The Principal Chiral Model and the Integrability Bridge

The principal chiral model places a group-valued field in $1+1$ dimensions and combines non-Abelian left–right symmetry, asymptotic freedom, and a classical Lax connection. It is therefore a natural bridge from sigma-model dynamics to exact scattering. The bridge has a required middle span: one must show that suitable conserved charges survive quantization and then verify factorization, the Yang–Baxter equation, unitarity, crossing, spectrum, and ultraviolet consistency. A classical Lax pair alone is not a quantum-integrability theorem.

Required background. Sigma-model target geometry and control supplies the group-manifold target, while quantum currents and improvements supplies the distinction between a classical current and a renormalized quantum charge. Helpful background. Lie groups, Lie algebras, and the adjoint action supplies Maurer–Cartan forms and invariant traces.

Action, symmetry, and currents

Let $U(x)\in SU(K)$ and use the fundamental trace. The Lorentzian action is

$$S =-\frac{1}{2g_0^2} \int\mathrm d^2x\, \operatorname{tr}(j_\mu j^\mu), \qquad j_\mu=U^{-1}\partial_\mu U .$$

Because $j_\mu$ is anti-Hermitian, the target metric $-\operatorname{tr}(j_\mu j_\nu)$ is positive on tangent vectors. The global transformation

$$U\longmapsto LUR^{-1}, \qquad (L,R)\in SU(K)_L\times SU(K)_R,$$

acts on the left-invariant Maurer–Cartan form and the right-invariant Maurer–Cartan form as

$$j_\mu\longmapsto Rj_\mu R^{-1}, \qquad \ell_\mu\equiv\partial_\mu U\,U^{-1} \longmapsto L\ell_\mu L^{-1}.$$

Here “left-invariant” and “right-invariant” describe invariance of the differential form under multiplication on that side; they need not coincide with a convention that names a Noether current by the symmetry it generates. The common central pair acts trivially, so the faithful global group is

$$\frac{SU(K)_L\times SU(K)_R}{\mathbb Z_K}.$$

Boundary conditions can further determine which transformations act as physical symmetries.

Varying $U\mapsto Ue^\epsilon$ gives the equation of motion

$$\partial^\mu j_\mu=0.$$

Independently of the action, $j=U^{-1}\mathrm dU$ obeys the Maurer–Cartan identity

$$\partial_\mu j_\nu-\partial_\nu j_\mu +[j_\mu,j_\nu]=0.$$

The first equation is dynamical; the second is geometric. Their coexistence produces a one-parameter flat connection.

Classical Lax connection

Define $x^\pm=x^0\pm x^1$, $\partial_\pm=\tfrac12(\partial_0\pm\partial_1)$, and the one-form components $j_\pm=\tfrac12(j_0\pm j_1)$. Then

$$\partial_+j_-+\partial_-j_+=0,$$

and

$$\partial_+j_- -\partial_-j_+ +[j_+,j_-]=0.$$

For complex spectral parameter $z$, set

$$\mathcal L_+(z) =\frac{j_+}{1-z}, \qquad \mathcal L_-(z) =\frac{j_-}{1+z}.$$

Its curvature is

$$\begin{aligned} (1-z^2)\mathcal F_{+-}(z) ={}& \partial_+j_- -\partial_-j_+ +[j_+,j_-] \\ &-z\left( \partial_+j_-+\partial_-j_+ \right). \end{aligned}$$

Thus $\mathcal F_{+-}(z)=0$ for every $z$ if and only if the equation of motion and Maurer–Cartan identity hold.

On a spatial circle, the monodromy

$$T(z) =\mathcal P\exp\!\left[ \int_0^L\mathrm dx^1\, \mathcal L_1(x^0,x^1;z) \right]$$

evolves by conjugation when the connection is flat and the fields obey compatible periodic boundary conditions. Consequently conjugation-invariant functions such as $\operatorname{tr}T(z)^r$ are classically conserved. Expansions of the monodromy around suitable spectral points generate nonlocal charges, including the classical precursor of the Yangian tower. Local higher-spin currents arise separately from invariant tensors. For example, the equations of motion and Maurer–Cartan identity imply

$$\partial_-\operatorname{tr}(j_+^m)=0, \qquad \partial_+\operatorname{tr}(j_-^m)=0$$

for invariant powers that survive the group-specific independence relations. The two towers can be related by the full integrable structure, but they are not produced by the same formal expansion. The displayed Lax identity is reproduced algebraically in the chapter’s benchmark.

From ultraviolet running to a mass scale

For a compact simple target group, the Ricci tensor is positive in the invariant metric, and the two-dimensional coupling is asymptotically free. The one-loop coefficient is proportional to the dual Coxeter number; its numerical value depends on the normalization of the invariant trace and coupling. The non-Abelian sigma-model running and mass-scale problem are developed in Polyakov and Wiegmann 1983, pp. 121–126. Running therefore generates a scale.

That ultraviolet statement does not yet determine the exact mass spectrum. In the accepted $SU(K)$ bootstrap and Bethe-ansatz description, the rank-$r$ particle multiplet transforms as

$$(R_r,\overline{R_r}) \quad\text{under}\quad SU(K)_L\times SU(K)_R, \qquad R_r=\bigwedge^r\mathbf K,$$

and has the sine-law ratio

$$\frac{M_r}{M_1} =\frac{\sin(\pi r/K)}{\sin(\pi/K)}, \qquad r=1,\ldots,K-1,$$

as developed in Ogievetsky, Reshetikhin, and Wiegmann 1987, §§ 3–5. This is quantum bootstrap data with ultraviolet checks, not a consequence of the classical Lax pair alone. A massive symmetric phase is also supported by large-$K$ methods and comparison with the equivalent $O(4)$ description at $K=2$. As in the other laboratories, a quoted mass must refer to a physical representation and a named RG scheme.

The model’s control and scope are summarized in the strong-coupling laboratory map and regime comparison.

Quantum integrability is an additional result

Quantization can spoil a classical conservation law through short-distance singularities, operator mixing, or an anomaly in the nonlocal current. Establishing quantum integrability requires at least:

1. renormalized conserved charges with no obstructing anomaly;
2. charges acting nontrivially on asymptotic particles;
3. elastic scattering with no particle production;
4. factorization compatible with the Yang–Baxter equation;
5. unitarity, crossing, and the correct global-symmetry tensor structure;
6. a spectrum and bootstrap that close without unexplained poles;
7. an ultraviolet check, such as thermodynamic Bethe ansatz or form-factor behavior, matching the sigma model.

Lüscher showed how quantum nonlocal charges constrain particle production in two-dimensional nonlinear sigma models Lüscher 1978, §§ 2–4. Faddeev and Reshetikhin construct an ultralocal Hamiltonian and lattice route with the same classical equations of motion Faddeev and Reshetikhin 1986, §§ 2–5. Identifying its quantum scaling limit with the original principal chiral model is an additional step rather than an automatic consequence of the classical equivalence.

The later integrability exact-data chain owns the passage from conserved charges to factorized scattering, Yang–Baxter consistency, bootstrap poles, and finite-volume tests. Here the conclusion is intentionally bounded: the principal chiral model supplies the classical structures and the physical questions; the exact-data chain supplies the quantum certification.

The SU(2) and O(4) match

Write

$$U=n^0\mathbf 1+i n^a\sigma^a, \qquad (n^0)^2+n^a n^a=1.$$

Then $U\in SU(2)$ and

$$\operatorname{tr} \left( \partial_\mu U^\dagger\partial^\mu U \right) =2\,\partial_\mu n^A\partial^\mu n^A, \qquad A=0,1,2,3.$$

In the normalization used here,

$$S_{\mathrm{PCM}} =\frac{1}{g_{\mathrm{PCM}}^2} \int\mathrm d^2x\, \partial_\mu n^A\partial^\mu n^A.$$

Comparing with $S_{O(4)}=(2g_O^2)^{-1}\int(\partial n)^2$ gives

$$g_O^2=\frac{g_{\mathrm{PCM}}^2}{2}.$$

After this explicit match, the $SU(2)$ principal chiral action is the $O(4)$ sigma-model action on $S^3$. Moreover,

$$\frac{SU(2)_L\times SU(2)_R}{\mathbb Z_2} \simeq SO(4),$$

so the faithful connected global symmetries agree. This is a stringent special-case cross-check of beta functions, spectrum representations, and scattering. It does not turn the $SU(K>2)$ group manifold into a sphere or prove quantum integrability for general $K$ by analogy.

Common pitfalls

Equating flatness with quantization. A classically flat spectral connection generates classical charges. Composite-operator renormalization and anomalies must be checked before those charges constrain the quantum S-matrix.

Suppressing boundary conditions. Conservation of monodromy invariants uses a circle or sufficiently decaying fields on a line. Boundaries generally require reflection data and a modified integrability condition.

Comparing couplings without trace normalization. Rescaling the invariant trace rescales $g^2$ and beta-function coefficients. Physical mass ratios are safer until schemes and normalizations are matched.

Exercises

1. Starting from the two light-cone equations, verify the displayed curvature identity for $\mathcal L_\pm(z)$.

Solution

Compute

$$\mathcal F_{+-} =\partial_+\mathcal L_- -\partial_-\mathcal L_+ +[\mathcal L_+,\mathcal L_-].$$

Multiplication by $(1-z)(1+z)$ gives

$$(1-z)\partial_+j_- -(1+z)\partial_-j_+ +[j_+,j_-].$$

Separating the $z^0$ and $z^1$ terms produces the Maurer–Cartan combination and minus $z$ times the equation of motion. Both vanish on shell.

2. Show directly that $U=n^0\mathbf1+i n^a\sigma^a$ is unitary with unit determinant when $(n^0)^2+n^a n^a=1$.

Solution

Using $\sigma^a\sigma^b=\delta^{ab}\mathbf1+i\epsilon^{abc}\sigma^c$,

$$U^\dagger U =\left[ (n^0)^2+n^a n^a \right]\mathbf1 =\mathbf1,$$

because the antisymmetric term contracts the symmetric product $n^a n^b$. The eigenvalues are $n^0\pm i\lvert\mathbf n\rvert$, whose product is one, so $\det U=1$.

3. Combine the light-cone equation of motion and Maurer–Cartan identity to derive the local conserved currents $\operatorname{tr}(j_\pm^m)$. Why are they distinct from the monodromy charges?

Solution

Adding and subtracting

$$\partial_+j_-+\partial_-j_+=0, \qquad \partial_+j_- -\partial_-j_+ +[j_+,j_-]=0$$

gives

$$\partial_-j_+=\frac12[j_+,j_-], \qquad \partial_+j_-=-\frac12[j_+,j_-].$$

Therefore

$$\partial_-\operatorname{tr}(j_+^m) =\frac m2\operatorname{tr}\!\left(j_+^{m-1}[j_+,j_-]\right)=0,$$

by cyclicity of the trace; the $j_-$ equation is analogous. These are local invariant-tensor currents. Monodromy expansions contain path ordering and generate nonlocal charges, so conservation of one tower does not automatically establish the quantum conservation or algebra of the other.

4. Use the sine-law formula to show $M_r=M_{K-r}$. What does this check, and what does it not prove?

Solution

Since $\sin(\pi(K-r)/K)=\sin(\pi-\pi r/K)=\sin(\pi r/K)$, the two masses agree. This matches conjugate antisymmetric representations, a necessary global-symmetry and charge-conjugation check. It does not derive the bootstrap, prove absence of additional particles, or show that a classically integrable regulator flows to the required quantum theory; those conclusions need the rest of the exact-data chain.

Continue

Use the integrability exact-data chain for the quantum bootstrap. For a contrasting exact change of variables without a Lax construction, see The Schwinger Model, Screening, and Bosonization.

References

• Faddeev, L. D., and N. Yu. Reshetikhin. “Integrability of the Principal Chiral Field Model in (1+1)-Dimension.” Annals of Physics 167 (1986): 227–256. DOI.
• Lüscher, M. “Quantum Nonlocal Charges and Absence of Particle Production in the Two-Dimensional Nonlinear Sigma Model.” Nuclear Physics B 135 (1978): 1–19. DOI.
• Ogievetsky, E., N. Reshetikhin, and P. Wiegmann. “The Principal Chiral Field in Two Dimensions on Classical Lie Algebras: The Bethe-Ansatz Solution and Factorized Theory of Scattering.” Nuclear Physics B 280 (1987): 45–96. DOI.
• Polyakov, Alexander M., and Paul B. Wiegmann. “Theory of Nonabelian Goldstone Bosons in Two Dimensions.” Physics Letters B 131 (1983): 121–126. DOI.