Particles

Here’s a question: what are the laws that describe how molecules bounce? It sounds like a niche question, but a bounce is pretty much the smallest unit of everything that happens. Two things meet, something changes hands, and they go their separate ways. Air pressure, heat, friction, sound: zoom in far enough and all of it is just particles bumping into each other.

Molecules are far too small to watch, of course. But a bounce is a bounce, so let’s study bigger ones. All we need is a bunch of metal balls on strings (a Newton’s cradle will do), a ruler and a scale. With just that, we are going to reinvent some of the most fundamental concepts in physics, starting with momentum and force.

1. The law of the pendulum

Before we let anything collide, we need a way to measure how fast the balls are going. Collisions are over in a flash, and we don’t have a stopwatch that’s anywhere near fast enough. But we do have pendulums, and a pendulum turns speed into distance: the faster a ball passes the bottom, the farther it swings out. That’s a start, but “farther” isn’t good enough. Twice the speed might mean twice the swing, or four times, or something else entirely.

So let’s ask the pendulums themselves. Hang two identical balls on strings of equal length, a little apart from each other. Pull one back twice as far as the other, and release both at the same time. Something remarkable happens: they stay perfectly in step. They reach the bottom together, turn around together, and come back together. As long as you keep the swings small compared to the length of the string, a big swing takes exactly as long as a small one. (Galileo supposedly noticed this in church, watching a chandelier sway.)

Try it yourself below. Change the mass, the length of the string or how far the ball is pulled aside, and press Go.

Pendulum 1

Period

Pendulum 2

Period

Watch closely, or take a few photos, and you’ll see something even stronger: at every moment, the big ball is exactly twice as far from the bottom as the small one. The two motions are perfect copies of each other, one simply scaled up by a factor of two. So in any stretch of time, however short, the big ball covers twice the distance. That includes the moment both balls pass the bottom, where our collisions will happen. Pull back three times as far, and the ball passes the bottom three times as fast. That is the law of the pendulum:

speed at the bottom∝swing\text{speed at the bottom} \propto \text{swing}

This means a ruler is all we need to measure velocities. And for the masses, we have the scale.

2. The law of colliding pendulums

Now hang the two balls so that they just touch when they hang still. Pull ball A aside and let go. It swings down, hits ball B at the lowest point, and then both balls swing out. Mark how far each one gets. Thanks to the law of the pendulum, these swings tell us the velocities of both balls right after the collision.

Ball 1
Ball 2

(This is no toy experiment, by the way. In the 1660s, Christopher Wren, John Wallis and Christiaan Huygens worked out the rules of collisions for the Royal Society in London, and Isaac Newton checked them with exactly this setup, using pendulums ten feet long.)

Same balls, different swings

Let’s start with a 1 oz ball A and a 3 oz ball B. Swinging to the right counts as positive, and swinging to the left as negative. We write Δv\Delta v for the change in velocity: the swing after the collision minus the swing before. Here is what you would measure:

A beforeB beforeA afterB afterΔvA\Delta v_AΔvB\Delta v_B
A released from 6 in60−2.12.7−8.1+2.7
A released from 12 in120−4.25.4−16.2+5.4
A released from 18 in180−6.38.1−24.3+8.1
Both released, meet in the middle10−6−11.61.2−21.6+7.2

Look at the last two columns. The harder the balls hit each other, the more both velocities change. But in every row, A’s change is exactly −3 times B’s. The changes always point in opposite directions, and they are proportional to each other:

ΔvA=−k ΔvB\Delta v_A = -k \, \Delta v_B

where kk depends only on which two balls collide, not on how they collide.

Different balls

So what decides kk? Let’s repeat the experiment with different pairs of balls. For each pair, we weigh both balls on the scale, release A from 12 inches, and work out kk from the swings:

Mass AMass BmB/mAm_B / m_A (scale)ΔvA\Delta v_AΔvB\Delta v_Bkk (pendulum)
1 oz1 oz1−10.8+10.81
1 oz2 oz2−14.4+7.22
1 oz3 oz3−16.2+5.43
2 oz3 oz1.5−13.0+8.61.5
3 oz1 oz0.33−5.4+16.20.33
1 oz (clay)2 oz (clay)2−8.0+4.02

Two completely different instruments, and they give the same number every time. The scale knows nothing about collisions, and the pendulums know nothing about weighing, yet kk is exactly the mass ratio. Even the clay balls, which stick together instead of bouncing, obey it. That is the law of colliding pendulums:

ΔvAΔvB=−mBmA\frac{\Delta v_A}{\Delta v_B} = -\frac{m_B}{m_A}

It is the precise version of something everybody already knows. Throw a ping pong ball at a bowling ball and the ping pong ball bounces right back, while the bowling ball barely notices. A ball three times as heavy changes its velocity three times as little.

Momentum: the currency of collisions

Multiply both sides by mAΔvBm_A \Delta v_B, and the law becomes

mAΔvA=−mBΔvBm_A \Delta v_A = -m_B \Delta v_B

Read it like a transaction. Whatever amount of mΔvm \Delta v ball A loses, ball B gains. The velocities don’t change by equal amounts, but the product of mass and velocity does. That product is clearly what’s being traded here, so it deserves a name: momentum,

p=mvp = m v

Momentum is the currency of collisions. In every bounce, some amount of momentum changes hands. Paid to a light ball, that amount buys a big change in velocity. Paid to a heavy ball, the same amount barely makes a dent. But no momentum is ever created or destroyed. It only moves from one ball to the other:

ΔpA+ΔpB=0\Delta p_A + \Delta p_B = 0

Go ahead and check it in the first table: 1×12+3×0=121 \times 12 + 3 \times 0 = 12 before, and 1×(−4.2)+3×5.4=121 \times (-4.2) + 3 \times 5.4 = 12 after. And because every collision in a box full of molecules is such a transaction, the total momentum of the box never changes, however chaotic things get inside.

3. Force

So far we have only looked at before and after. But zoom in on a collision, and it takes some time: the balls squish into each other, and momentum flows from one to the other over a few milliseconds. Our bookkeeping must hold at every instant during that squish, so the total momentum is constant all the way through:

mAvA+mBvB=constantm_A v_A + m_B v_B = \text{constant}

This is true, but a bit annoying to work with. Which constant? That depends on how fast the balls were going before they met. Worse, it even depends on who’s watching: walk alongside the pendulums, and all velocities change, so the constant changes too. The constant tells us nothing about the collision itself.

So let’s get rid of it. A constant doesn’t change over time, so if we take the derivative with respect to time, it simply disappears. And since the masses don’t change either, the derivative of each velocity is all that’s left, which is the acceleration aa:

mAaA+mBaB=0m_A a_A + m_B a_B = 0

Much cleaner. No unknown constant, no dependence on where the balls came from or who’s watching. Just a statement about what happens right now, at every instant of the collision. The product mam a is the rate at which a ball gains momentum, and we call it force:

F=ma=dpdtF = m a = \frac{dp}{dt}

In that language, the law of colliding pendulums becomes as simple as it gets:

FA+FB=0F_A + F_B = 0

Whatever force A feels, B feels the same force in the opposite direction. Force isn’t a mysterious push that comes out of nowhere. It is bookkeeping: the rate at which momentum changes hands.

Why bother with force?

If force is just momentum per second, why not stick with momentum? Because many interactions are not neat little bounces with a clear before and after.

And with that, all three of Newton’s laws come for free:

Two colliding particles

a) Elastic b) Inelastic c) Hyperelastic

Here’s a question: how do molecules bounce? Like perfect tennisballs, or do they lose a bit of velocity on every collision?

Turns out they

Still to come: