The third room · Inside the lampChapter 16 of 21
The arrows of the last chapter were arrows of something. Ask what, and physics gives the strangest and best-tested answer it has: light is a ripple in a field that fills all of space, and so is the electron.
Chapter 16 · Inside the lamp
The Field
What light is a ripple in, and what happens where it meets matter, with the pictures Feynman drew to keep count.
Look first
What is light travelling through?
A wave on a pond is a movement of water. Sound is a movement of air. For most of the nineteenth century physicists took it for granted that light was a movement of something too, and called the something the aether. In 1887 Albert Michelson and Edward Morley looked for the Earth’s motion through it, with an instrument easily fine enough to find it. They found nothing.
The answer that replaced the aether is stranger. Light does not travel through a substance. There is a field: at every point of space a quantity that can be set trembling, and the trembling passes from each point to the next. Space with no light in it is that field at rest.
The notes of chapter 13’s oven are the field’s own ways of trembling, one for every wavelength and direction. The lumps Planck found belong to them: each way of trembling can hold energy only in whole steps of size hf. One step, in one of those ways of trembling, is one photon.
The energies one way of trembling can have, for three colours. The lowest rung is the field with no photons in it, and each rung up is one more photon of that colour.
Commit to a prediction
An empty box
Take a box and pump out every atom. Shield it from all light, and cool its walls to absolute zero, so that not one photon is left in it. Inside, two flat metal plates with no electric charge face each other, a ten-thousandth of a millimetre apart.
Do the plates feel anything?
The pressure drawing two ideal plates together, against the gap between them. Both scales are logarithmic.
The plates are drawn together with a pressure of 13 pascals, about what a film of water 1.3 mm deep presses on a table. Halve the gap and it is sixteen times stronger. Hendrik Casimir predicted it in 1948, at the Philips laboratory in Eindhoven, in a paper of three pages. An experimental confirmation, he wrote at its end, “might be of a certain interest”. In 1997 Steven Lamoreaux measured the force and found it within 5% of the prediction.
The reason Casimir gave is the bottom rung. Even with no photons in it, every way the field can tremble keeps half a step of energy that cannot be taken out. Metal plates forbid some of those ways in the gap between them, and the field outside the plates presses them together. The force, he wrote, “may be interpreted as a zero point pressure of electromagnetic waves”.
Why the first guesses made sense
Empty is a state, not an absence
If you said nothing, you were using the word empty as everyone does: no atoms, no light, so nothing. In this physics, empty is the lowest of the field’s possible states. It has properties, and they depend on what surrounds it.
If you said pushed apart, you pictured something trapped in the gap. It is the other way about. The gap holds less than the space outside it, and it is the outside that wins.
How light meets matter
Feynman’s pictures
The electron is a ripple too, in a field of its own. That is why every electron is exactly like every other: they are all steps of the same field. And the whole of how light meets matter comes down to one event, drawn here as a corner: an electron gives out a photon, or takes one in.
Chapter 15’s rule still holds: list every way a thing could happen, give each an arrow, add the arrows, square the sum. What Feynman supplied in 1948 was a way to list the ways. Each is a small picture built of corners, and each picture stands for one arrow. He called it making diagrams to help analyse the calculation quicker.
Every corner shortens the arrow by the same factor, about 0.085, a number set by the electron’s charge. Squared, it is 1/137. So a picture with two more corners has an arrow 137 times shorter: simple pictures matter most, and each more elaborate one matters less.
Left: the one event, a corner. Right: the two ways to build a photon bouncing off an electron out of two corners. Time runs up the page.
A case you have not seen
Two pictures, one answer
Go back to chapter 12: a gamma ray from caesium, 662 keV, bounces off an electron. Two pictures can be drawn with two corners. In A the electron takes in the incoming photon and later gives out the outgoing one. In B it gives out the outgoing photon first.
Work out picture A alone, and it says that a certain number of gamma rays turn through a right angle: 0.164, in chapter 12’s units. Picture B alone says exactly the same: 0.164.
What do the two pictures say together?
How often a 662 keV gamma ray is turned through each angle: from picture A alone, from B alone, and from both. The dots are chapter 12’s formula, which was checked against measurement.
The arrows are added before anything is squared, and at a right angle the two partly oppose each other, by just enough that the sum is no longer than either. At other angles they oppose each other more or less. Over all directions, adding what each picture gives alone would overstate the true rate by 49%.
Only the curve for both matches what detectors count. There is no experiment that could tell you which picture a particular gamma ray followed, because neither picture is a thing that happened. They are two terms in a sum.
The bench
A photon meets an electron
Press a picture to leave it out of the sum. The curve shows how often the photon is turned through each angle.
The pictures in the sum
Things to try
- Set the energy to green light. What does each picture give alone, at every angle? What do they give together?
- Raise the energy to thallium’s. Which picture sends almost nothing straight back?
- Change the bookkeeping. Which of the three curves stays where it was?
What you can now say
Light is a ripple in a field that fills all of space, and it comes in whole steps, photons. Matter is made of ripples in fields too. To predict what happens when they meet, draw every way it could happen, give each picture an arrow, add the arrows, and square the sum. No single picture is what happened.
The field is not made of anything further that we know of. It is where the explanation, for now, stops.
For the determined reader: the sum, the bookkeeping, and the limits
What is computed
For each choice of the electron’s spin before and after, and of the two photons’ polarisations, each picture is a complex number. In the frame where the electron starts at rest, with p and p′ the electron’s energy and momentum before and after, k and k′ the photons’, and ε and ε′ their polarisations, each standing inside the brackets for the four-by-four matrix built from it:
M = ū′ [ ε′(p + k + m)ε / 2p·k − ε(p − k′ + m)ε′ / 2p·k′ ] u
The first term is picture A and the second is B. The page multiplies out the four-by-four matrices that Dirac introduced for the electron, adds the two terms, squares, and averages over the spins and polarisations. The result equals the formula of Klein and Nishina, which chapter 12 used, to nine figures. A second calculation in the tests, done by a different method, agrees with both.
The bookkeeping
A photon’s polarisation can be described in more than one way without changing anything that can be measured. Physicists call the freedom a choice of gauge. Each picture alone depends on the choice. Their sum does not, and that is a strict test of the arithmetic: the tests change the gauge and require the sum to stay put.
Where this stops
These are only the two simplest pictures. There are infinitely many more, with more corners, and every further pair of corners shortens the arrow 137 times. So the next pictures can be expected to shift the answer for a photon bouncing off an electron by something of the order of one percent. They are not computed here. The electron is treated as free and at rest, as in chapter 12. And the theory covers light and electrons only: the fields of the nucleus, and gravity, are other stories.
One more case
The electron’s magnetism
An electron is a tiny magnet. The simplest picture of an electron in a magnetic field, a single corner, gives the strength of that magnet as exactly 1 in the natural units. The measured strength is 1.00116.
The difference comes from more elaborate pictures, in which the electron gives out a photon and takes it back again. Julian Schwinger worked out the first of them in 1948. Since then physicists have gone on, order after order: 891 pictures at the fourth order, and 12,672 at the fifth.
How closely does the sum of the pictures match the measurement?
Schwinger’s one picture is already right to 2 parts in a million. Five orders are right to 3 parts in a million million. The sum here leaves out heavier particles, which also flicker in and out of the pictures: their known share is about 4 parts in a million million. With them included, the published prediction and the measurement differ by about 1 part in a million million.
It is often called the most exact agreement between a calculation and a measurement in science. Feynman called the theory the jewel of physics. Schwinger’s first term is carved on his gravestone.
The jewel
The first quantum theory of light and matter together was written by Paul Dirac in 1927. It described an atom giving out a photon, and it worked. But pushed beyond the simplest pictures it gave the same answer to every question: infinity.
For twenty years that stood. Then in June 1947, at a small conference on Shelter Island, off New York, Willis Lamb reported that one energy level of the hydrogen atom sat very slightly above where Dirac’s theory put it. Here was a small, finite number, measured, that the theory ought to give and could not.
Within two years it could. Julian Schwinger and, in Japan, Shin’ichirō Tomonaga found how to absorb the infinities into the measured mass and charge of the electron, and Schwinger computed the electron’s magnetism. Richard Feynman reached the same theory by his own road, the sum over paths of chapter 15, and drew its terms as pictures. Freeman Dyson showed that Feynman’s pictures and the others’ equations were the same theory. Schwinger, Tomonaga and Feynman shared the Nobel Prize of 1965.
In his Nobel lecture Feynman described what he had made as “a kind of handbook on how to do quantum electrodynamics”.
Where this explanation stops
The weight of nothing
The half step that every way of trembling keeps cannot be taken out, and there are ways of trembling at every wavelength, however short. Add up all those half steps and empty space ought to hold an enormous energy. Energy has weight, and the weight of empty space would show in how the universe expands. It does show, faintly. But stop the sum at the shortest length present theories can speak of, and it comes out too large by about 120 factors of ten.
The theory that predicts an electron’s magnetism to a part in a million million gives, for the weight of empty space, what has been called the worst prediction in physics. Nobody knows why.
End of the third room
Added to your notebook
Light is a ripple in a field that fills all of space, and it comes in whole steps, photons. Matter is made of ripples in fields too. To predict what happens when they meet, draw every way it could happen, give each picture an arrow, add the arrows, and square the sum. No single picture is what happened.
And when light lands in an eye, what is made of it?
Three rooms have followed light: out of the lamp, past the edges of sight, and back inside. The fourth follows what happens after it arrives.