PETROLEUM FUELS WHEELER G. LOVELL Ethyl Corp., Defroif, Mich.
A
commonplace but nevertheless optimistic observation in the field of technology i s that we do the best with what we have. By "what we have" is meant our collective knowledge of the mass and energy relationships of materials. But to the extent that we may expect to learn more about these relationships, so may the technology improve to produce more human satisfactions from material per manhour expended in working on the material. Consequently,
it is worth while to direct attention to what we should learn more about with specific techno!ogical objectives in mind. The specific objective of making better use of petroleum requires definition o f the limiting factors that prevent us from doing better than we presently do. And so some classification o f the limiting factors may be made b y considering the methods now used to get power or useful work from petroleum.
OST of the petroleum obtained from the ground is used t o produce power Because vie have not learned horn to produce power directly from petroleum we use a somewhat round-about way and first burn the petroleum with air. In fact, almost all the petroleum produced is oxidized or combined with the oxygen of the air t o form hot products of combustion. These hot products are used in three general ways:
All these methods involve the liberation of chemical energy in the form of heat. The amount of heat liberated from a definite amount of petroleum is something we think we cannot do much about, because, if there is enough oxygen present, the amount of heat depends on the amount and kind of petroleum or hydrocarbons. This heat of combustion is considered t o be a fixed thing, and as long as the atoms that make up the hydrocarbon molecules are riot disturbed, not much can be done about changing it. Consequently, only the working fluid, or the mechanisms used to convert its heat into work, or the mechanisms for getting its heat into some other material, can be improved
anisms or engines must be built to stand the higher temperatures resulting from less nitrogen in the air so as to produce more power. In fact, engine temperatures are now close to current mechanical, metallurgical, and lubrication limits. Consequently, t o make much of a change in a working fluid, there must be simultaneously great advances in the separation of nitrogen from air and also in engine design, metallurgy, and lubrication. As a result, the practical potentialities of changing the working fluid do not seem very great in the immediate future. As long as the working fluids or the products of combustion are used in a heat engine, Tye are up against the second la% of thermodynamics a t least as a statistical matter. This law says essentially that the availability of heat energy in the medium is greatest when the relative temperature drop is the greatest. Therefore, it is desirable to devise mechanisms to heat the fluid as hot aspossible in order to get the most work out of it. A reciprocating engine is a fairly simple or cheap mechanism for doing this; it compresses and then expands a gaseous charge Rith a reciprocating piston. In the Diesel-cycle engine the fuel is injected into the hot compressed air and it burns. In the Ottocycle engine the fuel and air are mixed before compression and ignited with a spark. The two may sometimes be combined to have fuel injection and spark ignition. As yet, however, no other practical cycle has been designed, and t>heengines and cycles now proved will probably continue in use for some time to come.
The Working Fluid
Spark Ignition Engines
As a result, the limiting factors of the production of heat and power from petroleum constituents and products are concerned with the relationships between the resulting working fluids and the mechanisms. For this reason it is possibly simplest to consider the currently used mechanisms, what their limiting factors may be, and what can be done about the limitations now believed to be insurmountable natural laws. The products of combustion of petroleum constituents and air constitute the working fluid in an mgine. hIost of it is nitrogen from the air, and removing some of the nitrogen from the air would allow use of higher temperaturesin anenginewithgreater efficiency and also the use of smaller engines. However, this separation cannot be accomplished economically and surely not with small equipment. The equipment is important because most of the installed horsepower of the \Torld is in automobiles and trucks and is portable. If the nitrogen is removed from the air in a central station, then in addition t o the fuel, about three times its weight of oxygen must be carried in the vehicle. If the nitrogen is removed from the air in the moving vehicle, then the separation apparatus must be very portable and light. Furthermore, mech-
There are two main avenues of approach toward improved engine performance. The first is to squeeze the cylinder charge of fuel and air mixture by increasing the engine compression ratio; this results in increased thermal efficiency. The second is by supercharging the fuel-air mixture into the cylinder; this results in increased power output and the same thermal efficiency. Both of these methods encounter the important limitation called fuel knock. Knock limits both the compression ratio and the amount of supercharge. -4s the charge of gasoline vapor and air is compressed by the piston in the cylinder, chemical reactions occur in the mixture. Then, after the mixture has been ignited by the spark plug, the expanding gases compress the unburned mixture still further during which time it undergoes additional reactions. With many gasoline hydrocarbons the reactions have proceeded to a critical state in the last portion of the unburned charge (end-gas), and when it ignites it burns too rapidly. It knocks. The knock is thus a very definite limitation, because it is not practical mechanically to run an engine with any significant amount of knock. Attempts t o do something about it are worth while becawe if
M
1. As a working fluid in a mechanical device surh as an Ottocycle or Diesel-cycle engine or in a turbine. 2. As a working fluid, as in a jet, to take advantage of their temperature or pressure to push against the combustion chamber. 3. By transferring their heat to another working fluid, such as water or mercury, and using this latter fluid to run an engine or turbine.
1426
INDUSTRIAL AND ENGINEERING CHEMISTRY
Vol. 45, No. 7
knock could be completely eliminated present engine efficiencies would greatly increase. This is no small matter, for even a 10% increase in efficiency is like discovering whole new oil fields, decreasing engine size significantly, or extending petroleum reserves. There are a number of possibilities for lessening the limitation of knock. A good deal of thought has been given t o using the knock, but so far without much hope. Elimination of end-gas may mean modifying the cycle somewhat and putting the fuel in as fast as it burns, with a flame stationary with respect t o the engine cylinder but moving through swirling gas, as in the Texaco %ombustion process.” This ingenious mechanism is the result of straight-forward research, and there can be no doubt of its effectiveness for the complete elimination of knock. The other approach t o the same result of eliminating knock is by changing the fuel. Which method may be better can only be determined by experience and by new knowledge of how best t o go on both routes. Certainly, at present, modification of the cycle may be best used under operating conditions where fuel cost is high relative t o engine cost. But for most engine-propelled automobiles, the present engine and mechanism cost accounts for over three quarters of total operating cost. Other special attention may be given the end-gas by mechanical means. Keeping it cool, diluting it, or shortening its existence in the engine cycle are methods that have been used to reduce knock. These features of engine and combustion chamber design are sometimes termed “mechanical octane numbers.” They move the barrier of knock ahead, but they do not eliminate it fundamentally. So i t is worth while to look a t alternative methods for achieving the potentialities which can result from the elimination of knock. Knock can be eliminated by changing the molecular structure of the fuel. The importance of these changes is evident in the fact that the octane number scale of 0 t o 100 represents about a fivefold variation in the knack-limited power t o be obtained from a supercharged engine. The difference between the isomers, n-heptane and triptane (2,2,3-trimethylbutane),which have identical composition and the same heat of combustion within a third of a per cent, is so great that in an engine of varied knocklimited compression ratio the thermal efficiency may vary almost twofold. The relationship between molecular structure and incidence of knock has been worked out on an empirical basis for some 300 pure hydrocarbons by a n American Petroleum Institute Research project, so that firm fundamental data are available. The useful work t o be obtained from a gasoline depends not only on the heat content (which cannot be changed very much) but also on its molecular sbructure. because some structures can be used a t higher compression ratios without knock. This is due to reactions occurring in the end-gas ahead of the flame in the engine combustion chamber. If these reactions did not occur, there would be no knock. It is not the original fuel t h a t knocks, but some intermediate oxidation product of it. At present, knock in the end-gas can be controlled for the most part only by changing the original fuel or by the use of catalysts in the endgas. Changing the original fuel offers great potentialities in extracting mechanical energy from the fuel. It is established that changing the molecular structure of the fuel without changing significantly its heat content, volatility, or other important secondary properties can eliminate knock. B u t it is not known in general how to transform one definite molecular structure t o another by simple means. Perhaps this barrier can be eliminated by advancing our fundamental knowledge of hydrocarbon chemistry, but whether complete chemical knowledge will enable us to control the reactions in end-gas and completely eliminate knock cannot be anticipated. Certainly, however, such knowledge will not set us back. Knowledge of the physical control of these reactions will help also. Knowledge of how to influence the pressure and also the temperature of the reacting gas and that of the surrounding walls has resulted in great progress in reducing knock. Such progress goes by the name of “mechanical octane numbers,” as mentioned July 1953
%
previously. These gains are very real, but there is no evidence that they can go far enough in permitting higher compression ratios t o make us stop wanting better fuels, or more octane numbers, or more catalytic control of the chemical reactions of knock. Catalytic control of the knock reactions is gained by the use of so-called antiknocks metallic or nonmetallic, such as tetraethyllead or aromatic a d n e s which reduce knock when added t o t h e fuel in small amounts. The chemical mechanism of how antiknocks or their decomposition or oxidation products are effective can only be expressed in quite vague and general terms, but lead catalysts are used in various amounts in over 98% of t h e gasoline burned. If more were known about the chemical mechanism of antiknocks, they might be more effective, since it is known that their effectiveness varies severalfold with changes in fuel contaminants in the fuel, such as sulfur compounds, and engine operating conditions. Certainly new knowledge on this subject would seem t o offer great potentialities for practical improvement in the use of gasolines containing commercially known antiknocks or new ones. Another potential improvement in the use of fuels in spark ignition engines may be realized through the control of flame velocity. If knock-inducing reactions take place ahead of t h e flame there will be less time for them t o occur if the flame gets there sooner. So far we have not been able t o change flame velocity very much by reasonable addition agents or without gross fuel changes. We have been able t o influence i t importantly only by modifying physical factors such as turbulence. Of course, this reduces the knock, but too much turbulence results in too much heat loss to the walls, whereas an independent control of flame velocit would help out on knock and would also probably make a smalrbut important contribution t o engine thermal efficiency because of less heat transfer a t the walls. But judging by past experience, this is a lot to hope for in the immediate future. Secondary Properties of Fuels for Spark Ignition Engines There are hopes for improvement in engine fuel relationships based on changesin secondary fuel properties having to do not with the fundamentals of the cycle but rather with incidental relations. These may be classified as those due to incomplete burning, imperfectly adjusted physical properties, and destruction of the mechanism. Incomplete burning results in combustion products other than carbon dioxide and water. Other final products can be carbon monoxide and hydrogen. These are very undesirable because carbon monoxide is poisonous, and also because it represents potential heat in the gasoline which is not completely liberated. The reason “rich” mixtures are used or those without sufficient air is only partially because such slightly rich mixtures liberate more heat per unit volume of charge in the engine. They are used, with economic gain, in aircraft engines during take-off t o reduce knock and obtain more knock-free power. They are used in automobile engines sometimes because of imperfect mechanisms of carburetion and distribution t o the engine cylinders. However, the past record of improvement in the use of leaner mixtures in automobiles is outstanding, and this mechanical problem may be said t o be in good hands, with possibilities and probabilities for continued improvement, as indicated by the record of the past. Other results of incomplete burning are exhaust odor caused by aldehydes and other intermediate products of combustion caused by incomplete combustion of temporary very rich mixtures, nonfiring of a cylinder, or stratification of the fuel-air charge into localized volumes in the engine cylinder which are too rich or too lean to burn. Proper mechanical design can do much to eliminate this trouble. Then too, incomplete combustion of the fuel may lead to products which may deposit varnish on pistons or sludge in oil, and as we find out more about the constituents of gasoline responsible for this effect we may hope t o remove them and avoid this trouble. A most important result of incomplete combustion of the fuel and the oil and the antiknock catalyst is combustion chamber deposits. Unless the products of combustion of the fuel, of the
INDUSTRIAL AND ENGINEERING CHEMISTRY
1427
’
’
fuel additives, including antiknock compounds, of fuel impurities, and of the oil are scavenged from the Combustion space, they will remain there in many forms. Such deposits accumulate on the exposed engine parts (piston tops, combustion chamber walls, valves, spark plugs). Combustion chamber deposits are deleterious for a number of reasons. They interfere with normal heat transfer, thus increasing combustion chamber temperatures to a point a t which combustion or preignition may occur at the wrong time in the cycle. Deposits also increase the incidence of knock and reduce the amount of fuel-air mixture that can be inducted into the cylinder. thus tending t o reduce engine power output. Deposits which accumulate on the spark plug cause unsatisfactory operation by allowing the high voltage ignition pulses t o leak off to ground through the deposit rather than discharging across the spark gap. Many factors influence the rate at which deposits accumulate. Fuels, antiknock compounds, lubricating oils, and engine operating conditions are major factors in this respect However, the precise mechanism of deposit formation and the details of deposit composition are still not clearly understood. Here again the limitation is lack of knowledge and the potentialities for improvement are correspondingly great. The complete elimination of combustion chamber deposits would result in a large reduction in normal octane number requirement, the elimination of preignition due to deposits, and the elimination of spark plug troubles from this cause. The combustion of fuels containing tetraethyllead increases the weight of combustion chamber deposits but does not usually increase their knock-inducing effects. Under certain operating conditions the decomposition products of leaded fuels may have deleterious effects on exhaust valves and spark plugs which have not yet been completely overcome by mechanical design or metallurgical improvements. While halides are used with tetraethyllead to form more volatile lead halides, the resulting scavenging reactions are not completely effective. More knowledge is needed to improve lead scavenging or to develop an antiknock which would leave no ashes. A final important secondary relation of the fuels for use in engines lies in the ability of fuels to destroy the mechanism. Engines a-ear out; their rate of wearing out under some conditions is influenced by the presence of sulfur in the fuel, probably because of the eventual formation of condensed sulfuric acid by chemical mechanisms which are not presently understood. Here we are dealing with largely cmpirical observations and perhapa more knowledge may help us in avoiding such difficulties. Compression Ignition Engines
Turning now to the Diesel cycle which burns about 10% of our petroleum, another set of problems in engine fuel relationships or another set of places exists where more k n o ~ l e d g emay lead to improvement in fuel utilization. In the Diesel cycle, a charge of air is compressed in the engine cylinder and then the fuel is injected. The air is so hot because of the approximately adiabatic compression that the injected fuel ignites and burns. I n the spark ignition engine it is desirable to run a t high compression ratios, but this cannot be done because of a kind of ignition ahead of the flame. But in the compression ignition engine high compression ratios must be used in order to get the air hot enough to ignite the fuel in a reasonable time I n fact, the required compression ratios are possibly a little higher than the ideal, not from the standpoint of thermodynamic efficiency, but from the standpoint of mechanical and metallurgical design to withstand the high pressures. So, in the spark ignition engine a form of spontaneous ignition or knock is undesirable, but in the compression engine spontaneous ignition must occur. I n general, the combustion characteristics of fuels for spark ignition engines, which are measured by knock ratings or “octane
1428
numbers,” are about the opposite of combustion characteristics for compression ignition engines, which are measured and called “cetane numbers.” Usually, high cetane number in a fuel goes with low octane number, and vice versa. Generally, compact structural formulas for hydrocarbons go with low cetane numbers (or high octane numbers) and long chain formulas with high cetane numbers (low octane numbers). If the individual hydrocarbons of petroleum and its products could be separated into two classes of high and low cetane numbers, compression ignition engines could be operated on one, and the spark ignition engine on the other. The two engines would then run a t about the same compression ratios and about the same thermodynamic efficiencies. The choice between the two would depend upon the type of service and costs of manufacture. But we do not know how to make such a separation of hydrocarbons which is not based on simple physical properties but on combustion characteristics or molecular structure. A separation on the basis of combustion inevitably destroys the material; and a method of screening based on structure has not been devised. Perhaps in the future this can be done, even though it may mean processes of petroleuni refining which manufacture specific and fairly pure hydrocarbons of all fuels. Prcscnt fuels for compression ignition or Diesel engines are not perfect because of imperfections in their ignition. Better and faster ignition lvould be an improvement for many applications. The gains, however, from better ignition or higher cetane numbers are not comparable to those from higher octane numbers. Higher octane numbers offer potentialities of higher compression ratio. with better efficiency and more poner. Higher cetane numbers fundamentally enable low-er compression ratios to be used, with a loss of thermodynamic efficiency. And so, the gains of superior Diesel fuels are gains in the operation of engines and not those of fundamental thermodynamics. But the word “ignition,” as used here. covers a multitude of ignorance3 about how soon how much Vaporization and how fast the defined chemical reactions occur after injection of the fuel into the hot air. But ignition is greatly influenced by fuel composition and by the addition of almost catalytic amounts of material such as volatile organic nitrates. An early and definite ignition would ensure control of the q c l e as too much fuel could not be injected before burning. Also, if all the air is burned, the engine, which may sometimes be thought of as an air compressor, mill not be compressing unused air. The reason all the air is not used is that in the time available in the engine cycle it is impossible to ignite the fuel and mix it with the air well enough to burn all the air without having some pockets of too-rich mixtures which produce smoke. So the sorting out of all the simultaneous and consecutive matters of injection, dispersion, vaporization, ignition, and burning is a big job. But the reuards of doing it are in better control of the process. Potentially, this mould result in a cheaper and more reliable mechanism as a consequence of increased knowledge about the fundamentals of ignition. Secondary combustion properties of Diesel fuel of potential importance for improvement are thqse concerned TTith smoke, vihich limits the amount of fuel that can be burned in an engine cylinder, This limitation arises because the fuel does not get well mixed with the air because of mechanical or fuel vaporization limitations. Fuel properties, which are difficult to alter, may be a factor in this. Therefore, anticipated improvements in Diesel fuel-engine relations directed toward higher specific outputs will probablg be made in the mechanism of injection and atomization. Turbines and Jets
The power output from petroleum in internal combustion engines is almost all accounted for by reciprocating engines of the Otto or Diesel cycle. But a fundamentally simpler engine of great promise is the turbine or jet, which will without doubt in the future burn an increasing amount of petroleum products. How great an amount is possibly a matter of conjecture since it is the mechanism itself and its metallurgical composition which is now undergoing primary development at an increasingly rapid rate. The gas turbine has a major similarity t o the piston engine in that both produce useful power by means of the expansion of
INDUSTRIAL AND ENGINEERING CHEMISTRY
Vol. 45. No. 7
m
vr
heated air or combustion products. All gas turbines consist basically of an air compressor, a combustion section, and a turbine. The gas turbine functions by taking in atmospheric air and compressing it; fuel is then burned in the compressed air which expands through a turbine which drives the compressor. I n the turbojet the heated air and combustion products are expanded in the turbine so that only sufficient energy is extracted from the gases t o drive the compressor, the remaining energy being retained t o eject the gases in jet form and thus produce thrust. The combustion problem in the gas turbine or jet appears to be very simple, but this apparent simplicity is very deceptive. It would seem t h a t the problem of merely mixing fuel in air is vastly simpler than the combustion problem of the piston engine. The essential difficulty with combustion in the gas turbine is due to the fact that the fuel-air ratio is so low that if the fuel and air were uniformly mixed, the mixture would not ignite. The low fuel-air ratio is required in order to keep the temperature of the gases delivered t o the turbine a t a value t h a t the turbine wheel can tolerate. The present fundamental combustion problem of turbines in relation to petroleum utilization seems t o be how to obtain a large rate of heat evolution per unit volume and how to keep it under control. This means fundamentally obtaining more basic knowledge of the controlling factors in flame velocity and the importance of turbulence in the reactions of the burning gases. Sometimes jets and turbines become involved in all sorts of stationary, dynamic, and vibrating phenomena which lead to confusion. B u t judging by the large amount of research t h a t is going on, we may expect new knowledge of great potential value in the future. Here again, not only rapid combustion is desirable but also complete combustion or the absence of incomplete combustion products, which result in smoke and odor. As optimists, or as researchers, we may expect knowledge of controlling factors to permit the elimination of existing barriers t o progress, and better use can be made of petroleum in this application also.
Direct Heat
So far in considering petroleum uses for power, the products of combustion have been the working fluids, but about 30% of the petroleum burned is used for the production of heat which is trans-
ferred to other material, and a portio,i of this heat is transferred t o another working fluid, as in steam boilers. Here the problems may seem a little simpler because of extensive past developments. The limiting factor is not so much the volume needed for the combustion itself as it is heat transfer through walls to the other working fluid, wliich may be a matter of maintaining the youthfulness of the surfaces. Of paramount importance, consequently, are the reactions of the products of combustion on or with the surfaces of boiler tubes and refractories. We do know that, in addition to the chemical elements of carbon, hydrogen, oxygen, sulfur, and nitrogen, petroleum contains trace elements or very small amounts of vanadium, nickel, and magnesium, which are possibly fossil remains of the precursors of petroleum. Vanadium, according to present thinking, can be deleterious for boiler tubes. However, more knowledge of these elements and how they react in gases and on metal surfaces will undoubtedly produce constructive results.
Electricity Direct All conversions of petroleum into power or useful work so far discussed have involved thermodynamically irreversible combustion. Why not do the conversion reversibly? Why not obtain a fuel cell with reversible carbon-oxygen electrodes from which we may obtain electricity directly? This big area may have potentialities greater than anything that has been mentioned so far in this discussion of molecular as contrasted with nuclear energies. Where the Research Research on all these problems dealing with the utilization of petroleum for power production is going on. It goes on in industrial laboratories on problems where the solution may seem near a t hand or where there is need for new basic knowledge, in government laboratories, in universities with government and private sponsorship, and also under cooperative industrial sponsorship such a8 that of the American Petroleum Institute. It goes on for we need t o know more because we can use the knowledge and because we want t o know. We are optimistic enough t o have faith in the utility of such research and to expect better use of petroleum in the future. RECEIVED for review October 15, 1952.
ACCEPTED February 2, 1953.
Petroleum Processing PAUL F. SWANSON AND NELLSON R. ADAMS The M. W . Kellogg Co., New York, N.
OOKING back over the last half century of steady growth within the petroleum refining industry, it is difficult t o see how there can be anything but a bright future in this field. Without going into statistical details relating to world crude oil reserves and future demands for petroleum products, the authors feel i t is reasonable to assume t h a t for the next 20 years world-wide demands for petroleum derivatives will generally increase at a fairly steady rate and that crude oil requirements will be adequately met. The current picture of the petroleum and associated industries is summarized in Figure l and Table I ( 1 ) . Fundamentally, petroleum, and in a broad sense this includes crude oils, natural gas, shale oils, and even coal, represents a potential source of fuel or energy. It is true t h a t a rapidly expanding petrochemical industry now produces chemicals having a vast aggregate value, but examination of available statistical figures shows t h a t the total quantity and value of liquid fuels July 1953
Y.
produced from “petroleum,” as defined, greatly exceeds that of the chemicals, lubes, and specialty products. Figure 2 ( 2 , 6) which gives a comparison of the total annual value of petroleum products, synthetic organic chemicals, and petrochemicals serves t o clarify this point. Having broken down the over-all problem with respect to the possible types of petroleum raw materials and products, the analysis can be carried further to examine the future prospects of various types of processing. The following is a very broad outline of probable future petroleum processing developments: Liquid fuels from crude oil Lube oils and specialty products from crude oil Petrochemicals from crude oil and natural gas Petrochemicals from coal Liquid fuels from shale oils Liquid fuels from natural gas and coal
INDUSTRIAL AND ENGINEERING CHEMISTRY
1429
