June 11, 2026
When matter fields are integrated out in a large \(N\) approximation, the conformal anomaly induces an effective action up to conformally invariant correction terms. In the present work, we consider the implications on black hole evaporation of such a term involving a conformal invariant found by Fefferman and Graham. Working in an approximation where the spacetime is static, we compute the induced stress tensor around a Schwarzschild black hole. The boundary conditions select an Unruh-like asymptotic sector, but not a unique stress tensor. A new one-parameter family of quantum hair emerges which changes the stress tensor in the near-horizon region. This suggests a new semiclassical mechanism by which information about black hole formation could be encoded outside the horizon.
The semiclassical description of black hole evaporation in four spacetime dimensions remains an important test of effective field theory in gravity. Hawking radiation is well understood as quantum field theory on a fixed black hole background [1], [2], but incorporating the corresponding back-reaction on the geometry in a controlled and covariant approximation is considerably more difficult. In two-dimensional models, anomaly-induced actions lead to tractable local formulations in which the formation and evaporation of black holes can be followed explicitly [3], [4]. In four dimensions, the analogous problem is more intricate, but the trace anomaly again provides a natural organizing principle for the leading large-\(N\) quantum effects of conformally coupled matter.
A useful starting point is Riegert’s non-local effective action [5], which reproduces the four-dimensional trace anomaly in a curved background. In earlier work, and in particular in the generalized effective field theory treatment of four-dimensional black hole evaporation, this action was augmented by the simplest conformally invariant non-local correction involving the square of the Weyl tensor [6]–[11], see also [12], [13] for related work. After localization by auxiliary scalar fields, the resulting model permitted an analytic computation of the induced stress tensor in a Schwarzschild background [14]. With physically motivated boundary conditions—regularity on the future horizon and vanishing incoming flux at past null infinity—the static solution was uniquely selected and had the expected qualitative properties of the Unruh state. However, the model also exhibited two important limitations. First, the Riegert contribution by itself gives the wrong sign for the outgoing energy flux for ordinary conformally coupled matter of spin \(\leq1\). The Weyl-squared conformally invariant correction introduced in the generalized model restored a positive luminosity. Second, within that truncation, the physical boundary conditions removed all integration constants in the stress tensor, leaving no analog of quantum hair.
The purpose of the present paper is to investigate the next natural step in this effective field theory expansion. The complete one-loop effective action is expected to contain, in addition to the anomaly-determined Riegert term, an infinite sequence of conformally invariant non-local terms. These terms do not modify the trace anomaly, but they can affect the semiclassical stress tensor and therefore the physics of black hole evaporation. Here we include a new conformally invariant contribution built from a weight-six invariant found by Fefferman and Graham [15]. This term provides a controlled extension of the previous four-dimensional effective field theory while preserving covariance and analytic tractability in the static Schwarzschild problem.
We localize the extended non-local action by introducing two additional auxiliary scalar fields, in addition to the two scalars appearing in the earlier generalized model [11]. The resulting four-scalar formulation yields higher-derivative but local equations of motion, which simplify substantially on the Schwarzschild background. We solve these equations for static, spherically symmetric configurations and compute the full anomaly-induced stress tensor. As in the earlier work, we impose boundary conditions appropriate to an Unruh-like state: absence of incoming radiation from past null infinity, finite outgoing luminosity at future null infinity, and regularity for freely falling observers crossing the future horizon.
The main result is that the Fefferman–Graham conformal correction leaves the asymptotic Hawking flux under analytic control while introducing a qualitatively new feature. The physical boundary conditions no longer fix all parameters in the induced stress tensor. Instead, a one-parameter family of regular solutions remains, with the free parameter affecting the stress tensor near the black hole while falling off rapidly at infinity. This parameter is not associated with the classical black hole mass and does not change the imposed asymptotic flux conditions. We therefore interpret it as a form of quantum hair generated by conformally invariant terms in the anomaly-induced effective action.
This quantum hair provides a new mechanism by which information about black hole formation may be encoded in semiclassical observables outside the horizon. In the static single-center setting considered here, one may choose a particular value of the hair parameter as part of the vacuum specification. In more general time-dependent or multi-centered geometries, however, the same parameter may carry physical information that is not visible in the anomaly alone. The present analysis therefore suggests that conformally invariant sectors of the one-loop effective action can play a more important role in four-dimensional evaporation than is apparent from the trace anomaly by itself.
The paper is organized as follows. In Section II we construct the anomaly-induced effective action supplemented by the Fefferman–Graham conformal invariant and present its local scalar-tensor form. In Section III we solve the auxiliary scalar equations on a Schwarzschild background. In Section IV we impose the physical boundary conditions and derive the resulting induced stress tensor and outgoing flux. In Section V we analyze the remaining free parameter and interpret it as quantum hair. We conclude in Section VI with a discussion of the implications for semiclassical black hole evaporation and future time-dependent studies.
Our starting point is the following general expression for the leading order trace anomaly of the stress tensor in curved 3+1-dimensional spacetime, for classically conformally coupled fields, \[g^{ab}\left\langle T_{ab}\right\rangle =\frac{1}{16\pi^{2}}\left(a'C^{2}+b'E-c'\nabla^{2}R+d'F^{2}\right)\,,\label{eq:traceanom}\tag{1}\] where \(a',b',c',d'\) are coefficients that depend on the matter content of the theory1 and \[\begin{align} C^{2} & =R^{abcd}R_{abcd}-2R^{ab}R_{ab}+\frac{1}{3}R^{2},\label{eq:c2e}\\ E & =R^{abcd}R_{abcd}-4R^{ab}R_{ab}+R^{2},\nonumber \end{align}\tag{2}\] are the square of the Weyl tensor and the Euler density, respectively. The constant \(d'\) is proportional to the beta function of the gauge theory. In the present work we do not include gauge fields so set \(d'=0\). The \(c'\) term may be removed by a local counterterm, however we retain it in the present analysis (though as we will see, it does not contribute to the induced stress tensor in a Schwarzschild background).
The anomaly coefficients, in terms of the number of matter species with spins \(\leq1\), are given by \[\begin{align} a' & =\frac{1}{120}\left(n_{s}+6n_{f}+12n_{V}\right),\nonumber \\ b' & =-\frac{1}{360}\left(n_{s}+11n_{f}+62n_{V}\right),\label{eq:abcd}\\ c' & =-\frac{1}{180}\left(n_{s}+6n_{f}+12n_{V}\right),\nonumber \end{align}\tag{3}\] where \(n_{s},n_{f}\) and \(n_{V}\) are the number of scalars, Dirac fermions and vectors respectively [17]. In order for the semiclassical approximation to be under control the non-vanishing coefficients are taken to be of order \(N\gg1\) (while scaling \(\hbar\sim1/N\)) so that the matter field contribution to the effective action is dominant compared to that of the metric sector.
The trace anomaly (with \(d'=0\)) is reproduced by a scalar-tensor theory with two auxiliary scalar fields \(\phi\) and \(\chi\) with conformal weight 0, as previously considered in [6]–[10] and studied in the context of static black holes in [11]. In the following we include two additional scalar fields \(\rho\) (weight -2) and \(\psi\) (weight 0) to include a new conformally invariant term involving a weight 6 conformal invariant found by Fefferman and Graham [15], and a new weight 6 derivative term acting on \(\psi\) (a closely related discussion in the literature can be found in [18])
\[\mathcal{K}[\psi]=\tfrac{1}{3}R\nabla_{a}\psi\nabla^{a}\psi+\tfrac{1}{2}\nabla_{a}\nabla^{a}\psi\nabla_{b}\nabla^{b}\psi-R_{ab}\nabla^{a}\psi\nabla^{b}\psi\label{eq:paneitz}\tag{4}\]
\[\begin{equation} F_{6}=8R_{a}{}^{c}R^{ab}R_{bc}-\tfrac{28}{3}R_{ab}R^{ab}R+\tfrac{4}{3}R^{3}+8R^{ab}R^{cd}R_{acbd}-8R^{ab}R^{cd}W_{acbd}+\tfrac{2}{3}RW_{abcd}W^{abcd}-\nabla_{a}R\nabla^{a}R-12\nabla_{b}R_{ac}\nabla^{c}R^{ab}+12\nabla_{c}R_{ab}\nabla^{c}R^{ab}+16W_{acbd}\nabla^{d}\nabla^{c}R^{ab}+\nabla_{e}W_{abcd}\nabla^{e}W^{abcd}\label{eq:fgterm} \end{equation}\tag{5}\]
\[\begin{equation} \Delta_{6}[\psi]=-\tfrac{1}{24}R\nabla_{a}\psi\nabla^{a}R+\tfrac{1}{4}R^{bc}\nabla_{a}R_{bc}\nabla^{a}\psi+\tfrac{1}{24}R_{ab}\nabla^{a}R\nabla^{b}\psi-\tfrac{1}{2}W_{a}{}^{cde}W_{bcde}\nabla^{b}\nabla^{a}\psi-\tfrac{1}{4}R^{bc}\nabla^{a}\psi\nabla_{c}R_{ab}+\tfrac{1}{2}W^{bcde}\nabla^{a}\psi\nabla_{c}W_{abde}+\tfrac{1}{2}R^{bc}\nabla^{a}\psi\nabla_{d}W_{abc}{}^{d}+\tfrac{3}{2}W_{abcd}\nabla^{a}\psi\nabla^{d}R^{bc}\label{eq:rachwal} \end{equation}\tag{6}\]
\[\begin{equation} S=\int d^{4}x\,(-g)^{1/2}\left[\frac{1}{16\pi}R+\frac{1}{192\pi^{2}}(c'-\tfrac{2}{3}b')R^{2}-b'\mathcal{K}[\phi]-\mathcal{K}[\chi]+f_{2\psi}\mathcal{K}[\psi]-\frac{\phi}{8\pi}\left((a'+b')C^{2}+\frac{2b'}{3}\left(R^{2}-3R_{ab}R^{ab}-\nabla^{2}R\right)\right)+f_{4\chi}C^{2}\chi+f_{4\psi}C^{2}\psi+\rho\,F_{6}-\frac{1}{2}\rho\Delta_{6}[\psi]\right].\label{eq:action} \end{equation}\tag{7}\] The terms involving the \(\chi,\rho\) and \(\psi\) fields are conformally invariant and do not contribute to the trace anomaly. The \(\phi\)-dependent Riegert term is fixed by the anomaly. The Weyl-squared nonlocal correction dependent on \(\chi\) is the lowest weight conformally invariant term considered previously. The new terms involving \(\psi\) and \(\rho\) represent the next higher conformal-weight correction at second order in the curvature.
Varying this action with respect to the metric leads to the Einstein equations sourced by additional terms which we refer to as the induced stress tensor. In the following we view the parameters \(a',b',c'\) as fixed by the matter content of the theory, while the new parameters \(f_{2\psi}\), \(f_{4\psi}\) and \(f_{4\chi}\) are effective field theory parameters. In principle, these might be fixed by truncating the full one-loop action, though this is extremely complex, even for conformally coupled scalars [19], [20].
The derivative operator 4 is related to the Paneitz operator (by integration by parts) [21] and is the unique conformal weight 4 operator that acts on weight 0 scalars. The scalar invariant 5 was found in [15] and has conformal weight 6. It is a conformally covariant improvement of the term \(|\nabla_{e}W_{abcd}|^{2}\) which is second order in the Riemann tensor. It is unique up to the addition of terms third order in the Weyl tensor \(W^{\,\,cd}_{ab}W^{\,\,ef}_{cd}W^{\,\,ab}_{ef}\) and \(W_{abcd}W^{ac}_{\,\,ef}W^{bedf}\). We choose not to include these higher order terms, and focus on the most general weight 6 terms second order in the Riemann tensor. The derivative operator 6 is the unique weight 6 derivative operator acting on weight 0 scalar fields, and is a generalization of similar operators studied in [18]. Interestingly, it does not contain terms of the form \(\square^{3}\psi\) so would vanish in flat spacetime. The existence of apparently an infinite sequence of such operators with conformal weights \(\geq4\) is key to the construction of this extension to the non-local Riegert action proposed in the present work and in [11].
As a simple check on these expressions, one can take the trace of the induced stress tensor derived from this action \[g^{ab}\left\langle T_{ab}\right\rangle =\frac{b'}{2\pi}\left(\nabla^{2}\nabla^{2}\phi-\tfrac{2}{3}R\nabla^{2}\phi+2R^{ab}\nabla_{a}\nabla_{b}\phi+\tfrac{1}{3}\nabla^{a}R\:\nabla_{a}\phi\right)-\frac{1}{16\pi^{2}}\left(c'-\frac{2}{3}b'\right)\nabla^{2}R\,,\label{eq:stresstrace}\tag{8}\] and eliminate the scalar field using its equation of motion. Completing this computation requires the use of some nontrivial Riemann tensor identities which were carried out using the xAct MATHEMATICA package [22], [23]. After some straightforward algebra one then recovers the trace anomaly formula 1 with \(d'=0\).
The scalar equations of motion derived from the action 7 can be explicitly solved on the Schwarzschild black hole background, \[ds^{2}=-\left(1-\frac{2M}{r}\right)dt^{2}+\frac{dr^{2}}{1-\frac{2M}{r}}+r^{2}d\Omega^{2}.\label{eq:schwarzschild}\tag{9}\] Since \(R_{ab}=R=0\) in this background, the scalar equations reduce to \[\begin{align} \frac{(a'+b')R_{abcd}R^{abcd}}{8\pi}+b'\nabla_{b}\nabla^{b}\nabla_{a}\nabla^{a}\phi & =0\nonumber \\ f_{4\chi}{}R_{abcd}R^{abcd}-\nabla_{b}\nabla^{b}\nabla_{a}\nabla^{a}\chi & =0\nonumber \\ \tfrac{1}{16}R_{bcde}R^{bcde}\nabla_{a}\nabla^{a}\psi-\tfrac{1}{4}R^{bcde}\nabla^{a}\psi\nabla_{c}R_{abde}+\nabla_{e}R_{abcd}\nabla^{e}R^{abcd} & =0\nonumber \\ f_{4\psi}{}R_{abcd}R^{abcd}+\tfrac{1}{16}R_{bcde}R^{bcde}\nabla_{a}\nabla^{a}\rho+f_{2\psi}{}\nabla_{b}\nabla^{b}\nabla_{a}\nabla^{a}\psi-\tfrac{1}{4}R^{bcde}\nabla^{a}\rho\nabla_{c}R_{abde} & =0\,. \end{align}\] The general static spherically symmetric solution for each scalar field, \(\phi\) and \(\chi\), can be written as a linear combination of four independent solutions to the corresponding homogenous problem plus special solutions, \(\phi_{P}\) and \(\chi_{P}\), that satisfy the respective inhomogeneous equations, \[\begin{align} \phi & =a_{\phi1}\phi_{1}+a_{\phi2}\phi_{2}+a_{\phi3}\phi_{3}+a_{\phi4}+\phi_{P}\,,\nonumber \\ \chi & =a_{\chi1}\phi_{1}+a_{\chi2}\phi_{2}+a_{\chi3}\phi_{3}+a_{\chi4}+\chi_{P}\,,\label{eq:gensolution} \end{align}\tag{10}\] where \[\begin{align} \phi_{1} & =\log\left(1-\frac{2M}{r}\right)\,,\nonumber \\ \phi_{2} & =r^{2}+4Mr+8M^{2}\log r\,,\nonumber \\ \phi_{3} & =-\mathrm{Li}_{2}\left(\frac{2M}{r}\right)+\frac{r}{4M}+\frac{3}{2}\log r-\frac{1}{16M^{2}}\left(r^{2}+4Mr-8M^{2}\log r\right)\log\left(1-\frac{2M}{r}\right)\,,\nonumber \\ \phi_{P} & =\frac{(a'+b')}{8\pi b'}\Psi(r)\,,\nonumber \\ \chi_{P} & =-f_{4\chi}\,\Psi(r)\,,\nonumber \\ \Psi(r) & =\frac{2}{3M}r+2\log r+\frac{1}{12M^{2}}\left(r^{2}+4Mr+8M^{2}\log r\right)\log\left(1-\frac{2M}{r}\right)\,.\label{eq:solutions} \end{align}\tag{11}\]
The two new scalars satisfy second order equations with static solutions\[\begin{equation} \psi=a_{\psi2}{}+\tfrac{1}{30}(a_{\psi1}{}-\frac{1}{4M^{5}})r(480M^{4}+120M^{3}r+40M^{2}r^{2}+15Mr^{3}+6r^{4})+40\log(r)+32a_{\psi1}{}M^{5}\log(-2M+r)\,,\label{eq:newsol} \end{equation}\tag{12}\] and\[\[ \rho=a_{\rho2}{}-\frac{f_{4\psi}{}r^{3}(40M^{2}+15Mr+6r^{2})}{45M^{3}}+\tfrac{1}{30}a_{\rho1}{}r(480M^{4}+120M^{3}r+40M^{2}r^{2}+15Mr^{3}+6r^{4})+\frac{f_{2\psi}{}r^{4}\bigl(r^{5}+M^{5}(48-4a_{\psi1}{}r^{5})\bigr)}{6M^{7}}+32a_{\rho1}{}M^{5}\log(-2M+r)\,. \]\]The twelve constants \(a_{Xi}\) (with \(X=\phi,\chi,\rho,\psi)\) are to be fixed using physical boundary conditions.
Our goal is to evaluate the semiclassical stress tensor subject to suitable boundary conditions applied to the asymptotic form of the stress tensor as \(r\to2M\) and \(r\to\infty\) following the same logic as [11] and references therein.
We consider scalar fields of the form \[\begin{align} \phi(t,r) & =d_{\phi}\,t+\phi(r)\,,\label{eq:scalarsol}\\ \chi(t,r) & =d_{\chi}\,t+\chi(r)\,,\nonumber \\ \psi(t,r) & =d_{\psi}t+\psi(r)\,. \end{align}\tag{13}\] where \(\phi(r),\chi(r)\) and \(\psi(r)\) are static solutions of the form 10 and 12 and \(d_{\phi},d_{\chi}\) and \(d_{\psi}\) are additional free parameters to be fixed by boundary conditions. This simple ansatz (considered originally in [9]) gives a time-independent stress tensor. However the modification of the stress tensor ends up depending on only 2 combinations of these three parameters \[\begin{align} \bar{d_{1}} & =d_{\psi}{}^{2}f_{2\psi}{}-{b'}d_{\phi}{}^{2}-d_{\chi}{}^{2}\,,\\ \bar{d}_{2} & =({a'}+{b'})d_{\phi}{}+2(3{b'}a_{\phi3}{}d_{\phi}{}+3a_{\chi3}{}d_{\chi}{}-4f_{4\chi}{}d_{\chi}{}-8d_{\psi}{}f_{4\psi}{}+12a_{\rho1}{}d_{\psi}{}M^{3})\pi\,. \end{align}\] We then choose to solve for \(\bar{d}_{1}\) and \(\bar{d}_{2}\) to provide a non-redundant solution for the stress tensor. We require the combinations that enter the stress tensor to be real. Since the auxiliary fields are not themselves observables, we do not separately impose a reality condition on the constants \(d_{\phi},d_{\chi}\) and \(d_{\psi}\). For non-vanishing \(\bar{d}_{2}\) the solution breaks time-translation invariance and gives rise to a non-vanishing \(T_{rt}\) component.
The full stress tensor is given by rather lengthy expressions that we do not write out explicitly here (our MATHEMATICA expression contains 656 terms). The stress tensor is independent of \(a_{\phi4}\), \(a_{\chi4}\) and \(a_{\psi2}\) so we begin by setting these parameters to zero. In order that \(T^{\theta}_{\theta}\) not grow at large \(r\) we require \(a_{\psi1}=\frac{1}{4M^{5}}\). As we will see later, \(a_{\rho2}\) will be undetermined by our physical constraints, and represents a new quantum hair parameter. We will determine the impact of this in the next section. For now, we focus on the remaining parameters and set \(a_{\rho2}=0\). Apart from the appearance of the new parameter \(a_{\rho1}\), the determination of the remaining constants follows the same procedure as in the two-scalar model [11].
We are left with nine independent parameters to be determined. Some of these parameters are fixed by requiring a freely falling observer crossing the future horizon see a finite energy density. Some of the same conditions come from requiring the \(T_{\theta}{}^{\theta}\) component of the stress tensor to be finite at the future horizon, which amounts to a simpler calculation. Near the horizon, \[T^{\,\,\theta}_{\theta}=\frac{A_{1}}{(r-2M)^{2}}+\frac{2A_{1}}{M(r-2M)}+A_{2}\log^{2}\left(\frac{r}{2M}-1\right)+A_{3}\log\left(\frac{r}{2M}-1\right)+\cdots\,,\label{eq:tthetatheta}\tag{14}\] while near infinity, \[T^{\,\theta}_{\theta}=B_{1}+\frac{B_{2}}{r}+\frac{B_{3}}{r^{2}}+\frac{B_{4}}{r^{3}}+\cdots\,.\label{eq:tthth}\tag{15}\] where the \(A_{i}\) and \(B_{i}\) are quadratic functions of the \(a_{Xi}\) and \(\bar{d}_{1}\) and \(\bar{d}_{2}\). As in the two-scalar model there are 2 branches that produce the same stress tensor and are a candidate for the Unruh-like vacuum state. These 2 branches appear when \(a_{\rho1}=0\). In addition there appear to be 6 other branches arising from solving the set of simultaneous quadratic equations which have \(a_{\rho1}\neq0\). However these branches do not seem to be continuously connected to the branches found in the two-scalar model. In appendix 7 we present the reduced constraint equations for general \(a_{\rho1}\neq0\) so their solution is determined in terms of the remaining unknowns (which seems to be the most compact way of presenting the solution, and may readily be solved numerically). Depending on the range of the effective field theory parameters \(f_{2\psi}\) and \(f_{4\psi}\), the new branches give unphysical complex solutions for the \(a_{Xi}\). It is possible these extra branches are in general unstable, however proving that is beyond the scope of the present paper. For now we focus on the branches with \(a_{\rho1}=0\).
The conditions \(B_{1}=B_{2}=B_{3}=0\) are solved by \(a_{\phi2}=a_{\chi2}=0\). With \(a_{\phi2}\neq0\) and \(a_{\chi2}\neq0\) there are additional \(\frac{\log r}{r^{3}}\) terms. Finiteness near the horizon requires setting \(A_{1}=A_{2}=A_{3}=0\). The \(A_{2}\) coefficient involves a sum of two squares and demanding \(A_{2}=0\) on the space of real parameters leads to two conditions, \[a_{\phi3}=\frac{a'+b'}{6\pi b'}\,,\quad a_{\chi3}=-\frac{4f_{4\chi}}{3}\,.\label{eq:c3d3}\tag{16}\]
Let us define the null geodesic vectors \(n^{\mu}_{\pm}=\left(\frac{1}{1-\frac{2M}{r}},\pm1,0,0\right)\) in \((t,r,\theta,\phi)\) coordinates. The ingoing flux at past null infinity \(\mathscr{I}^{-}\) is \[\frac{1}{4}n^{\mu}_{+}T_{\mu\nu}n^{\nu}_{+}=\frac{C_{3}}{r^{2}}+\cdots\,.\label{eq:nullcontract}\tag{17}\] To find the analog of the Unruh state, with vanishing ingoing flux we must set \(C_{3}=0\). Our next requirement is \(B_{4}=0\) which ensures a faster than \(r^{-3}\) falloff of the angular components of the stress tensor, as is expected of the Unruh vacuum [24]. This fixes \[\begin{align} \bar{d}_{1} & =\frac{{a'}+{b'}-384(f_{4\psi}{}+40f_{2\psi}{})\pi^{2}}{32M^{2}\pi^{2}}\label{eq:dpvalue}\\ \bar{d}_{2} & =-\frac{({a'}+{b'})^{2}}{4{b'}M\pi}-\frac{16f_{4\chi}{}^{2}\pi}{M}\,. \end{align}\tag{18}\]
With these constraints, we return to the near-horizon limit. Regularity for a freely falling observer requires that \[\begin{align} n^{\mu}_{-}T_{\mu\nu}n^{\nu}_{-} & =\frac{A_{4}}{(r-2M)^{3}}+\frac{A_{5}}{(r-2M)^{2}}+\frac{A_{6}}{r-2M}+\frac{A_{7}}{r-2M}\log\left(r-2M\right)+\\ & A_{8}\log^{2}\left(r-2M\right)+A_{9}\log\left(r-2M\right)+\cdots\,,\nonumber \end{align}\] be finite on the horizon, where again the coefficients \(A_{4},\ldots,A_{9}\) are quadratic functions of \(a_{Xi}\).
Finally we solve the remaining near horizon conditions, \(A_{1}=A_{5}=0,\) to fix \(a_{\phi1}\) and \(a_{\chi1}\). This leads to doubling, with the solutions given by\[\[ a^{\pm}_{\phi1}=\frac{64(a'+b')f_{4\psi}\pi\pm2Z^{1/2}/{b'}^{2}}{(a'+b')^{2}+64b'f_{4\chi}{}^{2}\pi^{2}}-\frac{(a'+b')\bigl(3+2\log(2M)\bigr)}{12b'\pi} \]\]with\[\[ Z={b'}^{3}f_{4\chi}{}^{2}\Bigl(-({a'}+{b'})^{3}({a'}+3{b'})-128{b'}({a'}+{b'})\bigl({a'}f_{4\chi}{}^{2}+2{b'}f_{4\chi}{}^{2}-10{a'}f_{4\psi}{}-10{b'}f_{4\psi}{}-248({a'}+{b'})f_{2\psi}{}\bigr)\pi^{2}-4096{b'}^{2}\bigl(f_{4\chi}{}^{4}+16f_{4\psi}{}^{2}-4f_{4\chi}{}^{2}(5f_{4\psi}{}+124f_{2\psi}{})\bigr)\pi^{4}\Bigr) \]\]and \[\[ a^{\pm}_{\chi1}=f_{4\chi}{}\left(\tfrac{4}{3}\log(2M)+2\right)-\frac{512{b'}f_{4\chi}{}f_{4\psi}{}\pi^{2}\mp\left(({a'}+{b'})Z^{1/2}\right)/\left(4{b'}^{2}f_{4\chi}{}\pi\right)}{({a'}+{b'})^{2}+64{b'}f_{4\chi}{}^{2}\pi^{2}}\,. \]\] When the above solutions for \(a_{Xi}\) are inserted, all the coefficients \(A_{i},B_{i}\) and \(C_{i}\) vanish without imposing any additional conditions.
These two different branches give the same quantum induced stress-tensor. This indicates a degeneracy between the couplings of certain modes of the scalars \(\phi\) and \(\chi\) for the special case of the Schwarzschild background. In particular, they yield a prediction for the outgoing null flux at future null infinity \(\mathscr{I}^{+}\) , \[\frac{1}{4}n^{\mu}_{-}T_{\mu\nu}n^{\nu}_{-}=\frac{2f^{2}_{4\chi}}{M^{2}r^{2}}+\frac{\left(a'+b'\right)^{2}}{32\pi^{2}M^{2}b'\,r^{2}}+\cdots\,,\label{eq:outflux}\tag{19}\] corresponding to an object with finite outgoing luminosity. Recall the parameters \(a'\),\(b'\) are known and depend on the matter content of the theory while \(f_{4\chi}\) is an effective field theory parameter which we treat as a free parameter, but in principle is computable assuming a truncation of the complete effective action of [20].
The final result respects the asymptotic conditions described in table 1 of [24], assuming conformally coupled scalar matter.
| \(r\to\infty\) | \(r\to2M\) | |||
|---|---|---|---|---|
| \(T^{r}_{\,r}\) | \(T^{\theta}_{\,\theta}\) | \(T^{r}_{\,r}\) | \(T^{\theta}_{\,\theta}\) | |
| CF | \(\frac{1}{r^{2}}\) | \(\frac{1}{r^{4}}\) | \(-(1-\frac{2M}{r})^{-1}\) | \(-1\) |
| LL | \(\frac{1}{r^{2}}\) | \(\frac{1}{r^{4}}-\frac{\log r}{r^{4}}\) | \(-(1-\frac{2M}{r})^{-1}\) | \(-1\) |
In particular, with \(f_{4\chi}\) suitably chosen, the outgoing flux at infinity is positive, in line with Hawking’s prediction. Likewise, one has vanishing ingoing flux at past null infinity. However, the \(r^{-4}\log r\) behavior at large \(r\) indicates the four-scalar model does not give a perfect match to the expected result, as already noted in studies of the two-scalar model [11].
We interpret the scalars \(\phi,\rho,\chi\) and \(\psi\) as auxiliary fields, so the only observables associated with them are derived from the metric itself. A novel feature of the present construction is that physical asymptotic conditions on the metric and the induced stress tensor at infinity and on the future horizon do not fix the parameter \(a_{\rho2}\), other than requiring finiteness. Thus this parameter can be interpreted as a new form of quantum hair. Note in earlier discussions of the one-scalar and two-scalar models there were no analogous parameters, and the asymptotic conditions completely fixed the vacuum state. The observable hair is not the value of the auxiliary scalar itself, but the induced stress tensor deformation that remains after all asymptotic and horizon regularity conditions have been imposed. The stress tensor contribution due to \(a_{\rho2}\) takes the diagonal form \[\begin{align} \delta T_{tt} & =-\frac{96M(2M-r)(91M^{2}-85Mr+20r^{2})}{r^{10}}a_{\rho2}\\ \delta T_{rr} & =\frac{96M(21M^{2}-20Mr+5r^{2})}{(2M-r)r^{8}}a_{\rho2}\\ \delta T_{\theta\theta} & =\frac{48M(112M^{2}-105Mr+25r^{2})}{r^{7}}a_{\rho2}\,, \end{align}\] which falls off like a rapid power law near infinity, and is also smooth on the future horizon \[n^{\mu}_{-}\delta T_{\mu\nu}n^{\nu}_{-}=\frac{480M(-7M+3r)}{r^{8}}a_{\rho2}\,.\] Thus the addition of the weight-six conformally invariant term in the effective action provides an explicit semiclassical mechanism for evading the usual no-hair theorems [25]. For a single black hole, one might argue \(a_{\rho2}\) is simply fixed by demanding vacuum boundary conditions on past null infinity \(\mathcal{I}^{-}\). However the parameter only influences the induced stress energy in the vicinity of the black hole, so one could imagine a multi-centered black hole solution where the analog of \(a_{\rho2}\) approached different constant values near the different horizons. Thus we argue that \(a_{\rho2}\) represents a new type of quantum hair that can encode information about the formation of the black hole. Moreover, within the present static effective theory, the boundary conditions do not fix its magnitude, and there is no apparent parametric suppression that would force it to be negligible. This opens the possibility that conformally invariant sectors of the one-loop effective action can leave observable imprints in near-horizon black hole physics. We plan to study this phenomenon further in the context of time-dependent solutions in the future.
In this paper we have extended the generalized effective field theory description of four-dimensional black hole evaporation by adding a new conformally invariant term to the anomaly-induced effective action. The term is built from the Fefferman–Graham weight-six conformal invariant and represents a further contribution beyond the anomaly-determined Riegert action and the Weyl-squared correction studied previously [11]. Since conformally invariant terms do not affect the trace anomaly, their coefficients are not fixed by the anomaly coefficients. Nevertheless, they contribute nontrivially to the semiclassical stress tensor and therefore to the physical evaporation problem.
After localizing the extended non-local action with auxiliary scalar fields, we obtained a four-scalar model whose equations can be solved analytically in a Schwarzschild background. Imposing the usual Unruh-type boundary conditions—regularity on the future horizon, and vanishing incoming flux at past null infinity—again selects a controlled semiclassical state. The outgoing luminosity contains the anomaly-induced contribution together with a positive contribution governed by the coefficient of the conformally invariant Weyl-squared sector. Thus, as in the preceding generalized effective field theory model, the sign problem of the minimal Riegert action for ordinary spin \(\leq1\) conformal matter can be avoided without invoking exotic matter species.
The new feature of the present construction is that the physical boundary conditions do not uniquely fix the full induced stress tensor. A single parameter remains, producing a finite, diagonal correction to the stress tensor that is smooth on the future horizon and decays rapidly at large radius. This parameter does not alter the imposed asymptotic radiation conditions and is not determined by the black hole mass. It therefore represents a genuine form of quantum hair in the semiclassical effective theory. Its origin lies in the conformally invariant part of the one-loop effective action, rather than in the trace anomaly itself.
This result sharpens the lesson of the earlier work. The minimal anomaly-induced action, and the first conformally invariant extension, led to a unique Unruh-like stress tensor once the physical boundary conditions were imposed. The inclusion of the Fefferman–Graham invariant shows that this uniqueness is not generic in the full effective field theory. Conformally invariant terms that are invisible to the trace anomaly can leave finite, observable imprints near the horizon. In this sense, the trace anomaly fixes only part of the semiclassical physics; the remaining conformally invariant sector can carry additional state-dependent information.
The quantum hair found here is especially suggestive for the black hole information problem. In the static Schwarzschild calculation, the hair parameter may simply be chosen as part of the specification of the state. In a dynamical collapse geometry, however, its value could be determined by the history of the formation process. In multi-centered or time-dependent situations, different near-horizon regions may support different values of the analogous parameter while preserving the same asymptotic flux conditions. This raises the possibility that information about the initial state is encoded in semiclassical stress-energy outside the horizon through conformally invariant terms in the effective action. Since this stress energy is generated by quantum effects and is localized in a region that is potentially observable, the corresponding hair parameter may provide a channel through which near-horizon quantum gravity effects leave observable imprints on astrophysical black holes.
Several limitations of the present analysis should be emphasized. We have restricted attention to a static Schwarzschild background and have not included the back-reaction of the induced stress tensor on the geometry. The resulting Unruh-like state has a constant outgoing flux and therefore cannot by itself describe a finite-mass evaporating black hole for all time. Thus the present calculation should be viewed as a diagnostic of the local and asymptotic structure of the semiclassical stress tensor, not as a complete evaporation solution. In addition, the model studied here is only a finite truncation of the full conformally invariant part of the one-loop effective action.
The next step is to study the corresponding time-dependent semiclassical equations with back-reaction included. In the one-scalar model, this has already been accomplished in [26] and without back-reaction in [27]. Such solutions would determine whether the quantum hair identified in the static analysis is dynamically generated during collapse, whether it remains stable during evaporation, and whether it affects the late-time outgoing radiation. More generally, the results of this paper indicate that a complete effective field theory treatment of four-dimensional black hole evaporation must include not only the anomaly-determined Riegert action but also the conformally invariant sector of the non-local effective action. That sector may provide the additional semiclassical structure needed to understand how information is stored and released in four-dimensional black hole evaporation.
We thank J. Hudson and L. Thorlacius for helpful discussions.
Solving the general constraint equations imposes the conditions \[a_{\psi1}=\frac{1}{4M^{5}},\;a_{\psi2}=0\,.\] The interpretation of \(a_{\rho2}\) is detailed in section 5. The solutions then further break up into a set of solutions with \(a_{\phi2}=0,\:a_{\chi2}=0\) and another special branch with \(\;a_{\phi2}\neq0,\:a_{\chi2}\neq0\,,\) but with \(\bar{d}_{2}=0\). We present the reduced constraint equations for these two main cases which is the most compact way of presenting the solutions. The equations are arranged into the maximal set of linear equations together with the remaining irreducible quadratic constraints.
There remain 7 independent constraints for the 7 remaining variables \(a_{\phi1},a_{\chi1},a_{\phi3},\) \(a_{\chi3},\bar{d_{1}},\bar{d}_{2}\) and \(a_{\rho1}\). As described above, there are 4 branches to these solutions. Two have been given explicitly above, and are continuously connected to the branches of the two-scalar setup [11]. The remaining 2 branches lead to nested quadratic equations that generate rather unwieldy expressions that we do not present in this paper, and generate values for the Hawking radiation distinct from 19 . For completeness we present these 7 reduced constraints, which we have arranged into 4 linear constraints and 3 irreducible quadratic constraints: \[\[ 3({a'}+{b'})a_{\phi3}{}+2\bar{d}_{2}{}M-24a_{\chi3}{}f_{4\chi}{}\pi=0\,, \]\]\[\[ 55({a'}+{b'})^{2}-330{b'}({a'}+{b'})a_{\phi3}{}\pi+16{b'}\bigl(55f_{4\chi}{}(3a_{\chi3}{}+4f_{4\chi}{})-3618a_{\rho1}{}M^{3}\bigr)\pi^{2}=0\,, \]\]\[\[ -4({a'}+{b'})(7{a'}+10{b'})+24{b'}\bigl(12({a'}+{b'})a_{\phi1}{}+(7{a'}+10{b'})a_{\phi3}{}-12\bar{d}_{2}{}M\bigr)\pi-\tfrac{64}{5}{b'}\bigl(5f_{4\chi}{}(36a_{\chi1}{}+21a_{\chi3}{}+28f_{4\chi}{})+1440f_{4\psi}{}+3312a_{\rho1}{}M^{3}\bigr)\pi^{2}+24\bigl(({a'}+{b'})^{2}+6{b'}({a'}+{b'})a_{\phi3}{}\pi+16{b'}f_{4\chi}{}(-3a_{\chi3}{}+4f_{4\chi}{})\pi^{2}\bigr)\log(2M)=0\,, \]\]\[\[ -9{a'}+576\bigl(12f_{4\psi}{}+480f_{2\psi}{}+M^{2}(\bar{d}_{1}{}-18a_{\rho1}{}M)\bigr)\pi^{2}-9{b'}(1+6a_{\phi3}{}\pi)=0\,, \]\]\[\[ ({a'}+{b'})^{2}+4{b'}({a'}+{b'})(4a_{\phi1}{}-3a_{\phi3}{})\pi+4{b'}\Bigl({b'}(4a_{\phi1}{}-3a_{\phi3}{})^{2}+(-4a_{\chi1}{}+3a_{\chi3}{}+4f_{4\chi}{})^{2}+64\bigl(-16f_{2\psi}{}+M^{2}(\bar{d}_{1}{}-48a_{\rho1}{}M)\bigr)\Bigr)\pi^{2}\biggr)+4\log(2M)\biggl(3\Bigl(({a'}+{b'})^{2}+8{b'}({a'}+{b'})a_{\phi1}{}\pi+4{b'}\bigl(3{b'}(4a_{\phi1}{}-3a_{\phi3}{})a_{\phi3}{}+(4a_{\chi1}{}-3a_{\chi3}{}-4f_{4\chi}{})(3a_{\chi3}{}-4f_{4\chi}{})\bigr)\pi^{2}\Bigr)+\Bigl(({a'}+{b'})^{2}+12{b'}({a'}+{b'})a_{\phi3}{}\pi+4{b'}\bigl(9{b'}a_{\phi3}{}^{2}+(3a_{\chi3}{}-4f_{4\chi}{})^{2}\bigr)\pi^{2}\Bigr)\log(2M)=0\,, \]\]\[\[ -11({a'}+{b'})^{2}+12{b'}({a'}+{b'})(5a_{\phi1}{}+a_{\phi3}{})\pi+4{b'}\bigl(9{b'}a_{\phi3}{}(-10a_{\phi1}{}+9a_{\phi3}{})+(-30a_{\chi1}{}+27a_{\chi3}{}-44f_{4\chi}{})(3a_{\chi3}{}+4f_{4\chi}{})+1728a_{\rho1}{}M^{3}\bigr)\pi^{2}+5\bigl(({a'}+{b'})^{2}-4{b'}(9{b'}a_{\phi3}{}^{2}+9a_{\chi3}{}^{2}-16f_{4\chi}{}^{2})\pi^{2}\bigr)\log(2M)=0\,, \]\]\[\[ {a'}^{2}+2{a'}{b'}(1-6a_{\phi3}{}\pi)+{b'}\bigl(4(3a_{\chi3}{}+4f_{4\chi}{})^{2}\pi^{2}+{b'}(1-6a_{\phi3}{}\pi)^{2}\bigr)=0\,. \]\]The outgoing flux on this branch is given by 19 together with the two additional solutions\[\[ \frac{1}{4}n^{\mu}_{-}T_{\mu\nu}n^{\nu}_{-}=-\frac{1}{8{b'}M^{2}\pi^{2}\bigl((103460{a'}+75119{b'})^{2}+685054182400{b'}f_{4\chi}{}^{2}\pi^{2}\bigr)r^{2}}\left((1300{a'}-51161{b'})({a'}+{b'})^{2}(103460{a'}+75119{b'})+64{b'}\Bigl(413840{a'}^{2}\bigl(650f_{4\chi}{}^{2}+603(81f_{4\psi}{}+8308f_{2\psi}{})\bigr)+4{a'}{b'}\bigl(-1231616590f_{4\chi}{}^{2}+107683137(81f_{4\psi}{}+8308f_{2\psi}{})\bigr)+{b'}^{2}\bigl(-5998674641f_{4\chi}{}^{2}+181187028(81f_{4\psi}{}+8308f_{2\psi}{})\bigr)\Bigr)\pi^{2}+1695088640{b'}^{2}f_{4\chi}{}^{2}\bigl(325f_{4\chi}{}^{2}+603(81f_{4\psi}{}+8308f_{2\psi}{})\bigr)\pi^{4}\pm4081104(-{b'})^{3/2}\pi|f_{4\chi}{}|\left|5820{a'}^{2}+7151{a'}{b'}+{b'}\bigl(1331{b'}+128(2910f_{4\chi}{}^{2}+5427f_{4\psi}{}+556636f_{2\psi}{})\pi^{2}\bigr)\right|\right)\,. \]\]As one can see from this expression, while analytic solutions of these equations are relatively straightforward, they quickly lead to very long formulas.
The remaining branch has \(\bar{d}_{2}=0\) which implies it has no outgoing Hawking radiation. There remain 8 independent constraints for the 8 remaining variables \(a_{\phi1},a_{\chi1},a_{\phi2},a_{\chi2},a_{\phi3},\) \(a_{\chi3},\bar{d_{1}}\) and \(a_{\rho1}\). This is a regular zero-flux state, analogous in its asymptotic flux properties to the Boulware state [28] , that is allowed by the truncated effective action. The reduced constraint equations are 3 linear constraints and 5 irreducible quadratic constraints \[({a'}+{b'})a_{\phi3}{}-8a_{\chi3}{}f_{4\chi}{}\pi=0\,,\] \[\[ 55({a'}+{b'})^{2}-66{b'}({a'}+{b'})(5a_{\phi3}{}-24a_{\phi2}{}M^{2})\pi+16{b'}\bigl(55f_{4\chi}{}(3a_{\chi3}{}+4f_{4\chi}{})-792a_{\chi2}{}f_{4\chi}{}M^{2}-3618a_{\rho1}{}M^{3}\bigr)\pi^{2}=0\,, \]\]\[\[ -5({a'}+{b'})(7{a'}+10{b'})+6{b'}\bigl(60({a'}+{b'})a_{\phi1}{}+5(7{a'}+10{b'})a_{\phi3}{}-384({a'}+{b'})a_{\phi2}{}M^{2}\bigr)\pi-16{b'}\bigl(5f_{4\chi}{}(36a_{\chi1}{}+21a_{\chi3}{}+28f_{4\chi}{})+1440f_{4\psi}{}-1152a_{\chi2}{}f_{4\chi}{}M^{2}+3312a_{\rho1}{}M^{3}\bigr)\pi^{2}+30\bigl(({a'}+{b'})^{2}+6{b'}({a'}+{b'})a_{\phi3}{}\pi+16{b'}f_{4\chi}{}(-3a_{\chi3}{}+4f_{4\chi}{})\pi^{2}\bigr)\log(2M)=0\,, \]\]\[\[ ({a'}+{b'})^{2}+4{b'}({a'}+{b'})(4a_{\phi1}{}-3a_{\phi3}{})\pi+4{b'}\Bigl({b'}(4a_{\phi1}{}-3a_{\phi3}{})^{2}+(-4a_{\chi1}{}+3a_{\chi3}{}+4f_{4\chi}{})^{2}+64\bigl(-16f_{2\psi}{}+M^{2}(\bar{d}_{1}{}-48a_{\rho1}{}M)\bigr)\Bigr)\pi^{2}\biggr)+4\log(2M)\biggl(3\Bigl(({a'}+{b'})^{2}+8{b'}({a'}+{b'})a_{\phi1}{}\pi+4{b'}\bigl(3{b'}(4a_{\phi1}{}-3a_{\phi3}{})a_{\phi3}{}+(4a_{\chi1}{}-3a_{\chi3}{}-4f_{4\chi}{})(3a_{\chi3}{}-4f_{4\chi}{})\bigr)\pi^{2}\Bigr)+\Bigl(({a'}+{b'})^{2}+12{b'}({a'}+{b'})a_{\phi3}{}\pi+4{b'}\bigl(9{b'}a_{\phi3}{}^{2}+(3a_{\chi3}{}-4f_{4\chi}{})^{2}\bigr)\pi^{2}\Bigr)\log(2M)=0\,, \]\]\[\[ -11({a'}+{b'})^{2}+12{b'}({a'}+{b'})(5a_{\phi1}{}+a_{\phi3}{}-24a_{\phi2}{}M^{2})\pi+4{b'}\bigl((-30a_{\chi1}{}+27a_{\chi3}{}-44f_{4\chi}{})(3a_{\chi3}{}+4f_{4\chi}{})+144M^{2}(3a_{\chi2}{}a_{\chi3}{}+4a_{\chi2}{}f_{4\chi}{}+12a_{\rho1}{}M)+9{b'}a_{\phi3}{}(-10a_{\phi1}{}+9a_{\phi3}{}+48a_{\phi2}{}M^{2})\bigr)\pi^{2}+5\bigl(({a'}+{b'})^{2}-4{b'}(9{b'}a_{\phi3}{}^{2}+9a_{\chi3}{}^{2}-16f_{4\chi}{}^{2})\pi^{2}\bigr)\log(2M)=0\,, \]\]\[\[ {a'}^{2}+2{a'}{b'}(1-6a_{\phi3}{}\pi)+{b'}\bigl(4(3a_{\chi3}{}+4f_{4\chi}{})^{2}\pi^{2}+{b'}(1-6a_{\phi3}{}\pi)^{2}\bigr)=0\,, \]\] \[{b'}a_{\phi2}{}^{2}+a_{\chi2}{}^{2}=0\,,\] \[\[ {a'}a_{\phi2}{}+2a_{\chi2}{}(3a_{\chi3}{}-4f_{4\chi}{}+8a_{\chi2}{}M^{2})\pi+{b'}a_{\phi2}{}(1+6a_{\phi3}{}\pi+16a_{\phi2}{}M^{2}\pi)=0\,. \]\]