D1-D5 CFT and AdS3 Microstate Data

The D1–D5 system turns a difficult black-hole question into a precise comparison between two descriptions of the same quantum theory. Near a weakly coupled symmetric-product point, twisted sectors organize the states into effective long strings. In another region of the conformal manifold, a weakly curved $\mathrm{AdS}_3\times S^3\times X_4$ description becomes useful. Anomalies and carefully chosen supersymmetric traces can connect those regions; a generic spectrum or correlator cannot.

Required background. D-Brane Bound States and the Strominger–Vafa Count fixes the Page-charge conventions and the protected-counting ceiling. BTZ Black Holes and Modular CFT Thermodynamics supplies the general BTZ/Cardy dictionary.

Helpful background. Chiral Blocks, Sewing, and Modular Invariance develops modular constraints. Elliptic Genera, Anomaly Data, and c-Extremization explains why a supersymmetric trace can be deformation invariant even when individual states are not.

This page first chooses the K3 or $T^4$ compactification, constructs its symmetric-product sectors, and uses spectral flow to identify the D1–D5–momentum states. It then makes one protected comparison with extremal BTZ and one deliberately unprotected comparison between the orbifold and supergravity regions. Motion on the CFT conformal manifold is not the radial attractor flow discussed on the next page.

Choose K3 or T⁴ before counting

Let $Q_5$ be the D5 Page charge and $Q_1$ the D1 Page charge in the convention fixed on the preceding page. The compactification must be chosen before writing the copy number, central charge, or trace:

Branch-dependent data at the symmetric-product locus

Datum	K3 branch	T⁴ branch	Consequence
Copy number	NK3 = Q1Q5 + 1	NT⁴ = Q1Q5	The K3 shift is finite and must not be silently discarded.
Orbifold target	SymNK3(K3)	SymNT⁴(T⁴)	The cycle construction is shared, but the seed spectrum is not.
Compact sigma-model central charge	cL = cR = 6(Q1Q5 + 1) before removing the universal sector	cL = cR = 6Q1Q5	The raw K3 model and the center-of-mass-stripped AdS boundary theory are different Hilbert spaces.
Protected trace	Ordinary elliptic genus	Modified, zero-mode-saturating trace	The ordinary T⁴ elliptic genus vanishes.
Coefficient means	Signed BPS index	Signed helicity-weighted BPS index	Neither coefficient is automatically an absolute degeneracy.

The K3 shift follows from induced D1 charge: if $N_1$ is the literal wrapped D1-brane, or instanton, number, then $Q_1=N_1-Q_5$, and the compact instanton sigma model used in the protected count has $N_{K3}=Q_1Q_5+1$. For $T^4$, $Q_1=N_1$ and $N_{T^4}=Q_1Q_5$. Many AdS₃ boundary calculations remove the universal K3 sector and instead quote the interacting copy number $N_{\mathrm{int}}=Q_1Q_5$ and $c_{\mathrm{int}}=6Q_1Q_5$. These are not two formulas for one unchanged Hilbert space. This page retains the compact $+1$ convention for the K3 protected coefficient and specializes the exact BTZ and correlator calculations to $T^4$. The instanton-number convention is explicit in Dijkgraaf 1999, § 5, eq. (5.1), and § 7, eq. (7.7), Open PDF; the Page-charge shift and center-of-mass-stripped comparison are discussed in Seiberg and Witten 1999, introduction footnote 1 and § 2, “Comparison to Symmetric Product”.

For K3, a useful convention for the elliptic genus is

$$Z_{K3}(\tau,z) =\operatorname{Tr}_{RR} \!\left[ (-1)^{F_L+F_R}y^{2J_L^3} q^{L_0-c_L/24}\bar q^{\,\bar L_0-c_R/24} \right],$$

where $q=e^{2\pi i\tau}$ and $y=e^{2\pi iz}$. The $(-1)^{F_R}$ cancellation restricts the surviving contribution to right-moving Ramond ground states, so the answer is holomorphic in $q$ after the trace is evaluated. On $T^4$, unsaturated fermion zero modes make this ordinary trace vanish. One instead inserts enough right-moving charge to soak them; schematically, and up to an overall conventional normalization,

$$E_2^{T^4}(\tau,z) =\operatorname{Tr}_{RR} \!\left[ (-1)^{2J_L^3-2J_R^3}(2J_R^3)^2y^{2J_L^3} q^{L_0-c_L/24}\bar q^{\,\bar L_0-c_R/24} \right].$$

This branch distinction is not cosmetic: the two traces retain different finite data even though their large-charge entropy exponents agree. Index invariance assumes a compact, discrete spectrum and a nonsingular deformation path; a continuum or singular splitting locus can invalidate the simple cancellation argument. The zero-mode analysis and toroidal trace are given explicitly in Maldacena, Moore, and Strominger 1999, §§ 1 and 3, especially eqs. (3.7)–(3.8), Open PDF.

For the concrete correlator later on, we choose the $T^4$ branch and set $N=Q_1Q_5$. The K3 formulas remain visible where the protected entropy comparison differs at finite charge.

Symmetric products, twist sectors, and long strings

At the orbifold point, the target is

$$\operatorname{Sym}^{N_X}(X_4) =\frac{X_4^{N_X}}{S_{N_X}},$$

so physical states are invariant under permutations of the $N_X$ copies. Twisted boundary conditions are classified by conjugacy classes of $S_{N_X}$, equivalently by nonnegative integers $m_k$ obeying

$$\sum_{k\geq1}k\,m_k=N_X.$$

Here $m_k$ counts cycles of length $k$. A $k$-cycle joins $k$ seed copies into one component string whose spatial length is $k$ times the seed circle. A mode that was integer-moded on the covering string therefore appears in the base-space CFT with spacing $1/k$. This fractionation is the long-string mechanism: a fixed total energy can be distributed among much finer excitations than on a single copy.

For example, at $N_X=6$, the cycle type $6$ describes one component string with fractional spacing $1/6$. The type $3+2+1$ describes three components, with spacings $1/3$, $1/2$, and $1$. Both sectors belong to the same $c=36$ theory. The cycle partition reorganizes its states; it does not change the total central charge. The NS vacuum of a $k$-cycle has the twist contribution

$$h_k=\bar h_k =\frac{c_{\mathrm{seed}}}{24} \left(k-\frac1k\right), \qquad c_{\mathrm{seed}}=6,$$

which is another way to see that the cover and base descriptions must be kept distinct.

For a seed elliptic genus

$$Z_X(\tau,z) =\sum_{\substack{m\geq0\\ r\in\mathbb Z}} c_X(m,r)q^m y^r,$$

the symmetric-product indices are packaged by the DMVV product

$$\sum_{N\geq0}p^N Z_{\operatorname{Sym}^N(X)}(\tau,z) =\prod_{k\geq1} \prod_{\substack{m\geq0\\ r\in\mathbb Z}} \left(1-p^kq^m y^r\right)^{-c_X(km,r)}.$$

The factor with cycle length $k$ is the algebraic trace of the same long-string structure. For $X=K3$ this directly generates the nonzero elliptic genera. For $X=T^4$, the ordinary seed genus vanishes, so the modified trace must be treated with its zero-mode insertions rather than inferred by substituting zero into this formula. The cycle derivation and product are given in Dijkgraaf, Moore, Verlinde, and Verlinde 1997, §§ 1–2, especially eqs. (1.3) and (2.22); the D1–D5 target-space interpretation is reviewed in David, Mandal, and Wadia 2002, §§ 4.5–5.8.

Spectral flow identifies the Ramond sector

The small $(4,4)$ superconformal algebra has Neveu–Schwarz (NS) and Ramond (R) presentations. Spectral flow is an automorphism relating them. With $j$ the eigenvalue of $J_L^3$ and with the convention that $\alpha=1$ maps NS to R,

$$h_\alpha=h-\alpha j+\frac{\alpha^2c_L}{24}, \qquad j_\alpha=j-\frac{\alpha c_L}{12}.$$

An NS chiral primary satisfies $h=j$. It therefore flows to a Ramond ground state with $h_R=c_L/24$. The right-moving relation is analogous. In the bulk dictionary, the NS vacuum is associated with global AdS₃, while the spectral-flowed RR vacuum is associated with the massless-BTZ threshold. This explains why cohomological data of the symmetric-product target organize Ramond ground states, but it does not imply that every named state remains fixed in an arbitrary basis as moduli change.

Momentum along the common D1–D5 circle is carried by excitations above a Ramond ground state. Define shifted weights

$$n_L=L_0-\frac{c_L}{24}, \qquad n_R=\bar L_0-\frac{c_R}{24}.$$

For the BPS D1–D5–momentum sector used below,

$$(n_L,n_R)=(n,0),$$

so the right movers remain in their Ramond ground state while the left movers carry integer momentum $n$. We also keep the left spin label $\ell_{\mathrm{spin}}=2J_L^3$ fixed; the entropy comparison below uses $\ell_{\mathrm{spin}}=0$. Nonzero spin changes the available excitation level and must be included before applying a Cardy estimate.

The D1–D5 CFT/BTZ charge dictionary

Let $L_{\mathrm{AdS}}$ be the AdS₃ radius, $M_{\mathrm{BTZ}}$ the BTZ mass, and $J_{\mathrm{BTZ}}$ its angular momentum. In the conventions inherited from the BTZ prerequisite,

$$n_L=\frac{M_{\mathrm{BTZ}}L_{\mathrm{AdS}}+J_{\mathrm{BTZ}}}{2}, \qquad n_R=\frac{M_{\mathrm{BTZ}}L_{\mathrm{AdS}}-J_{\mathrm{BTZ}}}{2}.$$

Thus $(n_L,n_R)=(n,0)$ gives

$$M_{\mathrm{BTZ}}L_{\mathrm{AdS}}=J_{\mathrm{BTZ}}=n,$$

the extremal BTZ relation in this orientation. The shifts by $c/24$ matter: a Ramond ground state has $n_L=n_R=0$ and maps to the massless-BTZ threshold, not to global AdS₃.

On a boundary cylinder whose spatial circle has radius $R_y$, the same shifted weights give

$$E_{\mathrm{CFT}}-E_{RR}=\frac{n_L+n_R}{R_y}, \qquad P_y=\frac{n_L-n_R}{R_y}.$$

Thus the selected BPS chirality carries $P_y=n/R_y$. In terms of the outer and inner BTZ radii,

$$n_L=\frac{(r_++r_-)^2}{16G_3L_{\mathrm{AdS}}}, \qquad n_R=\frac{(r_+-r_-)^2}{16G_3L_{\mathrm{AdS}}},$$

so $n_R=0$ implies $r_+=r_-$ and the Bekenstein–Hawking entropy becomes the one-sided Cardy expression.

D1–D5 CFT data and their bulk interpretation

CFT datum	Bulk datum	Control or caveat
cL, cR	Brown–Henneaux asymptotic-symmetry central charges	Match anomaly conventions and decoupled sectors.
nL, nR	(MBTZLAdS ± JBTZ)/2	These are vacuum-shifted weights.
(n, 0)	Extremal BTZ with MBTZLAdS = JBTZ	Use a one-sided protected argument, not an unsupported two-sided saddle.
Protected coefficient ΩX(n, ℓspin)	Indexed microstate sector at fixed charges	An index is not automatically an absolute degeneracy.
Twist-cycle basis	Stringy organization of states	A basis vector need not have a smooth classical geometry.

At the two-derivative classical level, dimensional reduction gives the center-of-mass-stripped Brown–Henneaux result

$$c_{\mathrm{BH}}=\frac{3L_{\mathrm{AdS}}}{2G_3}=6Q_1Q_5.$$

This equals the toroidal CFT central charge. In the compact K3 counting convention the raw sigma-model value is $6(Q_1Q_5+1)$, whereas the interacting boundary convention removes the universal sector and gives $6Q_1Q_5$. One must choose the same sector on both sides before discussing the remaining order-one higher-derivative or quantum terms. Brown and Henneaux derive the asymptotic-symmetry central charge in Brown and Henneaux 1986, pp. 207–226.

First application: a protected coefficient and extremal BTZ entropy

For context, a nonextremal modular saddle with both shifted weights large has

$$S_{\mathrm{Cardy}} =2\pi\sqrt{\frac{c_Ln_L}{6}} +2\pi\sqrt{\frac{c_Rn_R}{6}}.$$

The BPS sector has $n_R=0$, so that two-sided high-temperature derivation cannot simply be evaluated at its boundary. The needed one-sided statement instead comes from the protected chiral generating function. For the K3 genus and the selected $T^4$ modified trace, the relevant polar data give the effective central charge $c_{\mathrm{eff}}=c_L$. If $\Omega_X(n,0)$ denotes the corresponding protected coefficient, then the modular asymptotic in the joint sequence

$$Q_1Q_5\longrightarrow\infty, \qquad \frac{n}{Q_1Q_5}\longrightarrow\infty$$

is

$$\log\left|\Omega_X(n,0)\right| =2\pi\sqrt{\frac{c_Ln}{6}} +o\!\left(\sqrt{c_Ln}\right),$$

provided the selected coefficient is nonzero. A concrete sufficient scaling is $Q_1\sim Q_5\sim\Lambda$ and $n\sim\Lambda^{2+\epsilon}$ with $\epsilon>0$. The modular result is an index statement. Only its further interpretation as the logarithm of an absolute microstate count requires the separate premise that boson–fermion cancellations are not exponentially large. Consequently,

$$\begin{aligned} T^4:\quad \log\left|\Omega_{T^4}(n,0)\right| &=2\pi\sqrt{Q_1Q_5n}+o(\sqrt{Q_1Q_5n}),\\ K3:\quad \log\left|\Omega_{K3}(n,0)\right| &=2\pi\sqrt{(Q_1Q_5+1)n}+o(\sqrt{Q_1Q_5n}). \end{aligned}$$

The extremal BTZ horizon gives the same leading two-derivative exponent,

$$S_{\mathrm{BTZ}}^{(0)}=2\pi\sqrt{Q_1Q_5n}.$$

For $T^4$ the displayed central charges coincide directly. For K3, the declared joint limit makes the $+1$ subleading relative to $Q_1Q_5$, precisely where the classical area formula is no longer the complete bulk calculation. The bulk comparison simultaneously requires $L_{\mathrm{AdS}}^2/\alpha'\gg1$, $G_3/L_{\mathrm{AdS}}\ll1$, suppressed string loops, and horizon curvature below the string scale.

This comparison is reproducible only when all five entries in its contract are kept visible:

1. Observable and ensemble: a fixed-$(Q_1,Q_5,n,\ell_{\mathrm{spin}}=0)$ protected coefficient, not the unrestricted Hilbert-space dimension.
2. Transport: an elliptic genus or modified trace that is invariant along the allowed conformal-manifold path.
3. Asymptotics: the joint large-charge regime $Q_1Q_5\to\infty$ and $n/(Q_1Q_5)\to\infty$, with $c_{\mathrm{eff}}=c_L$ fixed by the relevant polar data.
4. Bulk approximation: a same-charge extremal BTZ saddle with weak curvature and suppressed derivative and loop corrections.
5. Status and uncertainty: equality of the leading exponent; finite-charge, higher-derivative, and cancellation effects are not included.

The preceding page supplies the index-to-degeneracy caveat, while BPS Indices, Absolute Degeneracies, and Wall Crossing develops it systematically.

Which data survive motion in moduli space

The symmetric-product point and a weakly curved supergravity point are different regions of one D1–D5 SCFT conformal manifold. Exactly marginal deformations change the local presentation of the theory without changing $c_L$ or $c_R$. Protection, however, is selective:

Transport status of commonly used D1–D5 data

Datum	What protects or controls it	What can still change
cL, cR, and anomalies	The (4,4) anomaly data are constant along the conformal manifold.	A comparison can fail if decoupled sectors or charge conventions differ.
K3 elliptic genus or T⁴ modified trace	Supersymmetric cancellations make the chosen index deformation invariant.	Individual short multiplets may pair, lift, or mix while the signed trace stays fixed.
Chiral ring and selected BPS three-point data	They form a bundle with a covariantly constant product under the applicable nonrenormalization result.	Components rotate with the connection; a fixed orbifold basis need not remain fixed.
Individual quarter-BPS multiplicities	Only special combinations are fixed by an index.	Short multiplets can recombine into long multiplets.
Long-multiplet dimensions	No general protection theorem applies.	Anomalous dimensions and operator mixing vary with the coupling.
Generic four-point function	Crossing and symmetry constrain it but do not fix its moduli dependence.	OPE coefficients, exchanged dimensions, and contact contributions can all vary.

The bundle language matters. Saying that the chiral ring is covariantly constant is stronger and more accurate than saying that every coefficient in an arbitrarily chosen basis is numerically constant. Operator mixing supplies the parallel transport. This geometry is developed in de Boer, Manschot, Papadodimas, and Verlinde 2009, §§ 2–4.

An endpoint correlator that really changes

A claim that “unprotected correlators vary” becomes useful only after the state, operators, normalizations, regimes, and endpoint answers are specified. Galliani, Giusto, and Russo provide such a comparison in the $T^4$ D1–D5 theory. Their coherent heavy Ramond state is

$$|s_B\rangle =N^{-N/2}\sum_{p=0}^{N} A^{N-p}B^p\,|{++}\rangle^{N-p}|{00}\rangle^p, \qquad |A|^2+|B|^2=N,$$

with full symmetrization over the copies and

$$\left\|\,|{++}\rangle^{N-p}|{00}\rangle^p\right\|^2 =\binom Np.$$

These conventions give $\langle s_B|s_B\rangle=1$. The light chiral primary $O_L=O^{++}$ has $(h,\bar h)=(1/2,1/2)$ and unit two-point function $\langle O_L(z,\bar z)\bar O_L(0)\rangle=|z|^{-2}$; the heavy operator also has unit two-point normalization. Define the full and reduced heavy–heavy–light–light correlators by

$$\mathcal F(z,\bar z) \equiv\langle \bar O_H(\infty)O_H(0)O_L(1)\bar O_L(z,\bar z)\rangle =\frac{\mathcal G(z,\bar z)}{|1-z|^2}.$$

At the free orbifold point, direct copy combinatorics gives

$$\begin{aligned} \mathcal G_{\mathrm{orb}} ={}&\frac1{|z|} +\frac{|B|^2}{2N} \frac{|z|^2+|1-z|^2-1}{|z|}\\ &+\frac{|A|^2|B|^2}{N} \left(1-\frac1N\right) \frac{|1-z|^2}{|z|}. \end{aligned}$$

At the supergravity point, the same heavy and light quantum numbers are represented by a smooth D1–D5 geometry and a linearized bulk field. Let $\mathcal Q_1$ and $\mathcal Q_5$ denote the dimensionful supergravity charges and $R_y$ the boundary-circle radius. On the real slice used for the displayed geometry, phases are fixed so that $a,b\geq0$ and

$$\begin{aligned} |A|&=R_y\sqrt{\frac{N}{\mathcal Q_1\mathcal Q_5}}\,a, & |B|&=R_y\sqrt{\frac{N}{2\mathcal Q_1\mathcal Q_5}}\,b,\\ a_0^2&\equiv\frac{\mathcal Q_1\mathcal Q_5}{R_y^2} =a^2+\frac{b^2}{2}. \end{aligned}$$

Define the common perturbation parameter

$$\varepsilon\equiv\frac{b^2}{a_0^2} =\frac{2|B|^2}{N}.$$

After taking $N$ large and then expanding at $b^2\ll a^2$, the exact orbifold result and the tree-level bulk result at the same order are

$$\begin{aligned} \mathcal G_{\mathrm{orb}} &=\frac1{|z|} \left[ 1+\varepsilon \left( \frac N2|1-z|^2 +\frac{|z|^2-|1-z|^2-1}{4} \right) \right]+O(\varepsilon^2),\\ \mathcal G_{\mathrm{sugra}} &=\frac1{|z|} \left[ 1+\varepsilon \left( \frac N2|1-z|^2 +\frac{|z|^2}{\pi}\widehat D_{2211}(z,\bar z) -\frac12 \right) \right]+O(\varepsilon^2). \end{aligned}$$

Here $\widehat D_{2211}$ uses exactly the unnormalized Euclidean AdS₃ bulk-to-boundary-kernel convention of Galliani, Giusto, and Russo, eq. (3.56); that convention fixes the displayed factor $1/\pi$. The constraint $a_0^2=a^2+b^2/2$ shows that $a$ varies as $b$ is expanded at fixed charges.

The control $B=b=0$ is immediate: both answers reduce to $1/|z|$. At nonzero $\varepsilon$, the leading order-$N$ off-diagonal term agrees, while the order-$N^0$ dynamical term differs. In particular, the supergravity contact function produces a $|1-z|^2\log|1-z|^2$ term near $z=1$. In the crossed-channel OPE, that logarithm records a positive order-$1/N$ anomalous dimension for an unprotected double-trace operator; the free-orbifold answer has no corresponding logarithm. Equations, normalizations, and the parameter map are given in Galliani, Giusto, and Russo 2017, §§ 2–4, especially eqs. (2.2)–(2.3), (2.7), (2.9), (3.21)–(3.23), and (3.56)–(3.58), with the logarithm in Appendix D, eq. (D.11).

This is a genuine endpoint comparison, but its domain is narrow: a coherent RR heavy state, selected light BPS operators, large $N$, small $b^2/a_0^2$, and leading classical supergravity. The double-scaling limit keeps $|B|^2=O(N)$ even though $b^2/a_0^2$ is small. It is not a correlator of a typical BTZ microstate. The strongest surviving conclusion is therefore:

Protected external dimensions and selected protected OPE data can match while the full four-point function changes. No index or anomaly argument licenses transporting this unprotected correlator from the orbifold point to the supergravity point.

The twist-two deformation locates the failure

Near the orbifold point, the motion toward an interacting region begins with an exactly marginal twist-two descendant $\mathcal D_{(2)}$. In Euclidean conventions with $e^{-S}$,

$$S_\lambda=S_{\mathrm{orb}} +\lambda\int d^2w\,\mathcal D_{(2)}(w,\bar w).$$

Because the external RR states and chiral primaries themselves form bundles over the conformal manifold, use a covariant derivative. Equivalently, choose a parallel-transport frame along the path. For the full correlator $\mathcal F_\lambda=\mathcal G_\lambda/|1-z|^2$,

$$\left.\nabla_\lambda\mathcal F_\lambda\right|_{\lambda=0} =-\int d^2w\, \langle\mathcal D_{(2)}(w,\bar w) O_1O_2O_3O_4\rangle_{\mathrm{conn,ren}}.$$

“Renormalized” is essential. At first order, small disks regulate deformation–external collisions. At second order, deformation–deformation collisions and the partition-function subtraction also enter. Local counterterms preserve exact marginality, and all operators with the same charges and unperturbed dimension must be mixed and diagonalized. If a selection rule makes the first-order term vanish, the first nonzero change may occur at second or higher order,

$$\Delta_a(\lambda) =\Delta_a(0)+\lambda^2\gamma_a^{(2)}+O(\lambda^3),$$

where the eigenvalues of the logarithmic mixing matrix, rather than its basis-dependent entries, are the anomalous dimensions. The construction of the singlet twist deformation is given in Avery, Chowdhury, and Mathur 2010, §§ 2–4; explicit operator mixing near the orbifold is analyzed in Burrington, Peet, and Zadeh 2013, §§ 3–5.

A nonzero covariant derivative, physical logarithmic coefficient, or anomalous-dimension eigenvalue proves that the observable is not constant on moduli space. An ordinary derivative of components in a rotating basis does not. None of these local data computes the value at a distant supergravity point, proves monotonicity, or rules out an accidental equality there. That is why the explicit endpoint calculation above carries information that local conformal perturbation theory alone does not.

Common pitfalls

Using a generic $X_4$ before choosing the branch. K3 and $T^4$ share the long-string mechanism but not the same copy number, finite central charge, seed spectrum, or protected trace. Choose the compactification first and retain that choice throughout the calculation.

Calling the ordinary $T^4$ elliptic genus the black-hole count. It vanishes because of fermion zero modes. A modified zero-mode-saturating trace is the relevant protected observable.

Treating an index as a list of individually fixed states. Short multiplets may pair or mix while a signed coefficient remains invariant. Equality of an index and an absolute degeneracy requires a separate cancellation argument.

Equating the orbifold point with the supergravity point. They are different regions of the same conformal manifold. Protection transports only the observables covered by an anomaly or nonrenormalization theorem.

Using two-sided Cardy at an extremal boundary. The generic nonextremal derivation assumes both shifted weights are in their asymptotic regimes. The BPS result instead uses the protected one-sided generating function.

Assigning a smooth geometry to every twist-sector vector. Cycle data are a useful weak-coupling basis. Most basis states need not survive as individually recognizable semiclassical geometries.

Confusing conformal-manifold motion with attractor flow. The first changes exactly marginal CFT couplings. The second is radial evolution of bulk scalar fields in a fixed charged solution.

Exercises

1. Choose the compactification before the trace

A K3 configuration has literal wrapped D1-brane number $N_1=17$ and $Q_5=3$. Compute $Q_1$, $N_{K3}$, and the compact sigma-model value of $c_L=c_R$. Compare it with the $T^4$ theory at Page charges $Q_1=14$, $Q_5=3$, and name the appropriate protected trace in each branch.

Solution

For K3,

$$Q_1=N_1-Q_5=14, \qquad N_{K3}=Q_1Q_5+1=43, \qquad c_L=c_R=6N_{K3}=258.$$

For $T^4$, $N_{T^4}=Q_1Q_5=42$ and $c_L=c_R=252$. Use the ordinary elliptic genus for K3 and a modified zero-mode-saturating trace for $T^4$.

2. Read a twist-cycle state

At $N_X=6$, compare the sectors with cycle types $6$ and $3+2+1$. What are the component-string lengths and the smallest fractional mode spacing in each sector?

Solution

The $6$-cycle is one string six times the seed length, so its spacing is $1/6$. The $3+2+1$ sector contains strings of lengths three, two, and one, with spacings $1/3$, $1/2$, and $1$; its smallest spacing is $1/3$. Both sectors are parts of the same $c=6N_X=36$ theory.

3. Map a BPS CFT state to extremal BTZ

Work in the $T^4$ branch with $N=Q_1Q_5$. Set $(n_L,n_R)=(n,0)$ in the BTZ dictionary. Solve for $M_{\mathrm{BTZ}}$ and $J_{\mathrm{BTZ}}$, then give the leading protected entropy for $c_L=6N$.

Solution

Adding and subtracting the dictionary equations gives

$$M_{\mathrm{BTZ}}L_{\mathrm{AdS}}=n, \qquad J_{\mathrm{BTZ}}=n.$$

The state is extremal in these conventions, and the one-sided protected asymptotic is

$$\log|\Omega_{T^4}(n,0)| =2\pi\sqrt{Nn}+o(\sqrt{Nn}).$$

This is an indexed fixed-charge coefficient, not a state-to-geometry map.

4. Diagonalize an unprotected mixing problem

Two orbifold operators with the same global charges acquire the leading mixing matrix

$$\gamma(\lambda)=\lambda^2 \begin{pmatrix} 2&1\\ 1&2 \end{pmatrix}.$$

Find its anomalous dimensions and explain why a constant supersymmetric index does not forbid them.

Solution

The normalized eigenvectors are $(1,-1)/\sqrt2$ and $(1,1)/\sqrt2$, with eigenvalues $\lambda^2$ and $3\lambda^2$. An index constrains a signed trace over protected multiplets. It does not fix long-multiplet dimensions or prevent operators with identical quantum numbers from mixing.

5. Audit an endpoint correlator

For the coherent heavy state above, suppose $B\neq0$. Which feature of $\mathcal G_{\mathrm{sugra}}$ proves that the orbifold four-point function was not transported unchanged? What is the strongest conclusion one may draw?

Solution

The supergravity contact integral generates a $|1-z|^2\log|1-z|^2$ term near $z=1$, whereas the orbifold expression is algebraic in $|z|$ and $|1-z|$. The logarithm encodes an anomalous dimension for an unprotected exchanged operator. Therefore this correlator is moduli dependent and needs independent strong-coupling data. The calculation does not establish the behavior of a typical BTZ microstate or every D1–D5 correlator.

What the D1–D5 dictionary establishes

The D1–D5 CFT gives a microscopic Hilbert-space organization, a precise charge dictionary, and protected coefficients whose leading growth matches an extremal black-hole horizon. It also provides its own stress test: an explicitly normalized four-point function can agree in a protected limit yet differ once unprotected exchanges contribute. The success is therefore not “everything at the orbifold point equals supergravity.” It is the sharper statement that specified protected data survive a controlled change of description while dynamical data expose where that transport ends.

Continue to Attractor Mechanism and Charge-Only Entropy for radial bulk scalar flow, Higher-Derivative and Quantum Entropy Corrections for finite-charge bulk terms, Microstate Geometries and Fuzzball Proposals for the state-to-geometry question, and Absorption, Emission, and Dynamical Tests for observables beyond protected counting.

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

• Avery, Steven G., Borun D. Chowdhury, and Samir D. Mathur. “Deforming the D1D5 CFT Away from the Orbifold Point.” Journal of High Energy Physics 2010, no. 06, 031 (2010). DOI. Open PDF.
• Brown, J. David, and Marc Henneaux. “Central Charges in the Canonical Realization of Asymptotic Symmetries: An Example from Three-Dimensional Gravity.” Communications in Mathematical Physics 104, 207–226 (1986). DOI.
• Burrington, Benjamin A., Amanda W. Peet, and Ida G. Zadeh. “Operator Mixing for String States in the D1-D5 CFT Near the Orbifold Point.” Physical Review D 87, 106001 (2013). DOI. Open PDF.
• David, Justin R., Gautam Mandal, and Spenta R. Wadia. “Microscopic Formulation of Black Holes in String Theory.” Physics Reports 369, 549–686 (2002). DOI. Open PDF.
• de Boer, Jan, Jan Manschot, Kyriakos Papadodimas, and Erik Verlinde. “The Chiral Ring of AdS₃/CFT₂ and the Attractor Mechanism.” Journal of High Energy Physics 2009, no. 03, 030 (2009). DOI. Open PDF.
• Dijkgraaf, Robbert. “Instanton Strings and HyperKähler Geometry.” Nuclear Physics B 543, 545–571 (1999). DOI. Open PDF.
• Dijkgraaf, Robbert, Gregory Moore, Erik Verlinde, and Herman Verlinde. “Elliptic Genera of Symmetric Products and Second Quantized Strings.” Communications in Mathematical Physics 185, 197–209 (1997). DOI. Open PDF.
• Galliani, Andrea, Stefano Giusto, and Rodolfo Russo. “Holographic 4-Point Correlators with Heavy States.” Journal of High Energy Physics 2017, no. 10, 040 (2017). DOI. Open PDF.
• Maldacena, Juan, Gregory Moore, and Andrew Strominger. “Counting BPS Blackholes in Toroidal Type II String Theory.” arXiv:hep-th/9903163 (1999). Open PDF.
• Seiberg, Nathan, and Edward Witten. “The D1/D5 System and Singular CFT.” Journal of High Energy Physics 1999, no. 04, 017 (1999). DOI. Open PDF.