Extended Operators, Defects, and Brane Charges

An extended observable is not specified by the shape of its support alone. A Wilson loop, for example, also carries a representation, an electric or magnetic charge, a scalar coupling, a normalization, and a choice of whether it is a genuine line or the boundary of a surface. Holographically, those data select a bulk string, brane, flux configuration, or fully backreacted geometry whose worldvolume reaches the boundary support. This page develops that dimension-and-charge dictionary and works through a controlled fundamental-string example in $AdS_5\times S^5$.

Required background. Bulk fields and boundary operators supplies the local field/operator map. Helpful background. Conformal boundaries and defects supplies defect CFT kinematics, while BPS Wilson, ‘t Hooft, and dyonic lines supplies protected examples and charge data.

From boundary support to a bulk worldvolume

Let $\Sigma_p$ be the $p$-dimensional spacetime support of a boundary defect $\mathcal D_p(\Sigma_p)$. Here $p$ labels the support dimension; it is not the label in “D$p$-brane.” If $\overline X$ denotes the conformally compactified AdS factor and $K$ the compact internal space, a factorized bulk carrier has the schematic form

$$\mathcal W_{p+r+1}\simeq M_{p+1}\times Y_r, \qquad Y_r\subset\overline X\times K, \qquad \partial M_{p+1}\cap\partial\overline X=\Sigma_p.$$

The noncompact factor $M_{p+1}$ has one more dimension than the boundary support because it extends along the holographic radial direction. The additional compact factor $Y_r$ can lie in directions transverse to $M_{p+1}$ inside AdS, in the internal space, or partly in both; only when it is a closed internal cycle will we write $Y_r=\Gamma_r\subset K$. The full worldvolume has dimension $p+r+1$. Thus the same boundary line can anchor an F1 worldsheet with an $AdS_2$ factor, a D3-brane with $AdS_2\times S^2$ worldvolume whose $S^2$ lies in $AdS_5$, or a D5-brane with $AdS_2\times S^4$ worldvolume whose $S^4$ lies in $S^5$. Neither the dimension count nor the presence of a compact factor by itself implies a nonzero homology charge.

For an extended operator inserted in a fixed theory, the normalized path integral is approximated semiclassically by a sum over admissible saddles,

$$\langle\mathcal D_p(\Sigma_p)\rangle =\frac{Z[\mathcal D_p]}{Z[1]} \sim \sum_\alpha e^{-S_{E,\mathrm{ren}}^{(\alpha)}} Z_{\text{1-loop}}^{(\alpha)} \left(1+\text{higher corrections}\right).$$

Here $Z[1]$ is the defect-free reference partition function, $\alpha$ labels the admissible saddles, and $Z_{\text{1-loop}}^{(\alpha)}$ is the functional-determinant factor from small bosonic, fermionic, and ghost fluctuations about saddle $\alpha$, with zero modes treated separately. A boundary condition, or an interface that joins different theories or couplings, instead changes the boundary or gluing data of the partition function; its normalization must be defined against an explicitly chosen reference rather than assumed to equal an ordinary expectation value in one theory. In either case the saddle sum matters: different fillings can exchange dominance, and a saddle can disappear when its charge or boundary condition is changed. The action must include tension, couplings to bulk form fields, worldvolume gauge fields, internal embeddings, and the boundary terms or counterterms appropriate to the observable. A single exponential is a controlled approximation only when one saddle dominates and its fluctuations and backreaction are small.

A dimension-and-charge dictionary

The table separates the universal radial-dimension rule from model-dependent top-down realizations. On a narrow screen, scroll horizontally so each entry remains readable.

Boundary data and common bulk carriers. The examples in the carrier column are not determined by support dimension alone.

Boundary observable	AdS-reaching factor	Common top-down carrier	Additional data that select the saddle
Wilson, 't Hooft, or dyonic line	A two-dimensional filling ending on the one-dimensional contour	F1, D1, or an IIB (m, n) string carrying integer F1/D1 charges; fluxed D3 or D5 for suitable large representations	Representation, electric/magnetic lattice, scalar coupling, orientation, global form, and boundary terms
Codimension-two surface defect in four dimensions	A three-dimensional filling ending on the two-dimensional support	For a half-BPS example, a D3 with an AdS3 × S1 worldvolume; in other models, a different brane or form-field configuration	Monodromy, Levi subgroup, defect parameters, internal cycle, and worldvolume flux
Three-dimensional defect in a four-dimensional CFT	A four-dimensional hypersurface reaching the boundary defect	For the D3/D5 defect CFT, a probe D5 with an AdS4 × S2 worldvolume	Localized fields, rank jump, preserved symmetry, flux, and probe or backreacted regime
One-sided conformal boundary condition	A bulk hypersurface on which the spacetime can end	An end-of-the-world brane in models that admit such a description	Boundary condition, preserved symmetry, boundary degrees of freedom, tension, and consistency of the bulk endpoint
Two-sided interface or domain wall	A bulk hypersurface or field profile separating two asymptotic regions	A Janus-type geometry or an interface brane in suitable models	The two theories, gluing or transmission law, preserved symmetry, couplings on each side, and interface-localized data
Higher-form charged defect or junction	A filling, linking cycle, or junction worldvolume of the required dimension	A brane electrically or magnetically coupled to a bulk form; possibly with attached lower-dimensional branes	Charge lattice, linking phase, endpoint law, internal wrapping, flux quantization, and anomaly inflow

The F1, D1, and IIB $(m,n)$ strings carry electric, magnetic, and dyonic external charges in the standard $AdS_5\times S^5$ construction Witten 1998, “Baryons and Branes,” §2, printed p. 3, PDF. The $AdS_4\times S^2$ D5 example is a concrete defect-CFT construction, not a universal representation of every interface or boundary DeWolfe, Freedman, and Ooguri 2002, §§2–3, PDF. Likewise, half-BPS surface operators in $\mathcal N=4$ SYM admit probe-brane and fully backreacted descriptions, but those carriers use the detailed supersymmetry, charges, and compactification of that theory Drukker, Gomis, and Matsuura 2008, §3.1, printed pp. 20–23; §4.1, printed pp. 32–36; Appendix C, printed pp. 57–59, PDF.

The end-of-the-world brane row refers to the AdS/BCFT proposal, where a one-sided boundary is represented by a bulk end surface Takayanagi 2011, §§2–3, printed pp. 1–2, PDF. The Janus row instead describes a two-sided interface across which couplings can change Clark, Freedman, Karch, and Schnabl 2005, Introduction and §2, printed pp. 1–4, PDF. These are distinct constructions. In every row, the support dimension determines a necessary noncompact factor; it does not identify the brane.

A Wilson loop is more than its contour

In Euclidean $\mathcal N=4$ $SU(N)$ SYM, a locally supersymmetric Wilson loop can be normalized as

$$W_R[C,n] =\frac{1}{\dim R}\operatorname{Tr}_R\,\mathcal P \exp\!\left[ \oint_C ds\, \left(iA_\mu\dot x^\mu+|\dot x|\,n^I\Phi_I\right) \right], \qquad n^In_I=1.$$

Here $s$ parametrizes the contour, $\dot x^\mu=dx^\mu/ds$, $R$ is a representation of the gauge group, and $\mathcal P$ orders the noncommuting matrix factors along the path. The convention takes the gauge field and scalars to be Hermitian; moving factors of $i$ between fields and exponent changes the displayed formula but not the required data. The contour $C$ fixes the endpoint in the conformal boundary, while $n^I(s)$ specifies how the string endpoint sits in $S^5$. The representation $R$ and the global form of the gauge theory determine which electric charge is genuine. The factor $1/\dim R$ makes $W_R=1$ when the connection and scalar coupling vanish. Cusps, intersections, and non-BPS scalar couplings can require additional renormalization. These are distinct choices, so “the circle” is not by itself a complete observable Drukker, Gross, and Ooguri 1999, §2.1, eqs. (2.2)–(2.4), PDF.

For the fundamental representation, constant $n^I$, and the standard $AdS_5\times S^5$ duality,

$$\lambda=g_{\mathrm{YM}}^2N, \qquad L^4=\lambda\alpha'^2, \qquad g_s=\frac{\lambda}{4\pi N}, \qquad T_FL^2=\frac{\sqrt\lambda}{2\pi}.$$

Here $L$ is the common $AdS_5$ and $S^5$ curvature radius, $\alpha'=\ell_s^2$ sets the string-length scale, $g_s$ is the string coupling, and $T_F=1/(2\pi\alpha')$ is the fundamental-string tension. The radius and coupling map derives these conventions. A classical fundamental worldsheet requires

$$\frac{\alpha'}{L^2}=\lambda^{-1/2}\ll1, \qquad g_s\ll1,$$

together with induced worldsheet curvature and contour-driven gradients that remain small in string units, and a single string, or only a few strings, whose stress and charge do not appreciably change the background. Large ambient radius alone does not control a contour with string-scale cusps or other rapid structure. The clean logical order is the planar limit $N\to\infty$ at fixed $\lambda$, followed by $\lambda\gg1$. At finite large $N$, merely writing an $O(N^{-2})$ correction is unsafe because its coefficient can grow with $\lambda$; for the half-BPS circle the first nonplanar-to-planar ratio scales as $\lambda^{3/2}/(96N^2)$ Drukker and Gross 2001, §3.1, eq. (3.5), PDF; Drukker and Fiol 2005, §3.4, printed p. 17, eqs. (3.20)–(3.22), PDF.

In this domain the F1 prescription is

$$\log\langle W_{\mathrm{fund}}(C)\rangle =-T_FA_{\mathrm{ren}}(C) +\text{worldsheet fluctuations} +\text{string-topology corrections}.$$

The area term is the classical prescription Maldacena 1998, §§3–4, PDF. Worldsheet determinants are subleading to its $O(\sqrt\lambda)$ exponent, while additional handles carry powers of the string coupling. For the circular string, an explicit one-loop determinant calculation also shows that its overall constant depends on careful zero-mode normalization Kruczenski and Tirziu 2008, Introduction, eqs. (1.1)–(1.3), and §§3–4, PDF. This is a strong-coupling saddle expansion, not an operator identity valid at every $N$ and $\lambda$.

Worked example: the half-BPS circle

Use Euclidean Poincaré coordinates and restrict to the three directions needed by a rotationally symmetric worldsheet,

$$ds^2=\frac{L^2}{z^2} \left(dz^2+dr^2+r^2d\phi^2\right).$$

Let the boundary contour be the circle $r=a$ at $z=0$. The smooth disk filling it is the hemisphere

$$r^2+z^2=a^2.$$

At fixed azimuthal angle $\phi$, the profile in the half-plane $r\geq0$ is the quarter-circle arc from $(r,z)=(a,0)$ to the smooth tip $(0,a)$. Rotating that arc around the axis $r=0$ produces the hemisphere. The cutoff $z=\epsilon$ intersects it at $r_\epsilon=\sqrt{a^2-\epsilon^2}$, and the removed near-boundary strip runs around the full circle.

Writing $r(z)=\sqrt{a^2-z^2}$, the induced area density simplifies to $L^2a/z^2$. Cutting the surface off at $z=\epsilon$ gives

$$\begin{aligned} A(\epsilon) &=\int_0^{2\pi}d\phi \int_\epsilon^a dz\,\frac{L^2a}{z^2} \\ &=2\pi L^2\left(\frac{a}{\epsilon}-1\right). \end{aligned}$$

The first term is the universal near-boundary perimeter divergence. The positive subtraction magnitude is

$$A_{\mathrm{ct}}(\epsilon)=\frac{2\pi L^2a}{\epsilon}.$$

Adding the corresponding boundary term appropriate to the fixed Wilson-loop data—equivalently, performing the standard Legendre subtraction in this example—gives $A_{\mathrm{ren}}=\lim_{\epsilon\to0}[A(\epsilon)-A_{\mathrm{ct}}(\epsilon)]$ and leaves

$$A_{\mathrm{ren}}=-2\pi L^2, \qquad S_{F1,\mathrm{ren}}=T_FA_{\mathrm{ren}}=-\sqrt\lambda.$$

A negative renormalized area is not a negative bare area: it is the finite remainder after subtracting the positive divergent perimeter term. This is the unnormalized geometric area; the equivalent calculation in Drukker, Gross, and Ooguri uses a dimensionless area normalized by $1/(2\pi)$ and reports the finite part as $-1$ Drukker, Gross, and Ooguri 1999, §3.2, printed p. 15, eqs. (3.16)–(3.17), and §3.3, printed pp. 20–22, PDF. The leading prediction is therefore

$$\log\langle W_\circ\rangle_{\mathrm{cl}}=\sqrt\lambda.$$

Supersymmetric localization gives an independent boundary calculation. Let $I_1$ denote the modified Bessel function of the first kind of order one. For the normalized fundamental loop in the planar limit,

$$\begin{aligned} \langle W_\circ\rangle_{N=\infty} &=\frac{2I_1(\sqrt\lambda)}{\sqrt\lambda} \\ &=\sqrt{\frac{2}{\pi}}\, \lambda^{-3/4}e^{\sqrt\lambda} \left[1-\frac{3}{8\sqrt\lambda} +O(\lambda^{-1})\right], \end{aligned}$$

so

$$\log\langle W_\circ\rangle =\sqrt\lambda-\frac34\log\lambda +\frac12\log\!\left(\frac{2}{\pi}\right) -\frac{3}{8\sqrt\lambda} +O(\lambda^{-1}).$$

The leading exponent checks the string tension and subtraction convention; the logarithm shows why an undifferentiated $O(\lambda^0)$ remainder would be wrong. Localization proves the protected matrix-model reduction, while the Bessel expression above is its planar evaluation Erickson, Semenoff, and Zarembo 2000, §1.2.3, eqs. (4)–(6), PDF; Pestun 2012, §1, eqs. (1.1)–(1.5), PDF. This agreement is special evidence for a half-BPS observable, not a theorem about generic Wilson loops.

For comparison, a rectangle of Euclidean time extent $T$ and separation $R$, with $T\gg R$, extracts a heavy-probe potential through $\log\langle W_\square\rangle\simeq-TV(R)$. After subtracting the two straight-string masses,

$$V(R)=-\frac{4\pi^2}{\Gamma(1/4)^4}\frac{\sqrt\lambda}{R} \left[1+O(\lambda^{-1/2})\right]$$

at planar strong coupling, where $\Gamma$ is Euler’s gamma function. The renormalization prescription is not the circle subtraction: it removes the two cutoff-divergent straight-string masses. A finite rectangle also has cusp factors, but those do not grow in proportion to $T$ and hence do not change the coefficient defining $V(R)$ as $T/R\to\infty$. The coefficient displayed here has been translated into this page’s convention $L^4/\alpha'^2=\lambda$ Maldacena 1998, §4, eqs. (4.1)–(4.8), PDF. The general Wilson-loop and static-potential interpretation is developed in Wilson loops, worldlines, monopole plasma, and confinement.

Form charge makes the endpoint physical

The endpoint condition has a charge counterpart. Suppose an $(m+1)$-dimensional worldvolume $\mathcal W_{m+1}$ couples electrically to a bulk $(m+1)$-form potential,

$$S_{\mathrm{charge}} =iq\int_{\mathcal W_{m+1}}C_{m+1}.$$

Under $C_{m+1}\mapsto C_{m+1}+d\Lambda_m$,

$$\delta S_{\mathrm{charge}} =iq\int_{\partial\mathcal W_{m+1}}\Lambda_m.$$

For a closed worldvolume the variation vanishes. For an anchored or open worldvolume, the endpoint operator must transform oppositely, an attached object must absorb the charge, or the proposed configuration is not gauge invariant. This elementary Stokes-theorem calculation is the bulk reason that endpoint and junction rules carry physical information. Worldvolume flux can induce lower-brane charges, so the full conserved charge need not equal the geometric brane label.

After conformal compactification or introduction of a radial cutoff, the endpoint law can also be recorded by relative-cycle data. If $Q$ denotes another permitted brane or end locus, the AdS-reaching factor has a schematic relative class

$$[M_{p+1}] \in H_{p+1}(\overline X,\partial\overline X\cup Q), \qquad \delta[M_{p+1}] =[\Sigma_p]+[\Sigma_Q],$$

where $\delta$ is the boundary map, $\Sigma_Q=\partial M_{p+1}\cap Q$, and relative signs depend on orientation; if $Q$ is absent, $\delta[M_{p+1}]=[\Sigma_p]$. A genuinely closed internal factor can additionally define $[\Gamma_r]\in H_r(K)$, although that class can be zero—for example, $H_4(S^5)=0$ for the $S^4\subset S^5$ in the antisymmetric-loop D5 construction Yamaguchi 2006, §2, printed pp. 3–5, PDF.

Two bulk saddles can have the same endpoint $\Sigma_p$ yet differ by a closed cycle, internal embedding, worldvolume flux, or discrete torsion. Ordinary relative homology is therefore useful geometric bookkeeping, not a universal classification theorem for brane charge. K-theory can distinguish configurations that homology alone identifies Witten 1998, “D-Branes and K-Theory,” §§1–2, PDF, while the Freed–Witten anomaly can obstruct an otherwise plausible D-brane worldvolume in background flux Freed and Witten 1999, §1, especially eq. (1.12), PDF.

The $AdS_5\times S^5$ baryon vertex is the cleanest example. A D5-brane wrapped on $S^5$ couples to the background five-form flux. The induced worldvolume electric charge is canceled by $N$ fundamental strings ending on the D5, producing an $N$-pronged Wilson-line junction for external probes. Omitting the attached strings violates the worldvolume Gauss law Witten 1998, “Baryons and Branes,” §2, printed pp. 3–5, eqs. (2.1)–(2.2), PDF.

Probe string, probe brane, and backreacted geometry

Changing the representation tests which approximation actually carries the charge. The transition occurs in two stages, not one.

A regime ladder for half-BPS Wilson loops in the standard large-N AdS₅ × S⁵ example.

Source scaling	Useful carrier	What must remain small	What has failed relative to the previous row
Fundamental or a fixed number of boxes as N grows	One F1 or a few fundamental strings	λ−1/2 ≪ 1, gs ≪ 1, and sourced stress and charge small compared with the background scales	Nothing: this is the fundamental-string domain
One coherent representation with rank k proportional to N	A fluxed D3 for a large symmetric representation or a fluxed D5 for a large antisymmetric representation	Small curvature and worldvolume-loop corrections, Sprobe/Sbackground ≪ 1, and sourced-flux/background-flux ≪ 1 at the actual k, λ, and N	The independent-F1 description; the probe approximation can still be valid
A source sector with action, stress, or flux proportional to N2	A fully backreacted or bubbling solution	The probe/background action or flux ratio is O(1); no small probe parameter remains, so solve the coupled bulk equations	The fixed-background probe-brane description itself

For the electric D3 description, the useful combination is $\kappa=k\sqrt\lambda/(4N)$, and the brane carries $k$ units of dissolved F1 charge. Its expected semiclassical window includes, parametrically,

$$\kappa\gg\lambda^{-1/4}.$$

The independent backreaction condition can be written as

$$kg_s^2\ll1.$$

Using the definitions above, this is parametrically equivalent, up to numerical factors, to

$$\kappa\ll\frac{1}{g_s\sqrt\lambda}.$$

The lower condition keeps the curved D3 worldvolume large in string units; the upper condition suppresses its backreaction Drukker and Fiol 2005, §3.3, printed p. 16, PDF. A D5 with electric flux describes the large antisymmetric representation Yamaguchi 2006, §§2–3, PDF. More general Young diagrams can be organized by D3 and D5 configurations Gomis and Passerini 2006, §§1–3, PDF.

The important adversarial case is $k\sim N$. It often changes the correct carrier from F1 strings to a D3 or D5. In the ‘t Hooft limit at fixed $\lambda$, one such probe has an action of order $N$, subleading to the order-$N^2$ background. If $\lambda$ itself scales with $N$, that counting need not survive and the action, stress, and flux ratios must be evaluated explicitly. Backreaction becomes order one when the sourced stress, charge, or flux competes with the background—for example, in common half-BPS scalings with order $N$ large rows or columns and order $N^2$ boxes. That box-counting rule is illustrative rather than universal; the invariant test is the sourced bulk response. The probe comparison and the regular fully backreacted half-BPS Wilson-loop families are developed in D’Hoker, Estes, and Gutperle 2007, §2.2, printed pp. 6–7, and §§9–11, printed pp. 38–50, PDF.

Once backreaction is order one, the worldsheet formula has not merely acquired a large numerical error: its assumed background is no longer a solution with the source included. The strongest surviving claim is that the boundary support and charge impose asymptotic and flux boundary data. Which smooth bulk saddle realizes them must be solved anew.

A reusable defect-dictionary check

Before assigning a bulk carrier to an extended observable, record seven items.

1. Observable. Give the support, orientation, representation, scalar or matter couplings, normalization, and renormalization prescription.
2. Global charge. State the faithful gauge-group or symmetry form, charge lattice, genuineness or attached-surface rule, and possible anomaly.
3. Carrier. Name the string, brane, form-field configuration, or geometry, including every additional compact worldvolume factor, genuine internal cycle, and worldvolume flux.
4. Endpoint class. State the asymptotic boundary condition and the relevant relative-cycle or linking data.
5. Action. Include boundary terms, counterterms, ensemble choice for worldvolume gauge fields, and any saddle sum.
6. Controls. Exhibit the parameters suppressing curvature, quantum loops, string topology, and backreaction; do not replace them with the words “large $N$.”
7. Independent check. Compare with a protected calculation, Ward identity, charge quantization law, anomaly, or controlled limit, and state what uncertainty remains.

This checklist distinguishes the dictionary entry from evidence for it. A BPS localization match can test normalization in one protected sector without establishing the same carrier for an unprotected defect or a different compactification.

Common pitfalls

Assuming the same contour defines the same operator. The support is only one part of the definition. Changing the representation, scalar coupling, electric or magnetic charge, global-form data, normalization, or attached-surface rule can change both the boundary observable and its bulk carrier.

Confusing support dimension with brane dimension. A boundary line forces a two-dimensional AdS-reaching factor, not necessarily a two-dimensional full worldvolume. Additional compact factors account for the extra dimensions of D3- and D5-brane carriers, and those factors need not all lie in the internal space.

Treating $k\sim N$ as automatic backreaction. This scaling often marks the F1-to-D-brane transition. A single fluxed probe can remain parametrically smaller than the $O(N^2)$ background.

Calling a negative renormalized area unphysical. The regulated area is positive and divergent. Its finite remainder may be negative after the specified local subtraction; the observable depends on that renormalized action.

Identifying every anchored minimal surface with an entropy surface. A Wilson-loop string worldsheet extremizes the F1 action and carries string charge. A Ryu–Takayanagi surface computes entropy through a gravitational area functional. Their boundary anchors, dimensions, and variational problems are different.

Using homology as the complete charge classifier. Relative homology records geometric filling data. Flux quantization, torsion, worldvolume anomalies, and K-theory can refine or obstruct the physical brane charge.

Exercises

1. Count the dimensions

A line defect has $p=1$. Verify the full worldvolume dimensions of an F1 with $AdS_2$, a D3 with $AdS_2\times S^2$ (the $S^2$ lies in $AdS_5$), and a D5 with $AdS_2\times S^4$ (the $S^4$ lies in $S^5$). Repeat for a surface defect represented by a D3 with $AdS_3\times S^1$.

Solution

For the line, the AdS-reaching factor has dimension $p+1=2$. The F1 has no additional factor, so its worldvolume dimension is $2$. The D3 has an additional two-sphere inside $AdS_5$, giving $2+2=4$. The D5 has an internal four-sphere factor, giving $2+4=6$. For the surface, $p=2$, so the AdS-reaching factor has dimension $3$; adding the circle gives $3+1=4$, the dimension of a D3 worldvolume. Dimension counting alone does not determine whether that additional factor lies in AdS or in the internal space.

2. Reproduce the circular subtraction

For $r(z)=\sqrt{a^2-z^2}$, compute the induced metric determinant in the Poincaré metric and recover $A(\epsilon)=2\pi L^2(a/\epsilon-1)$. Why does the answer not depend on $a$ after renormalization?

Solution

Since $r'=-z/r$,

$$1+r'^2=1+\frac{z^2}{r^2}=\frac{a^2}{r^2}.$$

The induced components are $g_{zz}=L^2a^2/(z^2r^2)$ and $g_{\phi\phi}=L^2r^2/z^2$, so $\sqrt{\det g}=L^2a/z^2$. Integrating over $\phi$ and $z$ gives the stated result. The boundary term removes $2\pi L^2a/\epsilon$, leaving $-2\pi L^2$. A boundary dilation changes $a$ but is a conformal transformation; no finite radius scale can remain for the circular observable in a CFT.

3. Derive the endpoint phase

Starting from $S_{\mathrm{charge}}=iq\int_{\mathcal W}C_{m+1}$, derive its gauge variation when $\partial\mathcal W\ne\varnothing$. What must happen at the endpoint?

Solution

Under $C_{m+1}\mapsto C_{m+1}+d\Lambda_m$,

$$\delta S_{\mathrm{charge}} =iq\int_{\mathcal W}d\Lambda_m =iq\int_{\partial\mathcal W}\Lambda_m$$

by Stokes’s theorem. The path-integral weight therefore acquires a boundary phase. A boundary operator or another attached brane must carry the opposite transformation, or the sum of endpoint charges must vanish. Otherwise the proposed open worldvolume is not gauge invariant.

4. Stress-test the probe approximation

Compare a half-BPS representation with one row of $k=cN$ boxes to a family with $c'N$ such large rows, where $c,c'>0$ are fixed constants of order one. Which transition occurs in each case, and what statement survives after the second transition?

Solution

One row with $k=O(N)$ is naturally reorganized as a fluxed D3-brane rather than $k$ independent F1 strings. Its action is ordinarily $O(N)$ in the planar scaling, so it can remain a probe of an $O(N^2)$ background. With $O(N)$ large rows, the total source can contain $O(N^2)$ boxes and contribute stress or flux at the same order as the background. The probe-brane approximation then fails and a backreacted solution is required. What survives is the asymptotic endpoint, representation, and charge data; the fixed-background worldvolume and its on-shell action do not.

Scope and handoffs

General support, codimension, and operator labels belong to Support, Codimension, and Operator Data; the canonical field-theory treatments of Wilson lines, surface defects, and boundaries and interfaces live there. The preceding global-form page determines which charges are genuine.

The top-down string interfaces develop brane construction, decoupling, compact cycles, and flux quantization. Probe branes, flavor, and mesons develops a different controlled use of probe branes. Entropy surfaces are treated separately in the Ryu–Takayanagi formula. General conformal representations and defect OPE data remain in Conformal Field Theory and Bootstrap; fixed-background timelike-boundary dynamics remains in QFT in Curved Spacetime; theorem-first boundary nets and reconstruction questions remain in Mathematical QFT.

The chapter overview contains the structure diagram and validity and failure diagram. They are embedded there once so that their shared chapter-level context is not repeated on every article.

For the chapter-wide comparison of assumptions, counterevidence, falsifiers, and claim ceilings, see the claim-domain table.

References

• Clark, Andrew B., Daniel Z. Freedman, Andreas Karch, and Martin Schnabl. “The Dual of Janus: An Interface CFT.” Physical Review D 71 (2005): 066003. arXiv. DOI.
• D’Hoker, Eric, John Estes, and Michael Gutperle. “Gravity Duals of Half-BPS Wilson Loops.” Journal of High Energy Physics 2007, 021 (2007). arXiv. DOI.
• DeWolfe, Oliver, Daniel Z. Freedman, and Hirosi Ooguri. “Holography and Defect Conformal Field Theories.” Physical Review D 66 (2002): 025009. arXiv. DOI.
• Drukker, Nadav, and Bartomeu Fiol. “All-Genus Calculation of Wilson Loops Using D-Branes.” Journal of High Energy Physics 2005, 010 (2005). arXiv. DOI.
• Drukker, Nadav, Jaume Gomis, and Shunji Matsuura. “Probing $\mathcal N=4$ SYM with Surface Operators.” Journal of High Energy Physics 2008, 004 (2008). arXiv. DOI.
• Drukker, Nadav, and David J. Gross. “An Exact Prediction of $\mathcal N=4$ SUSYM Theory for String Theory.” Journal of Mathematical Physics 42 (2001): 2896–2914. arXiv. DOI.
• Drukker, Nadav, David J. Gross, and Hirosi Ooguri. “Wilson Loops and Minimal Surfaces.” Physical Review D 60 (1999): 125006. arXiv. DOI.
• Erickson, John K., Gordon W. Semenoff, and Konstantin Zarembo. “Wilson Loops in $\mathcal N=4$ Supersymmetric Yang–Mills Theory.” Nuclear Physics B 582 (2000): 155–175. arXiv. DOI.
• Freed, Daniel S., and Edward Witten. “Anomalies in String Theory with D-Branes.” Asian Journal of Mathematics 3 (1999): 819–851. arXiv. DOI.
• Gomis, Jaume, and Filippo Passerini. “Holographic Wilson Loops.” Journal of High Energy Physics 2006, 074 (2006). arXiv. DOI.
• Kruczenski, Martin, and Arkady Tirziu. “Matching the Circular Wilson Loop with Dual Open String Solution at 1-Loop in Strong Coupling.” Journal of High Energy Physics 2008, 064 (2008). arXiv. DOI.
• Maldacena, Juan M. “Wilson Loops in Large $N$ Field Theories.” Physical Review Letters 80 (1998): 4859–4862. arXiv. DOI.
• Pestun, Vasily. “Localization of Gauge Theory on a Four-Sphere and Supersymmetric Wilson Loops.” Communications in Mathematical Physics 313 (2012): 71–129. arXiv. DOI.
• Takayanagi, Tadashi. “Holographic Dual of BCFT.” Physical Review Letters 107 (2011): 101602. arXiv. DOI.
• Witten, Edward. “Baryons and Branes in Anti-de Sitter Space.” Journal of High Energy Physics 1998, 006 (1998). arXiv. DOI.
• Witten, Edward. “D-Branes and K-Theory.” Journal of High Energy Physics 1998, 019 (1998). arXiv. DOI.
• Yamaguchi, Satoshi. “Wilson Loops of Anti-Symmetric Representation and D5-Branes.” Journal of High Energy Physics 2006, 037 (2006). arXiv. DOI.