← Back to Cosmic Inflation

Problems of Standard Cosmology

The Hot Big Bang model successfully predicts primordial nucleosynthesis and the cosmic microwave background. Yet it harbors three deep puzzles — the horizon problem, the flatness problem, and the monopole problem — that demand an explanation beyond the standard framework. These problems motivated the theory of cosmic inflation.

1. Introduction: Successes and Puzzles of the Big Bang

The standard Hot Big Bang model is built on the Friedmann–Lemaître–Robertson–Walker (FLRW) metric and general relativity. Its major successes include:

  • Hubble expansion: The universe is expanding, with galaxies receding at velocities proportional to their distance, $v = H_0 d$.
  • Big Bang Nucleosynthesis (BBN): The predicted abundances of light elements ($^{2}\text{H}$, $^{3}\text{He}$, $^{4}\text{He}$, $^{7}\text{Li}$) match observations when $\Omega_b h^2 \approx 0.022$.
  • Cosmic Microwave Background: A near-perfect blackbody spectrum at $T_0 = 2.7255 \;\text{K}$, confirming the universe was once hot and dense.
  • Large-scale structure: The observed distribution of galaxies is consistent with gravitational growth from small initial perturbations.

Despite these triumphs, the standard model requires extremely fine-tuned initial conditions that it cannot explain. These unexplained initial conditions constitute three fundamental puzzles:

The Horizon Problem

Why is the CMB so uniform across regions that have never been in causal contact?

The Flatness Problem

Why is the spatial curvature so extraordinarily close to zero?

The Monopole Problem

Where are the topological defects predicted by grand unified theories?

Key point: These are not failures of the Big Bang model in its domain of validity. Rather, they reveal that the standard model is incomplete — it requires special initial conditions but provides no mechanism to produce them. The FLRW metric with scale factor $a(t)$ governed by the Friedmann equation $H^2 = \frac{8\pi G}{3}\rho - \frac{k}{a^2}$ simply takes the initial state as given.

2. Derivation: The Horizon Problem

2.1 The Particle Horizon

In an FLRW universe, light travels along null geodesics with $ds^2 = 0$. The maximum comoving distance a photon could have traveled since the Big Bang defines the particle horizon. Starting from the FLRW line element:

$$ds^2 = -c^2\,dt^2 + a^2(t)\left[\frac{dr^2}{1-kr^2} + r^2\,d\Omega^2\right]$$

For a radial null geodesic ($ds^2 = 0$, $d\Omega = 0$), we have $c\,dt = a(t)\,dr/\sqrt{1-kr^2}$. The comoving particle horizon distance is:

$$\chi_H(t) = \int_0^t \frac{c\,dt'}{a(t')}$$

The proper (physical) distance to the particle horizon at time $t$ is:

$$d_H(t) = a(t)\,\chi_H(t) = a(t)\int_0^t \frac{c\,dt'}{a(t')}$$

2.2 Horizon in the Radiation-Dominated Era

During radiation domination, the scale factor evolves as $a(t) \propto t^{1/2}$. Writing $a(t) = a_0\,(t/t_0)^{1/2}$, the particle horizon becomes:

$$d_H(t) = a_0\left(\frac{t}{t_0}\right)^{1/2} \int_0^t \frac{c\,dt'}{a_0(t'/t_0)^{1/2}} = \left(\frac{t}{t_0}\right)^{1/2} \cdot c\,t_0^{1/2} \int_0^t t'^{-1/2}\,dt'$$

Evaluating the integral:

$$\int_0^t t'^{-1/2}\,dt' = 2t^{1/2}$$

Substituting back and simplifying:

$$\boxed{d_H(t) = 2ct} \quad \text{(radiation domination)}$$

Similarly, during matter domination with $a \propto t^{2/3}$, one obtains $d_H = 3ct$. The key result is that the particle horizon is always of order $ct$ — the Hubble radius $c/H$ — in a decelerating universe.

2.3 Angular Size of the Horizon at Recombination

At recombination ($z_{\text{rec}} \approx 1100$, $t_{\text{rec}} \approx 380{,}000$ years), the particle horizon had a proper size:

$$d_H(t_{\text{rec}}) \approx 2\,c\,t_{\text{rec}} \approx 2 \times (3 \times 10^5\;\text{km/s}) \times (380{,}000 \times 3.15 \times 10^7\;\text{s}) \approx 0.24\;\text{Mpc}$$

The angular diameter distance to the last scattering surface is $d_A \approx 12.8$ Mpc (accounting for the expansion). The angular size that the particle horizon at recombination subtends on today's sky is:

$$\theta_H \approx \frac{d_H(t_{\text{rec}})}{d_A} \approx \frac{0.24}{12.8} \;\text{rad} \approx 1.1^\circ$$

More careful calculation using the full integral through both radiation and matter eras gives $\theta_H \approx 2^\circ$. The full sky subtends $4\pi$ steradians, so the number of causally disconnected patches on the CMB sky is:

$$N_{\text{patches}} \approx \frac{4\pi}{\pi\,\theta_H^2} \approx \frac{4}{\theta_H^2} \approx \frac{4}{(2\pi/180)^2} \sim 10^4$$

The Horizon Problem: There are approximately $10^4$ causally disconnected regions on the CMB sky. Yet the CMB temperature is uniform to one part in $10^5$:$\Delta T / T \sim 10^{-5}$. In the standard Big Bang, there is no physical mechanism that could have brought these regions to thermal equilibrium — they have never been in causal contact. This requires an extraordinary fine-tuning of the initial conditions.

2.4 Conformal Time Perspective

The horizon problem becomes most transparent in conformal time $\eta = \int dt/a(t)$. The FLRW metric becomes:

$$ds^2 = a^2(\eta)\left[-c^2\,d\eta^2 + d\chi^2 + \chi^2\,d\Omega^2\right]$$

Light cones are 45-degree lines in the $(\eta, \chi)$ plane, just as in Minkowski space. The particle horizon is simply $\chi_H = c\,\eta$. For radiation domination,$a \propto \eta$ and $\eta \propto t^{1/2}$, so conformal time begins at $\eta = 0$ at the Big Bang. There is simply not enough conformal time between $\eta = 0$ and $\eta_{\text{rec}}$ for light to cross the entire last scattering surface. Inflation solves this by extending conformal time to$\eta \to -\infty$ before the hot Big Bang phase.

3. Derivation: The Flatness Problem

3.1 The Friedmann Equation and the Density Parameter

The Friedmann equation governs the expansion of the universe:

$$H^2 = \frac{8\pi G}{3}\rho - \frac{k c^2}{a^2}$$

where $H = \dot{a}/a$ is the Hubble parameter, $\rho$ is the total energy density, and $k = -1, 0, +1$ is the spatial curvature constant. Define the critical density and density parameter:

$$\rho_c = \frac{3H^2}{8\pi G}, \qquad \Omega = \frac{\rho}{\rho_c}$$

Dividing the Friedmann equation by $H^2$:

$$1 = \Omega - \frac{kc^2}{a^2 H^2}$$

Rearranging, we obtain the key relation:

$$\boxed{|\Omega - 1| = \frac{|k|\,c^2}{a^2 H^2}}$$

3.2 Time Evolution of the Curvature Term

The quantity $a^2 H^2 = \dot{a}^2$ determines how $|\Omega - 1|$ evolves. For a power-law expansion $a \propto t^n$:

$$\dot{a} \propto t^{n-1}, \qquad \dot{a}^2 \propto t^{2(n-1)}$$

Therefore:

$$|\Omega - 1| = \frac{|k|c^2}{\dot{a}^2} \propto t^{2(1-n)}$$

For the two standard cosmological eras:

  • Radiation domination ($n = 1/2$): $|\Omega - 1| \propto t^{2(1-1/2)} = t$. The departure from flatness grows linearly with time.
  • Matter domination ($n = 2/3$): $|\Omega - 1| \propto t^{2(1-2/3)} = t^{2/3}$. The departure grows as $t^{2/3}$.

In both cases, $n < 1$ and the universe is decelerating, so $|\Omega - 1|$always increases with time. Flatness ($\Omega = 1$) is an unstable fixed point.

3.3 Extrapolation to the Planck Time

Today, observations constrain $|\Omega_0 - 1| \lesssim 0.01$. Let us extrapolate backward. The age of the universe is $t_0 \approx 4.3 \times 10^{17}$ s. The Planck time is$t_P = \sqrt{\hbar G / c^5} \approx 5.4 \times 10^{-44}$ s. During the radiation era:

$$\frac{|\Omega(t_P) - 1|}{|\Omega_0 - 1|} \sim \frac{t_P}{t_0} \sim \frac{5.4 \times 10^{-44}}{4.3 \times 10^{17}} \sim 10^{-61}$$

This means:

$$\boxed{|\Omega(t_P) - 1| \lesssim 10^{-60}}$$

The Flatness Problem: The initial density of the universe must have been tuned to the critical density with a precision of 60 decimal places. Any deviation of even$10^{-59}$ at the Planck time would have led to either a rapid recollapse (for $\Omega > 1$) or an empty, cold universe (for $\Omega < 1$) long before the present epoch. The standard Big Bang provides no explanation for this fine-tuning.

3.4 How Inflation Drives Flatness

During inflation, the scale factor grows exponentially: $a \propto e^{Ht}$. Then:

$$\dot{a} = H\,a \propto e^{Ht}, \qquad \dot{a}^2 \propto e^{2Ht}$$

Therefore:

$$|\Omega - 1| = \frac{|k|c^2}{\dot{a}^2} \propto e^{-2Ht} \xrightarrow{t \to \infty} 0$$

After $N$ e-folds of inflation ($a \to e^N a$):

$$|\Omega - 1|_{\text{after}} = e^{-2N}\,|\Omega - 1|_{\text{before}}$$

For $N \geq 60$ e-folds: $e^{-120} \sim 10^{-52}$. Even if $|\Omega - 1|$were of order unity before inflation, after 60+ e-folds it is driven exponentially close to zero, naturally producing the observed flatness.

4. Derivation: The Monopole Problem

4.1 GUT Phase Transitions and Topological Defects

Grand Unified Theories (GUTs) predict that at temperatures $T \sim T_{\text{GUT}} \sim 10^{16}$ GeV, the strong and electroweak forces are unified under a single gauge group, such as $SU(5)$or $SO(10)$. As the universe cools below $T_{\text{GUT}}$, a symmetry-breaking phase transition occurs. By the Kibble mechanism, topological defects form wherever the Higgs field takes different vacuum orientations in causally disconnected domains.

Magnetic monopoles are point-like topological defects that arise when the vacuum manifold $\mathcal{M} = G/H$ has nontrivial second homotopy group:$\pi_2(\mathcal{M}) \neq \mathbb{1}$. This is guaranteed by the 't Hooft–Polyakov mechanism whenever a semi-simple gauge group is broken to a subgroup containing $U(1)_{\text{EM}}$.

4.2 Monopole Production Rate

The Kibble mechanism predicts that at least one monopole forms per correlation volume. The correlation length $\xi$ is bounded above by the particle horizon at the time of the phase transition:

$$\xi \lesssim d_H(t_{\text{GUT}}) \sim 2\,c\,t_{\text{GUT}} \sim \frac{c}{T_{\text{GUT}}^2/M_P} \sim \frac{M_P\,c}{T_{\text{GUT}}^2}$$

where we used the relation $t \sim M_P / T^2$ (in natural units) for the radiation era. The monopole number density at production is:

$$n_M \sim \frac{1}{\xi^3} \sim \frac{T_{\text{GUT}}^6}{M_P^3}$$

4.3 Monopole-to-Photon Ratio

The photon number density at $T_{\text{GUT}}$ is $n_\gamma \sim T_{\text{GUT}}^3$. The monopole-to-photon ratio at production is:

$$\frac{n_M}{n_\gamma} \sim \frac{T_{\text{GUT}}^6 / M_P^3}{T_{\text{GUT}}^3} = \frac{T_{\text{GUT}}^3}{M_P^3} \sim \left(\frac{10^{16}\;\text{GeV}}{10^{19}\;\text{GeV}}\right)^3 \sim 10^{-9}$$

This ratio is preserved as the universe expands (both $n_M$ and $n_\gamma$dilute as $a^{-3}$). Each monopole has a mass of order the GUT scale:

$$m_M \sim \frac{T_{\text{GUT}}}{\alpha_{\text{GUT}}} \sim \frac{10^{16}\;\text{GeV}}{1/40} \sim 10^{18}\;\text{GeV}$$

The monopole energy density today would therefore be:

$$\Omega_M = \frac{\rho_M}{\rho_c} \sim \frac{n_M \cdot m_M}{\rho_c} \sim \frac{(n_\gamma \times 10^{-9}) \times 10^{18}\;\text{GeV}}{\rho_c} \sim 10^{11}$$

The Monopole Problem: GUT monopoles would "overclose" the universe by a factor of $\sim 10^{11}$, dominating the energy density and producing a universe radically different from what we observe. The predicted monopole density $n_M/n_\gamma \sim 10^{-9}$ vastly exceeds the experimental upper bound from monopole searches, which give $n_M/n_\gamma \lesssim 10^{-30}$.

4.4 Inflationary Dilution

If inflation occurs after the GUT phase transition, the exponential expansion dilutes the monopole density exponentially. During $N$ e-folds of inflation, the scale factor increases by $e^N$, and the monopole number density (which dilutes as $a^{-3}$) decreases by:

$$n_M \to n_M \cdot e^{-3N}$$

For $N > 60$ e-folds:

$$e^{-3 \times 60} = e^{-180} \sim 10^{-78}$$

This is more than sufficient to dilute the monopole density to undetectable levels. In fact, the dilution is so extreme that it predicts at most $\sim 1$ monopole in the entire observable universe — or fewer. This is why inflation was originally proposed: Guth's 1981 paper was explicitly motivated by the monopole problem.

5. Derivation: Inflation as the Unified Solution

5.1 Conditions for Accelerated Expansion

The second Friedmann equation (the acceleration equation) is:

$$\frac{\ddot{a}}{a} = -\frac{4\pi G}{3}\left(\rho + \frac{3p}{c^2}\right)$$

For accelerated expansion ($\ddot{a} > 0$), we need:

$$\rho + \frac{3p}{c^2} < 0 \quad \Longrightarrow \quad p < -\frac{\rho c^2}{3}$$

Defining the equation of state parameter $w = p/(\rho c^2)$, the condition becomes:

$$\boxed{w < -\frac{1}{3}}$$

Normal matter ($w = 0$) and radiation ($w = 1/3$) always decelerate expansion. Inflation requires a form of energy with strong negative pressure. The simplest realization is a cosmological constant or vacuum energy with $w = -1$, giving $p = -\rho c^2$.

5.2 De Sitter Expansion

With constant vacuum energy $\rho = \rho_\Lambda = \text{const}$, the Friedmann equation (neglecting curvature) gives:

$$H^2 = \frac{8\pi G}{3}\rho_\Lambda = \text{const} \equiv H_I^2$$

The solution is exponential (de Sitter) expansion:

$$\boxed{a(t) = a_0\,e^{H_I t}}$$

5.3 The Shrinking Comoving Hubble Radius

The comoving Hubble radius is defined as:

$$R_H^{\text{com}} = \frac{c}{aH} = (aH)^{-1} \cdot c$$

This is the characteristic comoving scale within which causal physics can operate at any given time. For a power-law expansion $a \propto t^n$:

$$aH = \dot{a} \propto t^{n-1}, \qquad (aH)^{-1} \propto t^{1-n}$$

  • Standard cosmology ($n < 1$): $(aH)^{-1}$ grows with time. New modes constantly enter the horizon.
  • Inflation ($a \propto e^{H_I t}$): $(aH)^{-1} \propto e^{-H_I t}$ shrinks with time. Modes exit the horizon during inflation.

This is the essential mechanism of inflation. During the inflationary epoch, the comoving Hubble radius decreases rapidly. Perturbation modes with comoving wavelength $\lambda$ that were initially sub-horizon ($\lambda < (aH)^{-1}$) get pushed outside the horizon ($\lambda > (aH)^{-1}$) as the comoving Hubble radius shrinks past them. After inflation ends and standard decelerated expansion resumes, $(aH)^{-1}$ begins growing again, and these modes re-enter the horizon.

5.4 Solving the Horizon Problem

The inflation solution to the horizon problem is elegant: the entire observable universe was within a single causal patch before inflation began. During inflation, this patch was stretched exponentially, far beyond what would become our observable universe. The comoving Hubble radius at the start of inflation was much larger than it is today:

$$(aH)^{-1}_{\text{start}} \gg (a_0 H_0)^{-1}$$

For this to work, we need the comoving Hubble radius at the start of inflation to exceed the current comoving Hubble radius. This requires at minimum:

$$N \geq \ln\left(\frac{a_0 H_0}{a_{\text{end}} H_I}\right) \approx 60\;\text{e-folds}$$

The exact number depends on the energy scale of inflation and the details of reheating, but $N \sim 50$$70$ is the standard requirement. With 60 e-folds, regions that appear widely separated on the CMB sky were once in close causal contact, naturally explaining the uniformity of the CMB temperature.

6. Observational Applications

6.1 CMB Isotropy: COBE, WMAP, Planck

The horizon problem is directly confronted by precision measurements of the CMB:

MissionYearAngular ResolutionKey Result
COBE/DMR1992$7°$First detection of CMB anisotropies: $\Delta T/T \sim 10^{-5}$
WMAP2003–2012$0.2°$Confirmed near-scale-invariant spectrum, $n_s = 0.963 \pm 0.012$
Planck2013–2018$5'$Precision: $n_s = 0.9649 \pm 0.0042$, $r < 0.10$

All three missions confirm the central mystery: the CMB is remarkably uniform across scales much larger than the particle horizon at recombination. The tiny anisotropies that are observed ($\Delta T / T \sim 10^{-5}$) are consistent with quantum fluctuations generated during inflation and stretched to superhorizon scales.

6.2 Curvature Constraints from CMB + BAO

The flatness problem predicts that inflation should have driven the spatial curvature extremely close to zero. Modern observations provide stringent tests:

  • Planck 2018 (CMB alone): $\Omega_k = 0.001 \pm 0.002$
  • Planck + BAO: $\Omega_k = 0.0007 \pm 0.0019$
  • Combined constraints: $|\Omega_k| < 0.005$ at 95% C.L.

These measurements are consistent with a spatially flat universe to sub-percent precision, exactly as predicted by inflation. The position of the first acoustic peak in the CMB power spectrum at $\ell \approx 220$ is the most direct probe, as it corresponds to the sound horizon at recombination and is sensitive to spatial curvature through the angular diameter distance.

6.3 Monopole Search Experiments

If inflation did not occur, the predicted monopole abundance would be easily detectable. The absence of monopoles provides indirect evidence for inflation:

ExperimentMethodUpper Bound on Flux
MACRO (1989–2000)Scintillators + streamer tubes$< 1.4 \times 10^{-16}\;\text{cm}^{-2}\text{s}^{-1}\text{sr}^{-1}$
MoEDAL (LHC)Nuclear track detectorsProduction cross-section limits at TeV scale
IceCubeCherenkov radiation in iceCatalysis cross-section bounds

The Parker bound from galactic magnetic fields provides a complementary constraint: too many monopoles would drain the Galactic magnetic field faster than it can regenerate, giving$F_M \lesssim 10^{-15}\;\text{cm}^{-2}\text{s}^{-1}\text{sr}^{-1}$. All observations are consistent with a monopole density so low that inflation (or an equivalent dilution mechanism) must have occurred.

7. Historical Context

The development of inflationary cosmology involved several key contributions, each addressing different aspects of the problems described above.

Alexei Starobinsky (1979–1980)

Starobinsky proposed the first model of cosmic inflation based on quantum gravity corrections to Einstein's equations. His model added an $R^2$ term to the Einstein–Hilbert action, leading to a de Sitter phase in the early universe. This model was motivated by trace anomaly considerations rather than the horizon or monopole problems, but it achieved the same effect. Remarkably, the Starobinsky model remains one of the best-fitting models to Planck data four decades later.

Alan Guth (1981) — Old Inflation

Guth's seminal paper "Inflationary Universe: A Possible Solution to the Horizon and Flatness Problems" introduced the term "inflation" and explicitly identified the three problems discussed in this lecture. His model used a first-order phase transition: the inflaton field was trapped in a false vacuum state with $V(\phi) > 0$, driving exponential expansion. The problem was the "graceful exit": the phase transition completed through bubble nucleation, but the bubbles never percolated to fill space and thermalize the universe.

Andrei Linde (1982) — New Inflation / Chaotic Inflation

Linde proposed "new inflation" using a Coleman–Weinberg potential with a slow roll from the top of the potential. Later (1983), he introduced "chaotic inflation" with the radical idea that inflation does not require a phase transition at all — a simple potential like $V = m^2\phi^2/2$ suffices if the field starts at large values. This dramatically simplified the inflationary paradigm.

Albrecht & Steinhardt (1982)

Independently of Linde, Albrecht and Steinhardt developed a "new inflation" model that solved the graceful exit problem of Guth's original proposal. Their model used a slowly rolling scalar field in a Coleman–Weinberg type potential, allowing inflation to end smoothly as the field rolls to its true vacuum, followed by reheating.

Mukhanov & Chibisov (1981)

Working with the Starobinsky model, Mukhanov and Chibisov made the crucial calculation showing that quantum fluctuations during inflation produce a nearly scale-invariant spectrum of perturbations. This connected inflation to observable predictions — the seeds of large-scale structure — making it a testable theory.

Timeline summary: Starobinsky (1980) $\to$ Guth (1981, old inflation) $\to$ Mukhanov & Chibisov (1981, quantum perturbations)$\to$ Linde, Albrecht & Steinhardt (1982, new inflation) $\to$ Linde (1983, chaotic inflation). Each step addressed limitations of the previous models while preserving the core insight: a period of accelerated expansion solves the horizon, flatness, and monopole problems.

8. Python Simulations

The following simulations visualize the key concepts discussed above: the evolution of the comoving Hubble radius through different cosmological eras, and the evolution of the curvature parameter showing how inflation drives flatness.

Comoving Hubble Radius: Horizon Exit and Re-entry During Inflation

Python
script.py87 lines

Click Run to execute the Python code

Code will be executed with Python 3 on the server

Flatness Evolution: |Omega-1| vs Scale Factor With and Without Inflation

Python
script.py139 lines

Click Run to execute the Python code

Code will be executed with Python 3 on the server

Summary of Key Results

Particle Horizon (radiation era)

$$d_H(t) = a(t)\int_0^t \frac{c\,dt'}{a(t')} = 2ct \quad (a \propto t^{1/2})$$

Horizon Problem

$$\theta_H \approx 2° \implies N_{\text{patches}} \sim \frac{4\pi}{\pi\theta_H^2} \sim 10^4 \quad \text{causally disconnected regions}$$

Flatness Problem

$$|\Omega - 1| = \frac{|k|c^2}{a^2 H^2} \propto \begin{cases} t & \text{(radiation)} \\ t^{2/3} & \text{(matter)} \end{cases} \implies |\Omega(t_P) - 1| \lesssim 10^{-60}$$

Monopole Problem

$$\frac{n_M}{n_\gamma} \sim \left(\frac{T_{\text{GUT}}}{M_P}\right)^3 \sim 10^{-9} \implies \Omega_M \sim 10^{11} \quad \text{(overclosure)}$$

Inflation Condition

$$\ddot{a} > 0 \iff \rho + 3p < 0 \iff w < -\frac{1}{3}$$

Inflationary Solutions

$$|\Omega - 1| \propto e^{-2N}, \qquad n_M \propto e^{-3N}, \qquad (aH)^{-1} \propto e^{-N}$$