A STUDY OF THE GARBON DEPOSITS FOUND IN THE CYLINDERS OF INTERNAL COMBUSTION ENGINES GLENN HOWE JOSEPH THESIS FOR THE DEGREE OF BACHELOR OF SCIENCE IN ' CHEMICAL ENGINEERING COLLEGE OF LIBERAL ARTS AND SCIENCES UNIVERSITY OF ILLINOIS 1922 ' / 92 2, J 77 UNIVERSITY OF ILLINOIS — .May 2& 1922. THIS IS TO CERTIFY THAT THE THESIS PREPARED UNDER MY SUPERVISION BY G on _ H q w e _ J o s eph ENTITLED A _ STUDY 0 F _THE_ CARBON _DEPOS I_TS_ THAT _ARE_ POUND IN _ _ t _HE _ CYLINDERS _ OJF_ 1 HTPRNAL _ COMBUST ION_ ENG IKES # IS APPROVED BY ME AS FULFILLING THIS PART OF THE REQUIREMENTS FOR THE DEGREE OF BACHELOR. _OP_S_CIENCE IN CHEMICAL ENGINEERING ACTING HEAD OF DEPARTMENT OF _CHMXSTRY Tiie writer desires sincerely to express his appreciation of the aid and encouragement given in connection with this thesis hy Dr. T.E. Layng. Digitized by the Internet Archive in 2016 https://archive.org/details/studyofcarbondepOOjose TABLE OF CONTENTS page 1. INTRODUCTION 1 2. HISTORICAL 4 3. THEORETICAL 7 4. EXPERIMENTAL... (a) The Sample.,.....,..,.....,. 11 (h) Loss at 105° C 11 (c) Proximate Analysis. ........ . 13 (d) Ultimate Analysis 17 (e) Acidity 18 (f) Variation of Results........ 18 5. DISCUSSION OF RESULTS 19 6 . SUMMARY 24 7. BIBLIOGRAPHY 25 A STUDY OF THE CARBON DEPOSITS FOUND IN THEJ CYLINDERS OF INTERNAL COMBUSTION ENGINES I . INTRODUCTION The gasoline engine has been more than a score of years in reaching its present state of development. From being a novelty, it has grown to be a necessity. Automobiles have increased from 1,000,000 in 1912, to 9,000,000 in 1920 --over one million yearly increase. For this nine fold increase in automobiles, there has been only a two fold increase in the production of crude oil.'^ The investigators of the United States Geological Survey, after very exhaustive tests, 2 say that in 1934, the 6,000,000,000 barrels of crude oil now underlying the United States, willbe exhausted, based on the 1920 rate of production. This increase since 1909 may be ex- . 1 pressed as : Crude oil 140$ Gasoline 800$ Automobiles 2570$ All owners of gasoline engines, whether in a vehicle or of the stationary type, have experienced trouble with the formation of a deposit in the engine, commonly called a "Car- bon deposit." This deposit is evidently a hinderance to the smooth and economical running of the engine, and hence must be taken into consideration when efforts are being taken to prolong our present supply of fuel. "The enormous number of internal combustion engines in use, particularly in automobiles, trucks, and motor boats, , . : ■ • 2 . has been the means of giving a new and special significance to the word "darbon” and of making it a household word. ”Car- ’ ton" has come to mean to most people, the deposit which forms in the cylinders of these engines. There is perhaps no greater trouble maker for our engines than this so-called carbon deposit. It is understood by the automobile owner that he must have it removed from his motor every six months or so, if he wants to prevent his motor from getting extravagant. Power will decrease and the fuel consumption will increase. There will be mis-firing due to the sooty spark plugs. Premature firing of the charge sometimes results from the incandescent scales and the motorist’s ’’knock” occurs. With all these common troubles due to it, the guilty little deposit goes on through the years, with no one inter- fering. Many efforts have been spent in trying to prevent it-- sfforts based mainly on misinformation. For instance, some people advocate screens for the intake manifold to prevent road dust from being drawn into the cylinders, believing it to be the prime factor in the formation of the deposits and to constitute a greater portion of them.'* It is certain that no efforts from an intelligent stand- point can ever be made to eliminate or remove these deposits, without first knowing their composition. Knowing the nature of the compounds that make up the deposit, it is possible to suppose a satisfactory remover for it could be prepared. Of course the ideal way to remove the deposits would be to cause complete combustion of all carbon in the fuel and to prevent 3 the oil from being carbonized. In reality these deposits are not pure carbon but are only about three- fourths pure carbon. The remainder is made up of hydrocarbons and metallic oxides, from the engine. Some silica gets in through the intake from the road dust and can be found in the deposits. The composition varies from car to car, for different parts of the cylinder of the same car,, and from season to season. Sometimes deposits will be found that are asphaltic in appearance and others will be hard and brittle. The carbon in these deposits is evidence that there is waste in the combustion chamber. Carbon is fuel' — it could be burned to give power, and yet it is found to be about three-fourths of the deposit. It is the most important item in the composition of the deposits, acting as an indicator of incomplete combustion of the fuel. The iron content is an important factor, showing the amount of solution or wear of the metal parts of the engine. The cylinder walls and the piston rings are attacked and leakage of cylinder oil is the result. In cold weather, and in starting when the motor isoold, the raw gasoline that is allowed to enter the cylinder and that is not decomposed, ♦ flows down the cylinder wall 3 and dilutes the crankcase oil. Hard crusty deposits are formed around the valves and prevent the proper seating, causing losses in compression. V ) . » . . . ' 4 * II. HISTORICAL There have been carbon deposits and carbon troubles, no doubt, as long as there have been engines using gasoline. Twenty-five years ago when there were only four automobiles in the United StateB there was not much use publishing articles concerning these troubles. It has been only since the motor vehicle has become more common and is being used so widely, that we hear of the carbon trouble. 4 In 1918 the Blatt-Washburn Company had some work done on carbon deposits. They stated that the carbon deposits were not pure carbon, but contained only percents varying from 5 to 75. The remainder was said to consist of variable per- centages of metallic oxides and inactive earthy material. There was some solid compound that they said was an asphaltic material depending on the type of oil used. They did not make a very thorough study of the carbon and did not give any specific data. Another authority^ states "that a chemical analysis of the cylinder incrustations from the combustion space of an automobile engine, especially if the deposit has been formed during the summer months when the roads are dusty, will very likely show a considerable percentage of silicates which have been introduced in the form of road dust.” § Mr. Kettering, . T. A. Boyd, and Mr. Midgley of the General Motors Research Corporation, have contributed some to the knowledge of this deposit. They divide the deposit into four distinct divisions: free carbon, a hydrocarbon, a material containing oxygen, and a varying amount of ash. • • . ' . . „ > • * * . , . * # • 5 . This material containing oxygen is believed by them to be the binding compound. They show how alkaline liquids disin- tegrate the deposits, and how this must be due to the acid character of the compound. They have never given any data as to the further division of the deposit for publication, if they have made any more determinations on it* Dr. C. E. Waters and Mr. W. S. James of the United States Bureau of Standards, have probably done more work on this subject than anyone else. They have issued a bulletin on “The Carbonization of Lubricating Oils" in which they dis- cuss some of the carbon problems. Practically all the work that they have done is based on the theory that the deposits are results of the lubricating oil. They have endeavored to devise tests that will give indications in the laboratory of how the oil will act in actual use toward the deposition of carbon. As a result of such endeavors, the Waters carbonization test has been devised, and also the Conradson "carbon residue" test. There is no place where they have stated any definite data as to the composition of the carbon residue. In Europe, the only published account of analytical 7 work on the deposits was made by Mr. J. Marcusson. He divides the deposits into four divisions: an oily portion sol- uble in benzene, a brittle asphaltic constitutent insoluble in benzene, a coaly portion, and a mineral ash. By using a very large amount of solvent he obtained a portion of the asphaltic material, and from it extracted "asphal to genic acids or anhydrides and found the residue to consist of car- . . * 6 benes and carboids. M Prom these three published accounts, it is not pos- sible to gather much specific data concerning the chemistry of the deposits. They are all general in their statements and more or less theoretical — none giving any analytical data. . . * 7 . III. THEORETICAL It can be seen from the discussion of the formation of carbon deposits that appeared in the work of Dr. Waters and Mr. James, that there are two theories for the cause of car- bon deposition, both based on the supposition that it is caused from the lubricating oil. The oxidation of the oil to form asphaltic matter, which hardens after collecting metal- lic oxides and dust, is one way the deposits are explained. The other theory is based on the fact that after oils are subjected to destructive distillation, or are cracked, there 14 is a deposit of actual carbon. One author states that carbon deposits are similar to coal, and suggests Professor Parr’s method of extraction by solvents to bear out his statements. It is very hard to collect data to support either theory, due to the difficulty of duplicating the exact condi- tions that exist within the cylinders during the combustion. There are so many different makes of engines, each with dif- ferent conditions as to fuel used, oil used, piston rings, pressure during explosion, carburetor setting and speed. These conditions can only be standardized on a block test, and then it will be impossible to get actual conditions such as road dust and temperature. Q One well known authority states, "The deposit is by no means due to the lubricating oil alone, but comes partly from residual products in the fuels and its formation is probably controlled in part by the interaction between the oil . . . 4 ; ' • * . , ' . . •) i • „ • ’ '■ • . . i. i. -j :: .:V 8 and the fuel residues.” Those supporting the oxidation theory of carbon formation show the effects of sunlight on 3 the oil, impurities in the oil and catalyzers of oxidation. Oils in the sunlight will oxidize causing a precipitate to settle. This precipitate is an acid substance containing carbon, oxygen and hydrogen. There are some "resinous” substances about which there is little known, that are pre- sent in the crude oils or are formed in the refining process, that are the most important of the catalyzers of oxidation of oils. By heating, they polymerize and form asphaltic mat- ter. Iron oxide and sulphur are very good catalyzers of oxidation also. Dust and dirt, when present in the oil, cut down the percent of asphalt formed by heating. Cracking of oils, or the breaking down of the heavy molecules, is favored by high temperatures and pressures. In the cylinders we have high temperatures and pressures. The thin films of oil exposed to the influences of the cylinder during the explosion easily "crack” and the volatile products that are formed dilute the oil or are burned. The heavy residues are left behind to build up the deposits. Asphaltenes 9 are found in carbon deposits. Holde showed that the amount of carbonization depended on the proportion of asphaltenes present in the oils. Compared to the very small amounts of lubricating oil that enter the combustion chambers, there are large amounts of gasoline vapor-air mixtures. A study of the conditions that are present in this vapor-air mixture in the operation of the . ' * * , . ■ . , ■ 9 engine would probably lead to the conclusion that the deposits are due, more from the gasoline than from the oil. The fundamental idea of all combustion practice is to furnish just enough air for the complete combustion of all the carbon. Where there is not enough air, there will be in- complete combustion and a resulting smoke of unburned carbon particles. The nearer the boiling point the fuel can be rais- ed, the more vaporization that will take place. In the engine the vaporization is increased by allowing the pressure to fall by the suction stroke of the piston. The charge of vapor and air is then put under pressure and ignited by the aid of an electric spark. The added pressure causes the temperature to raise and at the same time lower the temperature at which the charge will ignite. When the engine is first started, the temperature is not high enough to cause all the charge to ignite. The driver, on noticing that his engine is failing, entiches the mixture of gasoline and air by allowing less air to enter. The temperature will increase due to the smaller amount of air entering. As the temperature increases the viscosity of the gasoline decreases and the volume of the air increases. This causes a double concentration of gasoline vapor. As a result of this enrichment, there will not be enough air for complete combustion and the attendant smoke formation and carbon deposition will take place. As the temperature increases there will also be an increase in the decomposition of the hydrocarbons in the fuel. At a temperature of 1500°C, 0.036 seconds is the time required for the complete ‘ . . . . . . . , ' . . 10 decomposition of the hydrocarbons and a deposition of carbon. 1 ^ Mr. C. E. Sargent, 1 ^ a research engineer of Indiana- polis, has shown how the decomposition of hydrocarbons takes place even when the best carburetting devices known were used. He took down an engine and cleaned the combustion chamber per- fectly, reassembled it and drove the car slowly for five miles. The engine was again taken down and it was found that there was a smooth deposit of soot or carbon all over the inner surface of the chamber. Thermometers had been inserted in the most critical places during the test and it was found that there were no places where the temperature of the metal of the cylinder walls was over 150°C. At this tdmperature it is impossible to get any burning of the lubricating oil and a carbon deposit. Mr. Sargent also used as a fuel, fixed gases, as illuminating gas, and found after long runs, there was no carbon forming in the cylinders. In view of the contradictory opinions that are held concerning the formation of carbon in the cylinders of internal combustion engines, this thesis was undertaken in an attempt to show that the deposits were the result of incomplete com- bustion of the fuel and not results of lubricating oil pro- ducts. ■ V. 11 . IV. EXPERIMENTAL (a) The Sample The samples were collected from various garages in this city and in other cities. Garage men were supplied with bottles fitted with good stoppers. They saved the carbon they took from cars and were careful to keep it as free from foreign matter as they could. After several weeks the bottles were collected and the whole mass of material was carefully ground together and mixed well. This sample was then labelled "Number 1", and no more additions were made to it later on. Other samples were collected from time to time and were kept in the laboratory in tightly stoppered bottles. The results given in this thesis are from one of these samples. The variation ih composition of the different samples is shown in Table 9. b) Loss at 105°C. for one hour The sample was put in small moisture capsules, those used for moisture determinations in coal, and heated in an oven at a temperature of 105°C. for one hour. The loss was 3.68$ of the weight of the sample. In order to ascertain the material that was given off at these conditions, the following method was used: Fifteen (15) grams of the sample were put in a small Erlenmeyer flask fitted with two outlet tubes. In the flask was forced a slow stream of nitrogen, purified by passing through pyrogallol and then through CaClg. On the other side, the gases coming out of the flask, passed through a weighed U-tube filled with CaC^, through an ordinary cylindrical CaCl tube, and into a receiving bottle. The 6 * . . . . . . # < , . . J second drying tube prevented any moisture from the receiving aspirators from backing up into the weighed U-tube if the stream of nitrogen stopped. In heating the flask to 105°C, a beaker of glycerine-water solution boiling at this temperature was used. In making the determination, a liter or so of nitrogen was passed through the apparatus first in order to completely displace all the air, that is before the collecting bottle was attached. lOOcc of this gas, mainly nitrogen added, were analyz- ed in a modified Orsat apparatus, where the percentages of carbon dioxide, oxygen, ethylene, benzene, hydrogen, carbon monoxide, methane and ethane were determined. These percen- tages were then calculated to a nitrogen free basis so as to eliminate the low percentages due to the nitrogen added and to show the true composition of the gases that passed through the CaClg tube. The moisture collected was weighed and its percentage calculated. The results are shown in Table 1 and Table 2: TABLE 1. Analysis of material driven off at 105° C. 1 hour % original sample * % of the loss » i Moisture 1.51 41.0 Gases by wt. 2.17 59.0 lOO.O (Table 2 - see next page) 13 TABLE 2. Analysis of the gases from Table 1. Gas % by volume of the gas Oxygen 5.8 Benzene 5.8 CO 2 34.6 Paraffins 53.8 100 .0 (value of n - 1.03 (c) Proximate Analysis A proximate analysis was made showing the material to fall into three main groups, shown in Table 3. The ash is the term that is applied to the material that remained in a crucible after all the combustible matter had been driven off by ignit- ing over a blast lamp until no change in weight occurred. The dishes which were used were very shallow and wide. If deep dishes were used, like ordinary porcelain crucibles, it was found very difficult to completely oxidize all the carbon. By playing a small stream of oxygen over the crucible during the heating, it was possible to get checking figures comparing to the ash obtained in shallow dishes. The samples were usual- ly of one gram, although by carefully etirring with nichrome wires, and using oxygen, it was possible to get good results with samples as large as fifteen grams. TABLE 3. Proximate Analysis of Original Sample. Given off at 105°C. (1 hr.) 3.68# Ash 9.76 Combustible matter 86.56 TOW A further division of the sample was made by means 14 ef a Soxhlet Extraction apparatus. Samples from different automobiles, different localities, and from different seasons of the year, were found to contain varying amounts of oil. Some were hard and brittle, while others were very oily, or M wet M . These were treated in the ordinary Soxhlet with different solvents, benzene, toluene, carbon-tetrachloride, and pyridine. The percentage of extraction varied over a range from 19$ to 29$ with the average at 21.00$. TABLE 4. Division of Sample by Extraction, Extractable with Benzene 21.00$ Residue from extraction 79.00 166.66 The residue from the thimble was d(i|r|ed in an electric oven at a temperature sufficient to drive off all the sol- vent, for one hour. It was then treated to determine the com- position. Carbon was determined by the combustion method of Liebig, using lead chromate and copper oxide in the combus- tion tube, as sulphur was thought to be present. Hydrogen was also found by this method, absorption in 35$ KOH. Sulphur was determined by the use of the Parr Peroxide Bomb; at the same time checks were run on the carbon by treating the fusion in a Parr Total Carbon apparatus and later the sulphur precipitated from an acid solution by BaClg and weighed as BaSO^ . The ash was obtained in the manner stated for the original sample. The results for the analysis of the residue from the Soxhlet extraction are shown in Table 5. • ' ■ 15 TABLE 5. Analysis of the material remaining after washing with benzene. # original sample » % of the residue Carbon 60.13 76.11 Hydrogen 3.31 4.19 Sulphur 1.32 1.68 Oxygen 6.95 8.80 Ash 7.29 9.22 W.o6 106.06 The extract was distilled in a distilling flask to drive off the solvent and enable the extracted matter to be analyzed. The residue from the extraction was black and had the appearance of asphalt. It was placed in a boat and an analysis on the combustion furnace was run. This gave the percentages of carbon and hydrogen directly. The boat was weighed after the determination to get the percentage of the little red powder that remained. The powder was found to be iron oxide by qualitative tests. It had evidently been in the colloidal condition in the extract. Sulphur that pro- bably was present, was oxidized to SOg and absorbed by the lead chromate. The results are given in Table 6. TABLE 6. Analysis of the material extracted with benzene. % original sample • % of the extract Carbon 11,46 54.57 Hydrogen 1.94 9.25 Sulphur .06 .28 Iron .05 .27 Oxygen 7.49 35.63 21.00 100. 00 16 The percentages of carbon, hydrogen and oxygen as shown, give proportions for a compound of this type, (C 5 H g 0 2 ) n The percentage of sulphur as shown, was calculated in the following manner: The sulphur originally present in the sample had to go into the residue or into the extract. There was a percentage of 1.68 sulphur found in the residue, which figured back tothe basis of the original sample by multiplying the percent resid^u j 79.00, by the sulphur percent, 1.68, obtaining the amount that was not in the residue, which equals 0.06 percent of the total sulphur on the original basis. The original sulphur will be shown later to be 1.38$. This difference of 0.06 percent was figured back on the basis of the extract being 100# by dividing it by the percent extract, 21.00, this produced the amount shown above, 0.28 percent that must be in the matter extracted by the benzene. The ash obtained by the ashing dish method, was anal- yzed. It had the appearance of iron oxide, red powder, very fine and heavy* This material was put into solution by us- ing strong hydrochloric acid. The iron was determined by the permanganate method and by the gravimetric method of weighing as the oxide after precipitating as the hydroxide. For the permanganate method, the ash was treated with hydro- chloric acid and stannous chloride to hasten and to get a better solution of the oxide. In all cases, it was necessary to boil the acid. Silica was obtained by filtering off the residue that remained after all the iron had dissolved and washing carefully. The precipitate was ignited in a small 17 . crucible to constant weight and then hydrofluoric and sul- furic acids were added and the crucible carefully heated again. The loss in weight was found to be the percent silica. All the residue in the crucible disappeared by this treat- ment showing that there was no other elements present. The analysis of the ash gave values as shown in Table 7. TABLE 7. Analysis of the Ash % original sample » . \ % of the aBh Iron Oxide 8.19 84.01 Silica 1.54 15.88 5773 A qualitative analysis of the sample was made at the first. It showed that there were no metals present except iron. In one sample there was a test for nickel, probably that came from the nickel-steel valves in the engine. Silica, sulphur, iron, carbon, sulphuric acid and carbonic acid were found, (d) Ultimate Analysis . Starting with the original sample the percentages of the elements were determined, irrespective of the com- pounds in which they occurred. The sample was dried at 105°C. and the dried material run in a Parr Eomb for the estimation of the carbon by the Parr T 0 tal carbon apparatus and sulphur by gravimetric means. Samples were treated by the combus- tion method to get the carbon and hydrogen, also. The silica was determined in the ash and the percentages calculated . , , . » . 18 . back to the original basis by dividing by the percent ash. This gives for the ultimate analysis, the values as shown in Table 8. TABLE 8. Ultimate analysis of the original sample: Total carbon . . TSTfif? Hydrogen 4.57 Iron 5.74 Sulphur 1.38 Silica 1.55 Oxygen (diff.) 14.56 I00.06 1 (e) Acidity of Sample Several grams of the sample, not dried, were placed in a flask and digested with 200cc of distilled water at a tempera- ture under the boiling point of water, in order not to drive off any carbonic acid. After several hours, the solution was fil- tered and the precipitate washed until free from any acid that could be present. The solution was titrated by phenolphthalein in the cold. There was a distinctly acid property apparent. If the material was boiled several hours in a reflux, the acidity was lowered. Methyl orange did not give as good values as Phenolphthalein. The acidity was expressed as the number of milligrams of KOH necessary to neutralize one gram of the sample. The KOH used was 0.0787 normal. The acid numbers var- ied on different samples from 7.9 to 16.4 TABLE 9. Showing the range of variation in composition for dif- ferent samples. Total carbon 68.50$ Hydrogen 4.2 Iron 3.8 Sulphur 45 Silica 50 TT/OCT 5.5 6.8 3.50 2.00 V. DISCUSSION OF RESULTS 19 . Samples from small cars where poorly refined oils and fuels had been used, could not be distinguished from samples taken from the cylinders of engines where only the best of oil and gasoline had been used. It has been supposed that oils with low viscosity would give deposits more readily than oils with high viscosities, but no difference could be found in the deposits formed by the two oils. This adds to 12 the proof of F. C. Robinson's statement j "The amount of carbon averages, broadly, about the same for low viscosity and high viscosity oils and about the same for Texas field oil and Pennsylvania oil; and the consistency of the four samples of carbon is identical.” This statement was later confirmed by the United States Bureau of Standards, where it was found that, "The layers of carbon which formed on the piston were so much alike in their properties and appearance 3 that they gave no hint as to the nature of the oil used.” Both these statements were made from data based more on the physical property and general appearance than on the chemical constitution. Data obtained by using methods described in this thesis, proves from a chemical standpoint that there is no relation between the carbon deposits and the type of oil used in lubrication. The outstanding feature of the results of this thesis is the fact that all lubricating oils give identical carbon deposits, chemically. Of course, there are samples where the oil is held in combination more closely than in others caused by having too much gasoline and oil in the cylinders. After . . . , . . . . . 20 , this extra oil had been washed out, the residue was of nearly a fixed composition. If the oil is responsible for the deposits, how can this be explained? It was shown how the temperatures of the combustion chamber were so low that no oil could be carbonized, i.e. oil in contact with the cylinder walls, where the temperature was never over 150°C. It was shown how carbon results from improper combustion of the fuel, and that incomplete combustion took place in the engine. Experiments were described where fixed gases were used as fuel and where no carbon deposition resulted. These facts stand out as clear evidence that the cause of the carbon deposits lies in the incomplete combustion of the gasoline and not in the lubricating oil. Mechanical engineers will have to give better means of obtaining good combustion or the chemist will have to regulate the reactions taking place in the cylinders and so control the decomposition pro- cess. Iron was found to be present in amounts running up to six percent. This is rather hi^ier than could be expected from the actual wearing of the metal parts of the engine. The conditions are such in the combustion chamber that there could be formed the carbide of iron. Around the exhaust 9 valves the temperature is over the minimum required for the formation of the compound Fe^C, and it is here that the carbide probably first forms. It is extremely hard and when it gets on the valves, it acts as an abrasive and causes excess wear of the valves. , * . . . . . . . . . . - 21 . The silica content of the samples varied very much. Some samples with only 0.50$ were found while some went as high as 2.00$. The content of silica varied very little for samples taken from cars that had been in the city all the time. It is the constitutent that was expected to be found high, after reading some of the discussions of the carbon 4 question in the popular magazines. One author asserts, "There are much higher percentages of road dust than oil residues." This compound has been placed in the deposits by supposition by many of the authors. For instance, one 5 author states, "As long as the tractor engineer realizes that dust is one of the most destructive elements to engines, IT FOLLOWS that such road dust (silicates) WILL be found in the carbon deposits." The percentages of sulphur are very significant. In one case the sulphur ran up to 3.50$. This is surprising from the fact that there is such a low percentage of sulphur in the oil and gasoline. The men at the Bureau of Standards 1 2 report having found as high as 4.00$ sulphur in some deposits. It has been shown^ 4 that where oils have a low carbonization value by the Waters test, the sulphur content is low, falling under 0.10$. These facts show that the sulphur has a strong tendency to pile up in the deposits, rather than to pass out in the exhaust as might be expected. The relation of sulphur and the formation of asphalt in nature, may bear some relation to this high sulphur content. The acid values of the deposits are very important. There are two causes apparent. The sulphur in the form of a • • . . . , • . • , ♦ . c t . .. . 22 . soluble sulfate is able to cause a greater part of the acid- ity, but the CO must be the cause of some of the acidity. This is shown by the fact that the samples when boiled with water for several hours showed no greater, and in most cases, less acidity than those digested with warm water for a short t ime . These acid values will be of use in calculating the composition of a remover for carbon. Knowing the acid num- bers of the average samples, it is possible to determine the proper alkalinity to make the solvent. The relations of the elements in the binding compound are shown to be (C 5 Hg0r,) n from the percentages of the elements found in the extract from the Soxhlet extraction. It is this type of hydrocarbon that exerts the binding influence upon the metallic particles and the soot. Such a compound was predicted by Mr. T. A. Boyd g and Mr. C. W. Adams of the General Motors Corporation. One striking feature of the results is the fact that no metals of the alkaline earth family were present. It would be supposed that when silica is found in the deposits, some traces of calcium or like metal vtruld be present, but none w e it we* found. There are two ways of explaining this absence; one is to suppose that there was no metal with the silica; the other, that the silica entered as some silicate of iron. The analysis of the gases given off at 105°C. presents some interesting facts. The benzene found probably results from decomposition of the fuels. The paraffins can not be divided into percents of ethane and methane as is usually done . . . . ' . . : ; „ . . 23 in flue gas analysis. It is not known just what they are, "but it is to be expected that they are some higher homologues of the series C H . such as hexane, heptane, or octane, n 2n 4- 2, The absence of unsaturated compounds and carbon monoxide is contrary to what might be expected. The oxygen present is due to the adsorption by the carbon. The moisture content is evidently due to the hydroscopic character of the deposits. It could not be present in the original sample before taking from the engine, but has been taken up by the material while the garage man was finishing the removing operation, or while the bottle was uncorked in the garage. « . •• . . « . 24 VI .SUMMARY 1. Ultimate and proximate analyses were given for the deposit* 2. It was shown that there was less road dust present than was commonly supposed to he there. 3. The results for the percentages of carbon, hydrogen and ash were shown to agree with those obtained by other work- ers using samples from different states. 4. The sulphur content was shown to be higher than was to be expected, and its significance discussed. 5. The compound was shown to be distinctly acid and the sources of this acidity were given. 6. The importance of knowing the acidity was given. 7. It was shown that no relation existed between the character of the deposit and the type of clyinder oil used. 8. The absence of alkali metals and alkaline- earth metals was proven. 9. The presence of paraffin hydrocarbons and moisture was discussed. 10. The matter of combustion in engine cylinders was dis- cussed and proof was given that the carbon deposits were the result of incomplete combustion rather than oil residues. . . V . . . . . , * . ► . 25 . VII. BIBLIOGRAPHY 1. Boyd, T. A. J. Ind. Eng. Chem. 13» 836 (1921) , 2. White, David J. Soc. Automotive Eng. 12, 561 (1919) 3. U.S. Bureau of Standards, Bulletin 99 "Carbonization of Lubricating Oils. " (Nov. 1920.) 4. Ikert, B. M. Motor Age < 34, 24-5 (July 18, 1921) 5. Ikert, B. M. Unpublished communication (April 11, 1922) 6. Kettering, C. W. Unpublished communication (Nov. 5, 1921) 7. Marcusson, J. J. Soc. ^ham. Ind. 40. 289A (1921) 8. Dickinson, H. C. (National Research Council) Unpublished communication (Dec. 2, 1921) 9. Garner, F. H. Gas World 74 84-5 (1921) 10. Katz, S. H. U.S. Bureau of Mines. Technical Paper 183 (1918) 11. Sargent, C. E. Private Conversation (Dec. 4, 1921) 12. Robinson, F. C. Atlantic Lubricator 3, p.8 (Feb. 1920) 13. Boyd, T. A. Unpublished communication (April 6, 1922) 14. Waters, C. E. TJ.S. Bureau of Standards. Unpublished com- munication. (Dec. 8, 1921)