Past exam of the mathematics course of the University of Cambridge / 2001 / iii / Paper 66 (Advanced String Theory)

In conformal gauge, variation of the flat-background Polyakov action gives the bulk wave equation and the endpoint termFor each endpoint and target-space direction, two basic choices make this term vanish. A Neumann boundary condition allows arbitrary $\delta X^{\mu }$ and requires ${\partial }_{\sigma }{X}^{\mu }=0$. A Dirichlet boundary condition fixes $X^{\mu }$ at the endpoint, hence $\delta X^{\mu }=0$; for a fixed endpoint location its time derivative also vanishes. ThusBoth are consistent with the principle of stationary action and zero energy flux through a stationary endpoint. Different directions can use different conditions. Background two-form or boundary gauge fields can modify tangential conditions into mixed ones; the displayed alternatives are the basic flat, unforced choices.

The common Neumann boundary conditions define a $(p+1)$-dimensional timelike worldvolume: one time direction and $p$ spatial directions. The remaining Dirichlet boundary conditions fix its transverse position. An open string therefore ends on a D-brane of spatial dimension $p$, a D$p$-brane. For the critical bosonic string theory there are $25-p$ transverse directions.

If the Dirichlet constants agree at the two endpoints, both ends lie on the same planar D-brane. If they differ, the string stretches between two parallel D-branes with the same orientation but different positions. Neumann directions are tangent to the brane; Dirichlet directions are transverse to it. The brane is dynamical: tangential massless open-string polarizations supply a gauge field, while transverse ones describe its displacement, as in worldvolume fields from open-string massless states. The unprojected bosonic brane also has a tachyon; the boundary-condition interpretation does not make that ground state supersymmetrically stable.

A supersymmetric type-II D-brane carries a Ramond–Ramond potential charge and saturates the corresponding generalized BPS bound in supersymmetry: its brane tension equals the charge-dependent central/tensorial-charge bound in the relevant normalization. It is therefore a BPS state, preserving sixteen of the thirty-two spacetime supercharges. Identical parallel branes with the same orientation impose the same supersymmetry projector. This statement does not apply to arbitrary unstable non-BPS branes or to the bosonic branes of the preceding parts.

One direct demonstration of the parallel D-brane no-force identity is the annulus vacuum amplitude. For separation $r$, its oscillator-dependent part is proportional toHereThe first two terms describe the projected Neveu–Schwarz sector, and the last is the oppositely signed Ramond sector contribution. The Jacobi abstruse identity gives $f_3^8-f_4^8-f_2^8=0$ for every $t$. ThereforeIn the closed-string channel, attractive graviton and dilaton exchange cancel repulsive Ramond–Ramond potential exchange. The classical probe calculation in the D-brane supergravity solution gives the same cancellation. The common supersymmetry allows separation to remain a flat modulus; relative angles, reversed RR charge or supersymmetry-breaking backgrounds can remove this protection. For a brane–antibrane pair the RR interaction reverses and does not cancel the NS-NS attraction. The charge and half-supersymmetry interpretation is established in Dirichlet-Branes and Ramond-Ramond Charges.

For $N$ parallel type-II D-branes, an oriented open string has endpoint labels $i,j$, its Chan-Paton factors. At coincidence, the massless states form $N\times N$ matrices, giving adjoint fields of $U(N)$. Tangential polarizations produce the Yang-Mills theory gauge field; transverse polarizations produce $9-p$ adjoint scalar fields. The leading low-energy action is the dimensional reduction of ten-dimensional super Yang-Mills theory to $p+1$ dimensions. Expanding the Dirac-Born-Infeld action gives the gauge and scalar kinetic terms, while dimensional reduction supplies the Hermitian commutator potential in D-brane Yang-Mills theory:The sign is important: for Hermitian $\varphi$, the commutator is anti-Hermitian. Zero potential requires commuting scalars, whose simultaneous eigenvalues are brane positions through the D-brane scalar-position normalizationThis makes the Higgs mechanism geometrical. Separating branes gives masses to strings joining different positions; bringing them together gives Coincident-D-brane gauge enhancement.

Two parallel D3-branes in type IIB superstring theory give four-dimensional N=4 super Yang-Mills theory with gauge group $U(2)$, containing a gauge field, six real adjoint scalar fields and four adjoint Weyl fermions. On its Coulomb branch, chooseFor distinct position vectors, $U(2)$ breaks to $U(1)\times U(1)$. The off-diagonal vector multiplets are the two orientations of a string joining the branes, andThe same mass follows from the scalar covariant-derivative term, $m_W^2={\sum }_I({\varphi }_1^I-{\varphi }_2^I{)}^2$. The overall $U(1)$ describes center-of-mass motion; the relative $SU(2)$ sector is Higgsed to its Cartan $U(1)$. At coincidence the off-diagonal states become massless and the full $U(2)$ gauge symmetry is restored.

A D-brane is a localized RR-charged source whose long-distance fields solve supergravity. The extremal charged $p$-brane solutions carry the same RR charges, brane tension and preserved supersymmetry as the open-string construction. For the standard asymptotically flat cases $p<7$, the D-brane supergravity solution in the string frame isThe function $H$ is a positive harmonic function away from the source, in $9-p$ transverse dimensions. In an electric gauge, $C_{0\cdots p}=g_s^{-1}(H^{-1}-1)$, with its orientation chosen to match the source. Magnetic branes use the dual flux; for the D3-brane, the five-form includes both electric and magnetic pieces and is self-dual. Killing spinors obey one brane projector, leaving sixteen supersymmetries. Aligned multicenter harmonic functions describe several branes with no static binding force.

To check this identification directly, insert an equally oriented static probe into the solution. Put ${\mu }_p=(2\pi {)}^{-p}{\alpha }^{\prime -(p+1)/2}$, so the asymptotic tension is ${\mu }_p/g_s$. The induced-volume factor is $H^{-(p+1)/4}$; combined with $e^{-\Phi }$, its DBI action density is $-{\mu }_p{g}_s^{-1}{H}^{-1}$. Its Wess-Zumino brane coupling contributes $+{\mu }_p{g}_s^{-1}(H^{-1}-1)$. Hence the total static Lagrangian density is the constant $-{\mu }_p/g_s$, independently of position. This gives the parallel D-brane no-force identity in the spacetime description.

The supergravity background is the back-reacted, coarse long-distance description, while the boundary-condition description exposes the light open-string degrees of freedom. Controlled use of supergravity requires curvature small in string units and suitable local coupling. Singular cores and the special codimension-two or domain-wall cases are not automatically resolved by the elementary harmonic ansatz.

A $(p+1)$-form gauge potential $A_{p+1}$ couples electrically to a spatial $p$-brane through the Wess-Zumino brane couplingThe worldvolume dimension supplies exactly the degree required for integration. Under $A\mapsto A+d\Lambda$, the change is $e{\int }_{\partial \Sigma }{X}^{\ast }\Lambda$, so a closed worldvolume is gauge invariant; an open one needs suitable endpoint couplings. If $F_{p+2}=d{A}_{p+1}$ has an ordinary dual potential, $\ast F$ has degree $d-p-2$ and the dual potential degree $d-p-3$. Thus the electric and magnetic dimensions of a form-coupled brane areElectric charge is measured by ${\int }_{S^{d-p-2}}\ast F$, magnetic charge by ${\int }_{S^{p+2}}F$, on spheres linking the corresponding objects. Flux normalization, and possible Chern-Simons modifications of charge definitions, must be fixed in interacting backgrounds; this is the basic linear form-field dictionary.

In ten dimensions the common NS-NS fields are the graviton $g_{\mu \nu }$, Kalb–Ramond field $B_2$, and dilaton $\Phi$. Their on-shell bosonic polarizations are $35,28,1$, respectively. The graviton count is $10(10-3)/2=35$, and a massless $k$-form has $\left(\genfrac{}{}{0ex}{}{8}{k}\right)$ transverse polarizations. The RR fields distinguish the two theories:

Each gravitino has $56$ physical polarizations, and each dilatino $8$, giving $128$ fermions in either theory. In IIA the two supersymmetry sectors have opposite ten-dimensional chiralities, and the dilatino in each sector has chirality opposite to its gravitino. In IIB both gravitini can be designated positive-chirality and both dilatini negative-chirality. The self-duality of the IIB five-form halves the unconstrained four-form count $\left(\genfrac{}{}{0ex}{}{8}{4}\right)=70$ to $35$. These are the massless type IIA spectrum and massless type IIB spectrum, not raw component counts.

The Neveu–Schwarz two-form couples electrically to the fundamental string and magnetically to the NS5-brane in both theories. IIA $C_1$ couples electrically to a D0-brane and magnetically to a D6-brane; $C_3$ couples electrically to a D2-brane and magnetically to a D4-brane. The magnetic descriptions use dual seven-form and five-form potentials, not extra independent polarizations. In IIB, $C_2$ gives D1-brane and D5-brane charges, while self-dual $C_4$ gives the self-dual D3-brane. The scalar $C_0$ couples to a Euclidean D-instanton, whose magnetic partner is the D7-brane. Momentum and, after circle compactification, Kaluza-Klein monopole charge come from the metric sector.

The full brane inventory also involves nonpropagating top-form potentials. A D8-brane is an IIA domain-wall source for Romans mass $F_0$, described dually by a nine-form potential; it is not generated by the propagating $C_1,C_3$ massless spectrum alone. A space-filling IIB D9-brane has a ten-form coupling and requires appropriate global RR-charge consistency. Such fields add no ordinary massless polarizations. This distinction avoids confusing the list of propagating particles with all possible charged extended objects.

For T-duality, compactify a spatial coordinate with $X^9\sim X^9+2\pi R$. Closed strings have integer momentum $n$ and winding $w$, withThe mapleaves $p_L$ unchanged and reverses $p_R$. Equivalently, it reverses $X_R^9$ while keeping $X_L^9$, and reverses the corresponding right-moving worldsheet fermion to preserve worldsheet supersymmetry. Multiplication of the right Ramond ground-state spinor by ${\Gamma }^9$ reverses its ten-dimensional chirality. Left and right Ramond projections therefore change from opposite chirality to equal chirality, or conversely:The zero-mode mass contribution $(n/R{)}^2+(wR/{\alpha }^{\prime }{)}^2$ and the oscillator spectrum are unchanged. The Buscher dilaton shift gives $g_s^{\prime }=g_s\sqrt{{\alpha }^{\prime }}/R$, which preserves the nine-dimensional gravitational coupling proportional to $g_s^2/R$. The mixed metric and two-form components along the circle are exchanged; RR potentials with a circle leg map to one lower degree and those without it to one higher degree, in the corresponding gauge-covariant description.

For open strings, the derivative relations ${\partial }_{\tau }\tilde{X}={\partial }_{\sigma }X$ and ${\partial }_{\sigma }\tilde{X}={\partial }_{\tau }X$ exchange Neumann boundary conditions with Dirichlet boundary conditions. Consequently a wrapped D$p$-brane becomes a D$(p-1)$-brane, while a transverse one becomes a wrapped D$(p+1)$-brane. The even-dimensional IIA branes and odd-dimensional IIB branes therefore transform consistently with the closed-string chirality flip; transverse positions and compact gauge Wilson lines are exchanged.

Retain only fields independent of the circle coordinate $y$, with coordinate period $2\pi r_0$. Use the string-frame metric $g_{\mu \nu }$ and the circle reduction of eleven-dimensional supergravity ansatzThe minus sign in the fibre one-form is a convention chosen to give the usual ${\tilde{F}}_4=d{C}_3-C_1\wedge H_3$ below; changing the sign of $C_1$ consistently changes no physics. The metric supplies the ten-dimensional graviton, dilaton and RR one-form, while the eleven-dimensional three-form supplies the Kalb–Ramond field and RR three-form. DefineThen $F_4^{(11)}={\tilde{F}}_4+H_3\wedge \eta$. In particular $d{\tilde{F}}_4=-F_2\wedge H_3$, the reduced Bianchi identity. The field strength, not bare $d{C}_3$ alone, is the gauge-covariant horizontal component.

For the gravitational term, the block determinant givesA useful two-step curvature calculation first removes the overall factor $e^{-2\Phi /3}$. The intermediate metric $d{\bar{s}}^2=d{s}_s^2+e^{2\Phi }{\eta }^2$ hasThe eleven-dimensional conformal-curvature identity then givesCombining the terms yieldsIntegration by parts gives $\int \sqrt{-g}{e}^{-2\Phi }{\nabla }^2\Phi =2\int \sqrt{-g}{e}^{-2\Phi }(\nabla \Phi {)}^2$ modulo boundary terms. Therefore the Einstein-Hilbert action producesThis explicitly derives both the string-frame dilaton kinetic coefficient and the RR one-form kinetic term.

Use the normalized form norm $|F_k{|}^2=F_{{\mu }_1\cdots {\mu }_k}{F}^{{\mu }_1\cdots {\mu }_k}/k!$, interpreting the printed four-form square in this convention. The horizontal and fibre pieces are orthogonal in the $\eta$ basis. Four horizontal inverse metrics contribute $e^{8\Phi /3}$ to $|{\tilde{F}}_4{|}_{11}^2$. Three horizontal inverse metrics and one fibre inverse metric contribute $e^{2\Phi /3}$ to $|H_3\wedge \eta {|}_{11}^2$. ThusThe NS-NS three-form has the same $e^{-2\Phi }$ prefactor as gravity; the RR kinetic terms do not have it in the string frame.

Finally reduce the topological term using the coordinate decomposition, which makes its signs transparent. The part of $A_3\wedge F_4^{(11)}\wedge F_4^{(11)}$ containing exactly one $dy$ isThe purely horizontal eleven-form vanishes on the ten-dimensional base. Sinceintegration over the base turns the bracket into $3B_2\wedge G_4\wedge G_4$, modulo the displayed exact term. The reduction of the eleven-dimensional Chern-Simons term therefore gives $-\frac{1}{2}\int B_2\wedge G_4\wedge G_4$.

Absorbing the coordinate circumference into ${\kappa }_{10}^2={\kappa }_{11}^2/(2\pi r_0)$, the complete bosonic massless type IIA supergravity action in these conventions isDropping boundary terms assumes appropriate boundary conditions; nontrivial flux bundles may require patchwise potentials. Higher circle modes and higher-derivative corrections are excluded by this low-energy zero-mode reduction. No Romans-mass term arises from the ordinary ansatz.

The M-theory origin of type IIA D-branes follows by momentum, wrapping and magnetic geometry:

An M2-brane wrapped around the circle gives the fundamental string, and an unwrapped M5-brane gives the NS5-brane. These are NS objects rather than D-branes. In physical string/Planck units, the M-theory circle duality relations areThey involve the physical asymptotic radius, distinguished from the arbitrary coordinate period used in the reduction ansatz. For example, $m_{D0}=1/(g_s{\ell }_s)=1/R_{11}$ is one momentum quantum. With $T_{M2}=1/[(2\pi {)}^2{\ell }_p^3]$ and $T_{M5}=1/[(2\pi {)}^5{\ell }_p^6]$,The D6-brane is purely gravitational in eleven dimensions, rather than an elementary six-dimensional membrane; the circle-fibration magnetic charge gives its ten-dimensional interpretation.

A D8-brane is the important qualification. It sources the zero-form flux of massive type IIA supergravity and is not obtained as an ordinary wrapped M2/M5 or Kaluza-Klein object of the undeformed circle reduction above. Its extended/generalized lifts are outside this simple dictionary. The massless-circle interpretation of the other branes is developed in Four Lectures on M-theory.