MOMENT OF INERTIA

Set a solid sphere and a hollow sphere of the same mass and radius at the top of a ramp, and let go. The solid sphere reaches the bottom first. Every time. This is not a trick of materials — both are rolling without slipping, both are under the same gravity, both start with the same potential energy. And yet one wins.

The reason is moment of inertia. Both bodies convert their initial m·g·h of gravitational potential energy into motion. But "motion" includes two flavours: the translational KE of the center of mass (½·m·v²) and the rotational KE of the body about its center (½·I·ω²). The no-slip condition ties v to ω by v = ω·R, so ½·I·ω² = ½·(I/R²)·v². The total KE is

The factor I/(m·R²) is pure geometry — it's what we'll call the rolling coefficient. For a solid sphere it's 2/5; for a hollow sphere 2/3. Same mass, same radius, same energy budget — but the hollow sphere has to stash proportionally more of its KE in rotation, leaving less for translation, so its center of mass ends up moving slower.

That is this topic in miniature: moment of inertia is what decides how mass distributes its rotational effort, and how it distributes it depends on where the mass sits relative to the axis.

Five bodies, same mass, same radius, same ramp. The sliding block (red) doesn't rotate at all, so every joule of PE becomes translational KE — it wins. A solid sphere (2/5) is second. The thin hoop (1) is last: all its mass is at maximum radius, so I = m·R², and it spends half its KE on rotation. The ordering is strict: small rolling coefficient → fast; large rolling coefficient → slow. None of the bodies care about their total mass (the m cancels out of the acceleration), only about how that mass is distributed.

For a continuous rigid body, moment of inertia about an axis is

where r is the perpendicular distance from each mass element dm to the axis. The r² is what makes everything interesting. Double the distance from the axis and the contribution to I goes up by four. So moving mass from near the axis to far away costs you dearly — moment-of-inertia-wise.

A few key results for bodies about their axis of symmetry:

Each one is an integral. The sphere's 2/5 comes from integrating r²·dm in spherical coordinates for a uniform density ρ = m/(4πR³/3); the factor drops out after two pages of algebra. The factor 1/12 for a rod is from integrating x² from −L/2 to L/2 with a constant mass per length. The results are tabulated in engineering handbooks precisely because these integrals come up all the time and nobody wants to redo them.

Bodies don't always rotate about axes through their center of mass. A pendulum swings about its pivot; a wheel may spin about a point off its axle. How does I depend on the choice of axis?

The answer is the parallel-axis theorem, proved by the Swiss geometer Jacob Steiner in the 1840s (and so sometimes called Steiner's theorem): the moment of inertia about any axis equals the moment of inertia about the parallel axis through the center of mass, plus M·d² for the parallel displacement d.

It is astonishingly useful. A thin rod of mass m and length L has I_CM = m·L²/12 about its centre. About one end (d = L/2), the parallel-axis theorem gives I = m·L²/12 + m·(L/2)² = m·L²/3. That is the formula you need for a rod swinging as a pendulum from one end — and you get it from the centre-of-mass result in one line, without having to redo the integral.

The theorem is a direct consequence of the definition of the center of mass: the first moment of the mass distribution about the CM is zero, so when you expand ∫(r − d)² dm = ∫r² dm − 2·d·∫r dm + ∫d² dm, the middle term vanishes, and you're left with I_CM plus a pure translational piece M·d². The arithmetic is clean; the conceptual payoff is enormous.

The inner plate is rotating about its own centre of mass; the outer one is rotating about a parallel axis displaced by d. Same body, same ω. The difference in the paths' size is the M·d² term — a pure translational contribution tacked onto I_CM.

For every rigid body and axis pair, there exists a single distance k — the radius of gyration — such that

Equivalently, k = √(I/m). It is the distance from the axis at which a point mass of the body's total mass would produce the same moment of inertia as the actual distributed body.

For a solid sphere, k = R·√(2/5) ≈ 0.632·R. For a hollow sphere, k = R·√(2/3) ≈ 0.816·R. For a hoop, k = R exactly. The radius of gyration is a compact single-number description of how spread-out a body's mass is — useful shorthand in engineering for comparing flywheels, rotors, and structural beams. Two beams with the same mass but different k will deflect differently in bending; two flywheels with the same mass but different k will store different amounts of rotational energy at the same ω.

Both bodies above spin identically. The bar has its mass smeared along its length; the phantom ring has the same total mass concentrated at a single radius k. They have identical I, so they have identical angular momentum and identical rotational energy at any ω — the ring is literally what the bar "feels like" to the dynamics.

The radius of gyration is also a useful abstraction when you want to treat an extended body like a point particle. For a body rolling without slipping, the total KE is ½·m·v²·(1 + k²/R²). All the complicated details of the body's shape are bundled into a single number, k/R, the ratio of the radius of gyration to the outer radius.

Everything so far has been about rotation about a single fixed axis. For rotation about an arbitrary axis of a three-dimensional body, I is not a single number but a rank-2 tensor — the inertia tensor — with nine components:

Given an angular velocity vector ω, the angular momentum vector is L = I·ω — meaning L may not be parallel to ω. That's the striking generalisation: in general, spinning a 3-D body about an arbitrary axis produces angular momentum pointing in a different direction from the spin. You can feel this: spin a book about its long axis and it's smooth; spin it about its short axis and it's smooth; spin it about the intermediate axis and it wobbles wildly. The wobble is the body's response to the mismatch between ω and L.

For every rigid body there are three special directions called the principal axes — mutually perpendicular, fixed in the body — about which I is diagonal. Spin the body about a principal axis and L is parallel to ω; there is no wobble. The three diagonal entries of the inertia tensor in the principal frame are called the principal moments of inertia, often written I₁, I₂, I₃. For symmetric bodies (cube, sphere, cylinder) the principal axes line up with the obvious symmetry directions. For asymmetric bodies they are harder to find, but they always exist.

The ellipsoid's three semi-axes are scaled inversely by √I — the longest visual direction is the axis of smallest moment of inertia, the direction the body prefers to tumble about. This is the tennis-racket theorem in geometry: the intermediate axis is unstable, and the other two are stable, which is why a spinning book is smooth two ways out of three.

This is the business end of rigid-body dynamics. The Euler equations (1750s) — which we met briefly in the angular-momentum topic — are the rotational analogue of F = m·a written in terms of the principal moments, and they govern every spinning top, wobbling tennis racket, and tumbling Hubble Space Telescope. They are the tool we will use in FIG.11 to understand why gyroscopes behave the way they do, and in FIG.12 to explain why the Earth itself wobbles on its axis.

Moment of inertia is everywhere in engineering and physics.

Flywheels. A flywheel stores rotational kinetic energy ½·I·ω². For a given ω, bigger I means more stored energy. For a given stored energy, bigger I means lower ω, which reduces bearing stresses. Flywheel-battery engineers push mass out to the rim and use carbon-fibre rotors that can survive high tip speeds — a hundred-metres-per-second rim velocity is typical. A 100 kg cylinder flywheel 1 m in diameter at 10,000 rpm stores about 6 kWh — enough to run a hot water heater for a few hours.

Engines. A reciprocating engine's crankshaft has a large flywheel bolted to one end. Its purpose is to smooth out the torque pulses from the cylinders: combustion in cylinder 1 delivers a huge torque impulse, then nothing, then cylinder 2, then nothing. A heavy flywheel averages all this out and keeps the crankshaft's angular velocity nearly constant between power strokes. Without it, the engine would lurch instead of spin.

Sports equipment. A baseball bat with a long "sweet spot" has its moment of inertia designed around the bat's swing pivot (wrist or handle): too much I and the swing is sluggish; too little and the bat's response to off-centre hits is jarring. Golf clubs, tennis racquets, and hockey sticks are all sold partly on the basis of their MOI characteristics ("balance point", "polar moment") because feel in the hand depends on the ratio of I to the grip torque the player can apply.

Biomechanics. A runner bends their knees to reduce the moment of inertia of the leg about the hip. A diver tucks into a ball to decrease I and increase angular velocity. A tightrope walker extends a long balancing pole horizontally to dramatically increase the I of their system about the tightrope, so any pitching torque from a stumble produces only a small angular acceleration — giving them time to correct.

Astronomy. A star collapses into a neutron star 20 km across. I drops by a factor of about 10¹⁰; conservation of L forces ω up by the same factor; a 30-day rotation becomes a millisecond spin. These are pulsars. A planet's moment of inertia ratio I/(m·R²) can be measured from its polar flattening and its nutation period; Mars's ratio (about 0.366) is lower than Earth's (0.330), reflecting subtle differences in interior mass distribution.

Every place rotation happens, moment of inertia is the quantity that controls it. Learn the integrals once and you have half the toolkit for rigid-body mechanics for life.

We now have:

That is enough to tackle the strangest consequences of rotation. A gyroscope does not fall when pushed; it precesses. A spinning top stands up. The Earth's rotation axis itself moves on two different timescales — a 26,000-year precession and a 433-day wobble. These are consequences of τ = I·α applied to bodies with non-trivial inertia tensors, and they are our subjects for the next two topics. FIG.11 is gyroscopes; FIG.12 is the Earth itself.