Cosmic Inflation

The hot Big Bang, assembled from the FLRW metric and the Friedmann equations, is one of the great successes of physics from the first few seconds onward — yet run it backward and it demands initial conditions so special they read less like physics than like an accident. The microwave sky is uniform to one part in $10^5$ across patches that, in a universe that has only ever decelerated, were never in causal contact. The spatial geometry is flat to sub-percent accuracy today, even though flatness is an unstable fixed point that any deviation should have amplified by sixty orders of magnitude. And the grand unified theories that unify the forces at $10^{16}\,\mathrm{GeV}$ predict a flood of magnetic monopoles the universe simply does not contain. Inflation, proposed by Alan Guth in 1981 and rebuilt within a year by Linde, Albrecht, and Steinhardt, dissolves all three problems at once: a finite burst of accelerated expansion, $\ddot a>0$ , in the first fraction of a second. Stretch a microscopic, causally connected, smooth patch by a factor of at least $e^{60}$ and it swells to engulf everything we can now see — homogeneity, flatness, and the absence of relics all follow because the observable universe is the blown-up image of one tiny pre-inflationary domain. The engine is a scalar field, the inflaton, poised on a flat stretch of its potential where its energy behaves like a temporary cosmological constant, $w\approx-1$ . Its deepest legacy is not classical but quantum: zero-point fluctuations stretched beyond the horizon freeze into a nearly scale-invariant spectrum of curvature perturbations. Those frozen ripples become galaxies and the acoustic peaks of the CMB, and their measured slight red tilt, $n_s\simeq0.965$ , is inflation's sharpest fingerprint.

Key takeaways

Cosmic inflation is a hypothesized epoch of quasi-exponential expansion in the first instant of the universe, driven by a scalar field (the inflaton) whose potential energy dominates and mimics a large cosmological constant with $w\approx-1$ .
Accelerated expansion $\ddot a>0$ requires $\rho+3p<0$ ( $w<-\tfrac13$ ), violating only the strong energy condition; it makes the comoving Hubble radius $1/(aH)$ shrink, solving the horizon problem and driving $|\Omega-1|\propto(aH)^{-2}\propto e^{-2N}$ toward flatness — an unstable fixed point turned attractor.
At least $N\gtrsim60$ e-folds are needed; the simplest chaotic model $V=\tfrac12 m^2\phi^2$ demands super-Planckian field values ( $\phi_\star\approx15.6\,M_{\rm Pl}$ ) and predicts $n_s=0.967$ , $r=0.13$ — the tilt fits but the large $r$ is now excluded.
Quantum fluctuations frozen at horizon exit give a curvature spectrum $\mathcal{P}_{\mathcal R}=H^2/(8\pi^2\epsilon M_{\rm Pl}^2)$ , nearly scale-invariant with tilt $n_s-1=-6\epsilon+2\eta$ and tensor ratio $r=16\epsilon$ .
Planck measures $n_s=0.9649\pm0.0042$ (red-tilted, as slow roll predicts) and BICEP/Keck bounds $r<0.036$ , pinning the inflationary energy scale below $V^{1/4}\approx1.4\times10^{16}\,\mathrm{GeV}$ .
Key idea: A scalar field trapped in a flat region of its potential acts as a transient cosmological constant, driving exponential expansion that flattens and smooths spacetime while quantum fluctuations of that field, frozen on super-horizon scales, become the seeds of all cosmic structure.

Where this lives today

Inflation is now tested at the level of its quantum predictions, and the frontier is the hunt for primordial gravitational waves. Planck's 2018 release fixed the tilt at $n_s=0.9649\pm0.0042$ — a five-sigma detection of departure from exact scale invariance — and found the perturbations Gaussian and adiabatic, with non-Gaussianity $f_{\rm NL}$ consistent with zero as the simplest single-field models require. The remaining prize is the tensor ratio $r$ : BICEP/Keck at the South Pole, combined with Planck, has pushed the bound to $r<0.036$ , already excluding the simplest monomial potentials and cornering inflation onto plateau models such as Starobinsky $R^2$ and Higgs inflation. Next-generation efforts — the Simons Observatory, CMB-S4, and the LiteBIRD satellite — aim at $r\sim10^{-3}$ , deep enough to detect or exclude essentially any high-scale plateau model; a confirmed B-mode signal would be the first direct evidence of quantized gravity, fixing the energy scale near $10^{16}\,\mathrm{GeV}$ . On the theory side, eternal inflation and its multiverse, the measure problem, and the swampland conjectures questioning whether slow-roll plateaus arise in consistent quantum gravity keep inflation an open, contested frontier rather than a closed chapter.

The physical picture

Everything about inflation follows from one reversal of intuition that Newton never faced: in general relativity pressure gravitates, and the quantity that decides whether expansion speeds up or slows down is $\rho+3p$ , not $\rho$ alone. Ordinary matter and radiation have positive $\rho+3p$ and decelerate the cosmos, as gravity should. A field whose pressure is as negative as its energy density is positive, $p\approx-\rho$ , flips the sign: $\rho+3p<0$ , gravity effectively repels, and the scale factor curves upward into near-exponential growth. A scalar field sitting on a flat plateau of its potential does exactly this for a brief epoch and then switches off — its potential energy dominates, its pressure is a tension, and while it lingers there the universe inflates. Picture the inflaton as a ball rolling on a nearly horizontal shelf immersed in thick syrup. The Hubble expansion supplies a friction term $3H\dot\phi$ that overwhelms the acceleration, so the ball creeps at terminal velocity with its potential energy almost constant — and that constant density is the fuel for exponential expansion. Only when the shelf ends and the ball plunges down a steep wall does the kinetic energy catch up, the vacuum-like condition break, and inflation stop, the field oscillating about the minimum and decaying to reheat the universe into the hot radiation era. The first figure is that potential, with the slow-roll plateau, the endpoint $\epsilon=1$ , and the reheating well.

Figure

The inflaton potential V of φ: the field slow-rolls down a nearly flat plateau where potential energy dominates and w ≈ −1, so the slow-roll parameter ε = (M_Pl squared over 2)(V prime over V) squared stays much less than 1; inflation ends where the slope steepens and ε reaches 1, after which the field oscillates about the minimum and reheats the universe

The horizon and flatness problems are two faces of one quantity: the comoving Hubble radius $1/(aH)$ , the largest comoving scale over which physics can act coherently in a Hubble time. In every decelerating era it grows, so new regions constantly enter causal contact for the first time — which is why patches on opposite sides of the CMB seem never to have communicated. During inflation the opposite happens: $\ddot a>0$ makes $1/(aH)$ shrink, so a scale that starts inside the horizon is stretched until it exits, freezes, and only re-enters billions of years later. The CMB's uniformity is then no mystery — those regions were in contact before inflation and share one thermal history, blown apart and reunited by the expansion — and the same shrinking $1/(aH)$ drives $|\Omega-1|\propto(aH)^{-2}$ toward zero. The second figure is the observational verdict: the plane of tilt $n_s$ against tensor ratio $r$ , where slow-roll models make sharp predictions the data have begun to eliminate.

Figure

The observational n_s–r plane: slow-roll inflation predicts a slight red tilt and a definite tensor ratio r = 16ε. Planck plus BICEP/Keck measure n_s = 0.9649 ± 0.0042 and bound r below 0.036 at 95 percent. The simplest chaotic model V ∝ φ squared lands at n_s = 0.967, r = 0.13 and is excluded, while plateau models such as Starobinsky R squared at r ≈ 0.003 sit comfortably inside

The core equations

The entry point is the second Friedmann (acceleration) equation, the only place in cosmology where the sign of the expansion's second derivative is fixed by the source. Accelerated expansion is a statement not about how fast the universe grows but about whether that growth is speeding up.

\frac{\ddot a}{a}=-\frac{4\pi G}{3}\left(\rho+3p\right) \qquad (1)

where: $a$ — FLRW scale factor; $\rho$ — total energy density; $p$ — isotropic pressure; $G$ — Newton's constant. Reading: $\ddot a>0$ requires $\rho+3p<0$ , i.e. $w<-\tfrac13$ — the defining condition for inflation and a violation of the strong energy condition. A scalar field supplies exactly such a source: its energy density and pressure split into kinetic and potential pieces whose competition sets the equation of state.

\rho_\phi=\tfrac12\dot\phi^2+V(\phi),\qquad p_\phi=\tfrac12\dot\phi^2-V(\phi) \qquad (2)

where: $\phi$ — the inflaton field; $V(\phi)$ — its potential; $\dot\phi$ — field velocity. Reading: when $\dot\phi^2\ll V$ the pressure approaches $-\rho_\phi$ and $w\to-1$ , so a potential-dominated field behaves like a transient cosmological constant. The field evolves by the Klein–Gordon equation in an expanding background, which adds a friction term absent in flat space.

\ddot\phi+3H\dot\phi+V'(\phi)=0 \qquad (3)

where: $H=\dot a/a$ — Hubble parameter; $V'(\phi)=\mathrm{d}V/\mathrm{d}\phi$ — the slope, acting as a force; $3H\dot\phi$ — Hubble friction. Reading: for a flat enough potential the friction dominates the inertia, and the field descends at a slow terminal velocity. How flat is "flat enough" is set by two dimensionless slow-roll parameters built from the shape of $V$ ; inflation persists only while both are small.

\epsilon\equiv\frac{M_{\rm Pl}^2}{2}\left(\frac{V'}{V}\right)^2,\qquad \eta\equiv M_{\rm Pl}^2\,\frac{V''}{V} \qquad (4)

where: $M_{\rm Pl}=(8\pi G)^{-1/2}\approx2.4\times10^{18}\,\mathrm{GeV}$ — reduced Planck mass; $\epsilon$ — potential slope, needs $\epsilon\ll1$ ; $\eta$ — potential curvature, needs $|\eta|\ll1$ . Reading: $\epsilon=1$ marks the end of inflation ( $\ddot a=0$ ), and $w=-1+\tfrac23\epsilon$ , so small $\epsilon$ is small departure from de Sitter. In the slow-roll regime the second-order equations collapse to first-order working equations.

3H\dot\phi\approx-V'(\phi),\qquad H^2\approx\frac{V}{3M_{\rm Pl}^2} \qquad (5)

where: the acceleration term is dropped from (3) and the kinetic term from $\rho_\phi$ . Reading: $H$ is fixed by the height of the potential and $\dot\phi$ by its slope over the friction — the expansion rate is set almost entirely by $V$ . The total amount of inflation is measured in e-folds; solving horizon and flatness needs at least about sixty.

N\equiv\ln\frac{a_{\rm end}}{a_\star}=\int_{t_\star}^{t_{\rm end}}\!H\,\mathrm{d}t\approx\frac{1}{M_{\rm Pl}^2}\int_{\phi_{\rm end}}^{\phi_\star}\frac{V}{V'}\,\mathrm{d}\phi \qquad (6)

where: $N$ — e-folds between field values $\phi_\star$ and $\phi_{\rm end}$ . Reading: large $N$ needs a long, flat stretch of potential, since $V/V'$ is large where the slope is small. The frozen quantum fluctuations set the primordial spectrum; its tilt and tensor ratio are pure functions of the slow-roll parameters.

\mathcal{P}_{\mathcal R}(k)=\frac{1}{8\pi^2}\frac{1}{\epsilon}\frac{H^2}{M_{\rm Pl}^2}\bigg|_{k=aH},\qquad n_s-1=-6\epsilon+2\eta,\qquad r=16\epsilon \qquad (7)

where: $\mathcal{P}_{\mathcal R}$ — dimensionless spectrum of the comoving curvature perturbation $\mathcal R$ , at horizon crossing $k=aH$ ; $n_s$ — scalar spectral index; $r$ — tensor-to-scalar ratio. Reading: exact scale invariance is $n_s=1$ ; slow roll predicts a slight red tilt $n_s<1$ , and the same $\epsilon$ that ends inflation sets the gravitational-wave amplitude through $r=16\epsilon$ . Finally, the master diagnostic of the mechanism is the comoving Hubble radius, whose shrinkage solves the classic problems.

\frac{\mathrm{d}}{\mathrm{d}t}\!\left(\frac{1}{aH}\right)<0 \iff \ddot a>0 \qquad (8)

where: $1/(aH)$ — the comoving Hubble radius, the causal scale in comoving coordinates. Reading: because inflation makes it shrink, comoving scales exit the horizon and freeze — the horizon and flatness puzzles are both consequences of this one sign.

Derivation AThe flatness problem and how inflation turns instability into an attractor

The goal is to show, from the Friedmann constraint alone, that spatial flatness is a repulsive fixed point of ordinary cosmology — so a universe observed flat today must have been fine-tuned to flatness by roughly sixty decimal places at early times — and then to show that one ingredient, a growing $aH$ , reverses that flow and makes flatness an attractor inflation reaches automatically. It turns on the time dependence of a single quantity, $aH$ , and exposes why the mechanism that smooths the sky also flattens it.

Step 1. Take the Friedmann equation $H^2=\tfrac{8\pi G}{3}\rho-k/a^2$ , divide by $H^2$ , and use $\Omega=\rho/\rho_{\rm crit}$ with $\rho_{\rm crit}=3H^2/8\pi G$ to obtain the exact relation $\Omega-1=k/(a^2H^2)=k/(aH)^2$ . Why: this collapses the whole flatness question into the behaviour of a single quantity — the comoving Hubble rate $aH$ — since $k$ is a fixed constant.

Step 2. In a decelerating universe $a\propto t^{1/2}$ (radiation) or $t^{2/3}$ (matter), so $aH=\dot a$ decreases with time; therefore $|\Omega-1|\propto(aH)^{-2}$ grows. Flatness is an unstable fixed point: any tiny departure from $\Omega=1$ is amplified. Why: this is the crux — standard cosmology drives the universe away from flatness, so present near-flatness is not natural but demands explanation.

Step 3. Extrapolate back: for $|\Omega-1|\lesssim0.01$ today, the deviation at the Planck era must have satisfied $|\Omega-1|\lesssim10^{-60}$ . Why: matching a mildly flat present to the amplification factor accumulated over the whole expansion history yields an absurd initial fine-tuning, the quantitative statement of the flatness problem.

Step 4. During inflation $a\propto e^{Ht}$ with $H$ nearly constant, so $aH\propto e^{Ht}$ grows, and $|\Omega-1|\propto e^{-2Ht}=e^{-2N}$ is crushed toward zero. With $N\gtrsim60$ the factor $e^{-120}\sim10^{-52}$ erases any initial deviation. Why: the sign of $\mathrm{d}(aH)/\mathrm{d}t$ flips, converting the repulsive fixed point into an attractor — inflation does not require flatness, it produces it. *The key results: $\Omega-1=k/(aH)^2$ ; decelerating eras give growing $|\Omega-1|$ , requiring $|\Omega-1|\lesssim10^{-60}$ at the Planck scale; inflation gives $|\Omega-1|\propto e^{-2N}$ .*

Go deeper · Derivation B · 1 minSlow-roll dynamics and the e-fold integral

Derivation BSlow-roll dynamics and the e-fold integral

The aim is to reduce the coupled second-order system — the Klein–Gordon equation for the inflaton and the Friedmann equation for the metric — to a pair of first-order slow-roll equations, and from them extract a closed formula for the number of e-folds as an integral over the potential. This is the workhorse of inflationary phenomenology: given any candidate $V(\phi)$ it returns how much the universe inflates and at what field value the observable scales left the horizon — exactly what one needs to predict $n_s$ and $r$ .

Step 1. Impose the slow-roll conditions $|\ddot\phi|\ll|3H\dot\phi|$ and $\tfrac12\dot\phi^2\ll V$ , self-consistent exactly when $\epsilon\ll1$ and $|\eta|\ll1$ ; dropping the small terms leaves $3H\dot\phi\approx-V'$ and $H^2\approx V/3M_{\rm Pl}^2$ . Why: an overdamped field forgets its initial velocity and settles onto an attractor set by the local slope, so the dynamics become first order.

Step 2. Solve for the field velocity, $\dot\phi\approx-V'/(3H)$ — a terminal, viscosity-limited descent in which Hubble friction always wins on a flat plateau. Why: this keeps the potential energy nearly constant, sustaining the $w\approx-1$ inflation requires.

Step 3. Change variables from time to field, $N=\int H\,\mathrm{d}t=\int (H/\dot\phi)\,\mathrm{d}\phi$ , and substitute $\dot\phi=-V'/3H$ with $H^2=V/3M_{\rm Pl}^2$ to get $N\approx M_{\rm Pl}^{-2}\int_{\phi_{\rm end}}^{\phi_\star}(V/V')\,\mathrm{d}\phi$ . Why: trading $t$ for $\phi$ turns an implicit time integral into an explicit quadrature over the known potential.

Step 4. Read the consequence: since $V/V'\sim\phi/2$ for a monomial potential, $N\gtrsim60$ needs field excursions $\Delta\phi\gtrsim M_{\rm Pl}$ , i.e. super-Planckian values in the simplest models. Why: this field-range problem pushes model-builders toward plateau potentials, where a flat $V$ gives many e-folds without large $\phi$ . *The key results: $3H\dot\phi\approx-V'$ , $H^2\approx V/3M_{\rm Pl}^2$ , and $N\approx M_{\rm Pl}^{-2}\int V/V'\,\mathrm{d}\phi$ ; for $V=\tfrac12 m^2\phi^2$ this gives $N\approx\phi_\star^2/4M_{\rm Pl}^2$ .*

Go deeper · Derivation C · 1 minFreezing quantum fluctuations into the scale-invariant spectrum

Derivation CFreezing quantum fluctuations into the scale-invariant spectrum

This derivation follows a single Fourier mode of the inflaton fluctuation from deep inside the horizon, where it is an ordinary quantum oscillator in its vacuum, out through horizon crossing, where the expansion freezes it into a classical perturbation of fixed amplitude, and converts that frozen ripple into the gauge-invariant curvature perturbation that seeds the CMB. It is the step that turns inflation from a fix for classical fine-tunings into a predictive quantum theory of structure, and explains why the spectrum is nearly scale-invariant and slightly tilted.

Step 1. In conformal time the rescaled mode $v_k=a\,\delta\phi_k$ obeys $v_k''+(k^2-a''/a)v_k=0$ , a harmonic oscillator with a time-dependent frequency; for $k\gg aH$ the $k^2$ term dominates and $v_k$ oscillates in the Bunch–Davies vacuum. Why: on sub-horizon scales gravity is irrelevant and the field is just quantum vibration, giving an unambiguous initial state.

Step 2. As $a$ grows the wavelength stretches until $k=aH$ — horizon crossing — beyond which the $a''/a$ term dominates, the oscillation ceases, and the mode freezes with variance $\langle\delta\phi^2\rangle_k\simeq(H/2\pi)^2$ at crossing. Why: once a mode exceeds the Hubble radius, causal forces cannot restore it, so its amplitude locks in as a classical number.

Step 3. Because $H$ changes only slowly (by $\epsilon$ per e-fold), successive modes freeze with almost the same amplitude, giving a nearly scale-invariant spectrum; the slow drift of $H$ and $\epsilon$ produces the small tilt. Why: near-constancy of $H$ in quasi-de Sitter is the origin of the near scale invariance observations confirm.

Step 4. Convert to the comoving curvature perturbation $\mathcal R=-(H/\dot\phi)\,\delta\phi$ , conserved on super-horizon scales, to obtain $\mathcal{P}_{\mathcal R}=H^2/(8\pi^2\epsilon M_{\rm Pl}^2)$ , with $n_s-1=-6\epsilon+2\eta$ and $r=16\epsilon$ . Why: $\mathcal R$ survives unchanged until horizon re-entry, so it imprints the CMB, and its mild $k$ -dependence is measured as $n_s$ . *The key results: $\delta\phi\sim H/2\pi$ ; $\mathcal{P}_{\mathcal R}=H^2/8\pi^2\epsilon M_{\rm Pl}^2$ ; $n_s-1=-6\epsilon+2\eta$ ; $r=16\epsilon$ .*

Limiting cases

Inflation occupies a narrow corner of the equation-of-state and energy-condition landscape; the regimes below bracket it:

Case	Result	Meaning
Exact de Sitter	$w=-1,\ H=\text{const},\ a\propto e^{Ht}$	the idealized limit: pure vacuum energy gives perfectly exponential growth, but never ends — real inflation must depart from it
Slow-roll regime	$\epsilon,\|\eta\|\ll1,\ w\approx-1+\tfrac23\epsilon$	the physical regime: $H$ drifts slowly, the spectrum is nearly scale-invariant, and inflation lasts many e-folds
Acceleration threshold	$w=-\tfrac13\iff\ddot a=0$	the boundary of the strong energy condition; $w<-\tfrac13$ accelerates, $w>-\tfrac13$ decelerates
End of inflation	$\epsilon=1,\ \tfrac12\dot\phi^2\sim V$	slow roll fails, kinetic energy catches up, $w$ climbs above $-\tfrac13$ , and reheating begins
Eternal regime	quantum kick $>$ classical roll per e-fold	where $H^2/\dot\phi>2\pi$ , fluctuations drive $\phi$ up-potential faster than it rolls down, so some regions inflate forever

Common trapInflation is faster-than-light expansion that violates energy conditions and needs exotic matter

Three misreadings bundled together. First, inflation is not superluminal in any local sense — comoving objects never move through space; space itself stretches, and no signal outruns light in a local frame. Second, inflation violates only the strong energy condition ( $\rho+3p\geq0$ ), the same one a cosmological constant breaks; the null and weak conditions remain satisfied, since a slow-rolling scalar has $\rho+p=\dot\phi^2\geq0$ . No ghost, phantom, or negative-energy matter is required — an ordinary scalar on a flat potential suffices. Third, the famous "sixty e-folds" is a minimum tied to the reheating temperature and the scale exiting the horizon, not a universal number; the observationally relevant window is only the fifty-to-sixty e-folds before the end during which today's largest scales left the horizon.

Worked examples

1 · Chaotic inflation and its downfall (fully worked)

Take the chaotic potential $V=\tfrac12 m^2\phi^2$ . Find the slow-roll parameters, where inflation ends, the field value $N=60$ e-folds earlier, and the predicted $n_s$ and $r$ ; then compare with data.

Step 1. With $V'=m^2\phi$ and $V''=m^2$ , the slow-roll parameters are $\epsilon=\tfrac{M_{\rm Pl}^2}{2}(V'/V)^2=\tfrac{M_{\rm Pl}^2}{2}(2/\phi)^2=2M_{\rm Pl}^2/\phi^2$ and $\eta=M_{\rm Pl}^2 V''/V=2M_{\rm Pl}^2/\phi^2$ . So for this potential $\boldsymbol{\epsilon=\eta=2M_{\rm Pl}^2/\phi^2}$ .

Step 2. Inflation ends at $\epsilon=1$ : $2M_{\rm Pl}^2/\phi_{\rm end}^2=1$ gives $\boldsymbol{\phi_{\rm end}=\sqrt2\,M_{\rm Pl}}\approx1.41\,M_{\rm Pl}$ .

Step 3. The e-fold integral gives $N=M_{\rm Pl}^{-2}\int_{\phi_{\rm end}}^{\phi_\star}(V/V')\,\mathrm{d}\phi=M_{\rm Pl}^{-2}\int(\phi/2)\,\mathrm{d}\phi=(\phi_\star^2-\phi_{\rm end}^2)/4M_{\rm Pl}^2$ . Solving for $N=60$ : $\phi_\star^2=(4N+2)M_{\rm Pl}^2=242\,M_{\rm Pl}^2$ , so $\boldsymbol{\phi_\star\approx15.6\,M_{\rm Pl}}$ — comfortably super-Planckian.

Step 4. At $\phi_\star$ , $\epsilon=\eta=2/242=1/121\approx0.00826$ , so $\epsilon\approx1/(2N)$ . The spectrum is $n_s-1=-6\epsilon+2\eta=-4\epsilon=-2/N=-0.0333$ , giving $\boldsymbol{n_s\approx0.967}$ , and $r=16\epsilon=8/N=\boldsymbol{0.13}$ .

Step 5. Verdict: the red tilt $n_s=0.967$ is consistent with the measured $n_s=0.9649\pm0.0042$ , but the tensor ratio $r=0.13$ is far above the bound $r<0.036$ . The simplest chaotic model $V\propto\phi^2$ is observationally excluded — it predicts too many primordial gravitational waves — which is why plateau potentials (Starobinsky $R^2$ , $n_s=0.967$ , $r\approx0.003$ ) are now favoured.

2 · The energy scale of inflation from the tensor bound (fully worked)

The tensor-to-scalar ratio fixes the inflationary energy scale through $V=\tfrac32\pi^2 A_s\,r\,M_{\rm Pl}^4$ , with measured scalar amplitude $A_s=2.1\times10^{-9}$ . Evaluate $V^{1/4}$ and the Hubble rate $H$ during inflation at the current upper bound $r=0.036$ .

Step 1. Insert the numbers with $M_{\rm Pl}=2.435\times10^{18}\,\mathrm{GeV}$ , so $M_{\rm Pl}^4=3.52\times10^{73}\,\mathrm{GeV^4}$ : $V=1.5\times\pi^2\times(2.1\times10^{-9})\times0.036\times3.52\times10^{73}=3.9\times10^{64}\,\mathrm{GeV^4}$ .

Step 2. Take the fourth root: $\boldsymbol{V^{1/4}\approx1.4\times10^{16}\,\mathrm{GeV}}$ — essentially the grand-unification scale. A detection at $r=0.06$ would place it at $1.6\times10^{16}\,\mathrm{GeV}$ ; the bound already caps inflation just below the GUT scale.

Step 3. The Hubble rate follows from the tensor amplitude $\mathcal P_t=rA_s=2H^2/(\pi^2 M_{\rm Pl}^2)$ , so $H=M_{\rm Pl}\sqrt{\pi^2 rA_s/2}=2.435\times10^{18}\times\sqrt{\pi^2\times0.036\times2.1\times10^{-9}/2}$ , giving $\boldsymbol{H\approx4.7\times10^{13}\,\mathrm{GeV}}$ .

Step 4. Interpret: inflation, if it happened, ran at energies a trillion times beyond the LHC. Measuring $r$ is a direct probe of physics at $10^{16}\,\mathrm{GeV}$ — the reason primordial B-mode searches are treated as a window onto quantum gravity.

3 · How many e-folds solve the horizon problem

Estimate the minimum e-folds for a GUT-scale inflationary patch to grow into today's observable universe.

Step 1. The comoving Hubble radius at the start of inflation must have been at least the comoving radius of today's observable universe — equivalently, the horizon at the end of inflation, blown up to the size it has redshifted to today, must exceed today's Hubble length.

Step 2. With near-instant reheating to $T_R\sim10^{15}\,\mathrm{GeV}$ and $T_0\approx2.35\times10^{-13}\,\mathrm{GeV}$ , the scale factor grows by $T_R/T_0\sim4\times10^{27}$ , so $N_{\rm min}\approx\ln(T_R/T_0)\approx\boldsymbol{64}$ ; lower reheating relaxes this to the familiar $N\approx50$ – $60$ . The canonical "sixty e-folds" is this horizon-matching minimum, set by the reheating scale, not a fundamental constant.

Physics of the portrait

Inflation is one of the rare frameworks whose birth can be dated almost to the night it was conceived, in the notebooks of a young particle physicist chasing an unrelated problem about magnetic monopoles.

Origins · 1 minAlan Guth and how this was discovered

Alan Guth

born 1947 · the horizon, flatness, and monopole problems and the birth of inflationary cosmology

Alan Guth was a postdoc drifting between temporary positions in particle theory when, in December 1979, he wrote in his notebook the words "SPECTACULAR REALIZATION." Studying how grand unified theories overproduce magnetic monopoles, he saw that a scalar field trapped in a false vacuum would drive a phase of exponential expansion that diluted those monopoles to invisibility — and, he quickly realized, would flatten the geometry and smooth the horizon in the same motion. His 1981 paper "Inflationary universe: A possible solution to the horizon and flatness problems" named the three puzzles and the cure, and became one of the most cited papers in cosmology. Guth was candid that his original version had a fatal flaw: the false-vacuum phase ended by bubble nucleation that never completed — the "graceful exit" problem — leaving an inhomogeneous mess. Within a year Andrei Linde, and independently Andreas Albrecht and Paul Steinhardt, replaced the first-order transition with a field slowly rolling down a smooth potential ("new inflation") that ended gracefully, and Linde soon showed with "chaotic inflation" that no special pre-conditions are needed at all. The decisive turn came at the 1982 Nuffield workshop, where Guth, Hawking, Starobinsky, Bardeen, Steinhardt, and others worked out that quantum fluctuations during inflation produce a nearly scale-invariant spectrum of density perturbations — converting inflation from a repair of initial conditions into a quantitative theory of why galaxies exist. Guth, long at MIT, has watched that prediction confirmed to ever finer precision, most sharply in Planck's red-tilted spectrum.

History

1979

— Guth conceives inflation while working on the grand-unified monopole problem

1980

— Starobinsky proposes an R² curvature-driven early accelerating phase

1981

— Guth publishes the inflationary solution to the horizon, flatness, and monopole problems

1982

— Linde, and Albrecht and Steinhardt, introduce slow-roll new inflation, curing the graceful-exit flaw

1982

— the Nuffield workshop derives the near-scale-invariant spectrum of perturbations

1983

— Linde proposes chaotic inflation, freeing the mechanism from special initial conditions

2013

— Planck confirms a red-tilted spectrum with n_s ≈ 0.965

Going further

Texts. Misner, Thorne & Wheeler, Gravitation (1973) · Wald, General Relativity (1984), §5.2 · Carroll, Spacetime and Geometry (2004), Ch. 8 · Hartle, Gravity (2003), Ch. 18 · Schutz, A First Course in General Relativity, 2nd ed. (2009), Ch. 12 · Guth, Phys. Rev. D 23, 347 (1981) · Linde, Phys. Lett. B 108, 389 (1982) · Mukhanov, Feldman & Brandenberger, Phys. Rep. 215, 203 (1992) · Baumann, "TASI Lectures on Inflation," arXiv:0907.5424 · Planck 2018 Results X: Constraints on Inflation, A&A 641, A10 (2020); BICEP/Keck, Phys. Rev. Lett. 127, 151301 (2021). Key numerical results: for $V=\tfrac12 m^2\phi^2$ at $N=60$ , $\epsilon=\eta=1/121$ , $\phi_\star=15.6\,M_{\rm Pl}$ , $n_s=1-2/N=0.967$ , $r=8/N=0.13$ ; Starobinsky $R^2$ gives $n_s=0.967$ , $r=12/N^2=0.003$ ; energy scale $V^{1/4}=1.4\times10^{16}\,\mathrm{GeV}$ and $H=4.7\times10^{13}\,\mathrm{GeV}$ at $r=0.036$ from $V=\tfrac32\pi^2 A_s r M_{\rm Pl}^4$ with $A_s=2.1\times10^{-9}$ ; horizon $N_{\rm min}\approx\ln(T_R/T_0)\approx64$ ; flatness fine-tuning $|\Omega-1|\lesssim10^{-60}$ at the Planck era. Observed values: $n_s=0.9649\pm0.0042$ (Planck 2018), $r<0.036$ (BICEP/Keck 2021), $M_{\rm Pl}=2.435\times10^{18}\,\mathrm{GeV}$ .

Builds on: The FLRW Metric · The Friedmann Equations · The Cosmological Equation of State · The Scale Factor and Hubble Parameter · Critical Density and the Density Parameters. Leads to: The Radiation-Dominated Era · The Cosmological Constant and Dark Energy · Cosmological Redshift.

Ask the tutor: Ask me to walk from the acceleration equation to the microwave sky — why a flat potential makes gravity repel, how a shrinking horizon smooths and flattens the universe, and why the slight red tilt $n_s=0.965$ is inflation's fingerprint.

Why does inflation need w < −1/3 rather than just any negative pressure?

Because acceleration is decided by $\ddot a/a=-\tfrac{4\pi G}{3}(\rho+3p)$ , whose sign is set by $\rho+3p$ , not by $p$ alone. For $\ddot a>0$ you need $\rho+3p<0$ , i.e. $w<-1/3$ ; a merely negative pressure with $-1/3<w<0$ still leaves the expansion decelerating. The factor of three is the whole point: in relativity all three principal pressures gravitate alongside the energy density, so it takes pressure more negative than $-\rho/3$ to overcome the attraction of the energy density itself. A potential-dominated scalar field with $w\approx-1$ clears this bar with room to spare.

Does inflation require exotic negative-energy matter?

No. A single ordinary scalar field on a flat enough potential does everything. Its energy density $\rho_\phi=\tfrac12\dot\phi^2+V$ is positive, and its enthalpy $\rho_\phi+p_\phi=\dot\phi^2$ is also positive, so both the weak and null energy conditions hold throughout. What inflation breaks is only the strong energy condition, $\rho+3p\geq0$ — the same one a cosmological constant violates — a statement about pressure gravitating, not about negative-energy matter. There are no ghosts or phantoms in standard slow-roll inflation; the only strangeness is that a positive potential energy produces repulsive gravity, which is a feature of general relativity.

How can quantum fluctuations end up as galaxies?

Each mode of the inflaton field starts as an ordinary quantum vibration in its vacuum. While the mode is smaller than the Hubble radius it oscillates, but as inflation stretches its wavelength past the Hubble radius no causal force can act across it in one expansion time, so the oscillation freezes and the amplitude is locked in as a classical, definite value — different in different places. These frozen ripples become ripples in the local energy density and spatial curvature. Long after inflation ends they re-enter the horizon as the seed over-densities around which gas collapses, forming the CMB acoustic peaks and eventually galaxies. The spectrum is nearly scale-invariant because each mode freezes at horizon crossing, where $H$ is almost the same for every mode.

What sets the minimum of about sixty e-folds?

It is a horizon-matching requirement, not a fundamental number. The comoving scale corresponding to today's observable universe must have been inside the Hubble radius when inflation began, so everything we see shares a causal origin. Tracing that scale forward through inflation and reheating gives a minimum set mostly by the reheating temperature, $N\approx\ln(T_R/T_0)\approx60$ – $64$ for GUT-scale reheating and fewer for lower scales. So "sixty e-folds" is the minimum for the observable window; inflation may have lasted far longer, but only the last fifty-to-sixty e-folds imprint scales we can measure.

Is inflation the same as today's dark-energy-driven acceleration?

The same mechanism at wildly different scales. Both are epochs where a component with $w\approx-1$ dominates, driving $\ddot a>0$ toward de Sitter expansion. But inflation happened in the first $\sim10^{-34}$ seconds at an energy density around $(10^{16}\,\mathrm{GeV})^4$ , and it ended when the inflaton rolled off its plateau and reheated the universe. Today's dark energy has a density some $10^{110}$ times smaller — about $(2\,\mathrm{meV})^4$ — and shows no sign of ending. Whether they are related, for instance if dark energy is a second gentler inflationary phase, is an open question, but they are separate events at vastly different energy scales.

Has inflation actually been proven?

Not conclusively, though its generic predictions are strikingly confirmed. Inflation predicted a spatially flat universe, a nearly scale-invariant spectrum with a slight red tilt, Gaussian adiabatic perturbations, and correlations on scales larger than the naive causal horizon — and every one matches Planck data, most sharply $n_s=0.9649\pm0.0042$ . What is missing is a smoking gun: a detection of primordial gravitational waves through CMB B-mode polarization, which would confirm the tensor sector and fix the energy scale. Critics note too that inflation is a framework of many models rather than a unique theory, and that its initial conditions and the eternal-inflation measure problem remain unsettled. It is the leading, extensively tested paradigm, but not yet established beyond alternatives.