Four-Velocity, Four-Momentum, and Mass-Energy Equivalence

Every second, the Sun becomes about four million tonnes lighter. Nothing is being thrown overboard: the lost mass leaves as sunlight, because the energy radiated away and the mass that vanishes are the same physical quantity read in different units. That identity is not a poetic slogan — it is the time component of a four-vector, and this article builds that vector from scratch. The construction answers an awkward question left over from the Lorentz transformation: Newtonian velocity $\mathrm{d}\mathbf{x}/\mathrm{d}t$ and momentum $m\mathbf{v}$ transform badly, because the $t$ in the denominator is itself frame-dependent. The repair is almost embarrassingly simple — differentiate the spacetime position with respect to the proper time $\tau$ , the invariant reading of a clock riding with the particle. The resulting four-velocity inherits clean Lorentz behavior, and multiplying by the invariant mass $m$ produces the four-momentum $p^\mu = (E/c, \mathbf{p})$ : energy and momentum revealed as the time and space faces of a single geometric arrow, unified the way the next articles unify force and power. The payoff is threefold. The fixed Minkowski length of that arrow is the energy–momentum relation $E^2 = (pc)^2 + (mc^2)^2$ . Its low-speed expansion exposes the rest energy $E_0 = mc^2$ hiding beneath Newtonian kinetic energy. And its conservation — one four-component law valid in every inertial frame at once — is the master tool of collider physics, nuclear energetics, and particle decays, generalizing the momentum conservation you met back in the Conservation of Linear Momentum article to the full spacetime stage.

Key takeaways

The four-velocity $U^\mu = \mathrm{d}x^\mu/\mathrm{d}\tau = \gamma(c, \mathbf{v})$ differentiates position with respect to invariant proper time; its Minkowski norm is always $U^\mu U_\mu = c^2$ — every particle "moves through spacetime" at the same fixed rate.
The four-momentum $p^\mu = mU^\mu = (E/c, \mathbf{p})$ packages total energy $E = \gamma mc^2$ and relativistic three-momentum $\mathbf{p} = \gamma m\mathbf{v}$ into one Lorentz-covariant object.
Its invariant length is the mass: $p^\mu p_\mu = (mc)^2$ , equivalent to the energy–momentum relation $E^2 = (pc)^2 + (mc^2)^2$ — the mass-shell condition.
At low speed, $E = mc^2 + \tfrac12 mv^2 + \cdots$ : Newton's kinetic energy rides on top of an enormous constant, the rest energy $E_0 = mc^2$ , which becomes observable whenever mass is allowed to change.
Conservation of total four-momentum is one law that enforces energy and momentum conservation in every frame simultaneously; squaring total four-momenta (the invariant-mass trick) solves collision and decay problems without ever finding a velocity.
Massless particles have no four-velocity ( $\mathrm{d}\tau = 0$ ) but a perfectly good null four-momentum with $E = pc$ — which is why light pushes.
Key idea: Differentiate position with respect to invariant proper time to build the four-velocity $U^\mu$ ; multiply by invariant mass to get the four-momentum $p^\mu=(E/c,\mathbf p)$ , whose fixed Minkowski length $p^\mu p_\mu=(mc)^2$ is exactly the energy–momentum relation $E^2=(pc)^2+(mc^2)^2$ .

Where this lives today

Four-momentum conservation is the daily arithmetic of the Large Hadron Collider. Every recorded collision is reconstructed by summing the four-momenta of thousands of tracks and calorimeter clusters; a resonance is literally a bump in an invariant-mass histogram — the Higgs boson announced itself in 2012 as a peak at $\sqrt{(\sum E)^2 - |\sum \mathbf{p}c|^2} = 125\ \mathrm{GeV}$ , and the High-Luminosity LHC now being commissioned will repeat that squaring operation trillions of times over. Even "missing energy," the signature of neutrinos and dark-matter candidates, is defined as the four-momentum imbalance the detector cannot account for — Pauli's neutrino logic, industrialized. The equivalence itself has become metrology and medicine. PET scanners image tumors by detecting the back-to-back 511 keV photons from positron–electron annihilation — each detection event is Example 2 run in reverse, rest energy converting entirely to null four-momenta. Nuclear power stations monetize the binding-energy curve at about $200\ \mathrm{MeV}$ per fission, and stellar models balance the Sun's books at $4.3 \times 10^9\ \mathrm{kg}$ of mass radiated per second. At the precision frontier, comparing Penning-trap mass differences against gamma-ray energies has confirmed $E = mc^2$ to a few parts in $10^7$ — the best-tested equation a mechanics course will ever hand you.

The physical picture

Imagine a clock glued to the particle. Proper time $\tau$ is what that clock reads, and every observer agrees on it — it is carved into the worldline itself. Differentiating the spacetime position with respect to this universal bookkeeper produces an arrow whose "speed through spacetime" is always exactly $c$ . A particle at rest spends all of that speed moving through time; boosting it tilts the fixed-length arrow toward the spatial directions, trading motion-through-time (that is the time dilation of the previous article) for motion-through-space.

Figure

Everyone moves through spacetime at the same rate — the four-velocity is a fixed-length arrow on a hyperboloid; rest directs it all along the time axis, boosting tilts it spaceward while the Minkowski norm stays c

Energy and momentum feel like separate bookkeeping in Newtonian physics — one scalar guarded by work, one vector guarded by isolation. Relativity shows they are two faces of that single arrow scaled by mass. A Lorentz boost rotates (more precisely, hyperbolically tilts) the four-momentum, trading energy for momentum the way an ordinary rotation trades $x$ for $y$ components. What no boost can change is the arrow's Minkowski length — and that length is the rest mass. Mass is not one more conserved substance riding along; it is the shape constraint on the energy–momentum arrow, the radius of the hyperboloid the arrow must end on.

Figure

The mass shell — allowed energy and momentum states of a particle trace a hyperbola with rest energy as intercept and the light line E equals pc as asymptote; the massless case runs along the asymptote itself

The rest energy $mc^2$ appears because the time component of $p^\mu$ refuses to vanish when the particle stops. A motionless particle still moves through time at rate $c$ , so it carries energy $mc^2$ with zero kinetic energy. Mass is frozen energy: any process that lowers a system's internal energy — binding a proton to a neutron, letting an excited nucleus decay — lowers its mass by exactly $\Delta E/c^2$ . Newtonian experiments never noticed because they only measure energy differences, and a constant offset, however colossal, cancels out. Massless particles break the four-velocity construction — a photon has no rest frame and $\mathrm{d}\tau = 0$ along its path. Yet four-momentum survives as a perfectly good null vector with $E = pc$ . This is why light carries momentum despite having no mass, why photon pressure works, and why the massless limit sits on the asymptote of the mass-shell hyperbola rather than on the curve.

The core equations

Everything in relativistic dynamics grows out of one derivative done correctly. The Newtonian $\mathrm{d}\mathbf{x}/\mathrm{d}t$ divides a piece of a four-vector by a frame-dependent time; dividing the whole displacement $\mathrm{d}x^\mu$ by the invariant proper time instead gives an honest four-vector, and time dilation ( $\mathrm{d}t = \gamma\,\mathrm{d}\tau$ , from the Relativistic Kinematics article) fixes its components:

U^\mu = \frac{\mathrm{d}x^\mu}{\mathrm{d}\tau} = \gamma\,(c,\,\mathbf{v}) \qquad (1)

where: $x^\mu = (ct, \mathbf{x})$ — spacetime position; $\tau$ — proper time along the worldline; $\gamma = 1/\sqrt{1 - v^2/c^2}$ — Lorentz factor; $\mathbf{v}$ — ordinary three-velocity. At rest, $U^\mu = (c, \mathbf{0})$ : the particle still "moves," entirely through time. The $\gamma$ is not decoration — it is the chain-rule factor $\mathrm{d}t/\mathrm{d}\tau$ that repairs the bad transformation of $\mathbf{v}$ . Contract the four-velocity with itself (signature $(+,-,-,-)$ throughout this article) and the speed dependence cancels identically:

U^\mu U_\mu = \gamma^2(c^2 - v^2) = c^2 \qquad (2)

where: $U^\mu U_\mu$ — Minkowski norm. This is the "everyone travels through spacetime at $c$ " statement made exact, and it is the engine behind the energy–momentum relation below: a constant norm survives multiplication by the constant $m$ . Mass now enters as a Lorentz scalar — the same number in every frame, measured once in the rest frame. Scaling the four-velocity by it defines the central object of the article:

p^\mu = m\,U^\mu = \left(\frac{E}{c},\,\mathbf{p}\right) = (\gamma m c,\,\gamma m \mathbf{v}) \qquad (3)

where: $m$ — invariant rest mass; $E = \gamma mc^2$ — total energy; $\mathbf{p} = \gamma m\mathbf{v}$ — relativistic three-momentum. The identification of $cp^0$ with energy is not a definition pulled from a hat: Derivation B shows it is forced by demanding the conserved time component reduce to Newtonian energy at low speed. Note what became of the Conservation of Linear Momentum article's $m\mathbf{v}$ : it survives as the space part, promoted by a factor $\gamma$ . Reading off the components gives the two workhorse formulas of relativistic dynamics:

E = \gamma m c^2,\qquad \mathbf{p} = \gamma m \mathbf{v} \qquad (4)

where: $E$ grows without bound as $v \to c$ — the deep reason no massive particle reaches light speed, since each further increment of speed costs diverging energy; $\mathbf{p}$ reduces to $m\mathbf{v}$ when $\gamma \to 1$ . Dividing the two lines gives the often-useful $\mathbf{v}/c = \mathbf{p}c/E$ , valid for massive and massless particles alike. The invariant norm of $p^\mu$ follows immediately from equations (2) and (3), and it deserves its own name — the mass-shell condition:

p^\mu p_\mu = \frac{E^2}{c^2} - |\mathbf{p}|^2 = (mc)^2 \qquad (5)

where: the left side looks frame-dependent (both $E$ and $\mathbf{p}$ change under boosts) but the combination is a scalar; the right side is built from the invariant mass. Multiplying through by $c^2$ produces the most quoted equation of particle physics,

E^2 = (pc)^2 + (mc^2)^2 \qquad (6)

where: $p = |\mathbf{p}|$ . Geometrically this is a right triangle with legs $pc$ and $mc^2$ and hypotenuse $E$ — momentum and mass contribute to energy in quadrature. It contains $E_0 = mc^2$ at $p = 0$ and $E = pc$ at $m = 0$ as its two edges, and unlike $E = \gamma mc^2$ it never mentions velocity, which is why experimentalists — who measure momenta in magnets and energies in calorimeters — prefer it. At rest the triangle collapses to its vertical leg, leaving the most famous special case in physics:

E_0 = mc^2 \qquad (7)

where: $E_0$ — rest energy. The conversion rate is enormous: one gram of anything is $9 \times 10^{13}\ \mathrm{J}$ , about the energy released by twenty kilotons of TNT. Chemistry rearranges electron bonds and shifts mass at the $10^{-10}$ level — invisible; nuclear reactions shift it at the $10^{-3}$ level — power stations and stars. What Newton called kinetic energy is everything above the rest shelf:

T = E - mc^2 = (\gamma - 1)\,mc^2 \qquad (8)

where: $T$ — relativistic kinetic energy. Its binomial expansion $T = \tfrac12 mv^2 + \tfrac38 mv^4/c^2 + \cdots$ (Derivation B) recovers Newton exactly at leading order — the correspondence-principle check that certifies the whole identification. How the pieces divide the labor:

You are given	Reach for	Because
speed $v$ or $\gamma$	$E = \gamma mc^2$ , $\mathbf{p} = \gamma m\mathbf{v}$	direct kinematic evaluation
momentum from tracking, no velocity	$E^2 = (pc)^2 + (mc^2)^2$	velocity never appears
a collision or decay to analyze	conservation of total $p^\mu$ , then square	invariants kill the unknowns
energy and momentum, want speed	$\mathbf{v}/c = \mathbf{p}c/E$	works even for $m = 0$

Derivation AFixed norm of the four-velocity and the energy-momentum relation

We derive the two invariants that anchor relativistic dynamics: the constant norm of the four-velocity, and — by scaling with mass — the energy–momentum relation. The strategy is pure geometry: compute the Minkowski contraction once in explicit components, watch the speed dependence cancel, then let the scalar mass carry the result over to $p^\mu$ . The only care needed is the metric signature: we use $(+,-,-,-)$ , so spatial squares enter with a minus sign.

Step 1. Contract the four-velocity $U^\mu = \gamma(c, \mathbf{v})$ with itself using the Minkowski metric:

U^\mu U_\mu = \gamma^2 c^2 - \gamma^2 v^2 = \gamma^2\left(c^2 - v^2\right)

Why the minus sign: the metric weighs time and space oppositely; this relative sign is exactly what makes the result invariant under boosts, which mix the two.

Step 2. Insert the definition of the Lorentz factor, $\gamma^2 = 1/(1 - v^2/c^2)$ :

U^\mu U_\mu = \frac{c^2 - v^2}{1 - v^2/c^2} = \frac{c^2\,(1 - v^2/c^2)}{1 - v^2/c^2} = c^2

Why this matters: the norm is independent of $v$ — every massive particle, whatever its speed, has a four-velocity of the same invariant length. Speeding up tilts the arrow; nothing can stretch it.

Step 3. Scale by the invariant mass to get the norm of the four-momentum $p^\mu = mU^\mu$ :

p^\mu p_\mu = m^2\,U^\mu U_\mu = m^2 c^2

Why $m$ passes through the contraction: it is a Lorentz scalar, the same in every frame, so it simply factors out squared.

Step 4. Expand the same norm in components $p^\mu = (E/c, \mathbf{p})$ and equate:

\frac{E^2}{c^2} - |\mathbf{p}|^2 = m^2c^2 \quad\Longrightarrow\quad E^2 = (pc)^2 + (mc^2)^2

Step 5. Cross-check with the explicit components $E = \gamma mc^2$ , $p = \gamma m v$ :

\frac{E^2}{c^2} - p^2 = \gamma^2 m^2 c^2 - \gamma^2 m^2 v^2 = \gamma^2 m^2 (c^2 - v^2) = m^2 c^2

Two boundary checks confirm the relation: at $p = 0$ it returns $E = mc^2$ , and at $m = 0$ it returns $E = pc$ . The electron example at $v = 0.8c$ worked below gives $0.6813^2 + 0.511^2 = 0.7253 = 0.8517^2\ \mathrm{MeV^2}$ to four figures.

Go deeper · Derivation B · 1 minRest energy and the Newtonian limit of the energy

Derivation BRest energy and the Newtonian limit of the energy

We justify calling $cp^0$ the energy — the step that turns geometry into physics. The strategy is the correspondence principle: assume only that the time component of the conserved four-momentum is some conserved scalar function of speed, expand it at $v \ll c$ , and demand that it reproduce the Newtonian kinetic energy. The expansion then hands us, unasked, a constant term with no Newtonian counterpart: the rest energy. Watch for why the constant cannot be discarded the way constants usually are.

Step 1. Write the time component of $p^\mu = m\gamma(c, \mathbf{v})$ and scale by $c$ to give it energy units:

cp^0 = \gamma m c^2 \equiv E

Why this needs justifying: any monotonic function of $\gamma$ has the same invariance properties; only the low-speed limit can single out which one deserves the name energy.

Step 2. Expand the Lorentz factor binomially for $v \ll c$ :

\gamma = \left(1 - \frac{v^2}{c^2}\right)^{-1/2} = 1 + \frac{v^2}{2c^2} + \frac{3v^4}{8c^4} + \cdots

Step 3. Multiply through by $mc^2$ :

E = mc^2 + \frac12 mv^2 + \frac{3}{8}\frac{mv^4}{c^2} + \cdots

Why this settles the identification: the second term is exactly Newtonian kinetic energy, so the conserved time component is the relativistic upgrade of energy conservation — with a constant shelf $mc^2$ underneath and computable corrections above.

Step 4. Read off the rest energy by setting $v = 0$ :

E_0 = mc^2

Why the constant is physical here when constants are usually gauge: in Newtonian mechanics only energy differences at fixed mass are observable, so $mc^2$ hides. But in any process where $m$ changes — nuclear binding, particle creation, annihilation — the shelf height itself changes, and the difference is released as ordinary energy. The constant is only unobservable while mass is frozen.

Step 5. Quantify the first correction: the $\tfrac38 mv^4/c^2$ term reaches one percent of the Newtonian term when $\tfrac34 v^2/c^2 = 0.01$ , i.e. at $v \approx 0.115c \approx 3.5 \times 10^4\ \mathrm{km\,s^{-1}}$ . As a dimensional check, every term in the expansion is mass times velocity squared. The deuteron example confirms that the rest-energy shelf shifts by exactly the binding energy, $2.224\ \mathrm{MeV}/c^2$ , consistent with mass-spectrometry values.

Go deeper · Derivation C · 1 minThreshold energy from four-momentum invariants

Derivation CThreshold energy from four-momentum invariants

We compute the minimum beam energy that makes a reaction possible — the threshold — using nothing but the conservation and invariance of four-momentum. The strategy is the single most-used trick in relativistic kinematics: square the total four-momentum, because the square is a Lorentz scalar you may evaluate in whichever frame is most convenient — the lab before the collision, the center-of-momentum frame after. A projectile of mass $m$ and lab energy $E$ strikes an identical particle at rest, and the final state has total rest mass $M$ . Watch how no velocity is ever computed.

Step 1. Add the two initial four-momenta in the lab frame, where the target sits at rest:

P^\mu_{\mathrm{tot}} = \left(\frac{E}{c} + mc,\ \mathbf{p}\right)

Why add: four-momentum is conserved component by component, so the total before equals the total after — and the total is itself a four-vector we may square.

Step 2. Square the total, using the projectile's own mass-shell relation $E^2/c^2 - p^2 = m^2c^2$ :

P^\mu_{\mathrm{tot}} P_{\mathrm{tot}\,\mu} = \left(\frac{E}{c} + mc\right)^2 - p^2 = \frac{E^2}{c^2} - p^2 + 2Em + m^2c^2 = 2m^2c^2 + 2Em

Why the cross term survives: it couples the beam energy to the target mass — this is where the lab conditions enter the invariant.

Step 3. Evaluate the same invariant at threshold in the center-of-momentum frame, where at minimum energy all products are created at relative rest, moving as one lump of mass $M$ with zero total momentum:

P^\mu_{\mathrm{tot}} P_{\mathrm{tot}\,\mu} = \left(Mc\right)^2

Why "at rest together" is the minimum: any relative motion of the products adds kinetic energy in the frame where total momentum vanishes, raising the invariant above its floor $(Mc)^2$ .

Step 4. Equate the two evaluations and solve for the threshold lab energy:

2m^2c^2 + 2E_{\mathrm{thr}}m = M^2c^2 \quad\Longrightarrow\quad E_{\mathrm{thr}} = \frac{(M^2 - 2m^2)\,c^2}{2m} = \frac{M^2c^4 - 2m^2c^4}{2mc^2}

Step 5. Apply to antiproton production $p + p \to p + p + p + \bar{p}$ , where $M = 4m_p$ :

E_{\mathrm{thr}} = \frac{(16 - 2)\,m_p^2c^4}{2m_pc^2} = 7\,m_pc^2

so the beam must carry kinetic energy $T_{\mathrm{thr}} = 6m_pc^2 \approx 5.63\ \mathrm{GeV}$ — only $2m_pc^2 \approx 1.88\ \mathrm{GeV}$ of which becomes new particles; the rest is wasted keeping the center of momentum moving. *Two checks bracket the result: the Bevatron was designed to 6.2 GeV precisely to clear this 5.63 GeV threshold, and found the antiproton in 1955. Setting $M = 2m$ (nothing new produced) gives $E_{\mathrm{thr}} = mc^2$ , a beam at rest, as it must.*

Limiting cases

Case	Result	Meaning
Non-relativistic, $v \ll c$ (formally $c \to \infty$ )	$E \approx mc^2 + \tfrac12 mv^2$ , $\ \mathbf{p} \approx m\mathbf{v}$	Newtonian momentum and kinetic energy recovered exactly; the rest-energy shelf is invisible because non-relativistic physics only compares energies at fixed mass.
First correction, $v \lesssim 0.1c$	$T = \tfrac12 mv^2\left(1 + \tfrac34 v^2/c^2 + \cdots\right)$	Relativistic kinetic energy exceeds Newton's by 1% already at $v \approx 0.115c$ — the regime of inner-shell electrons and precision accelerator injectors.
Ultra-relativistic, $E \gg mc^2$	$E \approx pc\left(1 + \tfrac{(mc^2)^2}{2(pc)^2}\right) \to pc$	Energy and momentum become interchangeable; mass survives only as a tiny quadratic correction. LHC protons at 6.8 TeV sit within $10^{-8}$ of the light line.
Massless, $m = 0$	$E = pc$ , $\ p^\mu p_\mu = 0$ , $\ v = c$ always	Photons: no rest frame, no four-velocity, but a well-defined null four-momentum — light carries momentum $E/c$ and exerts pressure.
Single particle at rest	$p^\mu = (mc, \mathbf{0})$	The entire four-momentum is rest energy; this is the frame in which decay kinematics are simplest.

Common trapDifferentiating by coordinate time, and the ghost of relativistic mass

Two habits from older textbooks quietly corrupt calculations. First: $\mathrm{d}x^\mu/\mathrm{d}t$ is not a four-vector — the numerator transforms properly but the denominator is one frame's time; only proper time $\tau$ may sit under a four-vector derivative. Second: the "relativistic mass" $\gamma m$ that supposedly grows with speed. Modern usage keeps $m$ strictly invariant — it is the norm of $p^\mu$ , the same in every frame — and books all velocity dependence explicitly in $E = \gamma mc^2$ and $\mathbf{p} = \gamma m\mathbf{v}$ . Treating $\gamma m$ as "the mass" suggests, wrongly, that a fast particle gravitates like a heavier static one and breaks the clean scalar/vector bookkeeping of every formula in this article. If a factor of $\gamma$ seems missing, find it in the kinematics — never stuff it into $m$ .

Applications

Collider physics and the invariant-mass trick. Conservation of total four-momentum plus the invariance of its square is the entire kinematic toolkit of particle physics. The center-of-mass energy $\sqrt{s} = c\sqrt{P^\mu P_\mu}$ decides which reactions are open. The threshold derivation shows why fixed-target machines waste energy — most of the beam energy is locked into center-of-momentum motion — and therefore why discovery machines collide beams head-on: two 6.8 TeV protons give $\sqrt{s} = 13.6\ \mathrm{TeV}$ , while hitting a stationary proton with one such beam yields only $\sqrt{s} = \sqrt{2Emc^2 + 2m^2c^4} \approx 113\ \mathrm{GeV}$ — a factor of 120 loss.

Nuclear binding and the mass defect. Because $E_0 = mc^2$ , a bound system weighs less than its parts by $\Delta m = E_{\mathrm{bind}}/c^2$ . The deuteron is $2.224\ \mathrm{MeV}/c^2$ lighter than a free proton plus neutron; iron-56 sits at the bottom of the binding-energy valley at 8.79 MeV per nucleon, which is why fusion powers stars up to iron and fission pays only beyond it. Every reactor and every star is an $E = mc^2$ accounting exercise.

Compton scattering. A photon bouncing off an electron is solved in three lines by writing $p^\mu_\gamma + p^\mu_e = p'^\mu_\gamma + p'^\mu_e$ and squaring to eliminate the unmeasured recoil: the wavelength shift $\Delta\lambda = (h/m_ec)(1 - \cos\theta)$ follows, with the electron Compton wavelength $h/m_ec = 2.426\ \mathrm{pm}$ . Its 1923 measurement was the decisive proof that light quanta carry the null four-momentum $E = pc$ .

Two-body decay spectroscopy. A particle of mass $M$ decaying at rest into daughters of masses $m_1, m_2$ has uniquely fixed daughter energies, e.g. $E_1 = (M^2 + m_1^2 - m_2^2)c^2/2M$ — monoenergetic lines, not spectra. Seeing a continuous electron spectrum in nuclear beta decay is exactly how Pauli inferred a third, invisible participant: the neutrino, postulated from four-momentum bookkeeping alone.

Worked examples

Example 1 — electron at four-fifths of light speed (confidence builder). An electron ( $mc^2 = 0.511\ \mathrm{MeV}$ ) moves at $v = 0.8c$ . Find $\gamma$ , its total energy, kinetic energy, and momentum, and verify the energy–momentum relation.

Step 1. Lorentz factor: $\gamma = 1/\sqrt{1 - 0.64} = 1/\sqrt{0.36} = 1/0.6 = 5/3 \approx 1.667$ .

Step 2. Total energy: $E = \gamma mc^2 = (5/3)(0.511) = 0.8517\ \mathrm{MeV}$ .

Step 3. Kinetic energy: $T = (\gamma - 1)mc^2 = (2/3)(0.511) = 0.3407\ \mathrm{MeV}$ — already two-thirds of the rest energy, far above the Newtonian estimate $\tfrac12 mv^2 = \tfrac12(0.511)(0.64) = 0.164\ \mathrm{MeV}$ .

Step 4. Momentum, most cleanly via $pc = E\,(v/c) = 0.8517 \times 0.8 = 0.6813\ \mathrm{MeV}$ .

Step 5. Check the mass shell: $(pc)^2 + (mc^2)^2 = 0.6813^2 + 0.511^2 = 0.4642 + 0.2611 = 0.7253\ \mathrm{MeV^2}$ , and $E^2 = 0.8517^2 = 0.7254\ \mathrm{MeV^2}$ . Consistent to rounding: $E = 0.852\ \mathrm{MeV}$ , $T = 0.341\ \mathrm{MeV}$ , $pc = 0.681\ \mathrm{MeV}$ .

Example 2 — charged-pion decay at rest (exam level). A $\pi^+$ ( $m_\pi c^2 = 139.57\ \mathrm{MeV}$ ) at rest decays to a muon ( $m_\mu c^2 = 105.66\ \mathrm{MeV}$ ) and a neutrino (take $m_\nu = 0$ ). Find the muon and neutrino energies and the muon kinetic energy, and verify the shell condition.

Step 1. Conservation in the pion rest frame: $p^\mu_\pi = p^\mu_\mu + p^\mu_\nu$ with $p^\mu_\pi = (m_\pi c, \mathbf{0})$ . The daughters fly apart back-to-back with equal and opposite momenta of magnitude $p$ .

Step 2. Isolate the unmeasured neutrino and square: $p^\mu_\nu p_{\nu\,\mu} = (p_\pi - p_\mu)^2 = m_\pi^2c^2 + m_\mu^2c^2 - 2m_\pi c\,E_\mu/c = 0$ , since the neutrino is null.

Step 3. Solve for the muon energy: $E_\mu = \dfrac{(m_\pi^2 + m_\mu^2)c^2}{2m_\pi} = \dfrac{139.57^2 + 105.66^2}{2 \times 139.57}\ \mathrm{MeV} = \dfrac{19479.8 + 11164.0}{279.14} = \dfrac{30643.8}{279.14} = 109.78\ \mathrm{MeV}$ .

Step 4. The neutrino takes the rest: $E_\nu = m_\pi c^2 - E_\mu = 139.57 - 109.78 = 29.79\ \mathrm{MeV}$ , and since it is massless, $pc = E_\nu = 29.79\ \mathrm{MeV}$ for both daughters.

Step 5. Muon kinetic energy: $T_\mu = E_\mu - m_\mu c^2 = 109.78 - 105.66 = 4.12\ \mathrm{MeV}$ — tiny, which is why decay muons from stopped pions are nearly monoenergetic and easily identified.

Step 6. Shell check on the muon: $(pc)^2 + (m_\mu c^2)^2 = 29.79^2 + 105.66^2 = 887.4 + 11164.0 = 12051.4$ , and $E_\mu^2 = 109.78^2 = 12051.6\ \mathrm{MeV^2}$ . The muon emerges with $E_\mu = 109.78\ \mathrm{MeV}$ ( $T_\mu = 4.12\ \mathrm{MeV}$ ), the neutrino with $29.79\ \mathrm{MeV}$ — the monoenergetic muon line that tagged pion decays for decades.

Origins · 1 minAlbert Einstein and how this was discovered

Albert Einstein

1879–1955 · Bern patent office → Berlin → Princeton

Einstein produced the mass–energy equivalence in a three-page afterthought to his relativity paper, in September 1905, asking with characteristic directness: "Does the inertia of a body depend upon its energy content?" His argument — a body emitting two light pulses symmetrically must lose kinetic energy in a moving frame, and the books balance only if it loses mass $\Delta m = E/c^2$ — used nothing but the relativity principle and the Doppler shift. He suspected radium salts might test it; the decisive confirmations came from nuclear physics decades later, from Cockcroft and Walton's 1932 accelerator experiment to the mass-spectrometry ledgers of every nuclear reaction since. Einstein himself kept refining the derivation for forty years, publishing half a dozen increasingly general proofs — a reminder that the shortest equation in physics carried, for its author, the longest burden of justification.

History

1905

— Einstein's September paper derives mass–energy equivalence from the relativity principle: emission of energy E lowers a body's mass by E over c squared

1906

— Planck sharpens the dynamics of moving bodies, giving relativistic momentum its Lagrangian pedigree

1908

— Minkowski recasts relativity in four-dimensional geometry; energy and momentum fuse into a single four-vector

1932

— Cockcroft and Walton split lithium with accelerated protons and verify E equals mc squared in a nuclear reaction to about half a percent

1955

— the Bevatron, designed around the 5.63 GeV antiproton threshold, produces the antiproton and confirms invariant-mass kinematics at the design edge

2005

— Penning-trap mass spectrometry and gamma-ray wavelength metrology combine to test mass-energy equivalence to four parts in ten million

Going further

References. Morin, Introduction to Classical Mechanics, Ch. 12 (Relativistic Dynamics) · Goldstein, Poole & Safko, Classical Mechanics, 3rd ed., Ch. 7 · Landau & Lifshitz, The Classical Theory of Fields, Vol. 2, §9 · Taylor, Classical Mechanics, Ch. 15 · Rindler, Introduction to Special Relativity, Ch. 5 Key numerical anchors used throughout: $U^\mu U_\mu=c^2$ and $p^\mu p_\mu=(mc)^2$ with signature $(+,-,-,-)$ ; kinetic-energy expansion $T\approx\tfrac12 mv^2+\tfrac38 mv^4/c^2$ ; antiproton threshold $E_{\text{thr}}=7m_pc^2$ , kinetic threshold $6m_pc^2\approx5.63\,\mathrm{GeV}$ (Bevatron design energy); standard rest energies $m_ec^2=0.511\,\mathrm{MeV}$ , $m_pc^2=938.3\,\mathrm{MeV}$ , deuteron binding $2.224\,\mathrm{MeV}$ , electron Compton wavelength $h/m_ec=2.426\,\mathrm{pm}$ ; pion-decay muon energy $E_\mu=(m_\pi^2+m_\mu^2)c^2/2m_\pi=109.78\,\mathrm{MeV}$ .

Scope note. This is the mechanics-course introduction. The full story — relativistic collision kinematics in depth, four-tensors and covariant electrodynamics, spin and the relativistic angular-momentum tensor — lives in the Special Relativity track (migrates later as its own subject).

Builds on: Special Relativity Postulates and the Lorentz Transformation · Relativistic Kinematics: Time Dilation, Length Contraction, and Velocity Addition · Conservation of Linear Momentum Leads to: Relativistic Dynamics and the Four-Force · The Relativistic Lagrangian for a Free and Charged Particle · Lagrangian Density and Continuum Lagrangian Formalism · Hamiltonian Density and Canonical Field Momentum

If energy and momentum change from frame to frame, what exactly is conserved in a collision?

Both, in every frame — but frame by frame. In any single inertial frame, the total $E$ and total $\mathbf{p}$ before equal those after; that is conservation. What changes between frames is the value each observer assigns, exactly as boosting rotates a four-vector's components. The two statements coexist because conservation is about comparing before with after in one frame, while invariance is about comparing observers. Only special combinations, like $E^2 - (pc)^2$ of the total, are both conserved and invariant — which is precisely why squaring total four-momenta is the universal trick.

Where does the kinetic energy actually go when a particle is accelerated close to the speed of light?

Into the $\gamma$ factor. Speed saturates near $c$ , but $E = \gamma mc^2$ and $p = \gamma mv$ keep growing without bound — the particle becomes harder to deflect and carries more punch, even though it barely moves faster. An LHC proton at 6.8 TeV has $\gamma \approx 7250$ : it travels only about $3\ \mathrm{m\,s^{-1}}$ slower than light, yet carries 7000 times its rest energy. Thinking "acceleration means speeding up" is Newtonian; relativistically, late-stage acceleration means loading energy and momentum onto an almost-fixed velocity.

Is the mass of a system just the sum of the masses of its parts?

No — and this is the most commonly missed consequence. The system's invariant mass is the norm of the total four-momentum: $Mc^2 = \sqrt{(\sum E)^2 - |\sum \mathbf{p}c|^2}$ . Two photons, each massless, flying in opposite directions with energy $E$ each, form a system of mass $2E/c^2$ . A hot gas outweighs the same gas cold; a stretched spring outweighs a relaxed one, by $\Delta E/c^2$ . Conversely a bound state weighs less than its parts — the deuteron by 2.224 MeV. Mass is additive only when all constituents are mutually at rest and non-interacting; in general it is a property of the total arrow, not a sum over parts.

Why do people say relativistic mass is obsolete when older books use it everywhere?

Because it buys one tidy formula and corrupts everything else. Writing $p = (\gamma m)v$ lets momentum look Newtonian, but the same $\gamma m$ fails elsewhere: kinetic energy is not $\tfrac12(\gamma m)v^2$ , and a fast particle does not gravitate like a static mass $\gamma m$ . Worse, "mass increases with speed" suggests something happens to the particle, when $\gamma$ is really a statement about the observer's frame. Modern convention keeps $m$ invariant — the norm of $p^\mu$ , a property of the particle alone — and lets the kinematic factors sit in plain sight. Same physics, cleaner bookkeeping, and every formula in this article assumes it.

The rest energy of everyday objects is enormous. Why can we not extract it?

Conservation laws stand guard over it. Baryon and lepton number conservation mean a proton cannot simply dissolve into light; the only way to cash out rest energy completely is annihilation with antimatter, and the universe handed us essentially none. What we can do is rearrange binding: chemistry skims $10^{-10}$ of the rest energy, fission about $10^{-3}$ per nucleon of the fissioning fuel, fusion up to $7 \times 10^{-3}$ . The mass-energy shelf is fully real — it is just guarded by quantum numbers, not by engineering difficulty.

How can the neutrino have been discovered from kinematics before any detector ever saw one?

Two-body decays are monoenergetic: if beta decay were $n \to p + e$ , four-momentum conservation in the nucleus rest frame would fix the electron energy uniquely, as in the pion worked example. Experiments instead showed a continuous electron spectrum up to a maximum — which, taken at face value, meant energy and momentum simply failed to balance. Pauli's 1930 alternative was a third, unseen, (nearly) massless participant sharing the four-momentum event by event. Every "missing transverse momentum" search at the LHC — for neutrinos or dark matter — is the same inference automated: trust $\sum p^\mu = \mathrm{const}$ , and let the imbalance point at the invisible.

Does a photon in a box make the box heavier?

Yes. Trap a photon of energy $E$ between mirrors: the system box-plus-photon has total energy increased by $E$ at zero total momentum (the photon's momentum averages out as it bounces), so the system's invariant mass rises by $E/c^2$ . Weigh the box and it is heavier. This is the cleanest way to see that mass is a property of the total four-momentum, not a count of massive constituents — the photon contributes mass to the system while having none itself. The same accounting makes a charged battery heavier than a drained one, immeasurably but exactly.

In the threshold formula, why must the products be at rest in the center-of-momentum frame and not in the lab?

Because "minimum energy" means no energy wasted on relative motion, and relative motion is judged in the frame where the total momentum vanishes. In the lab the products can never be at rest — momentum conservation forbids it, since the beam brings momentum in. At threshold the products all move together, as one composite lump: at rest in the CM frame, but drifting through the lab with the conserved total momentum. That is why fixed-target energy scales only as $\sqrt{s} \sim \sqrt{2Emc^2}$ : the lump's mandatory drift eats the rest.