tests of a gasoline plant the plant is first operated for an hour or two to insure that everything is working well. The meter and all pressure gages must be in good order. The cooling coils should dip enough to drain readily the gasoline into the storage tank. The efficiency of the cooling coils can be ascertained fairly well by measuring the temperature at different places in the water of the tank. At the point where the coil enters the water it will be hot enough to warm the water appreciably, but if the tank is large and a sufficient length of pipe for cooling purposes is installed this warming of water is only local.

There follows a form used by the authors in testing wells on a given lease. The data shown represent an actual "plant" test:

Date, March 7, 1913.

Number of wells, 4.

Temperature of gas at meter, 58° F. (14° C.).

Pressure of gas at meter, Atmospheric.

Pressure in accumulator tank, 300 pounds per square inch.

Temperature of water in cooler, 54° F. (12° C.)

Gas used, 1,100 cubic feet.

Condensate produced, 0.83 gallon per 1,000 cubic feet.

Gravity of gasoline produced, 85° B.

Evaporation loss of condensate on exposure to air, 11 per cent in 3 hours.

Gravity of gasoline after evaporation loss, 82° B.

Table 4 following shows the quantity of gas issuing from 16 wells on the same lease. The gas all comes from the same sand, the Berea grit, in West Virginia. The wells had been drilled about 10 years and the oil production averages about 25 or 30 barrels a day from 52 wells. The gas from all of the wells was not tested by the authors.

TABLE 4.—Results of laboratory and field tests of 16 different wells on the same lease.

A gasoline plant capable of taking care of 100,000 cubic feet of gas daily was in operation on this lease and was connected to wells 1, 6, 14, and 15, mentioned in the above table. Some gasoline was produced, but the successful operation of the plant had not been assured when the authors visited the lease. About 1,200,000 cubic feet of gas was available. Certain wells, among them many of high capacity, notably Nos. 2, 3, 4, 5, 7, 8, 10, 11, 12, and 16, appeared to be too dry for consideration for use in the condensation of gasoline from the gas.

RESULTS OF ANALYSES OF NATURAL GASES OF OKLAHOMA.

Following are the results of a laboratory analysis of a natural gas that is used for the condensation of gasoline from natural gas at a plant in the Glenn pool district; 250,000 cubic feet of gas was being used from eight wells. The operators claimed that they obtained about 11 gallons of gasoline from each 1,000 cubic feet of gas treated in the plant.

Results of analysis of casing-head gas from Oklahoma.

The oil absorption was 20.0 per cent and the specific gravity (air 1) was 0.75.

The analytical results are interesting as showing a gas from which gasoline is condensed, although it is low in specific gravity and oil absorption as compared with similar values of other samples tabulated below.

a The oxygen is undoubtedly due to a slight leakage of air in the sample container.

The following table shows the specific gravity and the oil absorption of gases obtained from the casing heads of different wells on the same lease in the Glenn pool district of Oklahoma. A gasoline plant had not been installed at the time the authors collected the samples, but one was contemplated. The results show the variation in gases from the same lease and are in marked contrast to the values for the gas mentioned above, which was also collected in the Glenn pool district but from a different lease.

Specific-gravity and oil-absorption values of Glenn pool gas samples.

COMPRESSION AND LIQUEFACTION OF THE CONSTITUENTS OF NATURAL GAS IN PLANT OPERATION.

The condensation of gasoline from natural gas is essentially a physical process. If any chemical reactions take place, they are slight and inappreciable. The authors tested residual gases from 10 different plant operations to determine whether carbon monoxide or olefin hydrocarbons were produced. These gases with others are found when the higher paraffins are decomposed at high temperatures and pressures in the absence of air. Neither carbon monoxide nor olefin hydrocarbons were found.

THREE COMMERCIAL PROCESSES.

At present three processes for the extraction of gasoline from natural gas are used commercially. The one most generally used involves compressing the gas to a certain pressure and subsequently cooling it by means of water or air. A second consists in simply cooling the gas without compression by means of a refrigerant, such as liquid ammonia, evaporating under reduced pressure. A third is a combination of the other two.

RELATION OF PRESSURES EXERTED BY GAS MIXTURES.

In order to understand changes that take place in the gas mixture as it passes through the compressors, knowledge of pressures exerted by gas mixtures is essential. In a mixture of gases exerting a certain total pressure each individual constituent of the mixture exerts a part of the pressure. Atmospheric pressure at sea level is about 15 pounds per square inch, of which about 3 pounds is due to the oxygen and about 12 pounds to the nitrogen. In order to be liquefied, a gas

must be compressed or cooled or both. If it is to be liquefied by pressure alone, the pressure applied must be greater than the vapor pressure of the liquefied gas when boiling.

When a vapor is mixed with other gases or vapors, only a part of the total pressure is exerted on the vapor. If the vapor constitutes 10 per cent of the mixture, the pressure on the vapor is 10 per cent of the total pressure. For such a mixture, it being assumed that the vapor under consideration will be the first to condense out, a pressure of 150 pounds would be required to have a pressure of 15 pounds on the vapor. Under a pressure of 15 pounds the vapor would begin to condense to a liquid at the temperature at which the liquid would normally boil. Butane boils at 1° C. (34° F.), when its vapor exerts a pressure of 15 pounds per square inch. Hence, to condense butane vapor to liquid at 1° C. (34° F.), there would be required a pressure of at least 15 pounds on the vapor. If the butane constituted 20 per cent of a mixture, there would be needed a total pressure of 75 pounds in order to have a 15-pound pressure on the butane vapor, and to cause condensation to begin. As condensation took place, butane vapor would, of course, be removed from the mixture; that is, its partial pressure would diminish and a pressure greater than 75 pounds would have to be applied to cause the condensation to proceed. Hence, if one knew the exact quantity of butane vapor in a particular natural gas, a pressure greater than that theoretically required to start the condensation would have to be applied in practice. From the above discussion it will be seen why one gas may produce condensate under a pressure of 75 to 100 pounds, whereas another gas may need a pressure of 200 to 300 pounds to produce the same quantity of condensate of the same constituent.

A similar calculation can be made for pentane or other vapors of the liquid paraffin hydrocarbons. The vapor pressures of three of these liquids and of two of the gaseous paraffin hydrocarbons when liquefied are tabulated here.

Vapor pressures of three liquid hydrocarbons at different temperatures.a

a Landolt and Börnstein, Physikalisch-chemische Tabellen, 1905, 3d ed., p. 137.

• C.

0... 12.5.. 15.0.

Vapor pressures of two gaseous paraffin hydrocarbons at different temperatures.

Vapor pressure of—

22.0.

34.5 a

53.0..

102.0 a.

If one were to make the assumption that the vapor of pentane only was present in a natural-gas mixture it would be possible to ascertain the pressure necessary to condense this vapor at any desired temper

ature.

For example, suppose that gas available for gasoline production contains 3 per cent of pentane vapor, and suppose that determination of the pressure necessary to condense this vapor at 20° C. (68° F.) is desired. The above table shows that pentane has a vapor pressure of 8.1 pounds per square inch at 20° C. (68° F.); therefore the pressure required to start the condensation is

8.1

0.03,or 270 pounds per square inch. As shown on page 25, 1 gallon of pentane will produce when volatilized about 31 cubic feet of vapor at 0° F. (32° C.) and 30 inches of mercury. If a gas yields 1 gallon of pentane per 1,000

31 1000

cubic feet of gas, then there must be part, or about 3 per cent of

vapor in the mixture. Consequently, when less than 3 per cent of vapor occurs the pressure would have to be raised above 270 pounds to condense the vapor. If a pressure above 270 pounds per square inch and subsequent cooling below 20° C. were not employed, the vapor would not condense.

It has been stated that methane and ethane are never liquefied in plant operation. But propane and butane may be liquefied under certain conditions. At 22° C. (72° F.), to liquefy the vapors of pure propane, a pressure of 132.3 pounds per square inch is required. If 50 per cent of the vapor is present, a pressure of twice 132.3, or 264.6, pounds per square inch is required at a temperature of 22° C. (72° F.).. As far as the authors are aware, no vapor-pressure curves for butane have been determined.

But as the boiling point of propane (-45° C.; 49° F.) is much lower than that of butane, the latter will be liquefied more easily than propane. As no data are available, no figures can be given. That butane is a constituent of some natural-gas condensates is shown by