The Conformal Window and Interacting Fixed Points

Massless $SU(N_c)$ SQCD is proposed to flow to an interacting superconformal field theory when

$$\frac32N_c<N_f<3N_c.$$

Within this open interval, anomaly freedom and fixed-point supersymmetry determine exact R-charges and protected dimensions, conditional on the proposed fixed point and on the absence of an accidental current that changes the R-symmetry. Existence and uniqueness throughout the entire interval remain dynamical claims supported by electric–magnetic duality and many consistency checks, not theorems obtained from a beta-function zero. The phase proposal and its magnetic description are developed in Seiberg 1995, §§2–4, PDF pp. 4–12 and reviewed in Intriligator and Seiberg 1996, §§5.3–5.5.

Required background. SQCD infrared phases gives the phase proposal, ultraviolet and infrared fixed points gives the RG criteria, and Seiberg duality supplies the magnetic frame. Helpful background. Holomorphic and NSVZ running fixes the beta-function conventions.

The upper boundary from asymptotic freedom

The electric one-loop coefficient is

$$b_0=3N_c-N_f.$$

For $N_f<3N_c$, the gauge coupling grows toward the infrared. To obtain a parametrically controlled weak-coupling limit, use the Veneziano scaling

$$N_c,N_f\longrightarrow\infty, \qquad x\equiv\frac{N_f}{N_c}=3-\varepsilon, \qquad 0<\varepsilon\ll1.$$

Then the infrared zero occurs at weak electric coupling in a controlled Banks–Zaks expansion. Equivalently, one may take an integer displacement such as $N_f=3N_c-1$ while sending $N_c$ large. At fixed small rank the discrete theory nearest the boundary need not be parametrically weak. This perturbative limit anchors the existence of interacting fixed points near the upper edge; it does not by itself establish the entire window Seiberg 1995, §2, PDF pp. 4–6.

At $N_f=3N_c$, the leading coefficient vanishes and the proposed fixed point meets the free theory. For $N_f>3N_c$, the electric theory is infrared free and is not the same asymptotically free ultraviolet problem.

The exact fixed-point constraint

In the NSVZ convention, the gauge beta function has numerator

$$3N_c-N_f(1-\gamma_Q),$$

where charge conjugation gives $\gamma_Q=\gamma_{\widetilde Q}$. Here the convention is

$$\gamma_Q\equiv2(\Delta_Q-1).$$

Some sources instead write $\widehat\gamma_Q=\Delta_Q-1$; then $\widehat\gamma_Q=\gamma_Q/2$ and the same numerator contains $1-2\widehat\gamma_Q$. Keeping this translation explicit prevents a factor-of-two mismatch. At an interacting fixed point with nonzero denominator,

$$\gamma_Q=1-\frac{3N_c}{N_f}.$$

The scaling dimension of a chiral field in a gauge-fixed description is

$$\Delta_Q=1+\frac12\gamma_Q =\frac32\left(1-\frac{N_c}{N_f}\right).$$

This agrees with the anomaly-free R-charge through $\Delta=3R/2$:

$$R(Q)=R(\widetilde Q)=1-\frac{N_c}{N_f}.$$

The gauge-variant field $Q$ is not itself a local gauge-invariant CFT operator, so its dimension should not be tested directly against the scalar unitarity bound. Gauge-invariant chiral composites must satisfy it.

Mesons, baryons, and the lower boundary

The meson has

$$R(M)=2\left(1-\frac{N_c}{N_f}\right), \qquad \Delta(M)=3\left(1-\frac{N_c}{N_f}\right).$$

A scalar chiral primary obeys $\Delta\ge1$, equivalently $R\ge2/3$. Therefore

$$\Delta(M)\ge1 \quad\Longleftrightarrow\quad N_f\ge\frac32N_c.$$

At the equality, $M$ reaches the free-field bound. Below it, the interacting formula would violate unitarity, so the meson must become free and an accidental symmetry appears. This agrees with the magnetic beta-function coefficient

$$\widetilde b_0 =3(N_f-N_c)-N_f =2N_f-3N_c.$$

The coefficient $\widetilde b_0$ has the infrared-free sign when $N_f<3N_c/2$. In the range where a non-Abelian magnetic gauge dual applies, $N_c+2\le N_f<3N_c/2$, this gives the free-magnetic phase. The theory is instead s-confining at $N_f=N_c+1$, has a quantum-modified moduli space at $N_f=N_c$, and has an Affleck–Dine–Seiberg runaway for $0<N_f<N_c$; pure super-Yang–Mills at $N_f=0$ is gapped. These phases must not be inferred from the sign alone Seiberg 1994, §§3–5, PDF pp. 6–15. A parametrically weak magnetic description just above the lower edge again requires a Veneziano limit,

$$\frac{N_f}{N_c}=\frac32+\widetilde\varepsilon, \qquad 0<\widetilde\varepsilon\ll1,$$

so that both the magnetic gauge coupling and the meson–quark Yukawa coupling are perturbative. At a fixed small rank, proximity by one flavor does not provide such a parameter Seiberg 1995, §3, PDF pp. 6–8.

Electric baryons have

$$R(B)=N_c\left(1-\frac{N_c}{N_f}\right), \qquad \Delta(B)=\frac{3N_c}{2} \left(1-\frac{N_c}{N_f}\right).$$

They do not set the generic lower edge for $N_c>2$; the meson reaches unitarity first. The exceptional intersection $N_c=2$, $N_f=3=N_c+1=3N_c/2$ has enhanced $SU(6)$ flavor symmetry and an s-confining composite description, rather than the generic free-magnetic gauge description. It must be treated separately Seiberg 1994, §5, PDF pp. 13–15.

Protected central charges

For the candidate exact R-symmetry,

$$\operatorname{Tr}R=-N_c^2-1, \qquad \operatorname{Tr}R^3=N_c^2-1-\frac{2N_c^4}{N_f^2}.$$

The four-dimensional $\mathcal N=1$ anomaly formulas give

$$a=\frac{3}{16}\left( 2N_c^2-1-\frac{3N_c^4}{N_f^2} \right),$$

and

$$c=\frac{1}{16}\left( 7N_c^2-2-\frac{9N_c^4}{N_f^2} \right).$$

These are exact protected outputs conditional on the fixed point and R-symmetry identification. Computing them in magnetic variables, including the $N_f^2$ mesons, gives the same result. The nonperturbative relation between superconformal central functions and R-current anomalies is established in Anselmi, Freedman, Grisaru, and Johansen 1998, §§2–4.

The inequalities $a>0$, $c>0$, and the supersymmetric conformal-collider bounds provide consistency checks but do not alone prove that the SCFT exists Hofman and Maldacena 2008, §2.4, eqs. (2.39)–(2.40), PDF pp. 21–22. The $a$-theorem independently requires $a_{\mathrm{UV}}>a_{\mathrm{IR}}$ along a nontrivial flow Komargodski and Schwimmer 2011, §§3–4. With $x=N_f/N_c$, the SQCD formulas give the exact check

$$a_{\mathrm{UV}}-a_{\mathrm{IR}} =\frac{N_c^2}{48} \left(2x-9+\frac{27}{x^2}\right).$$

It is positive for $3/2\le x<3$ and vanishes at the free upper boundary $x=3$. This is a necessary consistency test, not an existence proof.

Electric and magnetic control

Region	Useful frame	Controlled statement
$N_f/N_c=3-\varepsilon$ in the Veneziano limit	Electric	Weak fixed-point coupling; perturbation theory controls unprotected quantities order by order.
Interior	Neither generically weak	R-charges, anomalies, chiral data, and protected traces remain exact; generic correlators are strongly coupled.
$N_f/N_c=3/2+\widetilde\varepsilon$ in the Veneziano limit	Magnetic	Weak magnetic gauge and Yukawa couplings; the meson is near free.
$N_c+2\le N_f<3N_c/2$ when nonempty	Non-Abelian free-magnetic phase	The interacting-window formula for $M$ is invalid; accidental free fields must be explicit.
$N_f\le N_c+1$	Other strong-coupling regimes	Confinement, a quantum-modified moduli space, an ADS runaway, or pure-SYM vacua must be distinguished separately.

The two weak edges provide complementary perturbative anchors. Continuity, duality, anomaly matching, and operator-map checks supply strong evidence for the interior, but the anchors do not logically exclude an intermediate change of phase. Algebraic interpolation is therefore evidence, not a proof of global continuation.

Scheme dependence and exact information

The NSVZ beta function is exact in a particular coupling convention. Analytic redefinitions change its denominator and the coordinate location of a zero. At a genuine fixed point, however, scaling dimensions, R-charges, anomalies, and critical exponents are physical after redundant directions are removed.

The numerator equation is useful because it is equivalent to the anomaly-free superconformal R condition. It does not prove that the denominator is nonsingular, that the zero is reached by the flow, or that no accidental symmetry changes $R$. Those are separate gates.

Endpoints and accidental fields

When $3N_c/2$ is an allowed integer and $N_c>2$, the generic lower endpoint has $M$ at $R=2/3$, and the magnetic gauge–Yukawa description approaches a logarithmically free limit. For odd $N_c$, the boundary lies between adjacent integer-flavor theories; for $N_c=2$, the $N_f=3$ exception described above applies. A chiral primary saturating the bound obeys a free equation and contributes the free-field values

$$a_{\mathrm{chiral}}=\frac{1}{48}, \qquad c_{\mathrm{chiral}}=\frac{1}{24}.$$

If a composite crosses the bound in a more general theory, its emergent accidental current changes the anomaly functional. Algebraically one replaces the trial assignment for that operator by its free-field contribution and then repeats a-maximization. This replacement does not mean that the ultraviolet fermion trace had counted the composite as an independent elementary field.

What is exact and what is inferred

Exact conditional statements: anomaly relations, $R(Q)=1-N_c/N_f$ for the anomaly-free candidate, $\Delta=3R/2$ for chiral primaries at an SCFT, the protected $a,c$ formulas, and the unitarity obstruction at $N_f<3N_c/2$.

Dynamical conclusions: existence of a unique interacting fixed point throughout the open window, absence of intermediate phases, and the complete unprotected spectrum.

Perturbatively controlled conclusions: detailed correlators and anomalous dimensions near either weakly coupled edge.

Keeping these categories separate makes the fixed-point data useful without overstating the phase diagram.

Common pitfalls

Applying the unitarity bound to $Q$. $Q$ is gauge variant. Test gauge-invariant chiral primaries such as $M$ and $B$.

Treating an NSVZ numerator zero as an existence proof. The flow, denominator, and accidental symmetries still require control.

Including endpoints in an open-window formula. The upper edge and the generic even-rank lower endpoint meet logarithmically free limits and require separate analysis. Odd rank has no integer theory exactly at $3N_c/2$, while $SU(2)$ with $N_f=3$ is an enhanced-symmetry s-confining exception.

Exercises

For $N_c=3$, $N_f=6$:

1. compute $R(Q)$, $\Delta(M)$, and $\Delta(B)$;
2. compute $a$ and $c$;
3. identify whether either electric or magnetic frame is parametrically weak.

Solution

$R(Q)=1/2$, so $\Delta(M)=3/2$ and $\Delta(B)=9/4$. The central charges are

$$a=\frac{3}{16}\left(18-1-\frac{243}{36}\right) =\frac{123}{64},$$ $$c=\frac{1}{16}\left(63-2-\frac{729}{36}\right) =\frac{163}{64}.$$

Here $N_f=2N_c$ lies in the interior of the window, not parametrically near either $3N_c$ or $3N_c/2$. The pair is self-rank—both gauge groups are $SU(3)$—but neither description is forced to be weakly coupled.

References

• Anselmi, Damiano, Daniel Z. Freedman, Marc T. Grisaru, and Andreas A. Johansen. “Nonperturbative Formulas for Central Functions of Supersymmetric Gauge Theories.” Nuclear Physics B 526 (1998): 543–571. arXiv:hep-th/9708042.
• Hofman, Diego M., and Juan Maldacena. “Conformal Collider Physics: Energy and Charge Correlations.” Journal of High Energy Physics 05 (2008): 012. arXiv:0803.1467.
• Intriligator, Kenneth, and Nathan Seiberg. “Lectures on Supersymmetric Gauge Theories and Electric–Magnetic Duality.” Nuclear Physics B Proceedings Supplements 45BC (1996): 1–28. arXiv:hep-th/9509066.
• Komargodski, Zohar, and Adam Schwimmer. “On Renormalization Group Flows in Four Dimensions.” Journal of High Energy Physics 12 (2011): 099. arXiv:1107.3987.
• Seiberg, Nathan. “Exact Results on the Space of Vacua of Four-Dimensional SUSY Gauge Theories.” Physical Review D 49 (1994): 6857–6863. arXiv:hep-th/9402044.
• Seiberg, Nathan. “Electric–Magnetic Duality in Supersymmetric Non-Abelian Gauge Theories.” Nuclear Physics B 435 (1995): 129–146. arXiv:hep-th/9411149.