Civilization’s toughest technical challenges are those that require extraordinary (and constantly improving) performance to be delivered at a low cost. High levels of performance often require complex, difficult-to-produce technology operating close to the limits of what’s possible. Any contractor could put up a five-story building using off-the-shelf building technology, but a 500-story building would be far more difficult to achieve and would require pushing the boundaries of building technology forward to even be possible.

Trying to make something cheap while you’re pushing the boundaries of performance makes things even more difficult. You need to worry about things like minimizing maintenance costs, eliminating expensive materials or components, and having a design that can be manufactured inexpensively and minimizes costly expert labor. (And if you do require expensive components or labor, you need to spread it as thinly as possible.)

Building and operating a leading-edge semiconductor fab is an example of this sort of intersection. The most advanced semiconductors have features that are 1/2000th the width of a human hair and are made from materials where even a few atoms in the wrong place can cause catastrophic defects, thus requiring incredible levels of control and precision during the production process. This control and precision needs to be achieved consistently and continuously, so that transistors can be made in enormous quantities at very low unit costs.

Developing a new commercial aircraft is another example in this category, as is building a cheap, reusable rocket. And so is building a commercial jet engine. With jet engines, performance and economy are closely bound together. To be attractive to airlines an engine needs to be as efficient as possible, minimizing fuel consumption and the amount of maintenance it requires. High fuel efficiency requires high compression ratios and engine temperatures, which in turn require extremely efficient compressors, components that are both incredibly strong and incredibly lightweight, and materials that can withstand extreme temperatures. And a commercial jet engine must successfully operate hour after hour, day after day, for tens of thousands of hours before being overhauled.

Only a small number of organizations are capable of executing such technically difficult projects. Only three companies in the world operate leading edge semiconductor fabs (Samsung, Intel, and TSMC). Depending on how you count, there are just two to four builders of large commercial aircraft (Airbus, Boeing, Embraer, and now COMAC). With reusable rockets, SpaceX is in a category of its own. The number of participants is small partly because of the inherent technical difficulty and partly it’s because success typically requires spending billions of dollars. It takes on the order of $20 billion to build a leading edge semiconductor fab and anywhere from $10 to $30 billion or more to develop a new commercial aircraft.

We see something similar with large commercial jet engines. Only a handful of companies produce them: GE (both independently and via CFM, its partnership with France’s Safran), Pratt and Whitney, and Rolls-Royce. Developing a new engine is a multi-billion dollar undertaking. Pratt and Whitney spent an estimated $10 billion (in ~2016 dollars) to develop its geared turbofan and CFM almost certainly spent billions developing its LEAP series of engines. (As with leading edge fabs and commercial aircraft, the technical and economic difficulty of building a commercial jet makes it one area of technology where China still lags. China is working on an engine for its C919, but hasn’t yet succeeded.)

It’s not that building a working commercial jet engine itself is so difficult. It’s that a new engine project is always pushing the boundaries of technological possibility, venturing into new domains — greater power, higher temperatures, higher pressures, new materials — where behaviors are less well understood. Building the understanding required to push jet engine capabilities forward takes time, effort, and expense.

The jet engine was invented in the 1930s, independently by Frank Whittle in the UK and Hans von Ohain in Germany. At the time, propeller-driven piston-powered aircraft were capable of traveling around 300-400 miles per hour, but performance gains were getting harder and harder to achieve. At higher speeds, the propeller tips would begin to exceed the speed of sound, creating shockwaves and increasing drag. It was also becoming harder to increase the power of piston engines — as engines got more powerful, they got heavier, offsetting gains in power. Both Whittle and Ohain hypothesized that with an alternative propulsion system, aircraft could achieve much higher speeds and altitudes, and they began to investigate alternatives, ultimately converging on a gas turbine-based propulsion system, or jet engine.

The jet engine is a type of heat engine: it converts heat into useful work. Like a steam turbine or an internal combustion engine, the jet engine works by taking some working fluid (in this case air), compressing it, heating it, and then expanding it, extracting work from the heated fluid in the process.

More specifically, a jet engine operates on the Brayton cycle. Air is taken into the front of the engine, then run through a compressor, increasing the air’s pressure. This compressed air flows into a combustion chamber, where it’s mixed with fuel and ignited, producing a stream of hot exhaust gas. This exhaust gas then drives a turbine, which extracts energy from the hot exhaust as it expands, converting it into mechanical energy in the form of the rotating turbine. This mechanical energy is then used to drive the compressor at the front of the turbine.

In a gas turbine power plant, all the useful work is done by the mechanical energy of the rotating turbine. Some mechanical energy drives the compressor, while the remaining energy drives an electric generator. In a jet engine, the energy is used differently: some energy drives the compressor via the turbine, but instead of using the remaining energy to generate electricity, a jet engine uses it to create thrust through hot exhaust gases, pushing the aircraft forward the same way an inflated balloon propels when air rushes out of it.

Building a functional jet engine requires several key supporting technologies. One such technology is the compressor. In a Brayton cycle engine, roughly 50% of the energy extracted from the hot exhaust gas must be used to drive the compressor (this fraction is known as the back work ratio). Because the back work ratio is so large (a steam turbines has a back work ratio closer to 1%), any losses from compressor inefficiencies are proportionally very large as well. This means that a functional jet engine needs turbines and compressors that transfer as much energy as possible without losses. Whittle was successful partly because he built a compressor that ran at 80% efficiency, far better than existing compressors. Many contemporaries believed Whittle would be lucky to get 65% efficiency — jet engine designer Stanley Hooker noted that he “never built a more efficient compressor than Whittle”.

Another important advance was in turbine materials. The fuel in a jet engine burns at thousands of degrees, and the turbine needs to be both strong and heat-resistant to withstand the rotational forces and temperatures. Whittle’s first engine used turbine blades of stainless steel, but these failed frequently and it was realized that stainless steel wasn’t good enough for a production engine. The first production engines used turbine blades made of Nimonic, a nickel-based “superalloy” with much higher temperature resistance. As we’ll see, the need to drive engine temperatures higher and higher has pushed for the development of increasingly elaborate temperature resistant materials and cooling systems.

Von Ohain’s jet engine first flew in August of 1939 (powering the Heinkel He-178), and Whittle’s first flew in May 1941 (on the E.28/39). By the end of WWII both the UK and Germany had fielded several jet powered aircraft including the Gloster Meteor, the de Havilland Vampire, the Messerschmitt Me-262, and the Arado Ar234.

The US initially lagged in jet engine development, and the first US jet engine efforts relied heavily on existing British designs. The first jet engine built in the US, the GE 1-A, was a copy of Whittle’s W.2B/23 engine; the first mass-produced US engine, the J31, was derived from it. The prototype of the first successful US jet fighter, Lockheed’s P-80 Shooting Star, used a British H-1 “Goblin” engine (though the production version used American J33 engines).

But the US military quickly recognized the potential jet engines had for aircraft performance and began to pour huge amounts of money into companies like GE, Westinghouse, and General Motors to both build improved versions of British engines and begin developing their own models. The first US-designed jet engine, the Westinghouse J30, ran in 1943, and by 1950 the military had funded more than a dozen jet engine projects from a variety of manufacturers.

As money flowed into jet engine development, engine performance and technology quickly improved. Britain’s first production engine, a Rolls-Royce version of Whittle’s W.2B known as the Welland, produced 1600 pounds of thrust when it first ran in 1942. By 1950, the Pratt and Whitney J57 was producing 17,000 pounds of thrust. Thanks to better temperature-resistant materials, the turbine inlet temperature in the J57 was hundreds of degrees higher than in the Welland. And while early jet engines used centrifugal compressors that used a rotating impeller to push the air out to the sides, there were limits to how much these could pressurize air. By 1950 jet engines, including the J57, had almost universally changed to axial compressors, which compress the air along the length of the engine through a series of compression stages.

In addition to using axial compressors, engines further improved with the adoption of two-spool designs. The core of a basic axial compressor jet engine has basically a single large moving part: a rotating shaft with a turbine mounted on one end and a series of compression stages mounted to the other. With this arrangement, every part of the turbine and compressor spins at the same speed, but it’s often beneficial to have different parts of the engine spinning at different speeds: The low pressure front and high pressure rear parts of the compressor work most efficiently at different speeds, and it’s easier to maintain smooth airflow through the engine across a range of different conditions (and avoid things like compressor stall) if different parts of the compressor can rotate at different speeds.

One way of allowing different parts of the engine to rotate at different speeds is to add a second, inner shaft to the engine that rotates independently. The J57 was the first two-spool turbojet, and today most commercial jet engines use two or even three spools. Two-spool engine designs made it possible to achieve higher compression ratios and thus higher engine efficiencies. The Welland had a pressure ratio of around 4:1, while the J57 had a ratio of around 11.5:1.

But while by the 1950s jet engine performance had greatly increased, this was a hard-won achievement. Jet engines required a far greater understanding and mastery of the motion and behavior of hot gasses, and it took much more time to design them than piston engines. An executive at Pratt and Whitney noted that the J57 took on the order of 1.3 million man hours of design time, roughly twice as much as a piston engine did. And while piston engines could be made from comparatively thick and sturdy castings and forgings, much of a jet engine was made from thin sheets of exotic alloys carefully bent into shape, which required novel and complex manufacturing techniques. From “Dependable Engines: The Story of Pratt & Whitney:”

…the jet only became possible because of the development of alloys for that sheet metal that could handle higher temperatures and pressures. And Pratt had to figure out how to weld sheet metal. ‘With the multiplicity of joints in sheet metal parts of a jet, the distribution of stresses is one of the most important considerations. A weld becomes an actual design factor rather than a mere fastening device,’ Horner said.

He referred to many of the issues in converting to jets: things like relatively large diameter parts with very thin walls and all of the compressor and turbine components and airfoils with ‘a great variety of aerodynamic shapes of such awkward dimensions that our designers often complain that they have neither a beginning nor an ending.’

He called these ‘odd horses and peculiar cats in strange contrast to the piston engine’s comfortable old forgings and castings which were heavy and sturdy and supplied their own rigidity for machining.’ And all that sheet metal and oddly shaped stuff needed a lot of tools. To build the little J30, Pratt needed 5250 tools. By 1952 when Horner spoke, the J57 had 20,000 tools.

This complexity made developing a new jet engine difficult and time consuming. Early, centrifugal-compressor-based jet engines came together comparatively easily. The Rolls-Royce Nene, the most powerful aircraft engine in the world when it debuted, took just 5 months to go from starting design to a working engine. Axial-compressor-based engines, though capable of higher performance, were far more difficult to design and build because it was hard to ensure that the air flowed properly through the many different compression stages. Early jet engine designer Stanley Hooker noted that Rolls-Royce’s axial compressor engine that followed the Nene, the Avon, took seven years to iron out the various problems in it. GE’s first axial compressor engine, the J35, required a similar amount of time. And difficulty and time consuming meant expensive. Pratt and Whitney’s J57 took on the order of $150 million to develop, roughly $2 billion in 2025 dollars.

Among the successes of engines like the Avon, the J35, and the J57 were numerous failures. Westinghouse, one of the largest manufacturers of steam turbines, received a huge fraction of initial jet engine contracts and was poised to become a major jet engine manufacturer, but the company was unwilling to invest in the R&D necessary to master the new technology, and it struggled