The power of “sometimes”

If you owned a computer that only “sometimes” turned on, placed Amazon orders that only “sometimes” arrived, and had friends who only “sometimes” kept their promises, you’d live in a state of frequent frustration. Fortunately for most us, we don’t live in such a world. All around us, products and processes have been designed to be reliable.

However, when it comes to electronic and magnetic materials, reliability - at its literal extreme - is not necessarily a good thing. Transistors that are always conducting won’t make a good computer chip. Magnetic bits that always point up won’t make a good hard drive. And photochromic lenses that always stay dark won’t make good transition eyeglasses. Common to each of these technologies are materials that “sometimes” work one way and “sometimes” work another way. This is the power of “sometimes.”

No class of materials is more emblematic of the power of “sometimes” than the semiconductor, which sometimes conducts electricity and sometimes does not. The controllable conductivity of semiconductors forms the foundation of modern electronics, from computer chips to photovoltaics to wireless communications to the laser.

However, the semiconductor was not always such a star material. It wasn’t until 1910 that the term was coined, and even decades afterward, many physicists questioned whether semiconductors even existed. In 1931, future Nobel laureate Wolfgang Pauli wrote to his colleague Rudolf Peierls:

“On semiconductors one should not do any work, that’s a mess, who knows whether there are semiconductors at all!”

His attitude was hardly uncommon at the time. Another physicist of that era complained:

“What are semiconductors good for? They are good for nothing. They are erratic and not reproducible.”

Among scientists, semiconductors had gained a nasty reputation for being difficult to study because their electrical properties seemed to vary uncontrollably from sample to sample, sometimes differing by as much as a factor of a million. Considering that reproducibility is a bedrock principle of the scientific method, it was extremely challenging to study materials that could not be reproduced reliably.

Progress in the field of semiconductors remained sluggish until World War II, when the military’s well-funded desire for radar systems stoked demand for semiconductor diode rectifiers. Around the same time, scientists finally began to reliably control the doping level of semiconductors, as they switched away from self-doping materials like copper oxide and instead focused on impurity-doped materials like germanium and a rather pedestrian element found in dirt and sand: silicon.

Once we had harnessed silicon, the rest was history.

The most interesting technologies are those that sometimes do one thing, and sometimes do another. The most interesting people are those who sometimes do one thing, and sometimes do another. The most interesting stories are those where the plot sometimes goes one way, and sometimes goes another.

Being the same way all the time is boring, and worse, not all that useful.

This is the power of “sometimes.”