Lewis Hydrogen Rocket Experiments, 1958-1959
 After their initial success in operating a hydrogen-cooled, hydrogen-fluorine rocket engine in November 1957 (p. 92), Howard Douglass, Glen Hennings, and Howard Price, Jr. continued the experiments until February 1959. Fourteen runs were made using the showerhead and triplet type of injectors with comparable results. A maximum exhaust velocity of 3455 meters per second was obtained at a flow rate that was 14 percent liquid hydrogen with a combustion pressure of 20 atmospheres. This was 97 percent of the maximum theoretical performance. The experimenters reported no problems relative to engine operation, starting, or stability of combustion. They did, however, have a number of minor problems with the injectors and with operating the thrust chamber beyond its design limits. Following this series of experiments, another team of researchers made 26 more runs with the same type of engine over a range of combustion pressures and exhaust nozzle expansion ratios. Earlier, Vearl Huff had suggested the technique of exhausting the rocket into a properly proportioned duct closed at the rocket end. The high-velocity rocket exhaust pumped the air from the duct, reducing the pressure in the immediate vicinity of the rocket nozzle and thereby simulated high altitude. The exhaust duct needed for silencing and for removing hydrogen fluoride from rockets using fluorine was ideal for the new purpose, so the one duct served three purposes. The nozzle altitude simulation  technique was used to test a rocket with a nozzle area ratio of 100 and the measured exhaust velocity was 4730 meters per second (at a combustion pressure of 49 atm), one of the highest performance values obtained by a chemical rocket engine.23
Cell 22, with its two parallel test stands capable of handling engines up to 22 kilonewtons, was the workhorse cell for high-energy propellants through 1957. A new, larger facility for high-energy propellants and engine thrusts up to 89 kilonewtons was ready for its first hydrogen tests on 14 November 1957. The initial run used gaseous hydrogen and a water-cooled chamber which leaked, causing ignition problems and a minor explosion, or a "hard start" in the rocket engineer's vocabulary. The chamber was repaired and five days later satisfactory starting was achieved, but other troubles arose. A malfunctioning indicator led the operator to increase propellant flows, and after a second of operation at a pressure of 30 atmospheres in: a chamber designed for 20, the chamber burst. It was not a very auspicious start for the new facility, which continued to be plagued with propellant system, control, and instrument problems for several months. By mid-March 1958, fluorine was being used in the new facility and in the first week of May, liquid hydrogen. The climax to the series of facility problems also came in May when an experiment with gaseous hydrogen and liquid fluorine was conducted. Three successful runs were made and during a pause, with the propellant tanks still at high pressure, fluorine demonstrated its reactivity. A slight leak in a flanged joint at the top of the tank allowed fluorine to escape, and it immediately found substances with which to react. These reactions quickly heated the heavy stainless steel flange and pipe until they also reacted with the fluorine and with a swoosh, a column of fluorine shot upward, reacting with everything in its path, including water vapor in the air.24 Fortunately, a wind quickly dispersed the fluorine compounds. The joint that leaked contained a soft aluminum seal that had been thermally cycled many times over a period of months with no leakage. It had been tested just prior to the ill-fated experiment and found satisfactory. These kinds of problems are normal in research where new ground is being plowed. The flange problem was solved by using welded joints, but the accident and the subsequent delay caused a shift in research plans. Work with hydrogen-fluorine at 22 kilonewtons in Cell 22 was proceeding well, and a decision was made to concentrate on hydrogen-oxygen at the new facility.
As regeneratively-cooled thrust chambers at 89 kilonewtons were not available, the first series of tests with gaseous hydrogen-liquid oxygen was made with uncooled chambers. The gaseous hydrogen was no handicap in this situation for it simulated the same physical state at the injector as liquid hydrogen after absorbing heat in a regeneratively-cooled jacket. Nineteen runs were made during 1958, with performance ranging between 94 and 99 percent of theoretical.25
Meanwhile, the 89 kilonewton, regeneratively-cooled engine became available and by June 1959, 32 runs were made with liquid hydrogen-oxygen. Run times varied up to 102 seconds and all showed satisfactory cooling and high performance. Later, an additional 14 runs, equally successful, completed the investigation and the results were reported in April 1960. Exhaust velocities up to nearly 3300 meters per second were obtained with a nozzle designed for sea-level operation, so even higher velocities were possible with a larger nozzle and operating at simulated altitudes. The investigators used a small quantity of gaseous fluorine flowing ahead of the liquid oxygen to spontaneously ignite with the hydrogen and provide a smooth start.26 Figure 50 shows  a comparison of the 22 and 89 kilonewton, rege ne rat ively-cooled engines used for the liquid hydrogen-fluorine and liquid hydrogen-oxygen experiments during the 1957-1959 period and later.
The injectors designed for both propellant combinations at both thrust levels followed the general concepts agreed upon at the August 1957 design conference (p.89). They generally employed a large number of elements to promote vaporization and intimately mix the propellants. Some were impinging jets in triplet or doublet arrays, some were showerheads with either parallel or converging jets. All gave high, satisfactory performance. Figure 51 shows one of the injectors used with the 89 kilonewton engine.
The greatest value of research of this type is in advancing technology and getting someone else interested in using it for further advances. In the latter, the Lewis rocket research on hydrogen during the 1950s made two contributions. First, it influenced the views of Abe Silverstein, who began planning the NASA spaceflight program in the spring of 1958. He was greatly interested in hydrogen as a fuel, not only for rockets but for other applications. Silverstein followed the rocket work even after he went to Washington, and the May 1958 fluorine accident, which occurred while he was still commuting to the laboratory on weekends, reinforced his conviction that the performance gain by using the denser hydrogen-fluorine combination over hydrogen-oxygen was not worth the additional problems. The regeneratively-cooled hydrogen-oxygen operations at 89 kilonewton thrust in 1959 further convinced him that hydrogen-oxygen was by far the most attractive of the several propellant candidates.....
....for high-performance rocket stages. His convictions were to have an important bearing on decisions made at the end of 1959, decisions that have determined the course of space vehicles to this day.
The second value of the Lewis hydrogen research was the influence it had on other rocket engineers. During 1959, 92 people from 42 organizations made 60 visits to the Lewis rocket laboratory. While not all were interested in hydrogen, the two major rocket engine manufacturers, Rocketdyne and Aerojet, each made three visits; Pratt & Whitney, with a go-ahead in August 1958 from the Air Force to develop a hydrogen-oxygen rocket engine for flight, made three visits during 1959. In fact, Pratt & Whitney representatives began visiting the Lewis rocket laboratory in 1957, much to the surprise of the laboratory officials who had previously found the company aloof when it came to exchanging information about aircraft engines.27