William David Coolidge didn’t want to leave his beloved Boston or the scientific hum of physics research at the Massachusetts Institute of Technology. But in 1905, his overwhelming school debt and meager instructor’s salary made it impossible to refuse a lucrative job offer at General Electric (GE). So that fall the 32-year-old Coolidge reluctantly boarded the Boston to Buffalo Special, Train #49, bound for Schenectady, New York, the gridded, riverbank town that Thomas Edison had transformed into an electric city. Within several years Coolidge would solve a metallurgical mystery that would leave a mark on this city and the world at large. Decades later, however, most Americans would not realize that the light bulbs burning brightly in their homes held Coolidge’s filaments, not Edison’s.
When Coolidge arrived in Schenectady in September 1905, GE’s research laboratory, Building 19, resembled a wood cabin—a far cry from MIT’s more modern facilities. Coolidge already had reservations about doing research for a corporation, with the ultimate goal of making money for stockholders. He wondered if GE could offer the unfettered choice of interesting problems he’d enjoyed at MIT and feared he would be stuck in the corner of a manufacturing plant doing menial work. As soon as he entered the building, however, he had no time to be wistful for academia. A crisis was at hand.
GE’s main consumer product was light bulbs, primarily carbon-filament bulbs that combined Edison’s discovery with complementary technology from English inventor Joseph Swan. In 1879, Edison brought light to the world with filaments: first from a cotton thread, then cardboard, and later bamboo. But these bulbs were riddled with problems: They glowed dimly, broke easily, and lived briefly. GE had improved the carbon filament, extending their average lifespan from 100 hours to more than 500. But they needed to create something groundbreaking because competitors were nipping at their heels. They needed a new filament that glowed brighter, lasted longer, and was more robust.
Fashioning Metal Filaments
In Europe, inventors had started to make filaments from other materials, each having different levels of success. Osmium, a hard and brittle element, was the first metal fashioned into filaments. With its high melting point, it glowed brighter and was more efficient than GE’s improved carbon filaments. But osmium’s rarity and expense meant that customers had to return used bulbs to recover the metal. European researchers also tried neighboring elements on the periodic table with similar heat-resistant properties, such as tantalum and tungsten.
As Europeans made progress with these metals, GE saw an urgent need to create brighter bulbs with new filaments for the large U.S. market. But to do so, they would have to come up with something unique and overcome huge technical challenges along the way. Coolidge was hired by Willis Whitney, his former MIT chemistry professor who was brought on in 1900 as the research laboratory’s first director. To find this new filament, Whitney planned a strategy like a military campaign.
Whitney gazed at osmium and its neighbors on the periodic table, using them as clues for where to start his search for filament materials. He gave each of his 30 scientists elements to appraise. Their assignment was simple, but not straightforward: Find a metal that could form a hairlike filament and withstand heat up to thousands of degrees Celsius as electricity passed through it, like the coils within a toaster. Ideally, the material should remain inert to the small amounts of oxygen within bulbs, to avoid a common reaction that dimmed light. Additionally, Whitney instructed his scientists to explore a family of elements known as refractory metals, which were notorious for being extremely difficult to machine because of their resistance to wear.
Coolidge was assigned to work on tantalum, located in one column of the periodic table. He also investigated two metals from the adjacent column, molybdenum and tungsten, that were twice and thrice as hard. For months, Coolidge focused on tantalum, a metal that researchers in Germany had fashioned into filaments. When supplied with direct current, tantalum light bulbs burned for more than 900 hours—more than a month. But their lifespan plummeted by 70 percent with alternating current, which was predominant on American electrical grids. Sections of the tantalum threads became brittle from the electricity and broke when pulsed by the alternating jolts. Coolidge made little progress with tantalum and worked briefly with molybdenum before he eventually focused on tungsten.
Tungsten’s promising characteristics made it a worthy ultimate goal. With the highest melting point of all the elements of the periodic table (3,422 degrees), tungsten glows white-hot without melting as electricity passes through it. It renders lifelike colors rather than the yellowish hue of earlier bulbs. But tungsten’s high melting point also limited how scientists could shape it. Scientists usually worked with metals by heating and softening them, but most tools could not sustain the temperatures needed to mold tungsten.
Tungsten is also an unforgiving, hard, and brittle metal. To harness tungsten’s strength, metal workers and manufacturers didn’t use it alone: Instead they incorporated it into steels and other alloys to make them harder. Early experiments by other scientists showed that tungsten could not be worked and that it couldn’t be drawn into wire. The project almost seemed hopeless, Coolidge noted at the time, but the team persevered.
Coolidge used the next best starting point for tungsten—its powder form. Using a hydraulic press, he squeezed it into a bricklike shape and fused together the powder with electricity by sintering it with a mercury-arc furnace. Then Coolidge attempted to extrude this tungsten chunk, which he called a rod, through a hole to form a filament, but the metal would not comply. When Coolidge tried to grind off tungsten’s surface with a metal file, he found that filing tungsten damaged the tool. After months of trying to muscle tungsten into a filament, he attempted another approach. Coolidge poured tungsten powder into a binder made from starch and other organic compounds, mixed it together, and then squirted out a filament. When he tested the bulb, the inside of the glass blackened from the scorched binder.
In March 1906, a breakthrough happened when a spongy rod of tungsten accidentally fell into a pool of liquid mercury from the sintering furnace, and the mercury filled the pores. Coolidge then recalled getting a tooth filled as a child. His dentist had prepared the amalgam by combining silver slivers, shaved off of a Mexican coin, into liquid mercury; the young Coolidge noticed that this sticky paste was moldable before it stiffened into a permanent shape. Coolidge realized that mercury could be mixed with tungsten and then squirted into wire. He mixed tungsten into an amalgam made of mercury and other soft metals, including bismuth and cadmium, extruded it, and melted off the amalgam’s softer metals to create a tungsten filament. The filament glowed stably inside a light bulb.
Delighted about the success of his process, Coolidge wrote to his parents in early 1907, “The outlook for my method is certainly very bright now.” GE put the tungsten filaments on the market and soon sold nearly 500,000 of the new bulbs. Coolidge’s mother wrote that fall, “Your lamps are already in town.” But the tungsten filaments were still fragile: The amalgam gave filaments their flexibility, yet after the soft metals melted away, the remaining brittle tungsten particles could easily snap. So Coolidge was tasked with making a more robust version that could withstand harsher conditions such as vibrations in cars and trains.
Lessons from Sewing Needles
To solve the challenge of making tungsten ductile and durable, Coolidge and his four technicians went back to pressed tungsten powder and started making filaments from tungsten rods. First, they hammered the blocks at a range of temperatures, but the resulting pieces were still brittle when they cooled. Coolidge then hired a blacksmith to heat and hammer the tungsten, but the rods developed cracks. He later tried a hot rolling mill, but the pieces were again too brittle. After that, he pressed the tungsten between hot blocks, but the piece was too thick. Finally, he drew tungsten through heated dies, a process that stretched the metal grains into fibers and twisted them together into microscopic, ropelike strands. Coolidge repeated the process through dies of ever smaller diameters. At last, the final filament was flexible even at room temperature.
They now had a promising filament, but Coolidge’s lab experiment was too slow and cumbersome to manufacture filaments on a massive scale. As he searched for a production-level process, Coolidge drew inspiration from sewing needles. When he was a child, his mother made dresses, and she taught him to sew while he sat at her knee.
Early filaments were riddled with problems: They glowed dimly, broke easily, and lived briefly.
In December 1908, Coolidge visited Charles A. Cowles, an owner of a wire- and needle-making company in Ansonia, Connecticut. Cowles built special swaging machines that hit hot copper wire with thousands of blows per minute and then reduced the wire’s diameter by pulling it through a tapered metal hole. Once the wire reached a specific size, the wires were drawn through separate diamond dies of ever smaller diameters.
Coolidge became convinced that Cowles’s approach offered a practical way to make his ductile tungsten filaments. But a swaging machine, along with the diamond dies, was expensive. Fortunately, Whitney knew that research wasn’t cheap and believed that Coolidge’s idea could solve their problem. So Whitney convinced GE management to take the gamble, and by spring 1909 Coolidge got his equipment and his diamonds.
Durable and Ductile
With these tools, Coolidge gradually transformed tungsten from a brittle, unworkable metal to a durable, fine, ductile filament, modifying the process along the way, often by trial and error. He heated clean tungsten oxide powder to around 400 degrees in a crucible made from clay originating from Battersea, England, before heating it to more than 1,000 degrees in a quartz tube. Then he flowed hydrogen over this heated powder to react with the oxygen to make the pure tungsten metal. Afterward, tungsten powder was compressed and sintered into a rod, and the rod was heated and then swaged several times. Finally, the thin tungsten wire was heated red hot and drawn through several hot diamond dies of smaller and smaller diameters to create a thin filament. Sometimes Coolidge included additional steps of rolling, drawing, and hammering to make his filament.
At the time, he couldn’t see why some steps improved the performance of his tungsten filaments. For example, Battersea clay contains potassium, which incorporated into the tungsten and improved its ductility. But by being watchful and patient with his chemical partner, Coolidge finally got the desired result.
By 1910, after years of work, Coolidge’s process could produce kilometers of ductile tungsten filament, just 6 micrometers in diameter, and millions of tungsten bulbs entered the market. (In his 1913 patent, the filaments were only 2 micrometers in diameter.) GE marketed all of their tungsten filaments under the brand name Mazda, after Ahura Mazda, the Persian god of light and creation. By 1916, tungsten bulbs eclipsed Edison’s carbon filament bulbs, and 85 percent of incandescent bulbs sold in the United States had tungsten filaments. Every home soon had Coolidge’s bulbs. Tungsten filaments made with Coolidge’s basic process lit all modern incandescent bulbs, which have been displaced by more efficient compact fluorescent and LED bulbs only within the past several years.
Coolidge’s contributions have been largely forgotten compared with Edison’s, in part because the shy and introverted engineer preferred to quietly unlock puzzles and remain out of the spotlight. But Coolidge had tenaciously tamed tungsten, and in a 1922 newspaper article even Edison admired his ability to wrangle “so rebellious a metal.”
- Briant, C. L., and B. P. Bewlay. 1995. The Coolidge process for making tungsten ductile: The foundation of incandescent lighting. MRS Bulletin 20(8):67–73.
- Coolidge, W. D. 1910. Ductile tungsten. Transactions of the American Institute of Electrical Engineers 29:961–965.
- Morison, E. E. 1975. From Know-How to Nowhere: The Development of American Technology. New York: Basic Books.
- Miller, J. A. 1963. Yankee Scientist: William David Coolidge. Schenectady, NY: Mohawk Development Service.
- Wolff, M. F. 1984. William D. Coolidge: Shirt-sleeves manager. IEEE Spectrum 2(5):81–5.