A four-dimensional nonsupersymmetric Type II Scherk–Schwarz orbifold

The earlier A-page item def-cmp-geometric-orbifold-compactification-data defines the relevant orbifold data as a complete finite-group action on full worldsheet CFT data, with a twisted sector and centralizer action for every group element. Its genus-zero local and genus-one trace conditions require separate level-matching, phase-lift, twisted-OPE and modular-covariance checks; higher-genus sewing is additional data for a complete string background. It does not assert that every finite-group action has such a completion.

Antoniadis, Diaz Avalos, and Faraggi report a four-dimensional Type II closed-string construction on a symmetric ${\mathbb{Z}}_2\times {\mathbb{Z}}_2$ orbifold of $T^6=(T^2{)}^3$, with complex moduli $T^{(i)},U^{(i)}$ for $i=1,2,3$, a source-reported choice of sector-projection phases, and an additional ${\mathbb{Z}}_2$ modification of the genus-one phase sum specified below. The following basis, projection matrix, shifted lattices, and phase are the source data. All binary labels and additions below are in ${\mathbb{Z}}_2$.

Here a free-fermionic basis is an ordered list of boundary-condition vectors for the worldsheet fermions. In each vector, the displayed fields are periodic and every other fermion in the basis system is antiperiodic; sums of basis vectors are componentwise sums of these boundary-condition labels in ${\mathbb{Z}}_2$. The vertical bar separates left-moving fields from right-moving fields, and a bar on a field symbol marks a right mover. The index $\mu =0,1,2,3$ labels four-dimensional fermions, and $IJ$ denotes an internal index pair. As in the source, ${\chi }^{IJ}={\chi }^I+i{\chi }^J$; $y^{IJ}$ and $w^{IJ}$ denote the complexified fermions for internal bosons $X^{IJ}=X^I+i{X}^J$, with $\mathrm{\partial }{X}^I\sim y^I{\omega }^I$. The displayed $w^I$ notation is the source's ${\omega }^I$ notation.

The ordered free-fermionic basis is $(1,S,\bar{S},b_1,b_2,T_1,T_2)$, where $\mu =0,1,2,3$ labels four-dimensional spacetime fermions and $IJ$ denotes the indicated internal pair: $\begin{array}{rl}{\displaystyle 1}& {\displaystyle =\{{\psi }^{\mu },{\chi }^{1,\dots ,6},y^{1,\dots ,6},w^{1,\dots ,6}\mid {\bar{\psi }}^{\mu },{\bar{\chi }}^{1,\dots ,6},{\bar{y}}^{1,\dots ,6},{\bar{w}}^{1,\dots ,6}\},}\\ {\displaystyle S}& {\displaystyle =\{{\psi }^{\mu },{\chi }^{1,\dots ,6}\},\bar{S}=\{\mid {\bar{\psi }}^{\mu },{\bar{\chi }}^{1,\dots ,6}\},}\\ {\displaystyle b_1}& {\displaystyle =\{{\psi }^{\mu },{\chi }^{12},y^{34},y^{56}\mid {\bar{\psi }}^{\mu },{\bar{\chi }}^{12},{\bar{y}}^{34},{\bar{y}}^{56}\},}\\ {\displaystyle b_2}& {\displaystyle =\{{\psi }^{\mu },{\chi }^{34},y^{12},y^{56}\mid {\bar{\psi }}^{\mu },{\bar{\chi }}^{34},{\bar{y}}^{12},{\bar{y}}^{56}\},}\\ {\displaystyle T_1}& {\displaystyle =\{y^{12},w^{12}\mid {\bar{y}}^{12},{\bar{w}}^{12}\},T_2=\{y^{34},w^{34}\mid {\bar{y}}^{34},{\bar{w}}^{34}\}\mathrm{.}}\end{array}$ The source also denotes the derived sector $T_3=1+S+\bar{S}+T_1+T_2$; it is not an eighth independent basis vector.

A Gliozzi–Scherk–Olive (GSO) phase matrix is, here, the ordered array of signs used for the source's projections on free-fermion sectors: $C[v_i,v_j]\in \{+1,-1\}$ is the phase in row $v_i$ and column $v_j$. In the row/column order $(1,S,\bar{S},b_1,b_2,T_1,T_2)$, the complete source matrix is $C[v_i,v_j]=\left(\begin{array}{ccccccc}1& 1& 1& 1& 1& 1& 1\\ 1& -1& 1& -1& -1& -1& -1\\ 1& 1& -1& -1& -1& -1& -1\\ 1& 1& 1& 1& 1& 1& 1\\ 1& 1& 1& 1& 1& 1& 1\\ 1& -1& -1& 1& 1& 1& 1\\ 1& -1& -1& 1& 1& 1& 1\end{array}\right)\mathrm{.}$ The item records this matrix as source data; it does not independently prove that the full model's partition function is modular invariant or that its twisted OPE data sew.

For each torus define $T^{(i)}=T_1^{(i)}+i{T}_2^{(i)}$ and $U^{(i)}=U_1^{(i)}+i{U}_2^{(i)}$, with $T_2^{(i)},U_2^{(i)}>0$. The labels $h_i,g_i\in {\mathbb{Z}}_2$ are the orbifold twists and $H_i,G_i\in {\mathbb{Z}}_2$ are the shifted-lattice labels. Let $\tau \in \mathbb{H}:=\{z\in \mathbb{C}:\mathrm{Im}z>0\}$ be the worldsheet-torus modulus, and put $q=e^{2\pi i\tau }$ and $\bar{q}=e^{-2\pi i\bar{\tau }}$. The left-moving spin-structure labels are $a,b\in {\mathbb{Z}}_2$ and the right-moving labels are $k,l\in {\mathbb{Z}}_2$, each pair specifying boundary conditions around the two non-contractible torus cycles. For $\alpha ,\beta \in {\mathbb{Z}}_2$, represented by $0$ or $1$, write $\vartheta \left[\begin{array}{c}\alpha \\ \beta \end{array}\right]:=\vartheta \left[\begin{array}{c}\alpha \\ \beta \end{array}\right](0\mathrm{\mid }\tau ),\vartheta \left[\begin{array}{c}\alpha \\ \beta \end{array}\right](z\mathrm{\mid }\tau )={\sum }_{r\in \mathbb{Z}}{q}^{\frac{1}{2}(r-\alpha \mathrm{/}2{)}^2}\exp \text{\hspace{0.17em}⁣}(2\pi i(z-\beta \mathrm{/}2)(r-\alpha \mathrm{/}2))\mathrm{.}$ Thus every theta symbol below with omitted arguments means the theta constant at $z=0$ and this same $\tau$; all additions in its characteristics are reduced in ${\mathbb{Z}}_2$. For lattice summation indices $m_1,m_2,n_1,n_2\in \mathbb{Z}$, set $\mathbf{m}=\left(\begin{array}{c}{m}_1\\ m_2\end{array}\right),\mathbf{n}=\left(\begin{array}{c}{n}_1\\ n_2+H_i\mathrm{/}2\end{array}\right),{\mathsf{G}}^{(i)}=\frac{T_2^{(i)}}{U_2^{(i)}}\left(\begin{array}{cc}1& U_1^{(i)}\\ U_1^{(i)}& \mathrm{\mid }{U}^{(i)}{\mathrm{\mid }}^2\end{array}\right),{\mathsf{B}}^{(i)}=T_1^{(i)}\left(\begin{array}{cc}0& -1\\ 1& 0\end{array}\right)\mathrm{.}$ The source-normalized left/right lattice momenta and their norms are $p_L=\frac{({\mathsf{G}}^{(i)}{)}^{-1}}{\sqrt{2}}(-\mathbf{m}+{\mathsf{B}}^{(i)}\mathbf{n}+{\mathsf{G}}^{(i)}\mathbf{n}),p_R=\frac{({\mathsf{G}}^{(i)}{)}^{-1}}{\sqrt{2}}(-\mathbf{m}+{\mathsf{B}}^{(i)}\mathbf{n}-{\mathsf{G}}^{(i)}\mathbf{n}),p_{L,R}^2=p_{L,R}^I{\mathsf{G}}_{IJ}^{(i)}{p}_{L,R}^J\mathrm{.}$ They are in the dimensionless normalization used in the source, so their squared values enter the exponents of $q$ and $\bar{q}$ below; no conversion to dimensionful target-space momenta is made here. The shifted untwisted lattice factor is ${\mathrm{\Gamma }}_{2,2}^{(i)}\left[\begin{array}{c}{H}_i\\ G_i\end{array}\right](T^{(i)},U^{(i)})={\sum }_{m_1,m_2,n_1,n_2\in \mathbb{Z}}(-1{)}^{G_i{m}_2}{q}^{p_L^2\mathrm{/}2}{\bar{q}}^{p_R^2\mathrm{/}2},q=e^{2\pi i\tau }\mathrm{.}$ The full lattice factor with the paired orbifold twist $(h_j,g_j)$ is source-defined by ${\mathrm{\Gamma }}_{2,2}^{(i)}\left[\begin{array}{c}{H}_i\\ G_i\end{array}\right]\text{\hspace{0.17em}⁣}\left[\begin{array}{c}{h}_j\\ g_j\end{array}\right]=\{\begin{array}{ll}{\mathrm{\Gamma }}_{2,2}^{(i)}\left[\begin{array}{c}{H}_i\\ G_i\end{array}\right],& h_j=g_j=0,\\ \frac{1}{2}{\displaystyle \sum _{{\zeta }_i,{\delta }_i=0}^1(-1{)}^{{\zeta }_i{G}_i+{\delta }_i{H}_i+H_i{G}_i}{\mid \vartheta \left[\begin{array}{c}{\zeta }_i\\ {\delta }_i\end{array}\right]\vartheta \left[\begin{array}{c}{\zeta }_i+h_j\\ {\delta }_i+g_j\end{array}\right]\mid }^2,}& \text{otherwise}\mathrm{.}\end{array}$ In the total partition function the paired twists are $(h_2,g_2)$ for $i=1$, $(h_1,g_1)$ for $i=2$, and $(h_1+h_2,g_1+g_2)$ for $i=3$. Define the base spin-structure phase $P_0:=(-1{)}^{a+b+k+l}$ and the Scherk–Schwarz insertion ${\mathrm{\Xi }}_{\mathrm{S}\mathrm{S}}:=(-1{)}^{a(G_1+G_3)+b(H_1+H_3)+H_1{G}_1+H_3{G}_3+k{G}_2+l{H}_2+H_2{G}_2}\mathrm{.}$ The source's phase replacement is $P_0\mapsto P_0{\mathrm{\Xi }}_{\mathrm{S}\mathrm{S}}$, so the full phase is $(-1{)}^{a+b+k+l+a(G_1+G_3)+b(H_1+H_3)+H_1{G}_1+H_3{G}_3+k{G}_2+l{H}_2+H_2{G}_2}\mathrm{.}$

Use $\mathrm{\hslash }=c=1$, ${\alpha }^{\mathrm{\prime }}>0$, ${\mathrm{\ell }}_s=\sqrt{{\alpha }^{\mathrm{\prime }}}$ and $M_s={\mathrm{\ell }}_s^{-1}$ as in def-supergravity-conventions-and-normalizations. Here $\mathsf{G}$ and $\mathsf{B}$ are dimensionless lattice data: the internal physical metric in angular coordinates is ${\alpha }^{\mathrm{\prime }}\mathsf{G}$. Thus $T_2,U_2$ and the displayed $p_{L,R}$ are dimensionless. To remove the ambiguity of “string units”, write ${\widehat{m}}^2:={\alpha }^{\mathrm{\prime }}{M}_{\mathrm{p}\mathrm{h}\mathrm{y}\mathrm{s},s}^2$, where $M_{\mathrm{p}\mathrm{h}\mathrm{y}\mathrm{s},s}$ is the four-dimensional string-frame pole mass. In the NS chiral sectors relevant to the source's tachyon test, set ${\mathcal{L}}_L=\frac{1}{2}{p}_L^2+N_L-\frac{1}{2},{\mathcal{L}}_R=\frac{1}{2}{p}_R^2+N_R-\frac{1}{2}\mathrm{.}$ The source denotes these chiral levels $M_L^2,M_R^2$ in equations (2.21)–(2.22). A physical level-matched NS–NS state obeys ${\widehat{m}}^2=4{\mathcal{L}}_L=4{\mathcal{L}}_R$ in the adopted ${\alpha }^{\mathrm{\prime }}$ restoration, since its noncompact chiral momentum contribution is $-{\alpha }^{\mathrm{\prime }}{M}_{\mathrm{p}\mathrm{h}\mathrm{y}\mathrm{s},s}^2\mathrm{/}4$. The Ramond intercept is zero rather than $1\mathrm{/}2$ for a Ramond chiral factor. This normalization also gives ${\widehat{m}}^2=n^2\mathrm{/}{\mathsf{G}}_{11}$ for a pure circle momentum mode: the displayed lattice normalization has $p_L^2=p_R^2=n^2\mathrm{/}(2{\mathsf{G}}_{11})$, so each chiral level shift is $n^2\mathrm{/}(4{\mathsf{G}}_{11})$. No chiral level is silently equated to the full pole mass.

For the tree-level pole-mass interpretation, fix four-dimensional Minkowski space $M_4={\mathbb{R}}^4$ with global coordinates $x^{\mu }$ of length dimension, $\mu =0,1,2,3$, and string-frame metric $g_{4,s}=\eta =\mathrm{diag}(-1,1,1,1)$. Use the trivial spin structure and its complex Dirac bundle $S_4=M_4\times {\mathrm{\Delta }}_4$, ${\mathrm{\Delta }}_4={\mathbb{C}}^4$, with constant Clifford matrices ${\gamma }^{\mu }\in {\mathrm{End}}_{\mathbb{C}}({\mathrm{\Delta }}_4)$ satisfying ${\gamma }^{\mu }{\gamma }^{\nu }+{\gamma }^{\nu }{\gamma }^{\mu }=2{\eta }^{\mu \nu }I$. A vector-spinor is a smooth section $\psi \in C^{\mathrm{\infty }}(M_4,T^{\ast }{M}_4\otimes S_4)$, written $\psi =d{x}^{\mu }\otimes {\psi }_{\mu }(x)$ in this parallel trivialization. Define ${\mathcal{D}}_4{\psi }_{\mu }={\gamma }^{\nu }{\mathrm{\partial }}_{\nu }{\psi }_{\mu }$ and $({\mathrm{\square }}_{4,s}\psi {)}_{\mu }={\eta }^{\nu \rho }{\mathrm{\partial }}_{\nu }{\mathrm{\partial }}_{\rho }{\psi }_{\mu }$ componentwise. For a positive constant pole mass $m=M_{\mathrm{p}\mathrm{h}\mathrm{y}\mathrm{s},s}$, use the free linear spin-$3\mathrm{/}2$ description $({\mathcal{D}}_4-m){\psi }_{\mu }=0$, ${\mathrm{\partial }}^{\mu }{\psi }_{\mu }=0$, and ${\gamma }^{\mu }{\psi }_{\mu }=0$, with indices raised using $\eta$. A complex Fourier mode is ${\psi }_{\mu }(x)=u_{\mu }{e}^{i{p}_{\nu }{x}^{\nu }}$, where $p\in ({\mathbb{R}}^4{)}^{\ast }$ is a constant momentum covector of inverse-length dimension and $u_{\mu }\in {\mathrm{\Delta }}_4$ is its constant polarization; its constraints are $p^{\mu }{u}_{\mu }={\gamma }^{\mu }{u}_{\mu }=0$ and $(i{\gamma }^{\nu }{p}_{\nu }-m)u_{\mu }=0$. Composing the kinetic equation with ${\mathcal{D}}_4+m$ gives $({\mathrm{\square }}_{4,s}-m^2){\psi }_{\mu }=0$, so a nonzero mode has ${\eta }^{\nu \rho }{p}_{\nu }{p}_{\rho }=-m^2$. Here “gravitino” denotes the fermionic mode with this linear description whose massless supersymmetric parent is the gauge field of local supersymmetry. This is a flat-background tree-level description of the reported pole masses, not an interacting field equation on the putative one-loop vacuum. The symbols $S,\bar{S}$ label the source's left and right sectors, not two new scalar fields. Its equation (2.14) reports the two would-be massless gravitinos' dimensionless squared pole-mass shifts ${\widehat{m}}_S^2=\frac{1}{T_2^{(1)}{U}_2^{(1)}}+\frac{1}{T_2^{(3)}{U}_2^{(3)}},{\widehat{m}}_{\bar{S}}^2=\frac{1}{T_2^{(2)}{U}_2^{(2)}}\mathrm{.}$ Thus $M_{S,\mathrm{p}\mathrm{h}\mathrm{y}\mathrm{s},s}^2=M_s^2{\widehat{m}}_S^2$ and similarly for $\bar{S}$; writing ${\alpha }^{\mathrm{\prime }}=1$ gives precisely the paper's numerical formulas. A tree-level tachyon means a physical, level-matched, projection-surviving state with $M_{\mathrm{p}\mathrm{h}\mathrm{y}\mathrm{s},s}^2<0$ in the genus-zero spectrum. It is not a generic negative term before level matching and GSO/orbifold projections, nor a negative one-loop Hessian eigenvalue. The source reports absence of such states throughout the allowed torus moduli after its sector analysis; this item does not rederive that analysis.

At genus one the source reports a minimum of the twelve real $T,U$ torus-moduli directions at $T^{(i)}=U^{(i)}=1+i$, with positive Hessian in those directions and negative vacuum-energy density ${\mathrm{\Lambda }}_s$ in the four-dimensional string-frame effective action term $-\int \sqrt{-g_{4,s}}{\mathrm{\Lambda }}_s$. To state its separate dilaton caveat, let $v_6=V_{6,s}\mathrm{/}(2\pi {\mathrm{\ell }}_s{)}^6$ and define the four-dimensional dilaton by $e^{-2{\varphi }_4}=v_6{e}^{-2{\mathrm{\Phi }}_{10}}$, up to a fixed reference-volume constant absorbed into the gravitational coefficient. The local four-dimensional coupling variable is $g_4=e^{{\varphi }_4}$, while the ten-dimensional string coupling at a constant background is $g_{\mathrm{s}\mathrm{t}\mathrm{r}}=e^{{\mathrm{\Phi }}_{10}}$. At fixed torus volume their weak-coupling limits agree. The four-dimensional Einstein metric is $g_{4,E}=e^{-2{\varphi }_4}{g}_{4,s},V_E({\varphi }_4)=e^{4{\varphi }_4}{\mathrm{\Lambda }}_s\mathrm{.}$ This is the four-dimensional Weyl rescaling, distinct from the ten-dimensional relation $G_{10,E}=e^{-{\mathrm{\Phi }}_{10}\mathrm{/}2}{G}_{10,s}$. At fixed torus moduli, ${\mathrm{\Lambda }}_s<0$ gives no finite dilaton stationary point and $V_E\to 0$ from below as $g_4\to 0$. “Runaway” here means the absence of a stationary dilaton in this truncated potential; no dynamical rolling direction toward weak coupling is inferred from that limit. The source proposes additional flux/orientifold mechanisms as possibilities outside this construction. The negative torus-moduli energy therefore does not itself establish a stabilized AdS vacuum.

Geometric finite-group orbifold compactification data

Supergravity Conventions And Normalizations

The paper is the sole primary source for this specific GSO choice, shifted lattices, tree spectrum, and genus-one calculation. Its relevant full text and exact equations were inspected. Ginsparg's orbifold discussion is used only for the general conditional vocabulary in [F1], not to prove the model-specific partition function. The source's own title uses “AdS vacuum,” but its conclusion states that the negative one-loop energy does not by itself imply an AdS vacuum because the Einstein-frame dilaton runs away; this carrier preserves that caveat.