Burton, Houdry, and the Refining of Oil
The 1870s and 1880s saw the rise of two great chemical process industries in America: the steel industry led by Andrew Carnegie and petroleum refining under John D. Rockefeller. The two industries followed very different paths. After his retirement in 1901, Carnegie sold his company to the banker J. P. Morgan, who merged it with some rivals to form the United States Steel Corporation, bringing 60 percent of the nation’s steel-making capacity under the control of one firm. But U.S. Steel faced no serious competitive or regulatory challenges. No other material threatened to take away its market; and the federal government, which tried to regulate or break up other industry-dominant companies, left the steel industry alone. Relying on nineteenth-century processes, U.S. Steel and its rivals did not innovate in any fundamental way between 1901 and 1939. When foreign competitors using more modern methods appeared after 1945, the American steel industry was unprepared and lost ground.1
The petroleum industry had to face early challenges. Its continued existence required the timely discovery of new oil reserves. By the turn of the century, the industry also faced the prospect of gradually losing its principal market, the demand for kerosene illuminating oil, as indoor electric light spread. Demand for gasoline to power automobiles gave the petroleum industry a huge new market. But nineteenth-century refining methods could not meet the need for gasoline, and the dominant refiner, Standard Oil, resisted innovation that chemical engineers such as William Burton wanted to make. The near monopoly of Standard Oil aroused public opposition, however, and the U.S. Supreme Court broke up the firm in 1911. The breakup of the company enabled Burton to develop a process that increased the yield of gasoline from a barrel of crude oil so that refining was dramatically less wasteful of petroleum.
The Rise and Fall of Standard Oil
During the nineteenth century, most Americans lived beyond the networks that supplied illuminating gas and later electricity to town dwellers. Until midcentury, Americans relied on candles for indoor light or on lamps that burned whale oil. As the supply of whale oil declined from the killing of too many whales, petroleum-based illuminants began to take its place. Crude petroleum that seeped into ponds and creeks could be refined by simple methods to produce kerosene, a good illuminating oil, and a small industry to refine kerosene arose in the 1850s. The growth of the industry was hampered, though, by the minuscule supply of naturally surfacing crude oil. To supply the market for kerosene, the petroleum industry needed to tap reservoirs of oil that were known to exist underground.2
In 1854 a group of Connecticut investors sent a sample of crude petroleum from western Pennsylvania to Benjamin Silliman Jr., a professor of chemistry at Yale University. Silliman reported in 1855 that half of the oil sample could be refined into a high-quality kerosene illuminant and that the rest could have other uses.3 To look for oil, the Connecticut group hired Edwin L. Drake, a man with no experience in oil or drilling. Drake’s previous career as a railroad conductor gave him a free pass to travel, however, and in 1857, with his pass and the title of “Colonel” that he gave himself, Drake went to Titusville in northwestern Pennsylvania. He found an experienced driller, and after nearly two years, the two struck oil from an underground reservoir on August 28, 1859 (figure 4.1).4
Drake’s example attracted entrepreneurs who soon covered western Pennsylvania in oil wells and simple refining works. Petroleum divides into a range of “fractions” defined by their boiling points (sidebar 4.1). The lighter fuel gases, such as propane, boil at low temperatures. Gasoline boils next, followed by kerosene, diesel and light fuel oils, lubricating oils, and finally the heavy fuel oils and asphaltenes that boil at high temperatures. Early refiners obtained kerosene by heating the crude oil in stills. In a typical still, kerosene vaporized between 400 and 500 degrees Fahrenheit (205 to 260 degrees Celsius) and went into a coiled copper tube, where it condensed into a separate receiving tank. Boiling it again, but at a lower temperature, vaporized out lighter “fractions” in the fluid. Treating the kerosene with an acid and then an alkali removed its odor and improved its color, and when the residues of treatment had been removed, the kerosene was ready for sale.5
Figure 4.1. Edwin Drake (in top hat) at his oil well. Courtesy of Pennsylvania Historical and Museum Commission, Drake Well Museum Collection, Titusville, PA. Mather Photograph No. 4.
Petroleum Fraction | Boiling Range °F | Boiling Range °C |
Light naphtha | 30–300 | −1–150 |
Heavy naphtha | 300–400 | 150–205 |
Gasoline | 30–355 | −1–180 |
Kerosene | 400–500 | 205–260 |
Stove oil | 400–550 | 205–290 |
Light gas oil | 500–600 | 260–315 |
Heavy gas oil | 600–800 | 315–425 |
Lubricating oil | >750 | >400 |
Residuum | >1100 | >600 |
Four years after Drake’s well, a new figure entered the growing business of refining oil. Born in upstate New York, John Davison Rockefeller (1839–1937) had moved to Cleveland and left high school two months short of graduation at age fifteen. After working for three years as an assistant bookkeeper in a produce firm, he started his own business shipping produce to the cities of the East Coast. By 1863 he was moderately wealthy and was able to finance an oil refinery in Cleveland. Recognizing the future of the oil industry, Rockefeller left the produce business in 1865 to concentrate on oil refining (figure 4.2).6
The high cost of building a steel manufacturing plant kept the number of competing steelmakers small. In oil, the start-up costs of drilling for oil and refining it were much lower, and a large number of small refiners soon competed with each other. By the late 1860s rapid growth in the supply of oil and in refining capacity had outpaced demand, causing prices and profits to fall. The kerosene was also uneven in quality, causing many oil lamps to explode when lit. Rockefeller became the dominant refiner in Cleveland by carefully managing his own refinery and its product and by becoming a reliable supplier to New York and other eastern markets. In 1870 Rockefeller formed the Standard Oil Company to impose a uniform product standard and price.
The railroads suffered from overcapacity themselves in the 1870s, and Rockefeller was able to negotiate preferential rates from them. He invested in new technology to widen his advantage, building his own pipelines from oil fields to railheads and shipping in tank cars rather than in wooden barrels. By purchasing property through veiled subsidiaries, by inviting selected competitors to become his partners, and by driving others out of business through price cutting, Rockefeller brought most of the American refining industry under his control within the span of a decade. In 1873 Standard Oil controlled 10 percent of the refining capacity in the United States; by 1880 the firm had control of 90 percent. Total production of refined oil in the United States grew from a yearly average of 7 million barrels in 1873–75 to an average of 18 million barrels in 1883–85. Four-fifths of this output went to supply kerosene for lamps. Other products, such as solvents and lubricating oils, accounted for the rest of the industry’s sales.7
Figure 4.2. John D. Rockefeller, circa 1884. Courtesy of Rockefeller Archive Center, Tarrytown, NY.
In the 1880s and 1890s, Rockfeller supplied kerosene to consumers at a price of between five and eight cents per gallon, compared with a kerosene price of about forty cents per gallon in 1870.8 But Standard Oil’s national operation brought it into conflict with state laws and taxes. Some state laws outlawed the ownership of a company in one state by a company outside it, and Pennsylvania taxed the entire firm even if only part of its revenues were earned in the state. To accommodate state laws, reduce tax liability, and veil their degree of industry dominance, Rockefeller and his partners signed the Standard Oil Trust Agreement in 1882. The agreement broke the company into subsidiaries local to each state, and shares in these subsidiaries were held by a trust. As officers of the trust, Rockefeller and his partners retained control, but legally none of the companies owned by the trust owned each other.9 The Sherman Antitrust Act of 1890 outlawed the Standard Oil trust, which dissolved in 1892, leaving the partners owning the various subsidiaries individually. But New Jersey had passed a law that allowed one company to hold the stock of others, and in 1897 the Standard Oil owners exchanged their shares for those of Standard Oil of New Jersey. Jersey Standard in turn held the stock of the various subsidiaries.
Rockefeller’s control of the oil industry had long been controversial and the reconstituted monopoly came under attack after the turn of the century. In a series of magazine articles in 1902–3, Ida Tarbell depicted the company as a threat to democracy.10 The administration of President Theodore Roosevelt began to enforce antitrust law more vigorously than its predecessors, and in 1906 the United States sued Standard of New Jersey under the Sherman Act. The Supreme Court in 1911 ordered the breakup of Standard Oil into its thirty-three subsidiaries (figure 4.3).11 A handful of these firms, joined by a small number of new refining companies in Texas and California, still dominated the oil industry. But there were many more refining firms now, and some turned to engineering innovation to gain a competitive advantage.
The Frasch Process and the Need for Gasoline
Despite its dominant market position after 1880, even Standard Oil was far from secure. The company depended on oil reserves in Pennsylvania that were running down; without new supplies, the company faced ruin. In 1885 prospectors discovered vast oil fields in northwest Ohio and Indiana. But this “Lima-Indiana” oil, as it was called, had a high sulfur content that gave its refined products an unpleasant smell, which existing treatment processes could not eliminate. Rockefeller believed that “the whole future of the Standard depended upon thrusting westward.” When his partners objected, he threatened to develop the new oil fields on his own. His partners gave way.12
Figure 4.3. Standard Oil companies after the 1911 breakup. Source: Joseph E. Pogue, The Economics of Petroleum (New York: John Wiley and Sons, 1921), p. 217.
To solve the problems of Lima oil, Rockefeller turned to Herman Frasch (1851–1914), a chemist who worked as a consultant to Standard Oil from 1876 to 1885. Frasch had left the company to buy an oil field near London, Ontario, in Canada whose owner had recently failed because of the oil’s sulfur content. After mixing kerosene from the oil with various metallic oxides, Frasch found that reusable copper and lead oxides removed the sulfur. News of his success reached Standard Oil, and Frasch returned to be its chief chemist.13 After further experimental work, Frasch removed the sulfur from Lima-Indiana oil. Annual U.S. refined oil production rose from 18 million barrels in 1885 to 49 million barrels in 1899, and the new midwestern oil fields accounted for about one-third of this total.14
The Frasch process relieved a bottleneck in the supply of illuminating oil, but it could not save the industry from a much worse fate, shrinking demand. Although rural areas continued to buy kerosene illuminants until electrification arrived later in the twentieth century, electric power began to replace oil as a lighting source in urban and suburban America in the 1880s and 1890s. Petroleum refining would have gone the way of the gaslight industry if the automobile, with its gasoline-burning engine, had not created a new and even greater demand for oil after 1900. The promise of a new market brought with it a new challenge.
Simple or “straight-run” distillation methods, in which crude oil was boiled and condensed, yielded about 30 to 50 percent kerosene from a batch of crude oil. But no more than about 20 percent of crude oil could be distilled into gasoline.15 To meet the need for more gasoline, drillers began to search for new supplies of crude oil, but refining only a fraction of new oil was wasteful. Early automobiles were luxuries that did not stretch gasoline supplies at first, but Henry Ford’s introduction in 1908 of a mass-produced affordable car, the Model T, caused demand for gasoline to soar. The oil industry needed a more efficient way to supply this market.
William Burton and Thermal Cracking
William M. Burton (1865–1954) found a way to obtain more gasoline from a given amount of crude oil (figure 4.4). After earning one of the first American Ph.D.s in chemistry, from Johns Hopkins University in 1889, he went to work as Frasch’s assistant and then moved in 1890 to a new refinery at Whiting’s Crossing, Indiana, seventeen miles east of Chicago. Burton did not feel challenged by routine laboratory work, though, and he moved into management, becoming a vice president of Standard Oil of Indiana, the principal midwestern subsidiary of the Standard Oil group, in 1903. The Whiting refinery reported to him, and his attention returned to research as the need for more gasoline became urgent after 1908. Dr. Robert E. Humphreys had taken charge of the laboratory at Whiting, and at Burton’s direction, Humphreys and an assistant, Dr. F. M. Rogers, began to study ways to obtain more gasoline from crude oil (figure 4.5).16
Figure 4.4. William M. Burton. Courtesy of BP America, Inc., Chicago, IL.
Crude petroleum consists mostly of hydrocarbons, molecules composed of hydrogen and carbon atoms. Burton and Humphreys knew that heavier molecules could break apart into lighter ones. They decided to see if they could get more gasoline out of a barrel of crude oil by breaking or “cracking” heavier petroleum molecules into lighter ones in the gasoline range (sidebar 4.2).17 Under Burton’s direction, Humphreys at first tried different ways of heating the heavier oil to obtain more of the lighter vapors, with poor results. Next, he tried mixing in aluminum chloride, which was known to break heavier oil molecules into lighter ones. This approach used too much oil, though, and the aluminum chloride was expensive and could not be reused. As Burton later wrote, “our efforts were not successful … we met failure in every direction.”18
There remained one last possibility. Simple distillation subjected crude oil to varying temperatures but not to varying pressures. “It had been known for a long time that distillation of petroleum products under pressure resulted in their disassociation and production of some low-boiling and some high-boiling fractions,” recalled Burton, “but this process had never been applied in a practical way … owing to the extreme hazard.”19 Under high heat and pressure, crude oil could explode in a terrific fireball. But there seemed no other way to get more gasoline than to take this risk. Humphreys conducted experiments in which he heated diesel oil in a fifty-gallon steel tank under rising pressures as well as rising temperatures. A condenser received the vapors and took them to a receiving tank. For safety, Humphreys raised the pressures gradually. Finally, at a pressure of seventy-five pounds per square inch (about five times atmospheric pressure) and a temperature of about 700 degrees Fahrenheit, the still produced gasoline of acceptable quality. The process doubled the amount of gasoline that a barrel of crude oil could produce, from about 20 to about 40 percent.20
Figure 4.5. Robert Humphreys in the Whiting laboratory, circa 1908. Courtesy of BP America, Inc., Chicago, IL.
Gasoline
Chemical Reactions
Burton asked Standard Oil’s headquarters in New York for $1 million to build full-scale pressure stills. After Rockefeller’s retirement in 1897, however, the monopoly had come under more conservative management that refused to consider such a dangerous-sounding scheme. One of the Standard Oil directors in New York was reported to have said: “Burton wants to blow the whole state of Indiana into Lake Michigan.”21 The breakup of Standard Oil in 1911 came just in time for Burton and his team. The Whiting refinery went to the newly independent Standard Oil Company of Indiana (later known as Amoco, now BP America). The new company had refining capacity but no oil reserves, and its leaders threw their support behind Burton’s work. On July 3, 1912, he filed a patent on the new process, assigning the rights to the company. The patent was granted on January 7, 1913.
The Burton Patent
The innovation in the Burton process was its mechanical arrangement of the boiler and condenser (sidebar 4.3). To heat crude oil under pressure, Burton at first placed a shut-off valve between the boiling tank and the condenser, so that all of the heat and pressure would be confined to the boiling tank. But the condensed vapors produced in this way had too many unwanted by-products and an unacceptable odor. Humphreys moved the shut-off valve to a different location, between the condensing tube and the final receiving tank. Heating the heavier fractions of oil then created pressure in both the boiler and the condenser, not just in the boiler. The effect of making this simple move was dramatic: heavier oil cracked into acceptable gasoline. In addition to breaking heavier molecules, the cracking also produced molecules lighter than gasoline. Some of these combined with each other to form gasoline molecules. The result was a substantial increase in the amount of gasoline produced from a barrel of crude oil.22
The Burton Process
Cracking Heavier Molecules into Gasoline
The two engineers did not achieve this breakthrough by applying a scientific theory that predicted the results they would find. As Burton stunningly observed in his patent, the results occurred for chemical reasons “which I do not attempt to explain.” In an address to the Chicago section of the American Chemical Society, at which he was awarded the Josiah Willard Gibbs medal in 1918, Burton credited science with a role in the oil industry that began with Frasch and his process: “This was the starting point for a better feeling between the chemical profession and the petroleum industry, and from that time, more and more chemists have been employed in the refining industry, until to-day the larger refineries depend almost entirely upon chemists to manage, not only the refinery as a whole, but the various departments of the same.”23 But, in fact, what Burton attributed to science was a new kind of engineering, chemical engineering.
The aeronautical historian Walter Vincenti has described a way of understanding flows, called control-volume analysis, that sheds light on what Burton and his team achieved. Developed in the nineteenth century, control-volume analysis in essence held that if a flow going into a box was known and the flow going out was also known, it did not matter to engineering exactly what went on inside the box. To scientists, in contrast, a proper understanding required knowing in a precise way what occurred inside the box. As Vincenti observed of fluids in motion,
All nontrivial flows … vary either in space or time (or both), and a physicist in his quest for knowledge characteristically wants to know the point-by-point details of this variation. For this end the overall results of control-volume analysis are not enough. In fluid mechanics physicists therefore concentrate on solution of the differential equations of motion [i.e., equations that reduce matter to points and calculate how the points move in space and time]. …
In engineering the situation is very different. Engineers, unlike physicists, are after useful artifacts and must predict the performance of the objects they design. … Moreover, the problems that arise often present serious difficulties in the underlying physics or in the solution of differential equations, so that more than overall results are not feasible.24
Burton did not refer in his patent to control-volume analysis but his innovation was a textbook example of it. The thousands of simultaneous chemical reactions that took place in the thermal cracking of petroleum would have been impossible to isolate and calculate individually. Burton controlled what went into the system, he knew what came out, and he did not try to explain the chemical reactions that occurred along the way. But his approach was not a matter of cutting corners. It expressed the wisdom of knowing the limits to sophisticated analysis as well as of knowing when detailed and exact understanding was necessary and proper.
Building and operating full-scale pressure stills challenged Burton and his colleagues yet again. When Burton asked mechanical engineers for help building a containment vessel, he noted drily, “we did not receive very much encouragement.”25 The steel plates had to be riveted together, and their strength was uncertain at high pressures and temperatures. Thermal cracking required great courage on the part of refinery workers, who had to plug flaming leaks. These dangers lessened with the development of electric welding in the late 1920s and other improvements in still design.
Standard Oil of Indiana began producing cracked gasoline in January 1913 (figure 4.6), at about the time that Henry Ford perfected his method of mass-producing automobiles. In 1899 the United States produced about 6 million barrels of gasoline, for use mainly as cleaning solvents. By 1919, mostly to supply motor vehicles, U.S. gasoline production had climbed to 99.7 million barrels, of which 15.5 million (16 percent) consisted of cracked gasoline. By 1929 U.S. gasoline production reached 441.8 million barrels, of which 143.7 million (33 percent) was cracked gasoline (figure 4.7).26
Eugene Houdry and Catalytic Cracking
The Burton process was limited in one important way: it produced gasoline only in batches and had to stop approximately every two days so that the stills could be cleaned of carbon residue that built up during the refining process. Engineers working for the Texas Company (now Texaco) and for Standard Oil of New Jersey (later Exxon, now ExxonMobil) soon devised thermal cracking processes that could go for longer periods between cleanings. The most notable of these was a process invented by Jesse Dubbs and his son, whom he gave the prophetic name Carbon Petroleum Dubbs. Patented between 1915 and 1919, the Dubbs process could feed crude oil in and out of pressure stills on a nearly continuous basis, removing carbon residue as it did so. Thermal cracking and processes to improve it soon became entangled in patent litigation until the major refining companies agreed to share their processes in 1931.27
Figure 4.6. Burton stills in operation. Courtesy of BP America, Inc., Chicago, IL.
During the early 1920s, however, the petroleum refining industry faced a new technical challenge. The gasoline produced for motor vehicles suffered from a problem called engine knock, in which the gasoline failed to ignite evenly. As automobiles developed stronger engines to drive faster on newly paved roads, this problem became more severe. In the early 1920s, Charles Kettering and Thomas Midgley of General Motors found that adding tetraethyl lead to gasoline could sharply reduce engine knock. Motor vehicles ran on leaded gasoline from the 1920s until the 1970s, when safety concerns over air pollution brought new emission standards that removed lead from gasolines used by newer vehicles.28
Figure 4.7. Trend in output of crude oil, 1899–1920. Source: Joseph E. Pogue, The Economics of Petroleum (New York: John Wiley and Sons, 1921), p. 84.
To measure the antiknock properties of gasoline, Graham Edgar of the Ethyl Corporation devised a scale in 1926. Edgar rated gasoline with a number between 0 and 100 according to how its performance compared with two benchmark gasolines, normal heptane (rated 0, severe knocking) and iso-octane (rated 100, no detectable knocking). The average of two different tests gave numbers on this scale called the octane rating of gasoline, with an average closer to 100 being better.29 By 1930 gasolines reached octane ratings between 60 and 70. But the best thermal cracking could not raise the octane above 82. Newer and more powerful automobile engines required higher-octane gasoline, and aircraft engines needed fuel with octane ratings of nearly 100.30 The design of an efficient catalytic cracking process by the French engineer Eugene Houdry produced higher-octane gasoline and also increased the supply (figure 4.8).
Born in France and trained as a mechanical engineer, Eugene J. Houdry (1892–1962) joined his father’s steel company after serving in the First World War (1914–18). To relieve the French need for imported oil, Houdry tried in the early 1920s to convert domestic coal into oil. This proved uneconomical. But his other passion was automobile racing, and he owned and raced a Bugatti. On a visit to the United States in 1922, he attended an Indianapolis 500 race as a spectator and then visited the Ford plant in Detroit. Houdry realized that the limitation on speed was not just in the mechanical design of cars but in the quality of the gasoline.31
During the nineteenth century, chemists had discovered that adding certain chemicals, called catalysts, to a mixture accelerated chemical reactions. Burton and Humphreys at first tried to use a catalyst, aluminum chloride, to crack petroleum; however, the material was uneconomical, and the two shifted to thermal cracking. Houdry knew that cheaper catalysts released carbon by-products that stuck to the catalyst and interfered with its action. He believed that finding a catalyst that was economical and easy to clean would be the next great advance in the making of gasoline. Like Edison in his search for a proper filament, Houdry tried hundreds of possible catalysts. Finally, in 1927, Houdry found that fluid petroleum molecules adhered to solid oxides of silicon and aluminum (in pellet form) and that the catalysts cracked heavier petroleum molecules into lighter ones. The process required higher temperatures but lower pressures than thermal cracking. A carbon residue accumulated on the catalysts with repeated use, but by regularly burning the catalysts with air and heat, the residue could be removed. Houdry proved that catalytically cracked gasoline was as good as the best gasoline on the market by using it to accelerate his Bugatti racecar up a steep hill.32
Unable to obtain funding in France, Houdry interested the Vacuum Oil Company of New Jersey in his process. In 1930 he set up a pilot plant at the Vacuum research center in Paulsboro, New Jersey. The acquisition of Vacuum by Standard Oil of New York (known as Socony, later Mobil, now part of ExxonMobil) stalled Houdry’s research, but Sun Oil of Philadelphia (Sunoco) stepped in to carry it on. In 1936, backed by Socony-Vacuum and Sun, Houdry began producing gasoline for sale. In the best thermal cracking processes, a charge of heavier oil could yield about 40 percent gasoline, with octane ratings between 72 and 82. With catalytic cracking, the same amount of heavier oil produced almost 60 percent gasoline, and if the process was repeated, the amount reached 75 percent. The octane ratings on the first “pass” were between 81 and 95 and on the second between 83 and 95.33
Figure 4.8. Eugene J. Houdry. Source: Harold F. Williamson et al., The American Petroleum Industry: The Age of Energy 1899–1959 (Evanston, IL: Northwestern University Press, 1963), p. 613. Courtesy of Air Products Corporation.
Catalytic cracking increased the octane rating mainly by rebranching atoms. Gasoline molecules with the same numbers of carbon and hydrogen atoms can differ in the way the atoms link together (sidebar 4.4). Gasoline molecules with their atoms arranged in straight chains have low octane ratings because they tend to break and combust prematurely, while molecules with atoms in branched chains performed much better. Catalytic cracking greatly increased the proportion of branched molecules in gasoline.34
Burton and Houdry worked within a framework of basic chemical ideas provided by science. But their designs were not predicted by these ideas, and if they had tried to follow an approach that sought an exact understanding in the manner of physics, their efforts to innovate would have failed. They took a broad view of engineering as well and looked beyond the specialized fields in which they had been trained. Burton, a chemist, created an innovation that was essentially mechanical; Houdry, a mechanical engineer, found the chemicals that made catalysis in petroleum refining practical. The modern oil refinery needed mechanical engineering to work and required the twentieth century’s greatest machine, the automobile, to provide a market for what it produced.
