Page images
PDF
EPUB

A PROPOSED METHOD FOR TESTING OILS AND RECLAIMING USED GAS ENGINE OILS.

By LIEUT. COMMANDER Z. W. WICKS, U. S. N., MEMBER, AND LIEUTENANT R. F. MCCALL, U. S. N.

In recent years some attention has been paid to recovery of used gas engine oils. In most instances these efforts have been put forth by individuals on a small scale. These attempts have been successful in that oil has been recovered, but they have paralleled to a great extent present methods of refining lubricating oils, with all of the attendant difficulties as well as a few additional ones.

Present practice in oil refining consists of distillation at atmospheric pressure, using a steam jet in the still to reduce the partial pressure of the oil vapor. This will reduce the boiling temperature below that obtained when distilling without steam, but the reduction is not sufficient to totally eliminate disintegration of the oils. As a result, the distillate must be "sweetened" to remove the undesirable products formed during exposure to the high temperature. Practice today includes the use of sulphuric acid and caustic soda consecutively. The resultant refined product shows no indications, according to the usual tests, that the oil is anything but 100 per cent perfect.

Conditions at the Helium Production Plant were such that an investigation of the oil situation was necessary. Consumption of oil in fourteen (14) natural gas 250 H.P., BruceMacbeth, engines was high and the used oil showed an unusually high percentage of soft carbonaceous residue. In an effort to determine the real character of the oil used, samples were distilled under 1/10 of a millimeter mercury pressure, using ordinary laboratory equipment. Sixteen different oils were dis

tilled, and in each case a curve was plotted to temperature and per cent distilled as coordinates. These oils varied from very light engine oil to heavy aviation oil, some were used and some unused.

This course of investigation revealed, among other things, the fact that an oil sold as a 900 second viscosity oil was composed of fractions registering from below 100 seconds to over 4000 seconds, and that the residue was pitchy in nature. Curiously enough, this oil passed standard acceptance tests. Out of the entire lot of oils tested, only two showed anything approaching uniformity of composition.

These oils when distilled under a vacuum, sensibly a perfect one, showed absolutely no color change over a period of several months exposure to light and air. Identical oils distilled at atmospheric pressure were clear yellow when distilled, but blackened over night. This difference is attributed to decomposition at the higher temperature. What happens to this unstable material upon treatment with acid and alkali? It is not removed with the acid at all. Olefines form saturation products with the acid. Diolefines polymerize to a tar. Under the action of heat, oil will decompose first, which has been in its present molecular condition for ages or an unsaturated hydrocarbon compounded with sulphuric acid several months old. It is the opinion of the writer, based on results of the oils in use and under test, that the heavy sludge formed in use is due to the breaking down of the "purification products." This is supported by the positive data showing unusual consumption of oil every time a barrel is emptied into the system, containing approximately 450 gallons, indicating the presence of some portion of the oil which can not withstand the high temperatures of natural gas engines.

It is the belief of the writer that a vacuum distillation would be most effective in testing an oil for acceptance. It will show the presence of gasoline, kerosene, etc., on the light end, and heavy cylinder stocks and tar on the heavy end. Above all,

these are shown quantitatively and more satisfactorily than with a flash and fire test.

It is possible to mix undesirable light oils with undesirable heavy oils and obtain an oil which will pass acceptance tests. The answer to that subterfuge is a more rigid analysis, and vacuum distillation will effect it. The difference between a poor and a good oil is apparent by the curves shown.

The apparatus used in obtaining the data for these curves consisted of a 250 cubic centimeter distillation flask, a Liebig condenser, graduated receiver and a laboratory sized vacuum pump. The suction of the pump was connected to the graduated receiver, a vacuum of about one-tenth of a millimeter absolute pressure was pulled through the whole apparatus. The condensed vapors were collected in the receiver and all noncondensible vapors and gases were drawn out through the pump. A copper wire gauze in the form of a rolled sheet, extending from the bottom of the flask up to and beyond the delivery tube, was found indispensible to prevent "puking." Distillation was carried out otherwise in accordance with A.S.T.M. standards.

The data for the temperature-percentage distilled curves was obtained from a 100 cubic centimeter sample placed in the distillation flask. The temperatures recorded were obtained from a thermometer set in the flask with the bulb in line with the

delivery tube. The oil was heated slowly and evenly and the temperature when the first drop of condensed vapor occurred was recorded. The temperature for each two cubic centimeters up to ten was recorded, then for every three cubic centimeters up to twenty-five and for every five cubic centimeters to the end of the distillation. From this data the curves were plotted and these curves used to determine the vapor temperature at which to make the fractionation cuts on larger samples.

The fractionation of large samples was made in the same apparatus with the exception that a three thousand cubic centi

meter distillation flask was substituted for the small one. A two liter sample was used and the various fractions collected in different receivers. The vapor temperature at the end of each fraction was recorded. The viscosity of each fraction was determined in a standard Saybolt viscosimeter at 100 degrees F., 130 degrees F., and 210 degrees F. These viscosities were plotted against the end temperature of each fraction. These curves clearly indicate whether or not the oil is a close cut one or a wide cut. In a paraffin base oil a wide viscosity range means a compounded oil, in the asphaltic base oils a wide range can mean either a compounded oil or one which is obtained over a wide distillation of the crude from which it is produced. There is every reason to believe that of two oils the same in all respects except the viscosity range, the one having the narrowest range is the better lubricant. The writer is also of the opinion, that with a narrow viscosity range oil, a much lower viscosity can be used than with an oil having a wide viscosity range. It is also more than likely that there will be much less carbonaceous sludge formed in an internal combustion engine by an oil which has the low and high ends removed than by an oil in which the dilution of the low boiling fraction is compensated for by the addition of heavy ends or cylinder stock.

Oil No. 2 was ordered as a 900 second oil, but was actually 1010 seconds. The distillation curves show a temperature range in excess of 200 degrees F. The family of curves for this oil (Figure 1) was obtained by cutting a 2000 cubic centimeter sample in six equal fractions and then plotting the viscosity against the end point of each cut. The poor quality and lack of homogeneity is shown by the wide divergence of the end curves.

Oil No. 8 shows an oil unusually perfect in character, very low in low ends and only a moderate amount of heavy ends (Figure 2). Note the approach to a family of vertical lines which would be the case in a perfect oil. It is to be appre

[ocr errors][merged small][merged small][merged small][merged small][merged small]

END TEMPERATURE OF CUT, (CURVES 1,243) - TEMPERATURE,(CURVE 4)

[merged small][merged small][merged small][ocr errors][merged small][merged small]
[merged small][ocr errors][merged small][merged small][merged small]
[blocks in formation]

Ο

70

[blocks in formation]
« PreviousContinue »