1916 lab accident: How a pen dipped in molten metal helped create modern computer chips

A century ago, a simple laboratory mistake involving a pen and molten metal led to a discovery that would eventually help make modern computers, smartphones and countless electronic devices possible. 

In 1916, Polish chemist Jan Czochralski was working at the AEG laboratory in Berlin, studying how metals crystallise. According to the well-known account, he accidentally dipped his pen into a crucible of molten tin instead of an inkwell. When he pulled it out, a thin metallic filament appeared to hang from the nib. 

That strange filament was not just solidified metal. It was a clue to a way of producing single crystals — materials in which atoms are arranged in a continuous, highly ordered structure. 

A laboratory mistake to a scientific method 

Czochralski began investigating what had happened. He found that the speed at which the object was pulled from the molten metal affected the resulting crystal. He replaced the pen with a capillary and eventually developed a controlled technique for pulling crystals from molten material. 

His original work was focused on measuring the crystallisation rate of metals, rather than making computer chips. The method was first described in a paper submitted in 1916 and published in 1918. 

The technique became known as the Czochralski method. 

At its simplest, the process involves melting a material in a crucible, touching the molten material with a small crystal seed and slowly pulling the seed upward. As it rises, material solidifies onto the seed, preserving its crystal orientation and gradually forming a larger single crystal. 

Rotation, temperature and pulling speed can be carefully controlled to determine the quality and dimensions of the crystal. 

But where do silicon chips enter the story? 

The connection came decades later. As electronics advanced after World War II, scientists needed large, high-quality semiconductor crystals to make transistors. The Czochralski technique was adapted to grow germanium crystals and was subsequently applied to silicon. 

This was a crucial step because semiconductor devices require extremely controlled and uniform material. 

Silicon produced through crystal growth is formed into a large cylindrical single-crystal ingot. That ingot is then sliced into extremely thin discs known as wafers. Electronic circuits are manufactured on the surface of these wafers. 

In simple terms, the journey is: Molten silicon → single-crystal ingot → thin silicon wafer → microscopic electronic circuits → chips 

Why a single crystal matters 

Silicon is useful in electronics because its electrical properties can be precisely controlled. But imperfections and irregularities in the underlying crystal can interfere with semiconductor manufacturing. 

A single-crystal structure provides a highly uniform starting material. That makes it possible to build increasingly complex electronic structures on silicon wafers with extraordinary precision. The Czochralski process eventually became one of the dominant techniques for producing large semiconductor crystals. 

From tin to today's silicon economy 

The irony of Czochralski's discovery is striking. He was not trying to invent a way to manufacture microprocessors. He was studying the behaviour of metals when a pen dipped into molten tin revealed a phenomenon worth investigating. 

More than a century later, the same basic principle — carefully pulling a crystal from molten material — is central to producing the large silicon crystals from which semiconductor wafers are made. 

Modern crystal-growing systems are vastly more sophisticated than Czochralski's early experiments. They precisely control temperature, rotation, pulling speed, atmosphere and other parameters. The resulting silicon crystals can be produced at industrial scale and then sliced into wafers for semiconductor manufacturing.