Kaluza–Klein Reduction

Around 1920 there were two great field theories in physics, and they looked nothing alike.

Einstein's General Relativity, then only four years old, described gravity as the curvature of spacetime. Maxwell's electromagnetism, half a century older, described electric and magnetic forces as a field living inside spacetime (see Maxwell's Equations). One was geometry. The other was something placed on top of geometry.

A little-known mathematician in Königsberg, Theodor Kaluza, noticed something odd. If you take Einstein's equations and simply write them in five dimensions instead of four, and then ask what a four-dimensional observer would see, Maxwell's equations fall out of the extra pieces of the geometry. Electromagnetism appears not as something added to spacetime, but as gravity in a direction we cannot see. ^[1]

That observation, and the machinery built around it, is called Kaluza–Klein reduction. It failed as a final theory of nature, yet it became one of the most reused ideas in theoretical physics.

The Hidden Direction

A garden hose seen from far away

The standard picture is a garden hose. From across a lawn it looks like a line: one-dimensional. Up close, an ant walking on it discovers a second direction, around the circumference. That second direction is small and closes on itself.

Kaluza–Klein theory proposes that every point of our three-dimensional space might be like the hose: there could be an extra direction, so small and so tightly curled that nothing we normally do can move along it.

What the extra direction does

Now imagine a particle moving in five dimensions. It can have momentum in the usual three directions, and it can also have momentum around the small circle.

To a four-dimensional observer who cannot resolve the circle, that hidden momentum does not look like motion. It looks like two other things:

mass, because energy of motion around the circle is energy the particle carries even while it sits still in ordinary space;
electric charge, because the geometry that mixes the circle with ordinary directions behaves exactly like an electromagnetic field, and the hidden momentum is what that field couples to.

In this picture, charge is not a separate substance. It is motion in a direction you cannot see.

Why a circle gives steps

Because the circle closes on itself, a quantum wave travelling around it must fit a whole number of wavelengths into the circumference, just like a vibrating guitar string must fit a whole number of half-wavelengths between its ends. So the hidden momentum comes in steps: $0, 1, 2, 3, \dots$ units. That means both the extra mass and the charge come in discrete steps. This was Oskar Klein's contribution in 1926, and it offered, for the first time, a geometric reason why electric charge is quantized. ^[2,3]

How It Works

The procedure has three stages. Each one is simple on its own.

Start with a higher-dimensional theory. Take pure gravity — Einstein's equations — in five spacetime dimensions, with no electromagnetism put in by hand.
Compactify. Assume the fifth dimension is a circle of radius $R$ , much smaller than anything we can probe.
Reduce. Expand every field in harmonics around the circle (like a Fourier series) and integrate over the circle. What is left is a four-dimensional theory, but with more fields than you started with.

The surprise is in which fields appear. The five-dimensional metric has 15 independent components. A four-dimensional observer sorts them into three groups:

Piece of the 5D metric	Number of components	What a 4D observer sees
$g_{\mu\nu}$ (the 4×4 block)	10	ordinary gravity (the graviton)
$g_{\mu 5}$ (the mixed column)	4	a vector field $A_\mu$ — electromagnetism (the photon)
$g_{55}$ (the corner)	1	a scalar field $\phi$ — the size of the circle

The scalar goes by several names — the dilaton, the radion, or in older literature the Jordan–Thiry scalar. It measures how large the extra dimension is at each point of ordinary spacetime.

The Formal Description

The metric ansatz

Write the five coordinates as $x^\mu$ ( $\mu = 0,1,2,3$ ) for ordinary spacetime and $y$ for the extra one, with $y$ identified periodically: $y \sim y + 2\pi R$ . The general five-dimensional line element can then be written as

d\hat{s}^2 = g_{\mu\nu}(x)\,dx^\mu dx^\nu + \phi^2(x)\,\big(dy + \kappa A_\mu(x)\,dx^\mu\big)^2

Kaluza–Klein metric

Here:

$d\hat{s}^2$ is the five-dimensional interval (hats mark 5D quantities);
$g_{\mu\nu}$ is the ordinary four-dimensional metric of General Relativity;
$A_\mu$ is a four-component field that will turn out to be the electromagnetic potential;
$\phi$ is the scalar that sets the local size of the circle;
$\kappa$ is a constant that fixes units, related to Newton's constant $G$ .

Kaluza's key assumption, the cylinder condition, is that none of these fields depend on $y$ . Klein later reinterpreted this as the statement that we only see the lowest, $y$ -independent harmonic of each field.

Reducing the action

Einstein's theory in five dimensions is defined by the Einstein–Hilbert action

S_5 = \frac{1}{16\pi G_5}\int d^4x\,dy\,\sqrt{-\hat g}\;\hat R ,

where $\hat R$ is the five-dimensional curvature scalar and $G_5$ is the five-dimensional gravitational constant. Substituting the ansatz and integrating over the circle gives, up to a total derivative,

S_4 = \frac{2\pi R}{16\pi G_5}\int d^4x\,\sqrt{-g}\;\phi\left(R - \frac{\kappa^2\phi^2}{4}\,F_{\mu\nu}F^{\mu\nu}\right), \qquad F_{\mu\nu} = \partial_\mu A_\nu - \partial_\nu A_\mu

Reduced 4D action

Read term by term:

$R$ is the four-dimensional curvature: this is General Relativity.
$F_{\mu\nu}F^{\mu\nu}$ is the standard Lagrangian of the electromagnetic field: this is Maxwell's theory.
$\phi$ multiplies both, acting as a field-dependent strength of gravity and of electromagnetism.
The prefactor shows that four-dimensional gravity is weaker than five-dimensional gravity by the size of the circle: $G_4 = G_5 / (2\pi R)$ .

If $\phi$ is held fixed at 1, the result is exactly Einstein–Maxwell theory. Nothing electromagnetic was put in. It was already inside five-dimensional geometry. ^[1,4]

The symmetry of electromagnetism also has a geometric origin. Shifting the extra coordinate by an amount that depends on position, $y \to y + \kappa\,\lambda(x)$ , changes $A_\mu \to A_\mu - \partial_\mu \lambda$ . That is precisely the gauge freedom of the electromagnetic potential. In this language, electromagnetic gauge symmetry is just the freedom to relabel points around the circle.

The Kaluza–Klein tower

Now let the fields depend on $y$ . Any field on a circle can be expanded in a Fourier series:

\Phi(x,y) = \sum_{n=-\infty}^{\infty} \Phi_n(x)\,e^{i n y / R}.

A field that is massless in five dimensions obeys $\hat\Box\,\Phi = 0$ , where $\hat\Box$ is the five-dimensional wave operator. Splitting it into the four-dimensional part and the $y$ part gives, for each mode,

\left(\Box_4 - \frac{n^2}{R^2}\right)\Phi_n(x) = 0 .

This is the wave equation of a particle with mass. So one five-dimensional field becomes an infinite family of four-dimensional particles, the Kaluza–Klein tower, with masses

m_n = \frac{|n|\,\hbar}{R\,c}, \qquad n = 0, 1, 2, \dots

Kaluza–Klein mass tower

In natural units ( $\hbar = c = 1$ ) this is simply $m_n = |n|/R$ . The $n = 0$ mode is massless and is the field we see at low energy. The others are heavier and heavier copies, spaced by $1/R$ . A small circle means widely spaced, very heavy copies — which is why they would be invisible at everyday energies. ^[4,5]

Charge quantization

The same integer $n$ also controls how strongly the mode couples to $A_\mu$ . With the normalization in which the Maxwell term has its standard form, the charge of mode $n$ is

q_n = n\,\frac{\sqrt{16\pi G}\;\hbar}{c\,R}

(in Gaussian units), where $G$ is the ordinary Newton constant. Charge comes in integer multiples of a basic unit set by the size of the circle. Klein turned this around: if the basic unit is the electron's charge $e$ , the circle's circumference must be about $10^{-30}$ cm, some seventeen orders of magnitude smaller than a proton. ^[2,3,4]

Historical Development

Kaluza (1921). Kaluza's paper, "On the Unification Problem in Physics," was presented to the Prussian Academy of Sciences by Einstein and published in its proceedings in 1921. It showed that five-dimensional gravity with the cylinder condition contains both Einstein's and Maxwell's equations. Kaluza treated the fifth dimension as a formal device and fixed $g_{55}$ to a constant. ^[1,4]

Klein (1926). Oskar Klein, working in the new quantum mechanics, gave the fifth dimension a physical form: a tiny circle. He showed that quantum waves on the circle explain why charge comes in discrete units, and estimated the circle's size from the value of $e$ . ^[2,3]

The scalar is noticed (1940s–1950s). Setting $g_{55}$ to a constant turned out to be inconsistent with the full field equations unless $F_{\mu\nu}F^{\mu\nu} = 0$ everywhere. Jordan, Thiry and others restored the scalar as a genuine dynamical field, which is why it is sometimes called the Jordan–Thiry scalar. ^[4]

Non-abelian generalizations (1960s–1980s). Replacing the circle with a more complicated compact space yields more complicated gauge fields. Compactifying on a space whose symmetries form a non-abelian group produces Yang–Mills fields of the type used in the Standard Model. In 1981 Edward Witten showed that eleven is the minimum total number of dimensions in which the symmetry group $SU(3)\times SU(2)\times U(1)$ of the Standard Model can arise this way — which intriguingly matches the maximum dimension allowed for supergravity. ^[6]

Strings and M-theory (1980s–1990s). Superstring theory is consistent in ten dimensions. To describe a four-dimensional world, six dimensions must be compactified, and Kaluza–Klein reduction is the standard tool for working out what a four-dimensional observer sees. In 1995 Witten argued that one type of ten-dimensional string theory (type IIA) is itself the Kaluza–Klein reduction of an eleven-dimensional theory, now called M-theory, on a circle whose radius grows with the string coupling. The Kaluza–Klein modes of the eleventh dimension correspond to known objects in the string theory. ^[7] (For the higher-dimensional and extra-time proposals built on this, see 13-Dimensional Models.)

Large and warped extra dimensions (1998–1999). Arkani-Hamed, Dimopoulos and Dvali (ADD) proposed that extra dimensions could be as large as a millimetre if only gravity travels through them, while ordinary matter is confined to a four-dimensional "brane". Randall and Sundrum proposed a single warped extra dimension. Both ideas were attempts to explain why gravity is so much weaker than the other forces, and both predicted Kaluza–Klein gravitons that could be seen at colliders. ^[8,9]

Evidence

It helps to separate what is mathematically established from what is known about nature.

Established mathematics

That five-dimensional General Relativity on a circle reduces to four-dimensional gravity, electromagnetism and a scalar is a mathematical fact. It is a derivation, not a hypothesis, and it is reproduced in standard reviews. ^[4] The same is true of the tower formula $m_n = |n|/R$ for fields on a circle.

What is known about nature

No Kaluza–Klein particle has ever been observed, and no extra dimension has been detected. Every experiment so far is consistent with a world of three space dimensions and one time, down to the smallest distances tested. The results are limits.

Short-range gravity tests. If gravity leaks into extra dimensions of size $R$ , Newton's inverse-square law should change at distances shorter than about $R$ . Such deviations are usually written as

V(r) = -\frac{G\,m_1 m_2}{r}\left(1 + \alpha\, e^{-r/\lambda}\right),

where $V$ is the potential energy of two masses $m_1, m_2$ at separation $r$ , $\alpha$ measures the strength of the extra force, and $\lambda$ its range. For two extra dimensions shaped like a torus, Kaluza–Klein gravitons give $\alpha = 16/3$ and $\lambda = R$ . ^[10]

The Eöt-Wash group at the University of Washington uses torsion balances — delicate rotating pendulums — to measure gravity between test masses separated by less than a tenth of a millimetre. In 2007 they found the inverse-square law holding down to 56 micrometres and concluded that an extra dimension must be smaller than about 44 μm. ^[11] A 2020 experiment probed separations down to 52 μm and found Newtonian gravity fitting the data perfectly, excluding gravitational-strength extra forces with ranges above 38.6 μm. ^[12] The Particle Data Group summary quotes $R < 30\ \mu\text{m}$ for two large extra dimensions. ^[10]

Collider searches. At the Large Hadron Collider, ADD-type gravitons would escape the detector, leaving events with an energetic jet and "missing" momentum. ATLAS found no excess in its full 13 TeV dataset, pushing the fundamental higher-dimensional gravity scale $M_D$ above roughly 6–11 TeV depending on the number of extra dimensions. ^[13,10] Searches for Randall–Sundrum Kaluza–Klein gravitons, which would appear as bumps in the mass spectra of photon or lepton pairs, have excluded the lightest such graviton below about 4.5 TeV for a benchmark coupling. ^[10]

Astrophysics. Light Kaluza–Klein gravitons would carry energy out of supernovae. The observed neutrino burst from Supernova 1987A therefore gives constraints on large extra dimensions that are, for two dimensions, even stronger than the collider limits. ^[10]

Summary of the evidence: the original Kaluza–Klein picture, with a circle near the Planck length, is untestable with any foreseeable technology. The newer "large" versions made testable predictions, and so far those predictions have not been seen.

Limitations and Open Problems

The masses are wrong

Klein's own numbers expose the first problem. If the circle is small enough to give the electron's charge, the first charged Kaluza–Klein mode has a mass $\hbar/(Rc)$ of order $10^{17}$ GeV — about $10^{20}$ times heavier than the electron. The simplest version of the theory cannot contain the electron as a charged Kaluza–Klein mode. ^[4]

The scalar will not sit still

The scalar $\phi$ is massless in the simplest reduction. A massless scalar coupled to matter would produce a long-range "fifth force" and would make the effective strengths of gravity and electromagnetism vary from place to place. Precision tests of gravity and of the constancy of fundamental constants strongly restrict such effects. A realistic model must therefore give $\phi$ a mass, fixing the size of the extra dimension. This task, called moduli stabilization in string theory, remains one of the hardest problems in the field. ^[4,10]

Chirality

The Standard Model is chiral: the weak force treats left-handed and right-handed particles differently. Pure Kaluza–Klein theories on smooth compact spaces have great difficulty producing this asymmetry. Witten's 1981 analysis found that the Standard Model gauge group could be obtained, but the right fermion quantum numbers could not. ^[6] String theory addresses this by adding ingredients beyond pure geometry — gauge fields already present in ten dimensions, singular compact spaces, and branes — so Kaluza–Klein reduction survives as a tool rather than as the whole explanation.

Which extra space?

Once one extra dimension is allowed, nothing obvious fixes how many there are or what shape they take. String theory offers an enormous number of candidate six-dimensional shapes, and no known principle selects the one that describes our world. This weakens the predictive power of the approach: many observations can be matched after choosing the shape, which is different from predicting them in advance.

What would count as a discovery

The idea is falsifiable in its large-dimension versions. A clear deviation from the inverse-square law at short range, with the right strength, or a regularly spaced series of heavy resonances at a collider — copies of a known particle at masses $m, 2m, 3m, \dots$ above a threshold — would be strong evidence. The absence of both, so far, is why large extra dimensions are now tightly constrained.

A Failed Theory's Long Afterlife

Kaluza–Klein reduction matters for three separate reasons, and they should not be confused.

As an established mathematical result, it shows that the forces of nature can, in principle, be different aspects of one geometry. Electromagnetism and its gauge symmetry can be derived from gravity in one more dimension. That is a real and lasting insight into how gauge theories are built.

As a working tool, it is used constantly. Any theory in more than four dimensions — supergravity, string theory, M-theory, brane-world models — relies on Kaluza–Klein reduction to connect its equations to the four-dimensional world. It even appears in condensed-matter physics and in ordinary engineering problems where one direction of a system is much smaller than the others.

As a claim about nature, it remains unconfirmed. Our universe is consistent with three space dimensions and one time down to tens of micrometres in gravity and to scales of roughly $10^{-19}$ m in particle collisions. Extra dimensions may exist below those scales, or they may not exist at all.

Five-dimensional geometry really does contain Maxwell's theory, and higher-dimensional physics can make forces appear as the shapes of hidden dimensions. The unresolved part is the physical one: whether our universe has any such dimension to hide.

References

[1] Kaluza, Th. (1921; English translation 2018). "On the Unification Problem in Physics." Original: Sitzungsberichte der Preussischen Akademie der Wissenschaften (Math. Phys.), 966–972. Translation: International Journal of Modern Physics D, 27(14), 1870001.
https://doi.org/10.1142/S0218271818700017 · arXiv:1803.08616

[2] Klein, O. (1926). "Quantentheorie und fünfdimensionale Relativitätstheorie." Zeitschrift für Physik, 37(12), 895–906.
https://doi.org/10.1007/BF01397481

[3] Klein, O. (1926). "The Atomicity of Electricity as a Quantum Theory Law." Nature, 118, 516.
https://doi.org/10.1038/118516a0

[4] Overduin, J. M., & Wesson, P. S. (1997). "Kaluza–Klein Gravity." Physics Reports, 283, 303–378.
https://doi.org/10.1016/S0370-1573(96)00046-4 · arXiv:gr-qc/9805018

[5] Duff, M. J. (1994). "Kaluza–Klein Theory in Perspective." Talk at the Oskar Klein Centenary Symposium. arXiv preprint.
https://arxiv.org/abs/hep-th/9410046

[6] Witten, E. (1981). "Search for a Realistic Kaluza–Klein Theory." Nuclear Physics B, 186, 412–428.
https://doi.org/10.1016/0550-3213(81)90021-3

[7] Witten, E. (1995). "String Theory Dynamics in Various Dimensions." Nuclear Physics B, 443, 85–126.
https://doi.org/10.1016/0550-3213(95)00158-O · arXiv:hep-th/9503124

[8] Arkani-Hamed, N., Dimopoulos, S., & Dvali, G. (1998). "The Hierarchy Problem and New Dimensions at a Millimeter." Physics Letters B, 429, 263–272.
https://doi.org/10.1016/S0370-2693(98)00466-3 · arXiv:hep-ph/9803315

[9] Randall, L., & Sundrum, R. (1999). "Large Mass Hierarchy from a Small Extra Dimension." Physical Review Letters, 83, 3370–3373.
https://doi.org/10.1103/PhysRevLett.83.3370

[10] Demiragli, Z., & Pomarol, A. (2024). "Extra Dimensions." In Navas, S., et al. (Particle Data Group), "Review of Particle Physics." Physical Review D, 110, 030001.
https://doi.org/10.1103/PhysRevD.110.030001 · Chapter PDF

[11] Kapner, D. J., Cook, T. S., Adelberger, E. G., Gundlach, J. H., Heckel, B. R., Hoyle, C. D., & Swanson, H. E. (2007). "Tests of the Gravitational Inverse-Square Law below the Dark-Energy Length Scale." Physical Review Letters, 98, 021101.
https://doi.org/10.1103/PhysRevLett.98.021101 · arXiv:hep-ph/0611184

[12] Lee, J. G., Adelberger, E. G., Cook, T. S., Fleischer, S. M., & Heckel, B. R. (2020). "New Test of the Gravitational $1/r^2$ Law at Separations down to 52 μm." Physical Review Letters, 124, 101101.
https://doi.org/10.1103/PhysRevLett.124.101101 · arXiv:2002.11761

[13] ATLAS Collaboration (2021). "Search for new phenomena in events with an energetic jet and missing transverse momentum in $pp$ collisions at $\sqrt{s} = 13$ TeV with the ATLAS detector." Physical Review D, 103, 112006.
https://doi.org/10.1103/PhysRevD.103.112006