1. Why Strings?
Point particles cannot consistently describe quantum gravity. Extended one-dimensional objects — strings — resolve the ultraviolet divergences that plague graviton scattering and naturally produce a massless spin-2 particle: the graviton itself.
The Graviton Problem
The Einstein-Hilbert action for general relativity takes the form:
$$S_{\text{EH}} = \frac{1}{16\pi G_N}\int d^4x\,\sqrt{-g}\,R$$
Newton's constant $G_N$ has dimensions of $[\text{length}]^2$ in natural units. This means the gravitational coupling constant is dimensionful, making the theory non-renormalizable as a quantum field theory. At each loop order, new counterterms with higher powers of curvature are required:
$$\Gamma_{\text{div}} \sim \int d^4x\,\sqrt{-g}\left(c_1 R^2 + c_2 R_{\mu\nu}R^{\mu\nu} + c_3 R_{\mu\nu\rho\sigma}R^{\mu\nu\rho\sigma} + \cdots\right)$$
The Planck length $\ell_P = \sqrt{\hbar G_N/c^3} \approx 1.6 \times 10^{-35}\,\text{m}$ sets the scale where quantum gravitational effects dominate. Graviton scattering amplitudes diverge as:
$$\mathcal{A}(s) \sim G_N s \left(1 + G_N s \log\frac{s}{\mu^2} + \cdots\right)$$
where $s$ is the Mandelstam variable. The series is uncontrollable at $\sqrt{s} \sim M_P$.
Strings Resolve UV Divergences
Replace point particles with one-dimensional extended objects: strings. A string has a characteristic length scale, the string length:
$$\ell_s = \sqrt{\alpha'} \qquad T = \frac{1}{2\pi\alpha'}$$
where $\alpha'$ is the Regge slope and $T$ is the string tension. In string theory, the interaction vertex is not a point but is smoothly spread over the string worldsheet. This smearing replaces the UV divergent graviton propagator:
$$\frac{1}{k^2} \longrightarrow \frac{1}{k^2}\,e^{-\alpha' k^2/2}$$
The exponential damping at high momenta $k \gg 1/\ell_s$ renders all loop amplitudes finite. The string scattering amplitude for four gravitons (the Virasoro-Shapiro amplitude) is:
$$\mathcal{A}_4 = \frac{\Gamma(-\alpha' s/4)\,\Gamma(-\alpha' t/4)\,\Gamma(-\alpha' u/4)}{\Gamma(1 + \alpha' s/4)\,\Gamma(1 + \alpha' t/4)\,\Gamma(1 + \alpha' u/4)}$$
This amplitude is manifestly UV finite and reduces to the Einstein gravity result at low energies$\alpha' s \ll 1$.
The Graviton from Closed Strings
Open strings have endpoints; closed strings form loops. The mass spectrum of a string is quantized:
$$M^2 = \frac{2}{\alpha'}\left(N + \tilde{N} - 2\right) \qquad (\text{closed bosonic string})$$
with the level-matching constraint $N = \tilde{N}$. At level$N = \tilde{N} = 1$, the massless states decompose into:
$$\alpha^{\mu}_{-1}\tilde{\alpha}^{\nu}_{-1}|0;p\rangle \quad\longrightarrow\quad g_{\mu\nu} \oplus B_{\mu\nu} \oplus \Phi$$
The symmetric traceless part is a massless spin-2 field: the graviton $g_{\mu\nu}$. The antisymmetric part is the Kalb-Ramond B-field $B_{\mu\nu}$, and the trace is the dilaton $\Phi$. By the Weinberg-Witten theorem, any consistent interacting massless spin-2 particle must be the graviton of general relativity.
Open vs Closed Strings
Open strings carry gauge degrees of freedom at their endpoints. The massless spectrum of the open bosonic string at level $N = 1$ is:
$$M^2_{\text{open}} = \frac{1}{\alpha'}(N - 1) \qquad N=1:\;\alpha^{\mu}_{-1}|0;p\rangle \sim A_\mu$$
This is a massless vector boson — a gauge field. An important topological fact: open string loops inevitably produce closed strings. If your theory contains open strings, it mustcontain closed strings and hence gravity. This is why string theory is automatically a theory of quantum gravity.
Why D = 26 and D = 10?
Quantum consistency of the string requires cancellation of the Weyl anomaly on the worldsheet. The total central charge must vanish:
$$c_{\text{total}} = c_{\text{matter}} + c_{\text{ghost}} = 0$$
For the bosonic string, each spacetime coordinate $X^\mu$ contributes $c = 1$ to the matter central charge, while the reparametrization ghosts contribute $c_{\text{ghost}} = -26$:
$$D \cdot 1 + (-26) = 0 \quad\Longrightarrow\quad D = 26$$
For the superstring, each dimension contributes $c = 3/2$ (bosonic + fermionic), and the superconformal ghosts give $c_{\text{ghost}} = -15$:
$$D \cdot \frac{3}{2} + (-15) = 0 \quad\Longrightarrow\quad D = 10$$
Equivalently, the critical dimension can be derived from requiring the vanishing of the one-loop partition function anomaly, or from the light-cone quantization condition that the Lorentz algebra closes only in $D = 26$ (bosonic) or $D = 10$ (super).
The Regge Trajectory
The leading Regge trajectory relates spin $J$ to mass:
$$J = \alpha' M^2 + a_0$$
with intercept $a_0 = 1$ for the open bosonic string. The number of states at level$N$ grows exponentially:
$$d(N) \sim N^{-\frac{D+1}{4}}\exp\left(4\pi\sqrt{\frac{N(D-2)}{24}}\right) \qquad (N \gg 1)$$
This exponential growth defines the Hagedorn temperature$T_H = 1/(4\pi\sqrt{\alpha'(D-2)/24})$, above which the canonical ensemble diverges — hinting at a phase transition to a new regime of string physics.
Simulation: Bosonic String Spectrum
The following simulation plots the mass spectrum of the bosonic open string, showing the tachyon at level $N = 0$, massless states at $N = 1$, and the massive tower. The exponential growth of degeneracies is shown on a logarithmic scale.
Click Run to execute the Python code
Code will be executed with Python 3 on the server