The Forgotten History of Vacuum: From Torricelli's Tube to the Space Between Atoms

For 2000 years, philosophers insisted nature abhors a vacuum. Aristotle laid down the doctrine in the fourth century BCE: a true void cannot exist, because if it did, motion through it would happen instantaneously, which was, on his physics, impossible. The doctrine survived through medieval scholasticism and into the early modern period as an article of faith more than a testable claim. When the apothecaries of Florence noticed in the 1630s that water pumps could not raise water more than about ten meters, the response of the Aristotelian establishment was to declare that nature's abhorrence of vacuum had a numerical limit, which it had reached. This is what the history of physics looks like when a paradigm is dying: epicycles, not retreat.

The man who broke the paradigm was Evangelista Torricelli, who had been Galileo's secretary in his last months and had inherited his interest in the limits of pumping. In 1643, Torricelli filled a glass tube about a meter long with mercury, sealed one end, inverted it into a dish of more mercury, and watched. The mercury fell to a height of about 760 millimeters and stopped. Above it was something that, by Aristotle's lights, could not exist: a region of emptiness, sealed against the air below.

The interpretation was the harder part. Torricelli's first instinct, which he wrote about in a letter to a colleague, was that the mercury was held up by the weight of the atmosphere pressing down on the dish. The void at the top of the tube was real, and what kept the mercury suspended was the same thing that lets you suck water through a straw. This was a complete inversion of Aristotle: nature was not abhorring the vacuum, the atmosphere was simply pushing the mercury up to balance its own weight. The vacuum was the absence of push.

Pascal sends his brother-in-law up a mountain

Blaise Pascal, who had been working on the same problem in France, saw that Torricelli's interpretation made a testable prediction. If atmospheric weight was holding the mercury up, then less atmosphere should hold it up less. Pascal could not climb a mountain himself — he was already in poor health — so he wrote to his brother-in-law Florin Périer, who lived in Clermont-Ferrand near the 1465-meter Puy de Dôme. In September 1648, Périer carried a barometer up the mountain and back down, taking measurements at each step, while a duplicate barometer remained at the base for reference. The mercury column fell by about 75 millimeters at the summit. The atmospheric weight hypothesis was confirmed in a single afternoon's hike.

This experiment did more than vindicate Torricelli. It established that vacuum was a quantitative phenomenon, that pressure was a real physical quantity, and that experimental physics could resolve metaphysical questions that 2000 years of armchair argument had not. The Pascal of the mountain experiment is the same Pascal who would later write the Pensées, and the unity is not coincidental: he was working out what could be known by reason alone and what required measurement, and the barometer was an early demonstration of the second category.

Otto von Guericke and the Magdeburg hemispheres

The vacuum was no longer impossible, but it was hard to make on demand. Otto von Guericke, the mayor of Magdeburg in central Germany, took up the problem in the 1650s as essentially a public spectacle project. He invented the first vacuum pump, a piston-and-cylinder device that could pump air out of a sealed vessel rather than just demonstrate vacuum at the top of a mercury column.

His most famous demonstration came in 1654 at the Reichstag in Regensburg. He fitted two large copper hemispheres together with a leather gasket, pumped the air out, and challenged teams of horses to pull them apart. Sixteen horses, eight on each side, could not separate the hemispheres held together by nothing but atmospheric pressure on the outside. The Magdeburg hemispheres became the most famous physics demonstration of the 17th century, illustrated in textbooks for the next two hundred years.

Guericke's pump was the first instrument that let experimenters work with vacuum the way they had previously worked with mercury. Robert Boyle, in England, took up the pump and built a much better version with Robert Hooke as his assistant. Boyle's air pump produced the experiments that would establish the gas laws — that volume and pressure are inversely related, that air is compressible, that breathing requires air, that sound does not travel through vacuum, that combustion stops without air. The vacuum, in other words, became the laboratory in which the nature of air itself could be studied by removing it.

The luminiferous ether and the second wrong vacuum

The 17th century settled the question of whether vacuum exists. The 19th century reopened the question of what vacuum is, in a way that turned out to be wrong on roughly the same scale as the Aristotelian original.

By the 1860s, James Clerk Maxwell had unified electricity, magnetism, and light into a single theory, with the speed of light falling out of the equations as the propagation speed of electromagnetic waves. But waves had to propagate through something. Sound waves needed air. Water waves needed water. What did light waves need? The answer, agreed by virtually every physicist of the late 19th century, was the luminiferous ether: an invisible, massless, frictionless medium that filled all of space and through which light propagated. The vacuum, in this picture, was full of ether — an absence of ordinary matter but not of the substrate that made physics possible.

The ether had to have strange properties. It had to be rigid enough to support the high-frequency oscillations of light (rigidity scales with stiffness, and light is very fast), yet so insubstantial that planets could move through it without slowing down. It had to be everywhere yet undetectable by any experiment that did not specifically look for it. The Michelson-Morley experiment of 1887 specifically looked for it, by attempting to measure the Earth's motion relative to the ether using the interference of light beams traveling in perpendicular directions. They found nothing. The Earth, against all expectation, was not moving relative to the ether.

The interpretation took eighteen more years. Einstein's 1905 special relativity dispensed with the ether entirely: the speed of light was the same in every reference frame, and there was no medium in which it propagated. The vacuum was, again, empty — but in a deeper way than Torricelli could have meant. It was empty even of the substrate that earlier physics had insisted must exist.

The quantum vacuum is full

The 20th century then discovered that even Einstein's empty vacuum is not really empty. Quantum field theory describes the vacuum as the ground state of all possible fields, with virtual particle pairs constantly appearing and annihilating, with measurable energy density (the cosmological constant), and with effects like the Casimir force in which two parallel plates in vacuum attract each other because the vacuum between them has fewer allowed electromagnetic modes than the vacuum outside.

The Casimir effect was predicted in 1948 and measured precisely in 1997. It is the most direct empirical evidence that the quantum vacuum has structure. Two metal plates a micrometer apart in a perfect vacuum experience a measurable inward force on the order of nanonewtons, caused entirely by the difference in vacuum fluctuations on the inside versus outside of the gap. The vacuum is not nothing. It is the lowest-energy configuration of fields that fill all of space, and that lowest energy is not zero.

This is the third revision in the history of vacuum. Aristotle said the vacuum could not exist. Torricelli showed it could and was simply the absence of air. Maxwell's contemporaries said it was full of ether. Einstein removed the ether. Quantum field theory put structure back into the empty space — not a substance, exactly, but a set of fields whose ground state has measurable consequences. We have come most of the way back to "the vacuum is full," but in a sense Aristotle could not have imagined.

What the history shows

The history of vacuum is the history of the slow recognition that emptiness is not what we thought it was, and that the question keeps reopening at finer scales. Each generation thinks it has the final answer. Aristotle's plenum was certain for two thousand years. Torricelli's empty space was certain for two hundred. The luminiferous ether was certain for fifty. The quantum vacuum has been certain for about seventy years and may yet have surprises.

The deeper lesson is about how physics actually advances. It does not advance by armchair argument from first principles, despite Aristotle's confident attempt. It advances when someone builds an instrument that lets a question be answered by measurement: a glass tube full of mercury, a vacuum pump, an interferometer, a parallel-plate Casimir apparatus. The instruments are the slow accumulation. The theory follows. And the theories, plural, are always provisional in ways that the next instrument will reveal. We are still finding out what nothing is.