Refining, History of
REFINING, HISTORY OF
The U.S. petroleum refining industry has played a central role in the expansion of this country's energy capacity over the last century. Currently, petroleum refiners generate products which account for approximately 40 percent of the total energy consumed in the United States (with respect to Btu units). The industry is characterized by a small number of large, integrated companies with multiple high-capacity refining facilities.
Between 1982 and 1997, the total number of U.S. refineries had declined from 300 to 164 operating companies. This contraction was due for the most part to the closing down of the smaller refining operations, i.e., refineries with less than 50,000 barrels of crude oil per day (BPD) capacity. While the smaller refineries still generally account for up to half of all U.S. facilities, in aggregate they control barely 14 percent of total U.S. crude refining capacity.
PROCESS DESCRIPTION: OVERVIEW
In general, refining consists of two major phases of production. The first phase of production acts on the crude oil once as soon as it enters the plant. It involves distilling or separating of the crude oil into various fractional components. Distillation involves the following procedures: heating, vaporization, fractionation, condensation, and cooling of feedstock.
Distillation essentially is a physical operation, i.e., the basic composition of the oil remains unchanged. The second phase of production, which follows distillation, is chemical in nature in that it fundamentally alters the molecular composition of the various oil fractions. These processes depend on heat and pressure and, virtually in all cases, the use of catalysts. These processes are designed to improve the octane number of fuel products. For example, in the process of isomerization, rearrangement of the molecules occurs, resulting in different chemical configurations (e.g., branched vis-à-vis linear structures) but with the atomic composition remaining constant. Polymerization involves the formation of larger molecules (polymers) from smaller ones. Reforming transforms hydrocarbons into other hydrocarbons, often to make aromatics from non-aromatic petroleum constituents (e.g., paraffins and naphthenes).
The most important of the second phase refinery operations is cracking. Cracking breaks down large hydrocarbon molecules into smaller and lighter molecular units. At first performed using high temperatures and pressures only, by the 1930s, catalytic cracking was begun. Catalytic cracking went further than earlier cracking technology to reduce the production of less valuable products such as heavy fuel oil and cutter stock.
Cracking represents a fundamental advance in refining process technology. Innovation here was mostly responsible for increasing the throughput capability of the industry and heightening octane numbers for both motor and aviation fuel. Cracking technology was central to the success of the U.S. petroleum industry in delivering needed fuel for both the marketplace and the strategic requirements of World War II.
Cracking process technology has been a distinctly American achievement. Although Europeans helped transfer some important cracking innovations to the United States, it was the U.S.-based companies and their engineers who converted these yet unrealized prototypes into commercially viable, full-scale plants.
Cracking technology evolved out of the need to solve a series of increasingly complex technical problems. Ultimate success in handling such difficult problems as carbon buildup on the catalyst surface, non-uniform distribution of thermal energy, and catalyst breakage and equipment failure came with development of the fluid catalytic cracking process in the early 1940s.
EARLY REFINING TECHNOLOGY
In the late nineteenth and early part of the twentieth centuries, refining operations were in essence distillation procedures. The refinery did little more than separate the petroleum into various fractions for commercial use. Prior to World War I, gasoline was not a dominant product of the refinery. The most commercially useful products resided in the lower (e.g., denser) range of the petroleum fractions. In order of importance, these products consisted of kerosene (for heating and light), and a variety of oils, waxes, and lubricants for industrial and home use.
The machinery employed in the early refineries was rather small in scale and operated inefficiently. Generally, equipment consisted of a series of shell tubes or stills. These were placed in the horizontal position and were connected one to another from the top through the use of vapor pipes. These pipes directed the vapors from the stills into condensers which cooled the gases and so caused the products to separate out. These products were then collected in sequence, often at one point in the plant, as liquids of varying densities and properties.
Despite various mechanical attempts to increase throughput, operations were at first conducted in batches, which required the plant to be shut down for the still and ancillary equipment to be cleaned out. Until cracking technology entered the picture, refinery operations were inefficient as they captured for use barely 50 percent of the available petroleum.
THERMAL CRACKING TECHNOLOGY
The initial impetus for development of a cracking process came in the years prior to World War I with the greater demand for gasoline products for the emerging automotive industry. Although early experimental work on the high-pressure cracking of petroleum was conducted in Europe in the late nineteenth century, the commercial breakthrough came in the years leading up to World War I by Standard Oil (Indiana). In 1913, William Burton, a petroleum chemist at Standard Indiana, conceived of commercially breaking down the molecular components comprising petroleum into smaller molecular units by the force of high pressure.
While his experiments yielded acceptable gasolines, serious problems arose as Burton attempted to scale up his process. The short cycles forced on producers by extensive coking worked against continuous operations. The carbon buildup interfered with heat transfer from the furnace to the petroleum and resulted in the formation of "hot spots" and in turn damage to the vessel. Pressure control was also a problem.
An assistant of Burton at Standard Indiana, E. M. Clark modified Burton's still in a fundamental way. Clark perceived that Burton's problems arose because heat and pressure were being applied to a static mass of petroleum within the autoclave. This situation was conducive to the formation of carbon at the bottom of the vessel and made it difficult to control pressure and obtain a uniform cracking throughout the charge. Clark retained Burton's use of high pressures but within the context of the tube still. In this design, oil flowed through, and was cracked within, banks of tubes. Each tube measured five feet in length. The hydrostatic pressures within the tubes produced uniform flow conditions. These banks of tubes were suspended in a furnace. Partially cracked charge was then directed upward to an overhead "soaking" tank. Here, cracking continued to completion under pressure but without the application of heat.
An important aspect of Clark's technology was that the oil being cracked, which flowed through the tubes and in a near vapor state, was maintained in a continually dynamic or turbulent condition. This meant that the coke particles which formed were prevented from adhering to the sides of the tubes. Further, as a result of the high surface/volume ratio of the still, heat transfer through the tubes was facilitated and higher pressures could be applied. Thus, whereas the Burton process operated at a maximum of 100 psi, Clark's tube still handled pressures of up to 1,000 psi.
The first commercial Tube-and-Tank cracking plant came on line in 1922. Overall, compared to the Burton Process, the Tube-and-Tank Process allowed larger volumes of petroleum to be processed under conditions of intense cracking and longer production cycles.
In the 1920s, a number of new thermal cracking technologies emerged that, in essence, were variations of the Burton-Clark designs. These included most notably the Cross and Holmes-Manley processes. These processes provided innovations in such parameters as operating conditions, methods of removing coke, and the number and configuration of the overhead soaking chambers.
These processes remained batch-type operations. As such they reached capacities in the range of 1,000–3,000 barrels per day (BPD). A truly novel approach to thermal cracking emerged from the facilities of the Universal Oil Products company in the late 1920s. To a greater extent than the other thermal cracking technologies, the so-called Dubbs Process approached continuous operations. Important elements of this technology were the continuous recycling of heavier byproducts back into the cracking section for further processing and the so-called bottom draw-off technique that continuously removed heavy liquid from the bottom portion of the soaking drum. This procedure reduced the rate of carbon buildup in the system and so increased the time over which the still could operate before being shut down for cleaning.
CATALYTIC CRACKING TECHNOLOGY
By the early 1930s, thermal cracking had achieved a fairly high level of operation. Both the Dubbs (UOP) and Tube-and-Tank (Jersey Standard) Processes represented the state of the art in the field. Between the end of World War I, when the Burton Process was still revolutionary, and the early 1930s, octane ratings of gasoline increased 36 percent. This improvement resulted from the existence of more advanced thermal plants and the increasing use of additives, especially tetraethyl lead. Moreover, the quantity of gasoline produced increased. Not only were the Dubbs and Tube-and-Tank technologies inherently more efficient than earlier thermal designs, but they were also readily scaled up in capacity. Through the 1930s, the size of thermal plants increased. The dimensions and amount of equipment used within a facility grew.
By the mid-1930s, thermal cracking reached its zenith in scale and sophistication. At this time the world's largest thermal cracking unit, built by Indiana Standard in Texas city, Texas, had a capacity of 24,000 BPD. This represents a capacity over 270 times larger than the first Burton stills and over forty-two times larger than the early Tube-and-Tank units.
Fixed Bed Processes: Houdry Catalytic Cracking
At this time, forward-looking companies understood that the ability of a refiner to control the highest octane fuels could corner critical and specialized niche automotive and aviation fuel markets. Refiners looked on catalytic cracking as a way to more finely tune their products for these market segments.
By the mid-1930s, catalytic technology entered into petroleum refining. To a greater extent than thermal cracking, catalysis permitted the close control of the rate and direction of reaction. It minimized the formation of unwanted side reactions, such as carbon formation, and overall improved the yield and quality of fuel output.
Within the United States, catalytic cracking was carried out on petroleum vapors. In contrast to thermal cracking, catalytic cracking did not require the application of ultra-high pressures.
The first attempt at a commercial catalytic process was the Macafee Process. In this design, cracking took place within a circular vessel packed with aluminum chloride catalyst. Cracked gasoline vapor products exited the reactor through a series of chimney-like structures. The Gulf Refining Company built a working plant at Port Arthur, Texas, just after World War I. However, the rapid accumulation of carbon deposits on the inside of the reactor and the high cost of the catalyst precluded further development in this direction.
Further catalytic cracking efforts did not take place in earnest in this country until the early 1930s. Sun Oil undertook the first significant commercial effort in catalytic cracking beginning in 1932. At this time, Sun Oil, based in Marcus Hook, Pennsylvania, was a relatively small company. It operated a number of oil tankers and had at its disposal a growing pipeline network in the Northeast and increasingly the South and West. In addition to a refinery at Marcus Hook, it operated a large refinery at Toledo, Ohio. Sun established its reputation as a producer of a variety of high-quality fuels and related oil products. Having taken thermal cracking to its limit through the use of high pressures, Sun embraced a promising new method developed by Eugene Houdry, a French mechanical and automotive engineer.
Through the 1920s Houdry had experimented in France on a number of possible catalytic routes to higher octane fuel. Finding little success in France, he came to the United States to further develop his process. After initial attempts at commercialization under the sponsorship of Sacony-Vacuum Company (currently Mobil) in Paulsboro, New Jersey, failed, Houdry and his development company, Houdry Process Corporation, moved to Sun.
Over the next four years, Houdry, working closely with Sun's engineering team headed by Clarence Thayer, worked to build a commercial plant. The limitations imposed by a static catalyst bed design imposed a major obstacle, particularly in the formation of carbon deposits that fouled the catalyst mass and impeded a continuous system of production.
The commercialization of the Houdry Process involved borrowing and integrating mechanical designs from the automotive, electrical, and metallurgical industries. The Houdry Process consisted of a series of catalyst cases containing the cracking catalyst. Each case went through a succession of operations: cracking (which over time resulted in carbon deposits on the catalyst), preparation of the catalyst for regeneration (via the purging of oil vapors), regeneration of the spent catalyst (burning off of carbon deposits from the catalyst surface), preparations for the next cracking operation (i.e., purging the newly regenerated catalyst of combustion or flue gases), cracking, and so on. To approximate continuous operation, the process placed the catalyst cases on different phases of the cycle (via a staggered arrangement) at any one time. In this way, cracking, purging, and regenerations were carried on simultaneously in different cases.
Internal tube components were an important part of the Houdry Process. Tubes were used to distribute oil vapors over the catalyst during cracking and to receive and direct cracked vapors and combustion gases (after regeneration) to various parts of the system. Tubes also played a central role in heat control. Fluid-filled, heat control tubes were placed deep within the catalyst bed so that they tapped and transferred heat which built up throughout the mass.
The hot fluid moved through the tubes to a boiler into which heat energy was transferred. Steam generated here was transported back through the tubes to heat the catalyst for the cracking operation. Initially, the heat transfer fluid used was water or superheated steam. Because these fluids tended to cause tubes to crack, a more advanced design was developed that utilized molten salts, chemically and physically more stable under the rigorous operating conditions.
By the early 1940s, a Houdry plant capacity ranged from 7,000 to 15,000 barrels per day. Houdry's contribution to the U.S. energy industries extends beyond fuels. The Houdry Process was the first industrial technology in the United States to employ on a large scale the gas turbine system, adapted from European designs, and the heart of the energy recycling process. Capturing and turning into useful mechanical energy the heat generated during combustion in the regeneration cycle was especially critical. The Houdry Process paved the way for energy recycling technology, including the process commonly called cogeneration, in such industries as power generation and iron and steel production.
But there were many problems in the operation of the Houdry Process that could not be resolved. Even with its sophisticated regeneration system, carbon continued to form on the catalyst over a period of time. The heat transfer system was not as efficient as it needed to be. The process never achieved fully continuous operations, limiting the quantity of oil that could be processed and the quality of the gasoline produced. At its height (late 1930s, early 1940s) the process did not produce gasoline with an octane rating above 87. It never captured more than 10 percent of the total U.S. cracking capacity, although in the early 1940s it controlled over 90 percent of the total catalytic cracking market.
Moving-Bed Processes: The Thermofor and Air-Lift Systems
Sun attempted to resolve these problems by improving upon the fixed-bed process. The Sacony-Vacuum Company (formerly the Standard Oil company of New York and the future Mobil Oil) was the first to develop a different type of catalytic cracking process. While involved with Sun Oil in development of the fixed-bed process, Sacony early on understood the limitations of that technology. Beginning in 1932, Sacony began conceiving of the moving-bed concept.
In its development, it adapted two existing technologies. In the agricultural sector, the mechanics of grain elevators provided a model for how to move solids vertical distances and in closed-loop flow arrangements. Sacony engineers modified the elevator bucket systems traditionally used by the grain industry to carry hot catalyst from the bottom to top of vessels and between vessels.
Then too, Sacony modeled its regenerator vessel after a certain type of metallurgical furnace (known as the Thermofor kiln). The vessel consisted of a series of semi-independent burning zones. Distribution channels delivered compressed oxygen to each zone to fuel combustion. A series of baffles and ducts within each combustion compartment produced a uniform distribution of air through the catalyst. Flue gases emitted by the regeneration process collected in common headers located between the burning zones.
The so-called Thermofor moving-bed process borrowed two important techniques from Houdry's technology: a molten salt cooling system for the kiln section and gas turbine technology that generated power to pressurize more air for the regenerator. The catalyst was stored in an overhead hopper, from which it was fed into the catalyst-filled reactor. The catalyst then traveled down the vessel under the influence of gravity. In the commercial plants, oil, injected as a liquid spray near the top of the reactor, moved down along with the catalyst. During this time, the latter transferred its heat, obtained and stored during the previous regeneration cycle, to the oil. The system conserved on energy by thus recycling heat units through the catalyst which simultaneously vaporized and cracked the oil particles that descended with it.
Following cracking, the spent catalyst and oil descended to a disengager that separated the gasoline from the catalyst. The catalyst, with oil residue entrained on its surface, then moved through a purging section where superheated steam thermally removed oil remnants. The oil-free catalyst, still laden with carbon deposits, was then lifted by elevator from the bottom of the reactor to the top of the regenerator.
Regeneration was carried out as the catalyst fell through various burning zones. The flue gases were recycled by being directed to the turbo-compressors that the pressurized air used as fuel in the combustion process.
The first Thermofor cracking unit came on line in late 1942. By March of 1943, twenty Thermofor units had been completed or were under construction. The larger Thermofor plants could circulate 100-150 tons of catalyst per hour and could process up to 20,000 barrels of petroleum per day.
But problems persisted. The catalyst, moving at rapid rates, tended to disintegrate as it impacted the inside surface of equipment . Dust particles formed, clogging pipes and transfer lines and disrupting the smooth flow of operation. This difficulty in turn prevented uniform heat distribution through the system and affected the rate and extent of both cracking and regeneration. Both the volume and quality of fuel produced suffered.
An improved design undertaken by Sacony used high-velocity gases to replace the mechanical elevator systems as catalyst carriers. These so-called "air-lift" units improved upon the Thermofor process both in terms of economies and octane numbers. It was, however, only with the fluid cracking process that catalytic technology realized fully continuous production.
Fluid Catalytic Cracking
Fluid catalytic technology addressed two major shortcomings of previous moving-bed systems: the slowness with which catalyst traveled through the vessels and the inability to sustain the cracking process continuously over an indefinite period of time. The central problem facing the industry was how to move the catalyst around the production circuit at rapid rates and at the same time fully control the intensity and duration of cracking (and regeneration). Jersey Standard (Exxon), up to this point not a major player in advanced catalytic cracking, addressed these issues as it continued to improve upon more traditional cracking processes.
By the time Jersey Standard moved into fluid research in the late 1930s, it possessed both cracking (thermal) and catalytic expertise. As noted above, Jersey developed its most important cracking technology up to that point with its Tube-and-Tank process, a thermal approach. And beginning in 1927, the company formed a patent-sharing agreement with the German giant, I. G. Farben, for development of a high-pressure catalytic hydrogenation process. Jersey's early catalytic work resulted in a number of reforming practices, such as hydroforming (1938) and steam reforming (1942) which supplemented cracking processes.
By the late 1930s, Jersey, associated with a group of companies looking for improved catalytic processes (the Catalytic Research Associates), combined its past cracking and catalytic experience in developing a fully continuous catalytic cracking system.
Prior to coming upon the notion of fluidization as a basic principle of continuous catalytic cracking, Jersey experimented with a series of fixed-and moving-bed systems. A unique approach emerging out of these efforts involved the design of "slurry" systems, by which a pumping mechanism propelled catalyst particles and oil fluid together in concurrent fashion along a horizontal reactor.
While problems with the circulatory (i.e., pumping) system limited the usefulness of this approach, it led to the next major leap for fluid cracking. Warren K. Lewis, head of the Department of Chemical Engineering at MIT and consultant to Jersey, was the leading creative force behind fluid cracking. He led the initial experiments (at MIT) identifying the commercial viability of, and establishing the preconditions for, fluidization phenomena, including the existing of the so-called turbulent bed. He also directed the design of the first semi-commercial units.
Fluid cracking retained the moving bed concept of the catalyst transported regularly between the reactor and regenerator. And as with the air-lift systems, the fluid plant rejected mechanical carrying devices (elevators) in favor of standpipes through which the catalyst fluid traveled.
Fluid technology exceeded gas-lift technique by incorporating two innovative concepts, that is, under certain conditions of velocity flow and solid/vapor concentration: (1) a catalyst-gas mixture traveling around the plant behaves in just the same way as a circulating liquid; and (2) within vessels, rather than the catalyst and oil falling under gravity, they closely intermingle indefinitely in a dense turbulent but stable, well-defined and carefully (and indefinitely) controlled cracking "bed."
The fact that a flowing catalyst-vapor mixture acted just as a traditional liquid was a crucial point. It meant that the cracking plant was in essence a hydrodynamic system readily controlled over a range of velocities and throughput by simple adjustment of pressure gradients and such variables as the height of the catalyst (i.e., the head of pressure) in the stand-pipe and the velocity of upward moving gas which carried the catalyst from the bottom of the standpipe to processing vessels. An important control element also was the "aeration" of the moving catalyst at strategic points along the standpipe (and transfer and carrier lines) by use of jets of air to adjust pressure gradients as required.
At the heart of the process was the turbulent bed. As catalyst rose with the carrier gas up the standpipe and into the reactor, it tended to "slip back" and concentrate itself into a boiling but well-delineated mass. The boiling action served to keep the catalyst particles tumbling about. Rather than falling via gravity, the boiling action in effect served to counteract the force of gravity and so maintain the catalyst in an indefinitely suspended state. Particles simply moved from one part of the bed—the turbulent mass—to another, all the while undergoing cracking. This design applied as well during combustion in the regenerator.
The plant was able to operate continuously. The continual and controlled state of turbulence in the bed assured close intermixing between solids and vapors and an even distribution of thermal energy throughout the bed, and the "liquid" catalyst flowed smoothly and rapidly from one vessel to the next.
An early fluid cracking unit removed spent catalyst from reactors (to be directed to the regenerator) using an overhead cyclone system. A more efficient technique, the so-called down-flow design, followed. It altered operating and flow conditions so that spent catalyst concentrated in the bottom part of vessels, where they could be removed, resulting in greater ease of catalyst recovery, simplification of plant layout, and improvement in operating flexibility.
Fluid catalytic cracking rapidly overtook its competitors as both a source of fuel and of critical organic intermediates. Prior to 1942, the Houdry Process controlled 90 percent of the catalytic fuel market. But only three years later, in 1945, fluid cracking led all other catalytic cracking processes in market share (40 percent). At this time Thermofor technology stood at 31 percent, and Houdry at less than 30 percent.
After 1942, fluid cracking technology increasingly dominated U.S. petrochemical production. It manufactured high-tonnage fuels for both motor vehicles and aircraft for the wartime effort. The quality of the gasoline was unprecedented. Octane ratings of fluid-produced fuels exceeded 95, an unheard-of figure only a few years before, and critical for aviation gasoline. Fluid cracking technology played a central role the U.S. synthetic rubber program: its byproduct gases supplied the butylenes which were essential in the making of butadiene, the essential rubber intermediate.
RECENT DEVELOPMENTS IN PETROLEUM REFINING
Since 1945, the fluid catalytic cracking process has rapidly overtaken fuel production and has become the central technology in the U.S. petrochemicals industry. With fluid cracking, the scale of petrochemical operations grew enormously. For the first time, refiners could process virtually any volume of oil rapidly and efficiently.
Accordingly, technological change in refining technology has centered on alterations and improvements made to the fluid cracking process. These modifications have included improvements made to catalysts, materials of construction, interior linings, and a variety of mechanical details. Since 1945, the process has operated on progressively higher temperatures for both cracking and regeneration. It has also incorporate newer generations of catalysts.
By the early 1960s, fluid cracking had become the workhorse of the refining industry. It was the central process in the production of over 70 percent of all high-octane fuel. From early 1940s to mid-1960s, capacities of fluid units have grown from less than 20,000 BPD to between 100,000–200,000 BPD.
The flexibility of the process was manifested in its ability to economically process smaller volumes of feedstock. Between the 1950s and 1980s, refiners and engineering firms developed smaller, lower investment units which could be readily scaled up as required. Recently, such fluid crackers have been simplified to the point that cracking and regeneration take place within a single, vertical standpipe or riser tube. The capacities of such units often do not exceed 1,000 BPD.
By the mid-1990s, the technology was virtually the only catalytic cracking process in operation in the major refineries. The technology accounted for over 95 percent of all high-octane fuel within the United States. By 1997, there were approximately 350 fluid cracking facilities in operation worldwide, most located within the United States. Between 8 and 9 percent of the fluid units existing worldwide (25-30 units) are owned and operated by Exxon, the original innovator of fluid cracking.
In the late 1990s, the growth of fluid cracking as a major petroleum refining process was about 2 percent per year. Total world fresh feed capacity for the fluid process now stands at more than 11 million barrels/day. Worldwide, FCC makes 80 billion gallons of high-grade gasoline annually. This represents nearly half of total world gasoline production.
A major recent development for the technology has been its closer integration with the large petrochemical complex. Essential petrochemical activity has been relying more on fluid technology and less on thermal units for their intermediate feedstock. The big development for the 1990s is increasing integration of FCC with the large petrochemical plant.
In the postwar period, fluidization influenced U.S. manufacture in other areas as well. Most critically, it has been applied in combustion processes for the making of metallurgical materials and by the utilities for generating heat and electricity for industrial and residential use. Moreover, the technology is a "cleaner" method for producing energy and so is an important means by power-based companies to comply with stricter environmental regulations.
Sanford L. Moskowitz
See also: Refineries.
Enos, J. (1962). Petroleum, Progress and Profits: A History of Process Innovation. Cambridge, MA: MIT Press.
Frankenburg, W. G. et al., eds. (1954). Advances in Catalysis and Related Subjects, Vol. 6. New York: Academic Press.
Grace, J. R., and Matsen, J. M., eds. (1980). Fluidization. New York: Plenum Publishing.
Giebelhaus, A. W. (1980). Business and Government in the Oil Industry: A Case Study of Sun Oil, 1876–1945. Greenwich, CT: JAI Press.
Moskowitz, S. L. (1999). "Science, Engineering and the American Technological Climate: The Extent and Limits of Technological Momentum in the Development of the U.S. Vapor-Phase Catalytic Reactor, 1916–1950." Diss. Columbia University.
Murphree, E. V. et al. (1945). "Improved Fluid Process for Catalytic Cracking." Transactions of the American Institute of Chemical Engineers. 41:19–20.
Nelson, W. L. (1949). Petroleum Refining Engineering. New York: McGraw-Hill.
Othmer, D. F. (1956). Fluidization. New York: Reinhold Publishing Corp.
Russell, R. P. (1944). "The Genesis of a Giant." Petroleum Refiner 23:92–93.
Spitz, P. (1988). Petrochemicals: The Rise of an Industry. New York: John Wiley and Sons.
Squires, A. M. (1986). "The Story of Fluid Catalytic Cracking: The First Circulating Fluid Bed." Proceedings of the First International Conference on Circulating Fluid Beds, November 18-20, 1985. New York: Pergamon Press.
Williamson, H. F., et al. (1963). The American Petroleum Industry, Vol. II: The Age of Energy, 1899–1959. Evanston, IL: Northwestern University Press.