Hydrogen and Fluorine
 Early in 1949 William L. Doyle, a chemist engaged in rocket propellant research at North American Aviation, made a deal with Herrick Johnston. Doyle would come to work at the Ohio State University Rocket Laboratory if given a free hand to investigate the performance of liquid hydrogen with his favorite oxidizer, liquid fluorine.*
In February 1949, Doyle reported for duty at the Ohio State laboratory. He did not like the experimental equipment, the operations, or the procedures, so he began to make changes.
William Doyle was a dynamic young man who knew what he wanted and just how to do it. The antithesis of the desk-bound supervisor and paper shuffler, he liked to be part of the action. He found his right environment at Ohio State where a senior engineer was responsible for his entire project-from inception, through design, fabrication, installation, operation, data analysis, and writing up the results. Doyle found this situation ideal and he made the most of it.
Doyle's interest in the hydrogen-fluorine combination was natural. It represented the combination of the ultimate fuel and the ultimate oxidizer, with a higher theoretical performance than hydrogen and oxygen. In addition, the mixture of 6 percent hydrogen and 94 percent fluorine by weight not only resulted in near-maximum performance, but also meant higher average propellant density for the combination. Doyle visited the men in the fuels and oil branch at Wright Field and convinced them to modify the Ohio State contract to include the work he wanted to do.
One of Ohio State's rocket test facilities was rebuilt to handle liquid hydrogen and liquid fluorine. The hydrogen flow system was encased in a vacuum jacket for insulation. A series of problems with maintaining the vacuum were solved. The flow of liquid hydrogen was measured by a dual system: the conventional way of measuring the pressure differential across a sharp-edged orifice as well as continuous measurement of the hydrogen tank mass. Once the hydrogen system was functioning, it gave little more trouble, but many problems were encountered in the fluorine system. The fluorine gas, procured commercially as a compressed gas, was condensed in the  propellant tank by immersing it in a liquid air bath. Liquid fluorine flow was measured by the same methods used for hydrogen.
Doyle made his first liquid hydrogen-liquid fluorine run on 15 June 1950. He first operated the injector alone to see if the hydrogen and fluorine would ignite readily and spontaneously, which they did. He followed this experiment with rocket engine tests. By the first part of August, nine runs had been made and Doyle felt confident enough to invite his sponsors from Wright Field to witness a test. Judging from the mishaps reported for the first eight runs, Doyle was displaying a considerable amount of confidence. Don Kennedy arrived in response to the invitation and witnessed the tenth test on 11 August 1950. The run was perfect in Doyle's view, with a measured exhaust velocity of about 3600 meters per second at 20 atmospheres. Kennedy was greatly impressed and reported the results to his boss, Weldon Worth. Doyle continued the experiments and in mid-January 1951, Kennedy informed him that a group of high officials at Wright Field would visit Ohio State to witness a run with hydrogen-fluorine. Soon after the call, Doyle made a run at a high pressure (38 atmospheres) and measured an exhaust velocity of over 4300 meters per second. On 29 January, 14 people from Wright Field's Power Plant Laboratory arrived in terrible weather-a sheet of ice compounded by mist and drizzle. Icing difficulties delayed the run for an hour, but it was a success, lasting over a minute. Performance, however, was lower than obtained in earlier runs.14
One measurement necessary to determine performance-fluorine flow-had bothered Doyle from the start. Whereas the two flow measurements for liquid hydrogen checked with each other, the fluorine flow as measured by the orifice was lower than that measured by weighing the propellant tank. The difference was consistent-about 18 percent lower for the orifice. Five design changes were made to improve the orifice measurement, but the discrepancy remained.
Doyle was not the only experimenter having difficulty measuring the flow rate of liquid fluorine. Aerojet was having the same difficulty and investigators there began to suspect that the density of fluorine might somehow be wrong. This was heresy, for a number of eminent scientists had measured the density of fluorine and they all agreed. James Dewar and Henri Moissan had first measured it in 1897 and found it to be close to 1.14 grams per cubic centimeter at 83 K. The value In use in the 1950s was 1.13 grams per cubic centimeter at 77 K, determined by E. Kanda in 1937.
Near the end of April 1951, Kennedy telephoned Doyle that Aerojet, using a hydrometer, found that the density of liquid fluorine was 1.55 grams per cubic centimeter, considerably higher than the published value. Doyle used the Aerojet value with his orifice measurements and found that the 18 percent discrepancy with the weighing measurement disappeared! The greater density of liquid fluorine was an exciting discovery to rocket engineers, for it meant the oxidizer was even more attractive than first realized.**
 Doyle made his 48th and last hydrogen-fluorine run in mid-April 1951 and turned his attention to the ammonia-fluorine combination. This ended the Ohio State rocket experiments with liquid hydrogen, although the properties work continued, as well as some small-scale combustion research of a fundamental nature.