Tetrad and Connection Formulations

Tetrad gravity replaces the metric as the primary gravitational field by a coframe, while first-order gravity also treats the Lorentz connection as independent. The enlargement exposes local Lorentz gauge symmetry, separates torsion from curvature, and admits spinorial matter and canonical connection variables naturally. It does not produce a new classical vacuum theory unless the hypotheses establishing equivalence with Einstein gravity are changed.

Helpful background. Frames, Coframes, and Anholonomy supplies the frame bundle and local Lorentz transformations, The Gibbons-Hawking-York Boundary Term supplies the metric Dirichlet comparison, Presymplectic Potentials and Currents supplies the surface variation and charge language, The Einstein Equation supplies the shared metric dynamics, Hypersurface Normals and Tangential Projectors supplies the geometry used in time gauge, and Formulations of General Relativity locates this chapter among the metric, canonical, and hyperbolic formulations.

Enter the tetrad description

Work on an oriented, time-oriented four-manifold with internal metric

$$\eta_{IJ}=\operatorname{diag}(-1,1,1,1).$$

Selected coframes belong to the proper orthochronous sector and agree with the spacetime orientation. With $\epsilon_{0123}=+1$, the coframe volume form is

$$\operatorname{vol}_e =\frac1{4!}\epsilon_{IJKL} e^I\wedge e^J\wedge e^K\wedge e^L,$$

and a nondegenerate coframe determines $g=\eta_{IJ}e^I\otimes e^J$. The same metric is represented by every $e'^I=\Lambda^I{}_J e^J$ with $\Lambda\in SO^+(1,3)$; this local Lorentz freedom is gauge, not an extra metric degree of freedom.

An antisymmetric connection $\omega^{IJ}=-\omega^{JI}$ defines

$$T^I=De^I=de^I+\omega^I{}_J\wedge e^J, \qquad F^{IJ}=d\omega^{IJ}+\omega^I{}_K\wedge\omega^{KJ}.$$

Under a local Lorentz transformation, $\omega'=\Lambda\omega\Lambda^{-1}-d\Lambda\,\Lambda^{-1}$, so neither the connection coefficients nor an anholonomic coframe alone measure curvature. The torsion and curvature transform covariantly. For internal bivectors use

$$(\star X)^{IJ}=\frac12\epsilon^{IJ}{}_{KL}X^{KL}, \qquad \star^2=-1,$$

where the final sign is specific to Lorentzian signature.

The conditional equivalence chain

Let $\Sigma_{IJ}=\star(e_I\wedge e_J)$ and $\kappa=8\pi G$. A convenient Palatini–Holst convention is

$$\begin{aligned} S_\gamma[e,\omega] ={}&\frac1{2\kappa}\int_{\mathcal M} \left(\Sigma_{IJ}-\frac1\gamma e_I\wedge e_J\right)\wedge F^{IJ}\\ &-\frac{\Lambda}{24\kappa}\int_{\mathcal M} \epsilon_{IJKL}e^I\wedge e^J\wedge e^K\wedge e^L +S_{\partial}. \end{aligned}$$

For a nondegenerate coframe and real, finite, nonzero $\gamma$, the connection equation applies the invertible bivector map $P_\gamma=\star-\gamma^{-1}$ to the torsion equation. It implies $T^I=0$, which fixes $\omega$ to the Levi-Civita spin connection $\omega(e)$. The coframe equation then becomes Einstein’s equation. The map ceases to be invertible at the complex self-dual values $\gamma=\pm i$, so those cases require a different reality-condition analysis. The derivation and its normalization are given in Holst 1996, equations (1)–(14), PDF and Corichi, Rubalcava-García, and Vukašinac 2016, §§IV.A–IV.B, pp. 22–27, PDF.

On the torsion-free vacuum branch, the Holst term does not change the bulk Einstein equation. This statement is conditional: degenerate coframes, torsionful matter couplings, complex variables, and unmatched boundary data can invalidate steps in the argument. The Immirzi parameter therefore labels a family of classically equivalent real vacuum descriptions under the stated hypotheses; that fact alone neither proves nor refutes equivalence of their quantizations.

From a spacetime connection to canonical variables

Choose time gauge and an oriented spatial triad. Write $q=\det(\gamma_{ij})$ here to distinguish the spatial determinant from the Immirzi parameter $\gamma$. With the site’s convention

$$K_{ij}=-\frac12\mathcal L_n\gamma_{ij},$$

define

$$E^i{}_a=\frac12\epsilon^{ijk}\epsilon_{abc}e^b{}_j e^c{}_k =\sqrt q\,e^i{}_a, \qquad \mathcal K_i{}^a=-K_{ij}e^{ja}.$$

The real Ashtekar–Barbero connection and its bracket are then

$$A_i{}^a=\Gamma_i{}^a+\gamma\mathcal K_i{}^a =\Gamma_i{}^a-\gamma K_{ij}e^{ja}, \qquad \{A_i{}^a(x),E^j{}_b(y)\} =\kappa\gamma\,\delta_i^j\delta^a_b\delta^{(3)}(x,y).$$

The minus sign in the last expression for $A$ is forced by the inherited sign of $K_{ij}$. Gauss, spatial-diffeomorphism, and scalar constraints still have to be imposed; writing a connection does not solve them. Barbero’s real Lorentzian construction is developed in Barbero 1995, equations (1)–(18), PDF.

A route through the chapter

The six pages should be read in order:

1. Einstein Gravity in Tetrad Variables reconstructs the metric and volume form, identifies local Lorentz redundancy, and varies the tetrad action.
2. Spin Connections, Torsion, and Compatibility distinguishes anholonomy, torsion, and curvature and derives the unique compatible torsion-free connection.
3. First-Order Palatini Gravity varies the coframe and connection independently and states every hypothesis needed to recover Einstein gravity.
4. The Holst Term and the Immirzi Parameter analyzes the Holst deformation, the Nieh–Yan identity, and the exceptional self-dual values.
5. Ashtekar–Barbero Connection Variables performs the time-gauge canonical change of variables with the chapter’s extrinsic-curvature sign.
6. Boundary Terms in Tetrad and Connection Gravity matches surface terms to fixed connection, fixed coframe, or fixed metric data and explains the role of Lorentz gauge at boundaries and corners.

The first three pages establish the covariant variables and equations. The next two explain the Immirzi family and its canonical coordinates. The final page shows why equality of bulk field equations is not enough to identify two boundary-value problems.

Boundaries and the meaning of equivalence

Varying the displayed bulk action produces the connection surface term

$$\frac1{2\kappa}\int_{\partial\mathcal M} \left(\Sigma_{IJ}-\frac1\gamma e_I\wedge e_J\right) \wedge\delta\omega^{IJ}.$$

It vanishes if the pullback of $\omega$ is fixed. Fixed-coframe or fixed-metric data instead require an appropriate boundary completion. Fixing the metric is weaker than fixing a coframe because boundary Lorentz rotations remain. On the torsion-free branch, in a boundary-adapted gauge and with matched orientations, the Palatini completion reproduces the Gibbons–Hawking–York problem. Corners, asymptotic reference terms, and allowed boundary gauge transformations must also match. These distinctions are reviewed in Corichi, Rubalcava-García, and Vukašinac 2016, §§IV–V, pp. 21–42, PDF.

Classical equivalence therefore means agreement of solutions and gauge orbits for a specified nondegenerate sector, matter content, boundary data, and surface symplectic structure. It does not mean that every field parametrization has the same off-shell action, that every exact boundary form is physically invisible, or that different quantum representations must be unitarily equivalent.

Review the chapter

Consider spatially flat FLRW spacetime $ds^2=-dt^2+a(t)^2\delta_{ij}dx^i dx^j$ with the oriented coframe $e^0=dt$, $e^a=a(t)\delta^a_i dx^i$, unit lapse, and zero shift.

1. Compute $K_{ij}$ using the site’s sign.
2. Find $E^i{}_a$, $\Gamma_i{}^a$, $\mathcal K_i{}^a$, and $A_i{}^a$.
3. Explain why changing real nonzero $\gamma$ changes $A_i{}^a$ but not the reconstructed metric or the torsion-free Einstein equation.
4. State what additional choice is required before varying the action on a finite time interval.

Solution

Writing $H=\dot a/a$, the spatial metric obeys $\dot\gamma_{ij}=2H\gamma_{ij}$. Therefore

$$K_{ij}=-\frac12\dot\gamma_{ij}=-H\gamma_{ij}.$$

The oriented diagonal triad gives

$$E^i{}_a=a^2\delta^i_a.$$

Its intrinsic spin connection vanishes in these Cartesian coordinates, $\Gamma_i{}^a=0$. Since $e^{ja}=a^{-1}\delta^{ja}$,

$$\mathcal K_i{}^a=-K_{ij}e^{ja}=Ha\,\delta_i^a, \qquad A_i{}^a=\gamma Ha\,\delta_i^a.$$

Thus $\gamma$ rescales the extrinsic-curvature contribution to the canonical connection, while $E$ still reconstructs the same spatial metric. For a nondegenerate coframe and real nonzero $\gamma$, the independent-connection equation still sets torsion to zero and the remaining equation is Einstein’s equation. This is a classical statement; it does not identify quantum spectra at different $\gamma$. On a finite time interval one must also specify the endpoint data—fixed connection, coframe, or metric—and add the boundary and corner terms that make that chosen variational problem differentiable.

References

• Barbero G., J. Fernando. “Real Ashtekar Variables for Lorentzian Signature Space-Times.” Physical Review D 51, no. 10 (1995): 5507–5510. DOI. Open PDF.
• Corichi, Alejandro, Irais Rubalcava-García, and Tatjana Vukašinac. “Actions, Topological Terms and Boundaries in First Order Gravity: A Review.” International Journal of Modern Physics D 25, no. 4 (2016): 1630011. DOI. Open PDF.
• Holst, Sören. “Barbero’s Hamiltonian Derived from a Generalized Hilbert–Palatini Action.” Physical Review D 53, no. 10 (1996): 5966–5969. DOI. Open PDF.