From scientific tables we find that S at 60°F. has a value of 0.9990, and from the Westphal balance Sm has been found to be 0.9673. By transforming the formula above it is seen that

If it is desirable to ascertain the Baumé reading for the dry oil, we next ascertain its value from the above relationship of specific gravity and the Baumé scale from equation (2).

According to formula (3) this Baumé reading would of course be computed as follows:

When a large quantity of oil is to be purchased and it is desirable to carry the Baumé reading to still further decimal points, the two formulas will not of course check; hence, one or the other of these formulas should be agreed upon prior to a purchase of any magnitude.

CHAPTER XXVII

MOISTURE CONTENTS OF OILS

From our previous discussion of steam generation in the modern central station it was found that something over a thousand heat units are necessary to convert one pound of water at ordinary temperatures into saturated steam. When moisture appears in the oil used for heat generating purposes in the furnace it is evident, then, that large heat losses may thereby be involved. For, not only must this moisture be converted into saturated steam, but this steam itself must be superheated to the temperature of the outgoing chimney gases, thus dissipating energies that should go toward steam generation in the boiler.

FIG. 151. An electrically driven oil centrifuge.

In this centrifuge the four arms-two plain and two grad

uated-are caused to rotate by

electric power and the water

thus caused to separate from urement of the moisture pres

the oil. The consequent meas

ent is then easily ascertained.

Hence the water involved in fuel oil composition is a dead loss which should be avoided as far as possible. Settling tanks accomplish much in drawing off the water content, but when the water appears in the oil as an emulsion it is almost impossible to commercially segregate it from the oil. Since, then, all fuel oils contain a certain amount of moisture, the careful determination of its exact proportions often becomes an important problem in efficient steam engineering performance. Summary of Methods Employed in Determining the Moisture Content. There are ten methods by which the moisture content of oil can be ascertained with approximate accuracy. For detailed information on this subject the reader is referred to Technical Paper No. 25 of the United States Bureau of Mines entitled, "Methods for the Determination of Water in Petroleum and its Products." These methods may be briefly summarized as follows:

The moisture content of heavy oils and greases may be approximately ascertained by the loss of weight due to heating.

The moisture content of oil may be approximately obtained by diluting a sample with a sulphate and then causing separation by action of gravity. A diluent is to be avoided in this process, as inaccuracies are liable to be introduced.

Again by diluting with a solvent and separating the moisture content by means of a centrifuge, the moisture content is determined with a slightly greater degree of accuracy than by either of the above methods.

By treating a sample with calcium carbide, another convenient method is also arrived at, and its accuracy is approximately within 3 per cent. of the water percentage if care is observed. The sample, too, may be treated with sodium and a convenient and accurate method results.

A color comparator is sometimes used, but the method is only approximate, as is also the method of treating a sample with normal acids. The electrical treatment, on the other hand, is successful in breaking up an emulsion on a commercial scale, or reducing the water content of an oil to such a condition that it may be successfully treated in some other manner. An emulsion is a physical condition of the oil and water wherein the water is held in such intimate contact with the oil ingredients as not to be readily separated by gravity or other ordinary means.

Again, too, distilling a sample mixed with a non-miscible liquid proves accurate to 0.033 grams of water per 100 cc. of benzine and oil in the distillate.

The most reliable method, however, is that accomplished by directly distilling off the water. This method is convenient and accurate to about 0.003 grams of water in the distillate, if the water is cooled to about 35°F.

The Approximate Method of Treatment.-The method hinted at above wherein the sample is treated by a foreign agent will now be briefly set forth, since such a preliminary determination often proves sufficiently accurate for the issues involved.

The method here outlined is especially applicable for the lighter oils. A burette graduated into 200 divisions is filled to the 100 mark with gasoline, and the remaining 100 divisions with the oil, which should be slightly warmed before mixing. The two are then shaken together and any shrinkage below the 200 mark filled up with the oil. The mixture should then be allowed to stand in a warm place for 24 hr., during which the water and silt will settle to the bottom. Their percentage by volume can then

be correctly read on the burette divisions, and the percentage by weight calculated from the specific gravities.

Details Involved in Determination of Distillation.-Since the method of determination by distillation is to be recommended above all others, we shall now proceed to the details of its accomplishment. Stated in simple words, the method consists in heating a sample slightly above the boiling point of water but not so high as to cause the vaporizing of other ingredients of the oil. As a consequence, the water passes over and leaves waterfree oil in the sample.

FIG. 152. A Goetz attachment for water determination.

By attaching the pipe shown in the lower part of the figure to a faucet, sufficient power is obtained from the city main to cause the rapid rotation of the two arms shown in the figure. This high rotative speed, due to the centrifugal force developed, causes the separation of the moisture from the oil.

The Apparatus Involved and Preliminary Proceedings.-To quantitatively determine the moisture content the sample is placed in a copper vessel known as a still, which is about 4 in. in diameter and 6 in. high. The still is then placed in an asbestos hood through which a projecting stem connects to a condenser and a burette where the condensate is measured in a graduated tube. The can from which the sample is to be drawn is first immersed in a water bath with its cover released. After the water constituting the bath has been raised to a temperature of 150° to 170°F., the cover to the can is fastened tightly, and the can agitated for several minutes in order that any water that may have settled at the bottom may be thoroughly mixed with the oil. For successful agitation the sample can should not be filled more than two-thirds with oil. 100 cc. of oil sample, measured in a graduated jar, are now poured into the still. The exact measurement of the oil is difficult without experience, as froth

collects on the surface of the oil and tends to obscure any definite meniscus.

The jar is next washed with 50 cc. of benzol and 50 cc. of toluene. The washings are poured into the still. Since toluene has a tendency to absorb small quantities of water, accurate results may be interfered with if the toluene is not previously

FIG. 153. Still with hood used for water determination.

Many methods are utilized in determining the water content of oil. The simplest and most accurate method for fuel oil tests is that of distillation. In this method a sample of the oil is poured into a still and raised sufficiently in temperature to evaporate the water and not the ingredients of the oil. By condensing the moisture and ascertaining its proportions the moisture content is easily ascertained.

saturated. In order to avoid such a possibility when opening a fresh bottle, 5 to 10 cc. of water should be added. The presence of water in the bottom of the bottle shows that the toluene is