Introduction
Understanding quantum gravity remains one of the central challenges in modern theoretical physics. Over the past decades, various approaches have been developed to address this problem, including string theory [1 , 2 , 3 ] , loop quantum gravity [4 , 5 , 6 , 7 , 8 ] , asymptotic safety [9 , 10 , 11 ] , causal dynamical triangulations [12 , 13 ] . Despite considerable progress, a complete theory of quantum gravity remains elusive. Complementary to these fundamental approaches, a pragmatic strategy is to construct an effective description of quantum gravity [14 , 15 , 16 , 17 , 18 ] . In such descriptions, spacetime is still described by a metric, while quantum gravity effects are incorporated through modifications of the classical gravitational dynamics. Effective models can, in principle, be understood as semiclassical approximations to underlying quantum gravity theories. A notable example is loop quantum cosmology, where the effective dynamics derived from the underlying quantum theory accurately describes the evolution of sharply peaked semiclassical states [19 , 20 , 21 , 22 , 23 , 24 , 25 , 26 , 27 , 28 ] .
Beyond cosmology, effective approaches have also been extensively employed to investigate quantum corrections to black-hole spacetimes [29 , 30 , 31 , 32 , 33 , 34 , 35 , 36 , 37 , 38 , 39 , 40 , 41 , 42 , 43 , 44 , 45 , 46 , 47 , 48 , 49 , 50 ] . However, unlike in loop quantum cosmology, a systematic derivation of effective dynamics from an underlying quantum theory remains unavailable for most existing black-hole models, despite some progress in this direction [51 , 52 , 53 ] . In the absence of such a derivation, it is therefore important to identify fundamental principles that can guide and constrain the construction of consistent effective dynamics. In our previous works [46 , 47 , 48 , 49 ] , we addressed this issue by developing a systematic framework, which we refer to as covariant Hamiltonian gravity in spherical symmetry, for constructing spherically symmetric effective gravitational dynamics based on the principle of general covariance. Indeed, general covariance requires particular care in the Hamiltonian formulation, where spacetime is foliated into a family of spatial hypersurfaces and general covariance is no longer manifest [54 , 55 , 56 , 57 , 58 , 59 ] . Requiring general covariance in this formulation therefore imposes nontrivial restrictions on the admissible forms of the Hamiltonian constraint.
The Hamiltonian formulation, however, presents certain practical limitations when studying matter coupling. In general relativity, matter influences the spacetime metric only through its energy-momentum tensor, so different matter models with the same energy-momentum tensor act as identical sources of gravity. In the Hamiltonian formulation, matter enters through its Hamiltonian and diffeomorphism constraints, whose explicit forms generally depend on the details of the matter fields. A formulation of effective gravitational dynamics directly in terms of the spacetime metric and the energy-momentum tensor would be highly desirable.
Another limitation of the standard Hamiltonian formulation is that it relies on a foliation of spacetime by spacelike hypersurfaces, making it unsuitable for directly working with null foliations. This is particularly inconvenient when studying dynamical spacetimes such as the Vaidya geometry, as well as physical processes such as Hawking radiation, for which null coordinates are particularly useful.
Both limitations can be addressed by constructing an effective Einstein tensor. In the present work, we construct such a tensor from the covariant Hamiltonian gravity in spherical symmetry, allowing the field equations of gravity coupled to matter to be expressed directly in terms of the spacetime metric and the energy-momentum tensor. Moreover, since the resulting equations are tensorial, they can be transformed into arbitrary coordinate systems, including null coordinates, without requiring a null Hamiltonian formulation.
As an application of the effective Einstein equation, we construct the generalized Vaidya solutions of the effective theory. We find that the corresponding energy-momentum tensor retains the standard Vaidya form of classical general relativity. This result demonstrates that null dust can serve as the matter source for dynamical black-hole geometries, including regular black holes. Furthermore, we show that mass loss in the ingoing Vaidya solutions necessarily requires a violation of the null energy condition. This suggests that, within this class of solutions, quantum-gravity corrections to the gravitational dynamics alone may be insufficient to describe Hawking-like evaporation without additional matter effects.
This paper is organized as follows. In Sec. II, we briefly review covariant Hamiltonian gravity in spherical symmetry and the associated reconstruction algorithm. In Sec. III, we construct the effective Einstein tensor and derive its explicit expression in terms of the spacetime metric. In Sec. IV, we consider general matter coupling, introduce the general dynamical metric in the diagonal areal gauge, and evaluate the corresponding effective Einstein tensor. In Sec. V, we investigate the generalized Vaidya extension of static, spherically symmetric vacuum geometries. Finally, we summarize our results in Sec. VI.
Covariant Hamiltonian gravity in spherical symmetry
Motivated by the expectation that such a framework can arise as a semiclassical approximation to an underlying quantum theory of spherically symmetric general relativity, we assume that the canonical phase-space structure retains its classical form. Therefore, in this framework, gravity is described by two canonical pairs, ( E I , K I ) (E^{I}, K_{I}) ( E I , K I ) with I = 1 , 2 I = 1, 2 I = 1 , 2 , satisfying the Poisson brackets
{ K 1 ( x ) , E 1 ( y ) } = 2 δ ( x , y ) , { K 2 ( x ) , E 2 ( y ) } = δ ( x , y ) , (1) \begin{aligned}
\{K_{1}(x), E^{1}(y)\} &= 2\delta(x,y),\\
\{K_{2}(x), E^{2}(y)\} &= \delta(x,y), \tag*{(1)}
\end{aligned} { K 1 ( x ) , E 1 ( y )} { K 2 ( x ) , E 2 ( y )} = 2 δ ( x , y ) , = δ ( x , y ) , (1)
in geometrized units G = c = 1 G = c = 1 G = c = 1 [29 , 52 ] .
As in classical general relativity, the dynamics is governed by a set of constraints: the diffeomorphism constraint H x H_x H x and the effective Hamiltonian constraint H e f f H_{\mathrm{eff}} H eff . The former is assumed to retain its classical expression, as it generates spatial diffeomorphisms, whereas the latter, as mentioned above, may differ from its classical counterpart. Explicitly, the diffeomorphism constraint is given by
H x = E 2 ∂ x K 2 − 1 2 K 1 ∂ x E 1 . (2) H_x = E^{2}\partial_x K_2 - \frac{1}{2}K_1\partial_x E^{1}. \tag*{(2)} H x = E 2 ∂ x K 2 − 2 1 K 1 ∂ x E 1 . (2)
The form of H e f f H_{\mathrm{eff}} H eff remains undetermined and will be fixed by imposing further requirements.
Following the framework developed in [46 , 47 ] , the first requirement concerns the constraint algebra. To ensure the consistency of the dynamics, the constraints must form a first-class algebra. Since the diffeomorphism constraint retains its classical form (2), the Poisson brackets involving H x H_x H x are already fixed. It therefore remains to specify the bracket between two effective Hamiltonian constraints, which we require to take the form
{ H e f f [ N ] , H e f f [ M ] } = H x [ S ( N ∂ x M − M ∂ x N ) ] , (3) \{H_{\mathrm{eff}}[N], H_{\mathrm{eff}}[M]\} = H_x\left[S(N\partial_x M-M\partial_x N)\right], \tag*{(3)} { H eff [ N ] , H eff [ M ]} = H x [ S ( N ∂ x M − M ∂ x N ) ] , (3)
where S = μ E 1 / ( E 2 ) 2 S = \mu E^{1}/(E^{2})^{2} S = μ E 1 / ( E 2 ) 2 is the structure function and G [ f ] ≡ ∫ G f d x G[f] \equiv\int Gf\,\mathrm{d}x G [ f ] ≡ ∫ G f d x denotes the smeared constraint. The deformation factor μ \mu μ captures possible quantum-gravity corrections. It equals unity in classical general relativity but may depend on the phase-space variables in the effective theory.
The second requirement concerns general covariance, which is not guaranteed by the closure of the constraint algebra alone [17 ] . The subtlety arises from the Hamiltonian formulation, in which four-dimensional spacetime is foliated into a family of three-dimensional spatial hypersurfaces, thereby obscuring the manifest covariance of the theory. To solve the equations of motion in the Hamiltonian framework, one must specify the Lagrange multipliers, namely the lapse function N N N and the shift vector N x N^{x} N x . General covariance then requires that different choices of N N N and N x N^{x} N x , starting from the same physical initial data, yield spacetime metrics related by spacetime diffeomorphisms. To formulate this requirement explicitly, we must first specify how the spacetime metric is constructed from the canonical variables, lapse function, and shift vector. Motivated by the geometric interpretation of the constraint algebra (3) in terms of hypersurface deformations [17 , 46 , 47 ] , the effective spacetime metric reads
d s 2 = − N 2 d t 2 + ( E 2 ) 2 μ E 1 ( d x + N x d t ) 2 + E 1 d Ω 2 . (4) \mathrm{d}s^{2}=-N^{2}\mathrm{d}t^{2}+\frac{(E^{2})^{2}}{\mu E^{1}}(\mathrm{d}x+N^{x}\mathrm{d}t)^{2}+E^{1}\mathrm{d}\Omega^{2}. \tag*{(4)} d s 2 = − N 2 d t 2 + μ E 1 ( E 2 ) 2 ( d x + N x d t ) 2 + E 1 d Ω 2 . (4)
The requirement of general covariance, namely that different choices of N N N and N x N^{x} N x yield spacetime metrics related by diffeomorphisms, imposes further restrictions on the form of H e f f H_{\mathrm{eff}} H eff . To express these restrictions explicitly, it is convenient to introduce the scalars
s 1 = E 1 , s 2 = K 2 , s 3 = K 1 E 2 , s_1=E^{1},\qquad s_2=K_2,\qquad s_3=\frac{K_1}{E^{2}}, s 1 = E 1 , s 2 = K 2 , s 3 = E 2 K 1 ,
s 4 = ∂ x s 1 E 2 , s 5 = ∂ x s 4 E 2 . s_4=\frac{\partial_xs_1}{E^{2}},\qquad s_5=\frac{\partial_xs_4}{E^{2}}. s 4 = E 2 ∂ x s 1 , s 5 = E 2 ∂ x s 4 .
In terms of these variables, the generally covariant effective Hamiltonian constraint takes the form [46 , 47 , 48 ]
H e f f = − 2 E 2 [ ∂ s 1 M e f f + ∂ s 2 M e f f 2 s 3 + ∂ s 4 M e f f s 4 s 5 + R ] , (5) H_{\mathrm{eff}}=-2E^{2}\left[\partial_{s_1}M_{\mathrm{eff}}+\frac{\partial_{s_2}M_{\mathrm{eff}}}{2}s_3+\frac{\partial_{s_4}M_{\mathrm{eff}}}{s_4}s_5+\mathcal{R}\right], \tag*{(5)} H eff = − 2 E 2 [ ∂ s 1 M eff + 2 ∂ s 2 M eff s 3 + s 4 ∂ s 4 M eff s 5 + R ] , (5)
where R \mathcal{R} R is a function of s 1 s_1 s 1 and M e f f M_{\mathrm{eff}} M eff , while M e f f M_{\mathrm{eff}} M eff is a function of s 1 s_1 s 1 , s 2 s_2 s 2 , and s 4 s_4 s 4 , implicitly defined by
( ∂ s 2 M e f f ) 2 = μ s 1 4 ( s 4 2 + Z ) . (6) \left(\partial_{s_2}M_{\mathrm{eff}}\right)^{2}=\frac{\mu s_1}{4}\left(s_4^{2}+\mathcal{Z}\right). \tag*{(6)} ( ∂ s 2 M eff ) 2 = 4 μ s 1 ( s 4 2 + Z ) . (6)
for some function Z ( s 1 , M e f f ) \mathcal{Z}(s_1,M_{\mathrm{eff}}) Z ( s 1 , M eff ) . Moreover, general covariance further requires μ \mu μ to be a function of s 1 s_1 s 1 and M e f f M_{\mathrm{eff}} M eff . Thus, different models in covariant Hamiltonian gravity in spherical symmetry are characterized by different choices of the three functions R \mathcal{R} R , μ \mu μ , and Z \mathcal{Z} Z .
For a given choice of the three functions R \mathcal{R} R , μ \mu μ , and Z \mathcal{Z} Z , the resulting theory admits a unique one-parameter family of static, spherically symmetric vacuum solutions, as shown in [48 ] . In Schwarzschild-like coordinates, the metric takes the form
d s 2 = 1 4 N o 2 Z d t 2 − 4 μ Z d x 2 + x 2 d Ω 2 , (7) \mathrm{d}s^2 = \frac{1}{4}N_o^2\mathcal{Z}\mathrm{d}t^2 - \frac{4}{\mu\mathcal{Z}}\mathrm{d}x^2 + x^2\mathrm{d}\Omega^2, \tag*{(7)} d s 2 = 4 1 N o 2 Z d t 2 − μ Z 4 d x 2 + x 2 d Ω 2 , (7)
where d Ω 2 = d θ 2 + sin 2 θ d ϕ 2 \mathrm{d}\Omega^2 = \mathrm{d}\theta^2 + \sin^2\theta\mathrm{d}\phi^2 d Ω 2 = d θ 2 + sin 2 θ d ϕ 2 is the line element on S 2 S^2 S 2 , and
N o = exp ( 2 ∫ x ∂ M e f f R d x ) . (8) N_o = \exp\left(2\int x\partial_{M_{\mathrm{eff}}}\mathcal{R}\,\mathrm{d}x\right). \tag*{(8)} N o = exp ( 2 ∫ x ∂ M eff R d x ) . (8)
Here, the functions ∂ M e f f R \partial_{M_{\mathrm{eff}}}\mathcal{R} ∂ M eff R , Z \mathcal{Z} Z , and μ \mu μ are evaluated at s 1 = x 2 s_1 = x^2 s 1 = x 2 and
M e f f = M ( x ; m ) , M_{\mathrm{eff}} = M(x;m), M eff = M ( x ; m ) ,
where M ( x ; m ) M(x;m) M ( x ; m ) satisfies
∂ x M + 2 x R ( x 2 , M ) = 0 , (9) \partial_x M + 2x\mathcal{R}(x^2,M) = 0, \tag*{(9)} ∂ x M + 2 x R ( x 2 , M ) = 0 , (9)
with m m m denoting the integration constant. This equation follows from the constraints H e f f = 0 = H x H_{\mathrm{eff}} = 0 = H_x H eff = 0 = H x . Moreover, as shown in [48 ] , ∂ t M e f f = { M e f f , H e f f [ N ] + H x [ N x ] } ≈ 0 \partial_t M_{\mathrm{eff}} = \{M_{\mathrm{eff}},H_{\mathrm{eff}}[N] + H_x[N^x]\} \approx0 ∂ t M eff = { M eff , H eff [ N ] + H x [ N x ]} ≈ 0 , which ensures that m m m is also independent of time. Here, ≈ \approx ≈ denotes equality on the constraint surface defined by H e f f = 0 = H x H_{\mathrm{eff}} = 0 = H_x H eff = 0 = H x .
It is worth emphasizing the distinction between M e f f M_{\mathrm{eff}} M eff , M ( x ; m ) M(x;m) M ( x ; m ) , and m m m . The function M e f f ( s 1 , s 2 , s 4 ) M_{\mathrm{eff}}(s_1,s_2,s_4) M eff ( s 1 , s 2 , s 4 ) is defined on the gravitational phase space and enters the construction of the effective Hamiltonian constraint. The explanation following Eq. (9) implies that, when evaluated on the vacuum solutions, M e f f M_{\mathrm{eff}} M eff becomes M ( x ; m ) M(x;m) M ( x ; m ) , where m m m is a constant labeling the one-parameter family of vacuum solutions.
Restricting the form of μ \mu μ , Z \mathcal{Z} Z and R \mathcal{R} R
The functions μ \mu μ , Z \mathcal{Z} Z , and R \mathcal{R} R can be further constrained by requiring that a prescribed metric be a vacuum solution of the theory. To this end, we consider an arbitrary static, spherically symmetric metric depending on a mass parameter m m m , expressed in Schwarzschild-like coordinates as
d s 2 = − F ( x ; m ) d t 2 + d x 2 H ( x ; m ) + x 2 d Ω 2 . (10) \mathrm{d}s^2 = -F(x;m)\mathrm{d}t^2 + \frac{\mathrm{d}x^2}{H(x;m)} + x^2\mathrm{d}\Omega^2. \tag*{(10)} d s 2 = − F ( x ; m ) d t 2 + H ( x ; m ) d x 2 + x 2 d Ω 2 . (10)
Comparing Eqs. (7) and (10) , we obtain
1 4 N o 2 Z = − F , μ Z 4 = − H , (11) \frac{1}{4}N_o^2\mathcal{Z} = -F,\qquad\frac{\mu\mathcal{Z}}{4} = -H, \tag*{(11)} 4 1 N o 2 Z = − F , 4 μ Z = − H , (11)
as the conditions for the theory to admit the metric (10) as a vacuum solution. Here, the substitutions E 1 = x 2 E^1 = x^2 E 1 = x 2 and M e f f = M ( x ; m ) M_{\mathrm{eff}} = M(x;m) M eff = M ( x ; m ) must be made on the left-hand sides, so that both sides of the equations are expressed as functions of x x x and m m m . Equation (11) imposes constraints on the three functions μ \mu μ , Z \mathcal{Z} Z , and R \mathcal{R} R . However, it provides only two equations for these three unknown functions. Consequently, one functional degree of freedom remains undetermined. In our previous work [48 ] , this remaining functional freedom was parametrized by the function R \mathcal{R} R . In the present work, we adopt an alternative parametrization in terms of M ( x ; m ) M(x;m) M ( x ; m ) , the on-shell value of the effective mass function M e f f M_{\mathrm{eff}} M eff .
Equation (9) has related R \mathcal{R} R to the mass function M ( x ; m ) M(x;m) M ( x ; m ) . To further express the function Z \mathcal{Z} Z and μ \mu μ in terms of M ( x ; m ) M(x;m) M ( x ; m ) , let us consider the derivative of (9) with respect to m m m , leading to
∂ x ∂ m M + 2 x ( ∂ M e f f R ) ∂ m M = 0. (12) \partial_x\partial_m M + 2x(\partial_{M_{\mathrm{eff}}}\mathcal{R})\partial_m M = 0. \tag*{(12)} ∂ x ∂ m M + 2 x ( ∂ M eff R ) ∂ m M = 0. (12)
In addition, by the expression of N o N_o N o in (8), it also satisfies the equation
∂ x ( 1 N o ) + 2 x ( ∂ M e f f R ) 1 N o = 0. (13) \partial_x\left(\frac{1}{N_o}\right) + 2x(\partial_{M_{\mathrm{eff}}}\mathcal{R})\frac{1}{N_o} = 0. \tag*{(13)} ∂ x ( N o 1 ) + 2 x ( ∂ M eff R ) N o 1 = 0. (13)
Combining (13) with (12), we find that
∂ m M N o = C ( m ) (14) \partial_m M N_o = C(m) \tag*{(14)} ∂ m M N o = C ( m ) (14)
for some function C ( m ) C(m) C ( m ) . We now show that C ( m ) C(m) C ( m ) can be set to 1 by an appropriate choice of normalization. Indeed, the expression for N o N_o N o in Eq. (8) is defined only up to a multiplicative factor independent of x x x , which may depend on m m m . Equivalently, we may write
N o = A ( m ) exp ( 2 ∫ x o x ξ ∂ M e f f R d ξ ) , (15) N_o = A(m)\exp\left(2\int_{x_o}^{x}\xi\partial_{M_{\mathrm{eff}}}\mathcal{R}\,\mathrm{d}\xi\right), \tag*{(15)} N o = A ( m ) exp ( 2 ∫ x o x ξ ∂ M eff R d ξ ) , (15)
where A ( m ) A(m) A ( m ) represents the normalization freedom. Provided ∂ m M ≠ 0 \partial_m M \ne0 ∂ m M = 0 , this freedom allows us to choose C ( m ) = 1 C(m) = 1 C ( m ) = 1 , so that
N o = 1 ∂ m M . (16) N_o = \frac{1}{\partial_m M}. \tag*{(16)} N o = ∂ m M 1 . (16)
Substituting this result into Eq. (11) and using Eq. (9), we obtain
R = − ∂ x M 2 x , μ = H F ( ∂ m M ) 2 , \mathcal{R} = -\frac{\partial_x M}{2x},\qquad\mu= \frac{H}{F(\partial_m M)^2}, R = − 2 x ∂ x M , μ = F ( ∂ m M ) 2 H ,
Z = − 4 F ( ∂ m M ) 2 . (17) \mathcal{Z} = -4F(\partial_m M)^2. \tag*{(17)} Z = − 4 F ( ∂ m M ) 2 . (17)
These equations express R \mathcal{R} R , μ \mu μ , and Z \mathcal{Z} Z as functions of x x x and m m m , which can then be reexpressed as functions of ( s 1 , M e f f ) (s_1,M_{\mathrm{eff}}) ( s 1 , M eff ) through the relations s 1 = x 2 s_1 = x^2 s 1 = x 2 and M e f f = M ( x ; m ) M_{\mathrm{eff}} = M(x;m) M eff = M ( x ; m ) .
The effective Einstein tensor
We now turn to the spacetime-tensor formulation of the effective gravitational equations. We will consider a general matter coupling. As shown in [49 ] , when matter couples only to the metric and not to its derivatives, the coupled system remains generally covariant if the classical metric in the matter action is replaced by the effective metric (4). Accordingly, in this work, we restrict our consideration to matter action S m S_m S m depending only on g μ ν g_{\mu\nu} g μν , matter fields and the derivatives of the matter fields. For the gravity field, applying the inverse Legendre transformation yields
S g = − ∫ d t d x ( 1 2 K 1 E ˙ 1 + K 2 E ˙ 2 + N H e f f + N x H x ) . (18) S_g=-\int dt dx\left(\frac{1}{2}K_1\dot{E}^1+K_2\dot{E}^2+NH_{\mathrm{eff}}+N^xH_x\right). \tag*{(18)} S g = − ∫ d t d x ( 2 1 K 1 E ˙ 1 + K 2 E ˙ 2 + N H eff + N x H x ) . (18)
Its variation reads
δ S g = ∫ d t d x [ 1 2 ( − E ˙ 1 + { E 1 , H e f f [ N ] + H x [ N x ] } ) δ K 1 + ( − E ˙ 2 + { E 2 , H e f f [ N ] + H x [ N x ] } ) δ K 2 + 1 2 ( K ˙ 1 − { K 1 , H e f f [ N ] + H x [ N x ] } ) δ E 1 + ( K ˙ 2 − { K 2 , H e f f [ N ] + H x [ N x ] } ) δ E 2 \begin{aligned}
\delta S_g=\int dt dx[\frac{1}{2}(-\dot{E}^1+\{E^1,H_{\mathrm{eff}}[N]+H_x[N^x]\})\delta K_1+(-\dot{E}^2+\{E^2,H_{\mathrm{eff}}[N]+H_x[N^x]\})\delta K_2 \\
&+\frac{1}{2}\left(\dot{K}_1-\{K_1,H_{\mathrm{eff}}[N]+H_x[N^x]\}\right)\delta E^1+\left(\dot{K}_2-\{K_2,H_{\mathrm{eff}}[N]+H_x[N^x]\}\right)\delta E^2
\end{aligned} δ S g = ∫ d t d x [ 2 1 ( − E ˙ 1 + { E 1 , H eff [ N ] + H x [ N x ]}) δ K 1 + ( − E ˙ 2 + { E 2 , H eff [ N ] + H x [ N x ]}) δ K 2 + 2 1 ( K ˙ 1 − { K 1 , H eff [ N ] + H x [ N x ]} ) δ E 1 + ( K ˙ 2 − { K 2 , H eff [ N ] + H x [ N x ]} ) δ E 2
− H e f f δ N − H x δ N x ] . (19) -H_{\mathrm{eff}}\delta N-H_x\delta N^x]. \tag*{(19)} − H eff δ N − H x δ N x ] . (19)
Although the following calculations can be carried out directly in terms of the effective metric g μ ν g_{\mu\nu} g μν , it is convenient to introduce an auxiliary two-dimensional metric g ˚ a b \mathring{g}_{ab} g ˚ ab . This metric is not an independent dynamical field, but is constructed from the canonical variables and the deformation factor μ \mu μ . Its introduction simplifies the subsequent calculations. Specifically, for a , b = 0 , 1 a,b=0,1 a , b = 0 , 1 , we define
g ˚ a b d x a d x b = − N ˚ 2 d t 2 + ( E 2 ) 2 E 1 ( d x + N x d t ) 2 . (20) \mathring{g}_{ab}dx^a dx^b=-\mathring{N}^{2}dt^2+\frac{(E^2)^2}{E^1}(dx+N^xdt)^2. \tag*{(20)} g ˚ ab d x a d x b = − N ˚ 2 d t 2 + E 1 ( E 2 ) 2 ( d x + N x d t ) 2 . (20)
with
N ˚ = μ N . (21) \mathring{N}=\sqrt{\mu}N. \tag*{(21)} N ˚ = μ N . (21)
The 4-dimensional metric can be written in terms of g ˚ a b \mathring{g}_{ab} g ˚ ab as
g μ ν d x μ d x ν = 1 μ g ˚ a b d x a d x b + E 1 d Ω 2 . (22) g_{\mu\nu}dx^\mu dx^\nu=\frac{1}{\mu}\mathring{g}_{ab}dx^a dx^b+E^1d\Omega^2. \tag*{(22)} g μν d x μ d x ν = μ 1 g ˚ ab d x a d x b + E 1 d Ω 2 . (22)
Since μ \mu μ and E 1 E^1 E 1 are both scalars, g ˚ a b \mathring{g}_{ab} g ˚ ab defines a tensor in the theory which are covariant with respect to the effective metric g μ ν g_{\mu\nu} g μν . In terms of g ˚ a b \mathring{g}_{ab} g ˚ ab , we can rewrite the matter action S m S_m S m as
S m = S m [ g ˚ a b , μ , E 1 , ψ , ⋯ ] (23) S_m=S_m[\mathring{g}_{ab},\mu,E^1,\psi,\cdots] \tag*{(23)} S m = S m [ g ˚ ab , μ , E 1 , ψ , ⋯ ] (23)
where ψ \psi ψ denotes the matter fields and ⋯ \cdots ⋯ represents the derivatives of ψ \psi ψ . We then define T ˚ a b \mathring{T}_{ab} T ˚ ab , E ( μ ) \mathcal{E}^{(\mu)} E ( μ ) , and T ( θ ) T^{(\theta)} T ( θ ) through the variation of the matter action:
δ S m = − ∫ d t d x − g ˚ ( T ˚ a b δ g ˚ a b + E ( μ ) δ μ + T ( θ ) δ E 1 + ⋯ ) . (24) \delta S_m=-\int dt dx\sqrt{-\mathring{g}}\left(\mathring{T}_{ab}\delta\mathring{g}^{ab}+\mathcal{E}^{(\mu)}\delta\mu+T^{(\theta)}\delta E^1+\cdots\right). \tag*{(24)} δ S m = − ∫ d t d x − g ˚ ( T ˚ ab δ g ˚ ab + E ( μ ) δ μ + T ( θ ) δ E 1 + ⋯ ) . (24)
where ⋯ \cdots ⋯ denotes the terms arising from variations of the matter fields ψ \psi ψ . Here, T ˚ a b \mathring{T}_{ab} T ˚ ab , E ( μ ) \mathcal{E}^{(\mu)} E ( μ ) , and T ( θ ) T^{(\theta)} T ( θ ) are related to the four-dimensional energy-momentum tensor T μ ν T_{\mu\nu} T μν through
1 2 ∫ d 4 x − g T μ ν δ g μ ν = ∫ d t d x − g ˚ ( T ˚ a b δ g ˚ a b + E ( μ ) δ μ + T ( θ ) δ E 1 ) . (25) \frac{1}{2}\int d^4x\sqrt{-g}T_{\mu\nu}\delta g^{\mu\nu}
=\int dt dx\sqrt{-\mathring{g}}\left(\mathring{T}_{ab}\delta\mathring{g}^{ab}+\mathcal{E}^{(\mu)}\delta\mu+T^{(\theta)}\delta E^1\right). \tag*{(25)} 2 1 ∫ d 4 x − g T μν δ g μν = ∫ d t d x − g ˚ ( T ˚ ab δ g ˚ ab + E ( μ ) δ μ + T ( θ ) δ E 1 ) . (25)
which yields
T a b = T ˚ a b 2 π E 1 , T θ θ = − μ E 1 4 π T ( θ ) = T ϕ ϕ sin 2 θ , T_{ab}=\frac{\mathring{T}_{ab}}{2\pi E^1},\qquad T_{\theta\theta}=-\frac{\mu E^1}{4\pi}T^{(\theta)}=\frac{T_{\phi\phi}}{\sin^2\theta}, T ab = 2 π E 1 T ˚ ab , T θ θ = − 4 π μ E 1 T ( θ ) = sin 2 θ T ϕϕ ,
E ( μ ) = 1 μ g a b T ˚ a b . (26) \mathcal{E}^{(\mu)}=\frac{1}{\mu}g^{ab}\mathring{T}_{ab}. \tag*{(26)} E ( μ ) = μ 1 g ab T ˚ ab . (26)
These relations show that, for matter actions depending on the physical metric g μ ν g_{\mu\nu} g μν but not on any additional independent coupling to μ \mu μ , the quantity E ( μ ) \mathcal{E}^{(\mu)} E ( μ ) is not independent. Instead, it is completely determined by the trace of the two-dimensional energy-momentum tensor T ˚ a b \mathring{T}_{ab} T ˚ ab .
By definition, g ˚ a b \mathring{g}^{ab} g ˚ ab reads
g ˚ a b ∂ a ∂ b = − 1 N ˚ 2 N ⊗ N + E 1 ( E 2 ) 2 ∂ x ⊗ ∂ x , (27) \mathring{g}^{ab}\partial_a\partial_b=-\frac{1}{\mathring{N}^2}\mathcal{N}\otimes\mathcal{N}+\frac{E^1}{(E^2)^2}\partial_x\otimes\partial_x, \tag*{(27)} g ˚ ab ∂ a ∂ b = − N ˚ 2 1 N ⊗ N + ( E 2 ) 2 E 1 ∂ x ⊗ ∂ x , (27)
with
N = ∂ t − N x ∂ x . \mathcal{N}=\partial_t-N^x\partial_x. N = ∂ t − N x ∂ x .
Then, the variation of g ˚ a b \mathring{g}^{ab} g ˚ ab can be calculated as
δ g ˚ a b = 2 N ˚ 3 ( μ δ N + N ˙ 2 μ δ μ ) N a N b + 2 N ˚ 2 N ( a δ b 1 δ N x + ( 1 ( E 2 ) 2 δ E 1 − 2 E 1 ( E 2 ) 3 δ E 2 ) δ a 1 δ b 1 . (28) \begin{aligned}
\delta\mathring{g}^{ab}=\frac{2}{\mathring{N}^3}\left(\sqrt{\mu}\delta N+\frac{\dot{N}}{2\mu}\delta\mu\right)\mathcal{N}^a\mathcal{N}^b+\frac{2}{\mathring{N}^2}\mathcal{N}^{(a}\delta^b{}_1\delta N^x \\
&+\left(\frac{1}{(E^2)^2}\delta E^1-\frac{2E^1}{(E^2)^3}\delta E^2\right)\delta^a{}_1\delta^b{}_1. \tag*{(28)}
\end{aligned} δ g ˚ ab = N ˚ 3 2 ( μ δ N + 2 μ N ˙ δ μ ) N a N b + N ˚ 2 2 N ( a δ b 1 δ N x + ( ( E 2 ) 2 1 δ E 1 − ( E 2 ) 3 2 E 1 δ E 2 ) δ a 1 δ b 1 . (28)
Substituting this result into (24) , we get
δ S m = − ∫ d t d x μ N E 2 E 1 [ 2 μ N 3 T ˚ N N δ N + 2 μ N 2 T ˚ N x δ N x \delta S_m=-\int dt dx\frac{\sqrt{\mu}NE^2}{\sqrt{E^1}}[\frac{2}{\mu N^3}\mathring{T}_{\mathcal{N}\mathcal{N}}\delta N+\frac{2}{\mu N^2}\mathring{T}_{\mathcal{N}x}\delta N^x δ S m = − ∫ d t d x E 1 μ N E 2 [ μ N 3 2 T ˚ NN δ N + μ N 2 2 T ˚ N x δ N x
− 2 E 1 ( E 2 ) 3 T ˚ x x δ E 2 + ( T ˚ x x ( E 2 ) 2 + T ( θ ) ) δ E 1 + ( E ( μ ) + T ˚ N N μ 2 N 2 ) δ μ + ⋯ ] . (29) -\frac{2E^1}{(E^2)^3}\mathring{T}_{xx}\delta E^2+(\frac{\mathring{T}_{xx}}{(E^2)^2}+T^{(\theta)})\delta E^1+(\mathcal{E}^{(\mu)}+\frac{\mathring{T}_{\mathcal{N}\mathcal{N}}}{\mu^2N^2})\delta\mu+\cdots]. \tag*{(29)} − ( E 2 ) 3 2 E 1 T ˚ xx δ E 2 + ( ( E 2 ) 2 T ˚ xx + T ( θ ) ) δ E 1 + ( E ( μ ) + μ 2 N 2 T ˚ NN ) δ μ + ⋯ ] . (29)
with
T ˚ N N = T ˚ a b N a N b , T ˚ N x = T ˚ a b N a δ b 1 , \mathring{T}_{\mathcal{N}\mathcal{N}}=\mathring{T}_{ab}\mathcal{N}^a\mathcal{N}^b,\qquad\mathring{T}_{\mathcal{N}x}=\mathring{T}_{ab}\mathcal{N}^a\delta^b{}_1, T ˚ NN = T ˚ ab N a N b , T ˚ N x = T ˚ ab N a δ b 1 ,
T ˚ x x = T ˚ a b δ a 1 δ b 1 . \mathring{T}_{xx}=\mathring{T}_{ab}\delta^a{}_1\delta^b{}_1. T ˚ xx = T ˚ ab δ a 1 δ b 1 .
Combining this result (29) with Eq. (19) and imposing the stationary-action principle δ S g + δ S m = 0 \delta S_g+\delta S_m=0 δ S g + δ S m = 0 , we obtain: (1) Varying N N N while keeping all other variables fixed yields
T ˙ N N = − μ E 1 N 2 2 E 2 H e f f . (30) \dot{T}_{NN}=-\frac{\sqrt{\mu E^{1}}N^{2}}{2E^{2}}H_{\mathrm{eff}}. \tag*{(30)} T ˙ N N = − 2 E 2 μ E 1 N 2 H eff . (30)
(2) Varying N x N^{x} N x while keeping all other variables fixed leads to
T ˙ N x = − N μ E 1 2 E 2 H x (31) \dot{T}_{Nx}=-\frac{N\sqrt{\mu E^{1}}}{2E^{2}}H_{x} \tag*{(31)} T ˙ N x = − 2 E 2 N μ E 1 H x (31)
(3) Varying E 2 E^{2} E 2 while keeping all other variables fixed, and noting that δ μ = ∂ E 2 μ δ E 2 \delta\mu=\partial_{E^{2}}\mu\,\delta E^{2} δ μ = ∂ E 2 μ δ E 2 in this case, we obtain
2 N μ E 1 ( E 2 ) 2 T ˙ x x + ∂ μ ∂ E 2 N 2 μ ( H e f f − 2 μ E 3 E 2 E 1 E ( μ ) ) = − ∂ t K 2 + { K 2 , H e f f [ N ] + H x [ N x ] } . (32) \frac{2N\sqrt{\mu E^{1}}}{(E^{2})^{2}}\dot{T}_{xx}+\frac{\partial\mu}{\partial E^{2}}\frac{N}{2\mu}\left(H_{\mathrm{eff}}-\frac{2\sqrt{\mu}E^{3}E^{2}}{\sqrt{E^{1}}}\mathcal{E}^{(\mu)}\right)
=-\partial_{t}K_{2}+\{K_{2},H_{\mathrm{eff}}[N]+H_{x}[N^{x}]\}. \tag*{(32)} ( E 2 ) 2 2 N μ E 1 T ˙ xx + ∂ E 2 ∂ μ 2 μ N ( H eff − E 1 2 μ E 3 E 2 E ( μ ) ) = − ∂ t K 2 + { K 2 , H eff [ N ] + H x [ N x ]} . (32)
where we used (30) to replace T ˙ N N \dot{T}_{NN} T ˙ N N by H e f f H_{\mathrm{eff}} H eff . To simplify the formulation, we observe that
{ K 2 ( x ) , H e f f [ N ] } = { K 2 ( x ) , ∫ d y H e f f ( y ) μ ( y ) λ ( y ) } λ = N ˙ + H e f f N 2 μ ∂ μ ∂ E 2 . (33) \{K_{2}(x),H_{\mathrm{eff}}[N]\}
=\left\{\!K_{2}(x),\int\mathrm{d}y\frac{H_{\mathrm{eff}}(y)}{\sqrt{\mu(y)}}\lambda(y)\!\right\}_{\lambda=\dot{N}}
+H_{\mathrm{eff}}\frac{N}{2\mu}\frac{\partial\mu}{\partial E^{2}}. \tag*{(33)} { K 2 ( x ) , H eff [ N ]} = { K 2 ( x ) , ∫ d y μ ( y ) H eff ( y ) λ ( y ) } λ = N ˙ + H eff 2 μ N ∂ E 2 ∂ μ . (33)
with which, we can simplify (32) to be
T ˙ x x + μ 4 E 2 ∂ x E 1 2 E 1 E ( μ ) = ( E 2 ) 2 2 N μ E 1 D ˙ K 2 (34) \dot{T}_{xx}+\frac{\mu_{4}E^{2}\partial_{x}E^{1}}{2E^{1}}\mathcal{E}^{(\mu)}
=\frac{(E^{2})^{2}}{2N\sqrt{\mu E^{1}}}\dot{\mathcal{D}}K_{2} \tag*{(34)} T ˙ xx + 2 E 1 μ 4 E 2 ∂ x E 1 E ( μ ) = 2 N μ E 1 ( E 2 ) 2 D ˙ K 2 (34)
where we have introduced the operator D ˙ \dot{\mathcal{D}} D ˙ as
D ˙ = − ∂ t + { ⋅ , H ˙ e f f [ λ ] + H x [ N x ] } λ = N ˙ , (35) \dot{\mathcal{D}}=-\partial_{t}+\{\cdot,\dot{H}_{\mathrm{eff}}[\lambda]+H_{x}[N^{x}]\}_{\lambda=\dot{N}}, \tag*{(35)} D ˙ = − ∂ t + { ⋅ , H ˙ eff [ λ ] + H x [ N x ] } λ = N ˙ , (35)
with
H ˙ e f f = H e f f μ . (36) \dot{H}_{\mathrm{eff}}=\frac{H_{\mathrm{eff}}}{\sqrt{\mu}}. \tag*{(36)} H ˙ eff = μ H eff . (36)
The notation ∣ ∣ λ = N ˙ \left.\vphantom{\big|}\right|_{\lambda=\dot{N}} λ = N ˙ indicates that the Poisson bracket is evaluated first, treating λ \lambda λ as a phase-space-independent smearing function, and only then is λ \lambda λ replaced by N ˙ \dot{N} N ˙ . Moreover, in (34) , we used
∂ μ ∂ E 2 = − μ 4 ∂ x E 1 ( E 2 ) 2 , μ 4 = ∂ M e f f μ ∂ s 4 M e f f . (37) \frac{\partial\mu}{\partial E^{2}}=-\mu_{4}\frac{\partial_{x}E^{1}}{(E^{2})^{2}},\qquad\mu_{4}=\partial_{M_{\mathrm{eff}}}\mu\,\partial_{s_{4}}M_{\mathrm{eff}}. \tag*{(37)} ∂ E 2 ∂ μ = − μ 4 ( E 2 ) 2 ∂ x E 1 , μ 4 = ∂ M eff μ ∂ s 4 M eff . (37)
(4) Varying E 1 E^{1} E 1 while keeping all other variables fixed,
and noting that μ \mu μ depends on ∂ x E 1 \partial_{x}E^{1} ∂ x E 1 and E 1 E^{1} E 1 , we get
E 2 2 E 1 K ˙ 2 − μ N μ 4 ∂ x E 1 2 E 1 3 E ( μ ) + μ N E 2 E 1 T ( θ ) + ( μ N E 2 E ( μ ) E 1 − N 2 μ H e f f ) ∂ μ ∂ E 1 − ∂ x [ ( μ N E 2 E ( μ ) E 1 − N 2 μ H e f f ) ∂ μ ∂ ( ∂ x E 1 ) ] = 1 2 ( ∂ t K 1 − { K 1 , H e f f [ N ] + H x [ N x ] } ) . (38) \frac{E^{2}}{2E^{1}}\dot{K}_{2}-\frac{\sqrt{\mu}N\mu_{4}\partial_{x}E^{1}}{2\sqrt{E^{1\,3}}}\mathcal{E}^{(\mu)}+\frac{\sqrt{\mu}NE^{2}}{\sqrt{E^{1}}}T^{(\theta)}
+\left(\frac{\sqrt{\mu}NE^{2}\mathcal{E}^{(\mu)}}{\sqrt{E^{1}}}-\frac{N}{2\mu}H_{\mathrm{eff}}\right)\frac{\partial\mu}{\partial E^{1}}
-\partial_{x}\left[\left(\frac{\sqrt{\mu}NE^{2}\mathcal{E}^{(\mu)}}{\sqrt{E^{1}}}-\frac{N}{2\mu}H_{\mathrm{eff}}\right)\frac{\partial\mu}{\partial(\partial_{x}E^{1})}\right]
=\frac{1}{2}\left(\partial_{t}K_{1}-\{K_{1},H_{\mathrm{eff}}[N]+H_{x}[N^{x}]\}\right). \tag*{(38)} 2 E 1 E 2 K ˙ 2 − 2 E 1 3 μ N μ 4 ∂ x E 1 E ( μ ) + E 1 μ N E 2 T ( θ ) + ( E 1 μ N E 2 E ( μ ) − 2 μ N H eff ) ∂ E 1 ∂ μ − ∂ x [ ( E 1 μ N E 2 E ( μ ) − 2 μ N H eff ) ∂ ( ∂ x E 1 ) ∂ μ ] = 2 1 ( ∂ t K 1 − { K 1 , H eff [ N ] + H x [ N x ]} ) . (38)
where (30) and (32) are applied. Similarly as (33) , the Poisson bracket between K 1 K_{1} K 1 and H e f f [ N ] H_{\mathrm{eff}}[N] H eff [ N ] can be rewritten as
∫ d t d x { K 1 , H e f f [ N ] } δ E 1 = ∫ d t d x N μ H e f f ( ∂ μ ∂ E 1 δ E 1 + ∂ μ ∂ ∂ x E 1 ∂ x δ E 1 ) + ∫ d t d x { K 1 , H ˙ e f f [ λ ] } ∣ λ = N ˙ δ E 1 . (39) \int\mathrm{d}t\mathrm{d}x\{K_{1},H_{\mathrm{eff}}[N]\}\delta E^{1}
=\int\mathrm{d}t\mathrm{d}x\frac{N}{\mu}H_{\mathrm{eff}}\left(\frac{\partial\mu}{\partial E^{1}}\delta E^{1}+\frac{\partial\mu}{\partial\partial_{x}E^{1}}\partial_{x}\delta E^{1}\right)
+\left.\int\mathrm{d}t\mathrm{d}x\{K_{1},\dot{H}_{\mathrm{eff}}[\lambda]\}\right|_{\lambda=\dot{N}}\delta E^{1}. \tag*{(39)} ∫ d t d x { K 1 , H eff [ N ]} δ E 1 = ∫ d t d x μ N H eff ( ∂ E 1 ∂ μ δ E 1 + ∂ ∂ x E 1 ∂ μ ∂ x δ E 1 ) + ∫ d t d x { K 1 , H ˙ eff [ λ ]} λ = N ˙ δ E 1 . (39)
With this result, (38) can be simplified to get
μ N E 2 2 E 1 E ( μ ) ( 2 E 1 μ 1 − μ 4 ∂ x E 1 E 2 ) − ∂ x ( μ N μ 4 E 1 E ( μ ) ) + μ N E 2 E 1 T ( θ ) = − E 1 D ˙ K 1 2 E 1 . (40) \frac{\sqrt{\mu}NE^{2}}{2\sqrt{E^{1}}}\mathcal{E}^{(\mu)}
\left(2E^{1}\mu_{1}-\frac{\mu_{4}\partial_{x}E^{1}}{E^{2}}\right)
-\partial_{x}\left(\frac{\sqrt{\mu}N\mu_{4}}{\sqrt{E^{1}}}\mathcal{E}^{(\mu)}\right)
+\frac{\sqrt{\mu}NE^{2}}{\sqrt{E^{1}}}T^{(\theta)}
=-\frac{E^{1}\dot{\mathcal{D}}K_{1}}{2E^{1}}. \tag*{(40)} 2 E 1 μ N E 2 E ( μ ) ( 2 E 1 μ 1 − E 2 μ 4 ∂ x E 1 ) − ∂ x ( E 1 μ N μ 4 E ( μ ) ) + E 1 μ N E 2 T ( θ ) = − 2 E 1 E 1 D ˙ K 1 . (40)
where we introduced
μ 1 = ∂ s 1 μ + ∂ M e f f μ ∂ s 1 M e f f (41) \mu_{1}=\partial_{s_{1}}\mu+\partial_{M_{\mathrm{eff}}}\mu\,\partial_{s_{1}}M_{\mathrm{eff}} \tag*{(41)} μ 1 = ∂ s 1 μ + ∂ M eff μ ∂ s 1 M eff (41)
and used
∂ μ ∂ ( ∂ x E 1 ) = μ 4 E 2 . (42) \frac{\partial\mu}{\partial(\partial_{x}E^{1})}=\frac{\mu_{4}}{E^{2}}. \tag*{(42)} ∂ ( ∂ x E 1 ) ∂ μ = E 2 μ 4 . (42)
(5) Varying K 1 K_{1} K 1 while keeping all other variables fixed yields
δ g ˙ a b = 0 , (43) \delta\dot{g}^{ab}=0, \tag*{(43)} δ g ˙ ab = 0 , (43)
leading to
D ˙ E 1 = 0. (44) \dot{\mathcal{D}}E^{1}=0. \tag*{(44)} D ˙ E 1 = 0. (44)
(6) Varying K 2 K_{2} K 2 while keeping all other variables fixed yields, and noting that μ \mu μ depends on K 2 K_{2} K 2 , we get
( μ N E 2 E 1 E ( μ ) − N 2 μ H e f f ) ∂ μ ∂ K 2 = − ∂ t E 2 + { E 2 , H e f f [ N ] + H x [ N x ] } (45) \left(\frac{\sqrt{\mu}NE^{2}}{\sqrt{E^{1}}}\mathcal{E}^{(\mu)}-\frac{N}{2\mu}H_{\mathrm{eff}}\right)\frac{\partial\mu}{\partial K_{2}}
=-\partial_{t}E^{2}+\{E^{2},H_{\mathrm{eff}}[N]+H_{x}[N^{x}]\} \tag*{(45)} ( E 1 μ N E 2 E ( μ ) − 2 μ N H eff ) ∂ K 2 ∂ μ = − ∂ t E 2 + { E 2 , H eff [ N ] + H x [ N x ]} (45)
where (30) is applied. Similar as Eq. (33) , the Poisson bracket between E 2 E^2 E 2 and H e f f [ N ] H_{\mathrm{eff}}[N] H eff [ N ] can be simplified to
{ E 2 ( x ) , H e f f [ N ] } = { E 2 ( x ) , ∫ d y H e f f ( y ) μ ( y ) λ ( y ) } λ = N ˙ − H e f f N 2 μ ∂ μ ∂ K 2 (46) \left\{E^2(x),H_{\mathrm{eff}}[N]\right\}
=
\left\{E^2(x),\int dy\,\frac{H_{\mathrm{eff}}(y)}{\sqrt{\mu(y)}}\lambda(y)\right\}_{\lambda=\dot{N}}
-
H_{\mathrm{eff}}\frac{N}{2\mu}\frac{\partial\mu}{\partial K_2} \tag*{(46)} { E 2 ( x ) , H eff [ N ] } = { E 2 ( x ) , ∫ d y μ ( y ) H eff ( y ) λ ( y ) } λ = N ˙ − H eff 2 μ N ∂ K 2 ∂ μ (46)
Substituting this result into Eq. (45) , we get
μ μ 2 N E 2 E 1 E ( μ ) = D ˙ E 2 (47) \frac{\sqrt{\mu}\mu_2NE^2}{\sqrt{E^1}}\mathcal{E}^{(\mu)}=\dot{\mathcal{D}}E^2 \tag*{(47)} E 1 μ μ 2 N E 2 E ( μ ) = D ˙ E 2 (47)
where we introduced
μ 2 = ∂ M e f f μ ∂ s 2 M e f f . (48) \mu_2=\partial_{M_{\mathrm{eff}}}\mu\partial_{s_2}M_{\mathrm{eff}}. \tag*{(48)} μ 2 = ∂ M eff μ ∂ s 2 M eff . (48)
The above equations (30), (31), (34), (40), (44), and (47) follow directly from the variation of the total action, without assuming any Einstein-like field equations. We now reorganize them into a covariant tensorial form by identifying the corresponding gravitational and matter contributions. To this end, we define the tensors G ˙ μ ν \dot{\mathcal{G}}_{\mu\nu} G ˙ μν and T μ ν T_{\mu\nu} T μν as follows:
G ˙ α β d x α d x β = − N 2 μ E 1 H e f f 2 E 2 d t 2 − N μ E 1 E 2 H x ( d x + N x d t ) d t + ( E 2 ) 2 2 N μ E 1 D ˙ K 2 ( d x + N x d t ) 2 + E 1 D ˙ K I 4 N μ E 1 E 2 d Ω 2 (49) \dot{\mathcal{G}}_{\alpha\beta}dx^\alpha dx^\beta
=
-\frac{N^2\sqrt{\mu E^1}H_{\mathrm{eff}}}{2E^2}dt^2
-\frac{N\sqrt{\mu E^1}}{E^2}H_x(dx+N^xdt)dt
+\frac{(E^2)^2}{2N\sqrt{\mu E^1}}\dot{\mathcal{D}}K_2(dx+N^xdt)^2
+\frac{E^1\dot{\mathcal{D}}K_I}{4N\sqrt{\mu E^1}E^2}d\Omega^2 \tag*{(49)} G ˙ α β d x α d x β = − 2 E 2 N 2 μ E 1 H eff d t 2 − E 2 N μ E 1 H x ( d x + N x d t ) d t + 2 N μ E 1 ( E 2 ) 2 D ˙ K 2 ( d x + N x d t ) 2 + 4 N μ E 1 E 2 E 1 D ˙ K I d Ω 2 (49)
and
T α β d x α d x β = T ˙ a b d x a d x b + μ 4 E 2 ∂ x E 1 2 E 1 E ( μ ) ( d x + N x d t ) 2 + [ − T ( θ ) 2 + ( μ 4 ∂ x E 1 E 2 − 2 E 1 μ 1 ) E ( μ ) 4 E 1 ] d Ω 2 + E 1 2 N μ E 2 ∂ x ( μ N μ 4 E 1 E ( μ ) ) d Ω 2 (50) T_{\alpha\beta}dx^\alpha dx^\beta
=
\dot{T}_{ab}dx^a dx^b
+\frac{\mu_4E^2\partial_xE^1}{2E^1}\mathcal{E}^{(\mu)}(dx+N^xdt)^2
+\left[-\frac{T^{(\theta)}}{2}+\left(\frac{\mu_4\partial_xE^1}{E^2}-2E^1\mu_1\right)\frac{\mathcal{E}^{(\mu)}}{4E^1}\right]d\Omega^2
+\frac{\sqrt{E^1}}{2N\sqrt{\mu E^2}}\partial_x\left(\frac{\sqrt{\mu}N\mu_4}{\sqrt{E^1}}\mathcal{E}^{(\mu)}\right)d\Omega^2 \tag*{(50)} T α β d x α d x β = T ˙ ab d x a d x b + 2 E 1 μ 4 E 2 ∂ x E 1 E ( μ ) ( d x + N x d t ) 2 + [ − 2 T ( θ ) + ( E 2 μ 4 ∂ x E 1 − 2 E 1 μ 1 ) 4 E 1 E ( μ ) ] d Ω 2 + 2 N μ E 2 E 1 ∂ x ( E 1 μ N μ 4 E ( μ ) ) d Ω 2 (50)
Then, all of equations of motion can be written as
G ˙ α β = T α β , \dot{\mathcal{G}}_{\alpha\beta}=T_{\alpha\beta}, G ˙ α β = T α β ,
D ˙ E 1 = 0 , \dot{\mathcal{D}}E^1=0, D ˙ E 1 = 0 ,
D ˙ E 2 = μ μ 2 N E 2 E 1 E ( μ ) , (51) \dot{\mathcal{D}}E^2=\frac{\sqrt{\mu}\mu_2NE^2}{\sqrt{E^1}}\mathcal{E}^{(\mu)}, \tag*{(51)} D ˙ E 2 = E 1 μ μ 2 N E 2 E ( μ ) , (51)
where the last two equations define K I K_I K I in terms of E I E^I E I , their derivatives and E ( μ ) \mathcal{E}^{(\mu)} E ( μ ) . In what follows, we will call G ˙ μ ν \dot{\mathcal{G}}_{\mu\nu} G ˙ μν the effective Einstein tensor.
Explicit expression of G ˙ a b \dot{\mathcal{G}}_{ab} G ˙ ab
Now let us calculate the explicit expression of G ˙ \dot{\mathcal{G}} G ˙ . Equation (5) implies the following relation between H e f f H_{\mathrm{eff}} H eff and H x H_x H x :
H e f f − 2 ∂ s 2 M e f f ∂ x E 1 H x = − 2 E 2 ∂ x E 1 ( R ∂ x E 1 + ∂ x M e f f ) . (52) H_{\mathrm{eff}}-2\frac{\partial_{s_2}M_{\mathrm{eff}}}{\partial_xE^1}H_x
=
-\frac{2E^2}{\partial_xE^1}\left(\mathcal{R}\partial_xE^1+\partial_xM_{\mathrm{eff}}\right). \tag*{(52)} H eff − 2 ∂ x E 1 ∂ s 2 M eff H x = − ∂ x E 1 2 E 2 ( R ∂ x E 1 + ∂ x M eff ) . (52)
Taking advantage of this result, we can simplify the last two equations in Eq. (51) , to obtain
L N E 1 = 2 N ∂ s 2 M e f f , \mathcal{L}_{\mathcal{N}}E^1=2N\partial_{s_2}M_{\mathrm{eff}}, L N E 1 = 2 N ∂ s 2 M eff ,
L N E 2 = N μ 2 ( H e f f 2 μ − μ E 2 E 1 E ( μ ) ) − 2 N ∂ s 2 M e f f ∂ x E 1 H x + 2 N E 2 ∂ M e f f R ∂ s 2 M e f f + 2 N E 2 ∂ x ∂ s 2 M e f f ∂ x E 1 (53) \mathcal{L}_{\mathcal{N}}E^2
=
N\mu_2\left(\frac{H_{\mathrm{eff}}}{2\mu}-\frac{\sqrt{\mu}E^2}{\sqrt{E^1}}\mathcal{E}^{(\mu)}\right)
-\frac{2N\partial_{s_2}M_{\mathrm{eff}}}{\partial_xE^1}H_x
+2NE^2\partial_{M_{\mathrm{eff}}}\mathcal{R}\partial_{s_2}M_{\mathrm{eff}}
+\frac{2NE^2\partial_x\partial_{s_2}M_{\mathrm{eff}}}{\partial_xE^1} \tag*{(53)} L N E 2 = N μ 2 ( 2 μ H eff − E 1 μ E 2 E ( μ ) ) − ∂ x E 1 2 N ∂ s 2 M eff H x + 2 N E 2 ∂ M eff R ∂ s 2 M eff + ∂ x E 1 2 N E 2 ∂ x ∂ s 2 M eff (53)
where we introduced
L N E 1 = ∂ t E 1 − N x ∂ x E 1 , \mathcal{L}_{\mathcal{N}}E^1=\partial_tE^1-N^x\partial_xE^1, L N E 1 = ∂ t E 1 − N x ∂ x E 1 ,
L N E 2 = ∂ t E 2 − ∂ x ( N x E 2 ) (54) \mathcal{L}_{\mathcal{N}}E^2=\partial_tE^2-\partial_x(N^xE^2) \tag*{(54)} L N E 2 = ∂ t E 2 − ∂ x ( N x E 2 ) (54)
Solving ∂ s 2 M e f f \partial_{s_2}M_{\mathrm{eff}} ∂ s 2 M eff from the first equation and substituting the result into the second, we obtain
∂ M e f f μ L N E 1 4 μ H e f f − 2 N ∂ s 2 2 M e f f ∂ x E 1 H x = L N E 2 + μ E 2 L N E 1 ∂ M e f f μ 2 E 1 E ( μ ) − E 2 L N E 1 ∂ M e f f R − N E 2 ∂ x E 1 ∂ x ( L N E 1 N ) (55) \frac{\partial_{M_{\mathrm{eff}}}\mu\mathcal{L}_{\mathcal{N}}E^1}{4\mu}H_{\mathrm{eff}}
-\frac{2N\partial_{s_2}^2M_{\mathrm{eff}}}{\partial_xE^1}H_x
=
\mathcal{L}_{\mathcal{N}}E^2
+\frac{\sqrt{\mu}E^2\mathcal{L}_{\mathcal{N}}E^1\partial_{M_{\mathrm{eff}}}\mu}{2\sqrt{E^1}}\mathcal{E}^{(\mu)}
-E^2\mathcal{L}_{\mathcal{N}}E^1\partial_{M_{\mathrm{eff}}}\mathcal{R}
-\frac{NE^2}{\partial_xE^1}\partial_x\left(\frac{\mathcal{L}_{\mathcal{N}}E^1}{N}\right) \tag*{(55)} 4 μ ∂ M eff μ L N E 1 H eff − ∂ x E 1 2 N ∂ s 2 2 M eff H x = L N E 2 + 2 E 1 μ E 2 L N E 1 ∂ M eff μ E ( μ ) − E 2 L N E 1 ∂ M eff R − ∂ x E 1 N E 2 ∂ x ( N L N E 1 ) (55)
with applying Eq. (48) . Equations (55) and (52) form a system of two linear equations for H e f f H_{\mathrm{eff}} H eff and H x H_x H x . Solving this system, we obtain
H e f f = 4 μ D ∂ x E 1 ( 4 N 2 E 2 ( R ∂ x E 1 + ∂ x M e f f ) ∂ s 2 2 M e f f + ∂ x E 1 L N E 1 L N E 2 − N E 2 L N E 1 ∂ x ( L N E 1 N ) − E 2 ∂ x E 1 ( L N E 1 ) 2 ∂ M e f f R + μ E 2 ∂ x E 1 ( L N E 1 ) 2 ∂ M e f f μ 2 E 1 E ( μ ) ) (56) H_{\mathrm{eff}}
=
\frac{4\mu}{D\partial_xE^1}
\left(
4N^2E^2\left(\mathcal{R}\partial_xE^1+\partial_xM_{\mathrm{eff}}\right)\partial_{s_2}^2M_{\mathrm{eff}}
+\partial_xE^1\mathcal{L}_{\mathcal{N}}E^1\mathcal{L}_{\mathcal{N}}E^2
-NE^2\mathcal{L}_{\mathcal{N}}E^1\partial_x\left(\frac{\mathcal{L}_{\mathcal{N}}E^1}{N}\right)
-E^2\partial_xE^1(\mathcal{L}_{\mathcal{N}}E^1)^2\partial_{M_{\mathrm{eff}}}\mathcal{R}
+\frac{\sqrt{\mu}E^2\partial_xE^1(\mathcal{L}_{\mathcal{N}}E^1)^2\partial_{M_{\mathrm{eff}}}\mu}{2\sqrt{E^1}}\mathcal{E}^{(\mu)}
\right) \tag*{(56)} H eff = D ∂ x E 1 4 μ ( 4 N 2 E 2 ( R ∂ x E 1 + ∂ x M eff ) ∂ s 2 2 M eff + ∂ x E 1 L N E 1 L N E 2 − N E 2 L N E 1 ∂ x ( N L N E 1 ) − E 2 ∂ x E 1 ( L N E 1 ) 2 ∂ M eff R + 2 E 1 μ E 2 ∂ x E 1 ( L N E 1 ) 2 ∂ M eff μ E ( μ ) ) (56)
and
H x = − 2 N D ( 2 μ E 2 L N E 1 ∂ x E 1 ∂ M e f f R − 2 μ ∂ x E 1 L N E 2 − L N E 1 E 2 ( R ∂ x E 1 + ∂ x M e f f ) ∂ M e f f μ + 2 μ E 2 N ∂ x ( L N E 1 N ) − μ E 2 ∂ x E 1 L N E 1 ∂ M e f f μ E 1 E ( μ ) ) (57) H_x
=
-\frac{2N}{D}
\left(
2\mu E^2\mathcal{L}_{\mathcal{N}}E^1\partial_xE^1\partial_{M_{\mathrm{eff}}}\mathcal{R}
-2\mu\partial_xE^1\mathcal{L}_{\mathcal{N}}E^2
-\mathcal{L}_{\mathcal{N}}E^1E^2\left(\mathcal{R}\partial_xE^1+\partial_xM_{\mathrm{eff}}\right)\partial_{M_{\mathrm{eff}}}\mu
+2\mu E^2N\partial_x\left(\frac{\mathcal{L}_{\mathcal{N}}E^1}{N}\right)
-\frac{\sqrt{\mu}E^2\partial_xE^1\mathcal{L}_{\mathcal{N}}E^1\partial_{M_{\mathrm{eff}}}\mu}{\sqrt{E^1}}\mathcal{E}^{(\mu)}
\right) \tag*{(57)} H x = − D 2 N ( 2 μ E 2 L N E 1 ∂ x E 1 ∂ M eff R − 2 μ ∂ x E 1 L N E 2 − L N E 1 E 2 ( R ∂ x E 1 + ∂ x M eff ) ∂ M eff μ + 2 μ E 2 N ∂ x ( N L N E 1 ) − E 1 μ E 2 ∂ x E 1 L N E 1 ∂ M eff μ E ( μ ) ) (57)
where we define
D = ( L N E 1 ) 2 ∂ M e f f μ − 8 μ N 2 ∂ s 2 2 M e f f . (58) D = (\mathcal{L}_{\mathcal N}E^{1})^{2}\partial_{M_{\mathrm{eff}}}\mu- 8\mu N^{2}\partial^{2}_{s_{2}}M_{\mathrm{eff}}. \tag*{(58)} D = ( L N E 1 ) 2 ∂ M eff μ − 8 μ N 2 ∂ s 2 2 M eff . (58)
To express H e f f H_{\mathrm{eff}} H eff and H x H_{x} H x entirely in terms of N N N , N x N^{x} N x , E 1 E^{1} E 1 , E 2 E^{2} E 2 and their derivatives, we first need to eliminate ∂ s 2 2 M e f f \partial^{2}_{s_{2}}M_{\mathrm{eff}} ∂ s 2 2 M eff and M e f f M_{\mathrm{eff}} M eff in favor of these variables. To this end, we return to the covariance equation (6). Taking derivatives with respect to s 2 s_{2} s 2 for both sides of Eq. (6), we obtain
∂ s 2 2 M e f f = ∂ M e f f μ 8 μ ( L N E 1 N ) 2 + 1 8 E 1 μ ∂ M e f f Z , (59) \partial^{2}_{s_{2}}M_{\mathrm{eff}} = \frac{\partial_{M_{\mathrm{eff}}}\mu}{8\mu}\left(\frac{\mathcal{L}_{\mathcal N}E^{1}}{N}\right)^{2} + \frac{1}{8}E^{1}\mu\partial_{M_{\mathrm{eff}}}\mathcal{Z}, \tag*{(59)} ∂ s 2 2 M eff = 8 μ ∂ M eff μ ( N L N E 1 ) 2 + 8 1 E 1 μ ∂ M eff Z , (59)
where the first equation in (53) is applied. Furthermore, combining the first equation in (53) and (6) , we obtain
( L N E 1 N ) 2 = μ E 1 ( ( ∂ x E 1 E 2 ) 2 + Z ) . (60) \left(\frac{\mathcal{L}_{\mathcal N}E^{1}}{N}\right)^{2} = \mu E^{1}\left(\left(\frac{\partial_{x}E^{1}}{E^{2}}\right)^{2}+\mathcal{Z}\right). \tag*{(60)} ( N L N E 1 ) 2 = μ E 1 ( ( E 2 ∂ x E 1 ) 2 + Z ) . (60)
This equation allows us to determine M e f f M_{\mathrm{eff}} M eff implicitly in terms of N N N , E 1 E^{1} E 1 , E 2 E^{2} E 2 , and their derivatives. Substituting the result and (59) into (56) and (57), we can finally express H e f f H_{\mathrm{eff}} H eff and H x H_{x} H x entirely in terms of N N N , E 1 E^{1} E 1 , E 2 E^{2} E 2 , and their derivatives.
Now let us consider { K 2 , H ˙ e f f [ λ ] + H x [ N x ] } ∣ λ = N ˙ \left\{K_{2},\dot{H}_{\mathrm{eff}}[\lambda]+H_{x}[N^{x}]\right\}\big|_{\lambda=\dot{N}} { K 2 , H ˙ eff [ λ ] + H x [ N x ] } λ = N ˙ . Using the Hamiltonian constraint and { K 2 , H x [ N x ] } = N x ∂ x K 2 \left\{K_{2},H_{x}[N^{x}]\right\}=N^{x}\partial_{x}K_{2} { K 2 , H x [ N x ] } = N x ∂ x K 2 , we obtain
{ K 2 , H ˙ e f f [ λ ] + H x [ N x ] } ∣ λ = N ˙ = N x ∂ x K 2 − 2 N R − 2 N ∂ s 1 M e f f − 2 N H x ( E 2 ) 2 ∂ s 2 s 4 M e f f + ∂ x E 1 E 2 ( − 2 ∂ x N ∂ x E 1 + 2 N ∂ M e f f R ) ∂ s 4 M e f f + N ∂ x E 1 μ 4 2 μ ( E 2 ) 2 H e f f . (61) \begin{aligned}
\left\{K_{2},\dot{H}_{\mathrm{eff}}[\lambda]+H_{x}[N^{x}]\right\}\big|_{\lambda=\dot{N}}
={}&N^{x}\partial_{x}K_{2}-2N\mathcal{R}-2N\partial_{s_{1}}M_{\mathrm{eff}}\\
&-\frac{2NH_{x}}{(E^{2})^{2}}\partial_{s_{2}s_{4}}M_{\mathrm{eff}}\\
&+\frac{\partial_{x}E^{1}}{E^{2}}\left(-\frac{2\partial_{x}N}{\partial_{x}E^{1}}+2N\partial_{M_{\mathrm{eff}}}\mathcal{R}\right)\partial_{s_{4}}M_{\mathrm{eff}}\\
&+\frac{N\partial_{x}E^{1}\mu_{4}}{2\mu(E^{2})^{2}}H_{\mathrm{eff}}. \tag*{(61)}
\end{aligned} { K 2 , H ˙ eff [ λ ] + H x [ N x ] } λ = N ˙ = N x ∂ x K 2 − 2 N R − 2 N ∂ s 1 M eff − ( E 2 ) 2 2 N H x ∂ s 2 s 4 M eff + E 2 ∂ x E 1 ( − ∂ x E 1 2 ∂ x N + 2 N ∂ M eff R ) ∂ s 4 M eff + 2 μ ( E 2 ) 2 N ∂ x E 1 μ 4 H eff . (61)
Moreover, from the first equation of (53) , we have
∂ t K 2 − N x ∂ x K 2 = L N 2 E 1 − 1 N ( L N E 1 ) ( ∂ t N − N x ∂ x N ) 2 N ∂ s 2 2 M e f f + ∂ x E 1 L N E 2 − E 2 ∂ x ( L N E 1 ) ( E 2 ) 2 ∂ s 2 2 M e f f ∂ s 2 s 4 M e f f − L N E 1 ∂ s 2 2 M e f f ∂ s 1 s 2 M e f f . (62) \begin{aligned}
\partial_{t}K_{2}-N^{x}\partial_{x}K_{2}
={}&\frac{\mathcal{L}_{\mathcal N}^{2}E^{1}-\frac{1}{N}(\mathcal{L}_{\mathcal N}E^{1})(\partial_{t}N-N^{x}\partial_{x}N)}{2N\partial^{2}_{s_{2}}M_{\mathrm{eff}}}\\
&+\frac{\partial_{x}E^{1}\mathcal{L}_{\mathcal N}E^{2}-E^{2}\partial_{x}(\mathcal{L}_{\mathcal N}E^{1})}{(E^{2})^{2}\partial^{2}_{s_{2}}M_{\mathrm{eff}}}\partial_{s_{2}s_{4}}M_{\mathrm{eff}}\\
&-\frac{\mathcal{L}_{\mathcal N}E^{1}}{\partial^{2}_{s_{2}}M_{\mathrm{eff}}}\partial_{s_{1}s_{2}}M_{\mathrm{eff}}. \tag*{(62)}
\end{aligned} ∂ t K 2 − N x ∂ x K 2 = 2 N ∂ s 2 2 M eff L N 2 E 1 − N 1 ( L N E 1 ) ( ∂ t N − N x ∂ x N ) + ( E 2 ) 2 ∂ s 2 2 M eff ∂ x E 1 L N E 2 − E 2 ∂ x ( L N E 1 ) ∂ s 2 s 4 M eff − ∂ s 2 2 M eff L N E 1 ∂ s 1 s 2 M eff . (62)
Here, we define
L N 2 E 1 : = L N ( L N E 1 ) . \mathcal{L}_{\mathcal N}^{2}E^{1}:=\mathcal{L}_{\mathcal N}(\mathcal{L}_{\mathcal N}E^{1}). L N 2 E 1 := L N ( L N E 1 ) .
Combining Eqs. (61) and (62) yields
D ˙ K 2 = − L N 2 E 1 − 1 N ( L N E 1 ) ( ∂ t N − N x ∂ x N ) 2 N ∂ s 2 2 M e f f − 2 N ∂ s 1 M e f f + L N E 1 ∂ s 2 2 M e f f ∂ s 1 s 2 M e f f − [ 2 N H x ( E 2 ) 2 + ∂ x E 1 L N E 2 − E 2 ∂ x ( L N E 1 ) ( E 2 ) 2 ∂ s 2 2 M e f f ] ∂ s 2 s 4 M e f f + ∂ x E 1 E 2 ( − 2 ∂ x N ∂ x E 1 + 2 N ∂ M e f f R ) ∂ s 4 M e f f + N ∂ x E 1 μ 4 2 μ ( E 2 ) 2 H e f f − 2 N R . (63) \begin{aligned}
\dot{\mathcal{D}}K_{2}
={}&-\frac{\mathcal{L}_{\mathcal N}^{2}E^{1}-\frac{1}{N}(\mathcal{L}_{\mathcal N}E^{1})(\partial_{t}N-N^{x}\partial_{x}N)}{2N\partial^{2}_{s_{2}}M_{\mathrm{eff}}}
-2N\partial_{s_{1}}M_{\mathrm{eff}}+\frac{\mathcal{L}_{\mathcal N}E^{1}}{\partial^{2}_{s_{2}}M_{\mathrm{eff}}}\partial_{s_{1}s_{2}}M_{\mathrm{eff}}\\
&-\left[\frac{2NH_{x}}{(E^{2})^{2}}+\frac{\partial_{x}E^{1}\mathcal{L}_{\mathcal N}E^{2}-E^{2}\partial_{x}(\mathcal{L}_{\mathcal N}E^{1})}{(E^{2})^{2}\partial^{2}_{s_{2}}M_{\mathrm{eff}}}\right]\partial_{s_{2}s_{4}}M_{\mathrm{eff}}\\
&+\frac{\partial_{x}E^{1}}{E^{2}}\left(-\frac{2\partial_{x}N}{\partial_{x}E^{1}}+2N\partial_{M_{\mathrm{eff}}}\mathcal{R}\right)\partial_{s_{4}}M_{\mathrm{eff}}+\frac{N\partial_{x}E^{1}\mu_{4}}{2\mu(E^{2})^{2}}H_{\mathrm{eff}}-2N\mathcal{R}. \tag*{(63)}
\end{aligned} D ˙ K 2 = − 2 N ∂ s 2 2 M eff L N 2 E 1 − N 1 ( L N E 1 ) ( ∂ t N − N x ∂ x N ) − 2 N ∂ s 1 M eff + ∂ s 2 2 M eff L N E 1 ∂ s 1 s 2 M eff − [ ( E 2 ) 2 2 N H x + ( E 2 ) 2 ∂ s 2 2 M eff ∂ x E 1 L N E 2 − E 2 ∂ x ( L N E 1 ) ] ∂ s 2 s 4 M eff + E 2 ∂ x E 1 ( − ∂ x E 1 2 ∂ x N + 2 N ∂ M eff R ) ∂ s 4 M eff + 2 μ ( E 2 ) 2 N ∂ x E 1 μ 4 H eff − 2 N R . (63)
Moreover, using (55) to solve H x H_{x} H x and substituting the result, we get
2 N H x ( E 2 ) 2 + ∂ x E 1 L N E 2 − E 2 ∂ x ( L N E 1 ) ( E 2 ) 2 ∂ s 2 2 M e f f = ∂ x E 1 μ ( E 2 ) 2 ∂ s 2 2 M e f f [ ∂ M e f f μ L N E 1 4 H e f f + μ E 2 L N E 1 ∂ M e f f R − μ E 2 L N E 1 ∂ x N N ∂ x E 1 − μ 3 / 2 E 2 L N E 1 ∂ M e f f μ 2 E 1 E ( μ ) ] . (64) \begin{aligned}
\frac{2NH_{x}}{(E^{2})^{2}}+\frac{\partial_{x}E^{1}\mathcal{L}_{\mathcal N}E^{2}-E^{2}\partial_{x}(\mathcal{L}_{\mathcal N}E^{1})}{(E^{2})^{2}\partial^{2}_{s_{2}}M_{\mathrm{eff}}}
={}&\frac{\partial_{x}E^{1}}{\mu(E^{2})^{2}\partial^{2}_{s_{2}}M_{\mathrm{eff}}}\left[\frac{\partial_{M_{\mathrm{eff}}}\mu\,\mathcal{L}_{\mathcal N}E^{1}}{4}H_{\mathrm{eff}}+\mu E^{2}\mathcal{L}_{\mathcal N}E^{1}\partial_{M_{\mathrm{eff}}}\mathcal{R}\right.\\
&\left.-\frac{\mu E^{2}\mathcal{L}_{\mathcal N}E^{1}\partial_{x}N}{N\partial_{x}E^{1}}-\frac{\mu^{3/2}E^{2}\mathcal{L}_{\mathcal N}E^{1}\partial_{M_{\mathrm{eff}}}\mu}{2\sqrt{E^{1}}}\mathcal{E}^{(\mu)}\right]. \tag*{(64)}
\end{aligned} ( E 2 ) 2 2 N H x + ( E 2 ) 2 ∂ s 2 2 M eff ∂ x E 1 L N E 2 − E 2 ∂ x ( L N E 1 ) = μ ( E 2 ) 2 ∂ s 2 2 M eff ∂ x E 1 [ 4 ∂ M eff μ L N E 1 H eff + μ E 2 L N E 1 ∂ M eff R − N ∂ x E 1 μ E 2 L N E 1 ∂ x N − 2 E 1 μ 3/2 E 2 L N E 1 ∂ M eff μ E ( μ ) ] . (64)
Taking advantage of (64) , we get
D ˙ K 2 = − 2 N R + 1 8 ∂ s 2 2 M e f f [ 4 μ E 1 ∂ x E 1 ∂ x N ( E 2 ) 2 − 4 μ N E 1 ( ∂ x E 1 ) 2 ( E 2 ) 2 ∂ M e f f R + 2 μ N E 1 ∂ s 1 Z − N E 1 ( ∂ x E 1 ) 2 ( E 2 ) 3 ∂ M e f f μ H e f f + 2 ( L N E 1 ) 2 μ N ∂ s 1 μ + 4 L N E 1 ( ∂ t N − N x ∂ x N ) N 2 − 4 L N 2 E 1 N + 2 ( L N E 1 ) 2 N E 1 + N μ ∂ x E 1 E 2 E 1 ( μ 4 + μ E 1 ∂ x E 1 4 E 2 ∂ s 2 2 M e f f ∂ M e f f μ ) E ( μ ) ] . (65) \begin{aligned}
\dot{\mathcal{D}}K_{2}
={}&-2N\mathcal{R}+\frac{1}{8\partial^{2}_{s_{2}}M_{\mathrm{eff}}}\left[\frac{4\mu E^{1}\partial_{x}E^{1}\partial_{x}N}{(E^{2})^{2}}-\frac{4\mu NE^{1}(\partial_{x}E^{1})^{2}}{(E^{2})^{2}}\partial_{M_{\mathrm{eff}}}\mathcal{R}+2\mu NE^{1}\partial_{s_{1}}\mathcal{Z}\right.\\
&\left.-\frac{NE^{1}(\partial_{x}E^{1})^{2}}{(E^{2})^{3}}\partial_{M_{\mathrm{eff}}}\mu H_{\mathrm{eff}}+\frac{2(\mathcal{L}_{\mathcal N}E^{1})^{2}}{\mu N}\partial_{s_{1}}\mu\right.\\
&\left.+\frac{4\mathcal{L}_{\mathcal N}E^{1}(\partial_{t}N-N^{x}\partial_{x}N)}{N^{2}}-\frac{4\mathcal{L}_{\mathcal N}^{2}E^{1}}{N}+\frac{2(\mathcal{L}_{\mathcal N}E^{1})^{2}}{NE^{1}}\right.\\
&\left.+\frac{N\sqrt{\mu}\partial_{x}E^{1}}{E^{2}\sqrt{E^{1}}}\left(\mu_{4}+\frac{\mu E^{1}\partial_{x}E^{1}}{4E^{2}\partial^{2}_{s_{2}}M_{\mathrm{eff}}}\partial_{M_{\mathrm{eff}}}\mu\right)\mathcal{E}^{(\mu)}\right]. \tag*{(65)}
\end{aligned} D ˙ K 2 = − 2 N R + 8 ∂ s 2 2 M eff 1 [ ( E 2 ) 2 4 μ E 1 ∂ x E 1 ∂ x N − ( E 2 ) 2 4 μ N E 1 ( ∂ x E 1 ) 2 ∂ M eff R + 2 μ N E 1 ∂ s 1 Z − ( E 2 ) 3 N E 1 ( ∂ x E 1 ) 2 ∂ M eff μ H eff + μ N 2 ( L N E 1 ) 2 ∂ s 1 μ + N 2 4 L N E 1 ( ∂ t N − N x ∂ x N ) − N 4 L N 2 E 1 + N E 1 2 ( L N E 1 ) 2 + E 2 E 1 N μ ∂ x E 1 ( μ 4 + 4 E 2 ∂ s 2 2 M eff μ E 1 ∂ x E 1 ∂ M eff μ ) E ( μ ) ] . (65)
where need to relations following from the covariance equation (6)
∂ s 2 s 4 M e f f = ∂ s 2 2 M e f f ∂ s 2 M e f f ∂ s 4 M e f f + s 1 s 4 μ 4 ∂ s 2 M e f f , \partial_{s_{2}s_{4}}M_{\mathrm{eff}}=\frac{\partial^{2}_{s_{2}}M_{\mathrm{eff}}}{\partial_{s_{2}}M_{\mathrm{eff}}}\partial_{s_{4}}M_{\mathrm{eff}}+\frac{s_{1}s_{4}\mu}{4\partial_{s_{2}}M_{\mathrm{eff}}}, ∂ s 2 s 4 M eff = ∂ s 2 M eff ∂ s 2 2 M eff ∂ s 4 M eff + 4 ∂ s 2 M eff s 1 s 4 μ ,
∂ s 1 s 2 M e f f = ∂ s 2 2 M e f f ∂ s 2 M e f f ∂ s 1 M e f f + μ s 1 ∂ s 1 Z 8 ∂ s 2 M e f f + ∂ s 2 M e f f ( s 1 ∂ s 1 μ + μ ) 2 μ s 1 . (66) \partial_{s_{1}s_{2}}M_{\mathrm{eff}}=\frac{\partial^{2}_{s_{2}}M_{\mathrm{eff}}}{\partial_{s_{2}}M_{\mathrm{eff}}}\partial_{s_{1}}M_{\mathrm{eff}}+\frac{\mu s_{1}\partial_{s_{1}}\mathcal{Z}}{8\partial_{s_{2}}M_{\mathrm{eff}}}+\frac{\partial_{s_{2}}M_{\mathrm{eff}}(s_{1}\partial_{s_{1}}\mu+\mu)}{2\mu s_{1}}. \tag*{(66)} ∂ s 1 s 2 M eff = ∂ s 2 M eff ∂ s 2 2 M eff ∂ s 1 M eff + 8 ∂ s 2 M eff μ s 1 ∂ s 1 Z + 2 μ s 1 ∂ s 2 M eff ( s 1 ∂ s 1 μ + μ ) . (66)
Define G ˙ N N : = G ˙ μ ν N μ N ν \dot{\mathcal{G}}_{NN}:=\dot{\mathcal{G}}_{\mu\nu}N^{\mu}N^{\nu} G ˙ N N := G ˙ μν N μ N ν , G ˙ N x : = G ˙ μ ν N μ δ x ν \dot{\mathcal{G}}_{Nx}:=\dot{\mathcal{G}}_{\mu\nu}N^{\mu}\delta^{\nu}_{x} G ˙ N x := G ˙ μν N μ δ x ν . (49) gives
G ˙ N N = − N 2 μ E 1 2 E 2 H e f f , G ˙ N x = − N μ E 1 2 E 2 H x , G ˙ x x = ( E 2 ) 2 2 N μ E 1 D ˙ K 2 . (67) \begin{aligned}
\dot{\mathcal{G}}_{NN} &= -\frac{N^2\sqrt{\mu E^1}}{2E^2}H_{\rm eff},\\
\dot{\mathcal{G}}_{Nx} &= -\frac{N\sqrt{\mu E^1}}{2E^2}H_x,\\
\dot{\mathcal{G}}_{xx} &= \frac{(E^2)^2}{2N\sqrt{\mu E^1}}\dot{\mathcal{D}}K_2. \tag*{(67)}
\end{aligned} G ˙ N N G ˙ N x G ˙ xx = − 2 E 2 N 2 μ E 1 H eff , = − 2 E 2 N μ E 1 H x , = 2 N μ E 1 ( E 2 ) 2 D ˙ K 2 . (67)
Substituting Eqs. (56), (57) and (65) we obtain
G ˙ N N = − 2 μ N 2 μ E 1 D E 2 ∂ x E 1 [ ∂ x E 1 L N E 1 L N E 2 + 4 N 2 E 2 ( R ∂ x E 1 + ∂ x M e f f ) ∂ s 2 2 M e f f − N E 2 L N E 1 ∂ x ( L N E 1 N ) − E 2 ∂ x E 1 ( L N E 1 ) 2 ∂ M e f f R ] − μ 2 N 2 ( L N E 1 ) 2 ∂ M e f f μ D E ( μ ) . (68) \begin{aligned}
\dot{\mathcal{G}}_{NN}={}&-\frac{2\mu N^2\sqrt{\mu E^1}}{DE^2\partial_xE^1}\bigg[
\partial_xE^1\mathcal{L}_{\mathcal{N}}E^1\mathcal{L}_{\mathcal{N}}E^2\\
&+4N^2E^2(\mathcal{R}\partial_xE^1+\partial_xM_{\rm eff})\partial_{s_2}^2M_{\rm eff}\\
&-NE^2\mathcal{L}_{\mathcal{N}}E^1\partial_x\left(\frac{\mathcal{L}_{\mathcal{N}}E^1}{N}\right)\\
&-E^2\partial_xE^1(\mathcal{L}_{\mathcal{N}}E^1)^2\partial_{M_{\rm eff}}\mathcal{R}\bigg]\\
&-\frac{\mu^2N^2(\mathcal{L}_{\mathcal{N}}E^1)^2\partial_{M_{\rm eff}}\mu}{D}\mathcal{E}^{(\mu)}. \tag*{(68)}
\end{aligned} G ˙ N N = − D E 2 ∂ x E 1 2 μ N 2 μ E 1 [ ∂ x E 1 L N E 1 L N E 2 + 4 N 2 E 2 ( R ∂ x E 1 + ∂ x M eff ) ∂ s 2 2 M eff − N E 2 L N E 1 ∂ x ( N L N E 1 ) − E 2 ∂ x E 1 ( L N E 1 ) 2 ∂ M eff R ] − D μ 2 N 2 ( L N E 1 ) 2 ∂ M eff μ E ( μ ) . (68)
and
G ˙ N x = N 2 μ E 1 D E 2 [ 2 μ E 2 L N E 1 ∂ x E 1 ∂ M e f f R − E 2 L N E 1 ( R ∂ x E 1 + ∂ x M e f f ) ∂ M e f f μ − 2 μ ∂ x E 1 L N E 2 + 2 μ N E 2 ∂ x ( L N E 1 N ) ] − μ 2 N 2 ∂ x E 1 L N E 1 ∂ M e f f μ D E ( μ ) . (69) \begin{aligned}
\dot{\mathcal{G}}_{Nx}={}&\frac{N^2\sqrt{\mu E^1}}{DE^2}\bigg[
2\mu E^2\mathcal{L}_{\mathcal{N}}E^1\partial_xE^1\partial_{M_{\rm eff}}\mathcal{R}\\
&-E^2\mathcal{L}_{\mathcal{N}}E^1(\mathcal{R}\partial_xE^1+\partial_xM_{\rm eff})\partial_{M_{\rm eff}}\mu\\
&-2\mu\partial_xE^1\mathcal{L}_{\mathcal{N}}E^2+2\mu NE^2\partial_x\left(\frac{\mathcal{L}_{\mathcal{N}}E^1}{N}\right)\bigg]\\
&-\frac{\mu^2N^2\partial_xE^1\mathcal{L}_{\mathcal{N}}E^1\partial_{M_{\rm eff}}\mu}{D}\mathcal{E}^{(\mu)}. \tag*{(69)}
\end{aligned} G ˙ N x = D E 2 N 2 μ E 1 [ 2 μ E 2 L N E 1 ∂ x E 1 ∂ M eff R − E 2 L N E 1 ( R ∂ x E 1 + ∂ x M eff ) ∂ M eff μ − 2 μ ∂ x E 1 L N E 2 + 2 μ N E 2 ∂ x ( N L N E 1 ) ] − D μ 2 N 2 ∂ x E 1 L N E 1 ∂ M eff μ E ( μ ) . (69)
and
G ˙ x x = ( E 2 ) 2 16 N μ E 1 ∂ s 2 2 M e f f [ 2 ( L N E 1 ) 2 N E 1 + 4 μ E 1 ∂ x E 1 ∂ x N ( E 2 ) 2 + 2 μ N E 1 ∂ s 1 Z − 4 μ N E 1 ( ∂ x E 1 ) 2 ( E 2 ) 2 ∂ M e f f R + 2 ( L N E 1 ) 2 μ N ∂ s 1 μ + 4 L N E 1 ( ∂ t N − N x ∂ x N ) N 2 − 4 L N 2 E 1 N ] − ( E 2 ) 2 μ E 1 R + E 2 ∂ x E 1 ∂ M e f f μ 2 E 1 ( μ E 1 ∂ x E 1 4 E 2 ∂ s 2 2 M e f f + ∂ s 4 M e f f ) E ( μ ) . (70) \begin{aligned}
\dot{\mathcal{G}}_{xx}={}&\frac{(E^2)^2}{16N\sqrt{\mu E^1}\partial_{s_2}^2M_{\rm eff}}\bigg[
\frac{2(\mathcal{L}_{\mathcal{N}}E^1)^2}{NE^1}
+\frac{4\mu E^1\partial_xE^1\partial_xN}{(E^2)^2}
+2\mu NE^1\partial_{s_1}\mathcal{Z}\\
&-\frac{4\mu NE^1(\partial_xE^1)^2}{(E^2)^2}\partial_{M_{\rm eff}}\mathcal{R}
+\frac{2(\mathcal{L}_{\mathcal{N}}E^1)^2}{\mu N}\partial_{s_1}\mu\\
&+\frac{4\mathcal{L}_{\mathcal{N}}E^1(\partial_tN-N^x\partial_xN)}{N^2}
-\frac{4\mathcal{L}_{\mathcal{N}}^2E^1}{N}\bigg]\\
&-\frac{(E^2)^2}{\sqrt{\mu E^1}}\mathcal{R}
+\frac{E^2\partial_xE^1\partial_{M_{\rm eff}}\mu}{2E^1}
\left(\frac{\mu E^1\partial_xE^1}{4E^2\partial_{s_2}^2M_{\rm eff}}+\partial_{s_4}M_{\rm eff}\right)\mathcal{E}^{(\mu)}. \tag*{(70)}
\end{aligned} G ˙ xx = 16 N μ E 1 ∂ s 2 2 M eff ( E 2 ) 2 [ N E 1 2 ( L N E 1 ) 2 + ( E 2 ) 2 4 μ E 1 ∂ x E 1 ∂ x N + 2 μ N E 1 ∂ s 1 Z − ( E 2 ) 2 4 μ N E 1 ( ∂ x E 1 ) 2 ∂ M eff R + μ N 2 ( L N E 1 ) 2 ∂ s 1 μ + N 2 4 L N E 1 ( ∂ t N − N x ∂ x N ) − N 4 L N 2 E 1 ] − μ E 1 ( E 2 ) 2 R + 2 E 1 E 2 ∂ x E 1 ∂ M eff μ ( 4 E 2 ∂ s 2 2 M eff μ E 1 ∂ x E 1 + ∂ s 4 M eff ) E ( μ ) . (70)
Here, D D D is defined in Eq. (58) , and the Hamiltonian constraint H e f f H_{\rm eff} H eff appearing in D ˙ K 2 \dot{\mathcal{D}}K_2 D ˙ K 2 has been expressed in terms of G ˙ N N \dot{\mathcal{G}}_{NN} G ˙ N N .
For a given metric, when E ( μ ) = 0 \mathcal{E}^{(\mu)}=0 E ( μ ) = 0 , the effective Einstein tensor G ˙ a b \dot{\mathcal{G}}_{ab} G ˙ ab can be calculated without knowing the explicit functional form of M e f f M_{\rm eff} M eff in terms of the canonical variables. Instead, we only need Eqs. (59) and (60). The former determines ∂ s 2 2 M e f f \partial_{s_2}^2M_{\rm eff} ∂ s 2 2 M eff , while the latter implicitly determines M e f f M_{\rm eff} M eff , both in terms of the metric components and their derivatives. However, when E ( μ ) ≠ 0 \mathcal{E}^{(\mu)}\ne0 E ( μ ) = 0 , the explicit functional form of M e f f ( s 1 , s 2 , s 4 ) M_{\rm eff}(s_1,s_2,s_4) M eff ( s 1 , s 2 , s 4 ) is required, since ∂ s 4 M e f f \partial_{s_4}M_{\rm eff} ∂ s 4 M eff appears in the resulting expressions. In this case, once M e f f ( s 1 , s 2 , s 4 ) M_{\rm eff}(s_1,s_2,s_4) M eff ( s 1 , s 2 , s 4 ) is known, we can solve Eq. (60) for s 2 = K 2 s_2=K_2 s 2 = K 2 in terms of the metric components and their derivatives, and then substitute the resulting expression for K 2 K_2 K 2 into ∂ s 4 M e f f \partial_{s_4}M_{\rm eff} ∂ s 4 M eff .
Explicit expression of G ˙ θ θ \dot{\mathcal{G}}_{\theta\theta} G ˙ θ θ
The most challenging part of calculating G ˙ θ θ \dot{\mathcal{G}}_{\theta\theta} G ˙ θ θ is evaluating D ˙ K 1 \dot{\mathcal{D}}K_1 D ˙ K 1 , including the shift contribution. The key observation used in our calculation is the identity
{ H x , H ˙ e f f [ λ ] + H x [ N x ] } = ∂ x λ H ˙ e f f + N x ∂ x H x + 2 ( ∂ x N x ) H x . (71) \{H_x,\dot{H}_{\rm eff}[\lambda]+H_x[N^x]\}
=\partial_x\lambda\,\dot{H}_{\rm eff}+N^x\partial_xH_x+2(\partial_xN^x)H_x. \tag*{(71)} { H x , H ˙ eff [ λ ] + H x [ N x ]} = ∂ x λ H ˙ eff + N x ∂ x H x + 2 ( ∂ x N x ) H x . (71)
It leads to
∂ x N H ˙ e f f = − 1 2 ∂ x E 1 D ˙ K 1 + E 2 ∂ x ( D ˙ K 2 ) + [ ∂ t − N x ∂ x − 2 ( ∂ x N x ) ] H x + μ μ 2 N E 2 ∂ x K 2 E 1 E ( μ ) . (72) \begin{aligned}
\partial_xN\dot{H}_{\rm eff}={}&-\frac{1}{2}\partial_xE^1\dot{\mathcal{D}}K_1+E^2\partial_x(\dot{\mathcal{D}}K_2)\\
&+\left[\partial_t-N^x\partial_x-2(\partial_xN^x)\right]H_x\\
&+\frac{\sqrt{\mu}\mu_2NE^2\partial_xK_2}{\sqrt{E^1}}\mathcal{E}^{(\mu)}. \tag*{(72)}
\end{aligned} ∂ x N H ˙ eff = − 2 1 ∂ x E 1 D ˙ K 1 + E 2 ∂ x ( D ˙ K 2 ) + [ ∂ t − N x ∂ x − 2 ( ∂ x N x ) ] H x + E 1 μ μ 2 N E 2 ∂ x K 2 E ( μ ) . (72)
where the last two equations in Eq. (51) must be applied. Using Eq. (49) to express H ˙ e f f \dot{H}_{\rm eff} H ˙ eff , H x H_x H x , and D ˙ K 2 \dot{\mathcal{D}}K_2 D ˙ K 2 in terms of G ˙ N N \dot{\mathcal{G}}_{NN} G ˙ N N , G ˙ N x \dot{\mathcal{G}}_{Nx} G ˙ N x , and G ˙ x x \dot{\mathcal{G}}_{xx} G ˙ xx , respectively, we obtain from Eq. (72)
D ˙ K 1 = 4 ( ∂ x N ) E 2 G ˙ N N N 2 μ E 1 ∂ x E 1 + 4 E 2 ∂ x E 1 ∂ x ( N μ E 1 G ˙ x x ( E 2 ) 2 ) − 4 ∂ x E 1 [ ∂ t − N x ∂ x − 2 ( ∂ x N x ) ] ( E 2 G ˙ N x N μ E 1 ) + 2 μ μ 2 N E 2 ∂ x K 2 ∂ x E 1 E 1 E ( μ ) . (73) \begin{aligned}
\dot{\mathcal{D}}K_1={}&\frac{4(\partial_xN)E^2\dot{\mathcal{G}}_{NN}}{N^2\mu\sqrt{E^1}\partial_xE^1}
+\frac{4E^2}{\partial_xE^1}\partial_x\left(\frac{N\sqrt{\mu E^1}\dot{\mathcal{G}}_{xx}}{(E^2)^2}\right)\\
&-\frac{4}{\partial_xE^1}\left[\partial_t-N^x\partial_x-2(\partial_xN^x)\right]\left(\frac{E^2\dot{\mathcal{G}}_{Nx}}{N\sqrt{\mu E^1}}\right)\\
&+\frac{2\sqrt{\mu}\mu_2NE^2\partial_xK_2}{\partial_xE^1\sqrt{E^1}}\mathcal{E}^{(\mu)}. \tag*{(73)}
\end{aligned} D ˙ K 1 = N 2 μ E 1 ∂ x E 1 4 ( ∂ x N ) E 2 G ˙ N N + ∂ x E 1 4 E 2 ∂ x ( ( E 2 ) 2 N μ E 1 G ˙ xx ) − ∂ x E 1 4 [ ∂ t − N x ∂ x − 2 ( ∂ x N x ) ] ( N μ E 1 E 2 G ˙ N x ) + ∂ x E 1 E 1 2 μ μ 2 N E 2 ∂ x K 2 E ( μ ) . (73)
Substituting this result into G ˙ θ θ = T θ θ \dot{\mathcal{G}}_{\theta\theta}=T_{\theta\theta} G ˙ θ θ = T θ θ , which is equal to
D ˙ K 1 + 2 N μ E 1 E 1 E 2 G ˙ x x = 4 N μ E 2 E 1 T θ θ , (74) \dot{\mathcal{D}}K_1+\frac{2N\sqrt{\mu E^1}}{E^1E^2}\dot{\mathcal{G}}_{xx}
=\frac{4N\sqrt{\mu E^2}}{\sqrt{E^1}}T_{\theta\theta}, \tag*{(74)} D ˙ K 1 + E 1 E 2 2 N μ E 1 G ˙ xx = E 1 4 N μ E 2 T θ θ , (74)
we finally have
N μ ∂ x E 1 T θ θ E 2 = − ∂ x ( 1 N ) G ˙ N N + ∂ x ( N μ E 1 ( E 2 ) 2 G ˙ x x ) − E 1 E 2 [ ∂ t − N x ∂ x − 2 ( ∂ x N x ) ] ( E 2 G ˙ N x N μ E 1 ) + μ μ 2 N ∂ x K 2 2 E ( μ ) . (75) \begin{aligned}
\frac{N\sqrt{\mu}\partial_xE^1T_{\theta\theta}}{E^2}={}&-\partial_x\left(\frac{1}{N}\right)\dot{\mathcal{G}}_{NN}
+\partial_x\left(\frac{N\sqrt{\mu E^1}}{(E^2)^2}\dot{\mathcal{G}}_{xx}\right)\\
&-\frac{\sqrt{E^1}}{E^2}\left[\partial_t-N^x\partial_x-2(\partial_xN^x)\right]
\left(\frac{E^2\dot{\mathcal{G}}_{Nx}}{N\sqrt{\mu E^1}}\right)\\
&+\frac{\sqrt{\mu}\mu_2N\partial_xK_2}{2}\mathcal{E}^{(\mu)}. \tag*{(75)}
\end{aligned} E 2 N μ ∂ x E 1 T θ θ = − ∂ x ( N 1 ) G ˙ N N + ∂ x ( ( E 2 ) 2 N μ E 1 G ˙ xx ) − E 2 E 1 [ ∂ t − N x ∂ x − 2 ( ∂ x N x ) ] ( N μ E 1 E 2 G ˙ N x ) + 2 μ μ 2 N ∂ x K 2 E ( μ ) . (75)
To simplify this equation, we note that
− g ˙ g ˙ a b = − E 2 N μ E 1 N a N b + N μ E 1 E 2 δ 1 a δ 1 b . (76) \sqrt{-\dot{g}}\dot{g}^{ab}=-\frac{E^{2}}{N\sqrt{\mu E^{1}}}\mathcal{N}^{a}\mathcal{N}^{b}+\frac{N\sqrt{\mu E^{1}}}{E^{2}}\delta^{a}_{1}\delta^{b}_{1}. \tag*{(76)} − g ˙ g ˙ ab = − N μ E 1 E 2 N a N b + E 2 N μ E 1 δ 1 a δ 1 b . (76)
(75) can therefore be rewritten as
N μ ∂ x E 1 T θ θ 2 = E 1 E 2 [ ∂ a ( − g ˙ G ˙ x a ) − − g ˙ Γ ˙ x c a G ˙ a c ] + μ μ N ∂ x K 2 2 E ( μ ) (77) \frac{N\sqrt{\mu}\partial_{x}E^{1}\mathcal{T}_{\theta\theta}}{2}
=\frac{\sqrt{E^{1}}}{E^{2}}\left[\partial_{a}\left(\sqrt{-\dot{g}}\dot{\mathcal{G}}^{a}_{x}\right)-\sqrt{-\dot{g}}\dot{\Gamma}^{a}_{xc}\dot{\mathcal{G}}^{c}_{a}\right]
+\frac{\sqrt{\mu}\mu N\partial_{x}K_{2}}{2}\mathcal{E}_{(\mu)} \tag*{(77)} 2 N μ ∂ x E 1 T θ θ = E 2 E 1 [ ∂ a ( − g ˙ G ˙ x a ) − − g ˙ Γ ˙ x c a G ˙ a c ] + 2 μ μ N ∂ x K 2 E ( μ ) (77)
where we define
G ˙ c a = g ˙ a b G ˙ b c . (78) \dot{\mathcal{G}}^{a}_{c}=\dot{g}^{ab}\dot{\mathcal{G}}_{bc}. \tag*{(78)} G ˙ c a = g ˙ ab G ˙ b c . (78)
and note that
− g ˙ E 1 E 2 Γ ˙ x c a G ˙ a c = ∂ x ( 1 N ) G ˙ N N − ∂ x N x N G ˙ N x − N ˙ E 1 E 2 ∂ x ( E 1 E 2 ) G ˙ x x . \frac{\sqrt{-\dot{g}}\sqrt{E^{1}}}{E^{2}}\dot{\Gamma}^{a}_{xc}\dot{\mathcal{G}}^{c}_{a}
=\partial_{x}\left(\frac{1}{N}\right)\dot{\mathcal{G}}_{\mathcal{N}\mathcal{N}}-\frac{\partial_{x}N^{x}}{N}\dot{\mathcal{G}}_{Nx}-\frac{\dot{N}\sqrt{E^{1}}}{E^{2}}\partial_{x}\left(\frac{\sqrt{E^{1}}}{E^{2}}\right)\dot{\mathcal{G}}_{xx}. E 2 − g ˙ E 1 Γ ˙ x c a G ˙ a c = ∂ x ( N 1 ) G ˙ NN − N ∂ x N x G ˙ N x − E 2 N ˙ E 1 ∂ x ( E 2 E 1 ) G ˙ xx .
Using the definition of the covariant divergence associated with g ˙ a b \dot{g}_{ab} g ˙ ab and − g ˙ = N μ E 2 / E 1 \sqrt{-\dot{g}}=N\sqrt{\mu E^{2}/E^{1}} − g ˙ = N μ E 2 / E 1 , (77) reduces to
∇ ˙ a G ˙ x a + μ 2 ∂ x K 2 2 E ( μ ) = ∂ x E 1 T θ θ . (79) \dot{\nabla}_{a}\dot{\mathcal{G}}^{a}_{x}+\frac{\mu_{2}\partial_{x}K_{2}}{2}\mathcal{E}_{(\mu)}=\partial_{x}E^{1}\mathcal{T}_{\theta\theta}. \tag*{(79)} ∇ ˙ a G ˙ x a + 2 μ 2 ∂ x K 2 E ( μ ) = ∂ x E 1 T θ θ . (79)
Here, ∇ ˙ a G ˙ x a \dot{\nabla}_{a}\dot{\mathcal{G}}^{a}_{x} ∇ ˙ a G ˙ x a denotes the x x x -component of ∇ ˙ a G ˙ b a \dot{\nabla}_{a}\dot{\mathcal{G}}^{a}_{b} ∇ ˙ a G ˙ b a . We thus get
G ˙ θ θ = 1 ∂ x E 1 ( ∇ ˙ a G ˙ x a + μ 2 ∂ x K 2 2 E ( μ ) ) . (80) \dot{\mathcal{G}}_{\theta\theta}=\frac{1}{\partial_{x}E^{1}}\left(\dot{\nabla}_{a}\dot{\mathcal{G}}^{a}_{x}+\frac{\mu_{2}\partial_{x}K_{2}}{2}\mathcal{E}_{(\mu)}\right). \tag*{(80)} G ˙ θ θ = ∂ x E 1 1 ( ∇ ˙ a G ˙ x a + 2 μ 2 ∂ x K 2 E ( μ ) ) . (80)
Gravitational dynamics with matter in diagonal areal gauge
Having established the effective Einstein equation in the preceding section, we now turn to the gravitational dynamics in the presence of matter. To describe general dynamical spherically symmetric geometries, we work in the diagonal areal gauge and promote the constant mass parameter m m m to a spacetime-dependent function m ( t , x ) m(t,x) m ( t , x ) . Since a general metric in this gauge contains two independent functions, we introduce an additional function β ( t , x ) \beta(t,x) β ( t , x ) and write
d s 2 = − β ( t , x ) 2 F ( x , m ) d t 2 + 1 H ( x , m ) d x 2 + x 2 d Ω 2 . (81) \mathrm{d}s^{2}=-\beta(t,x)^{2}F(x,m)\mathrm{d}t^{2}+\frac{1}{H(x,m)}\mathrm{d}x^{2}+x^{2}\mathrm{d}\Omega^{2}. \tag*{(81)} d s 2 = − β ( t , x ) 2 F ( x , m ) d t 2 + H ( x , m ) 1 d x 2 + x 2 d Ω 2 . (81)
Provided that H ( x , m ) H(x,m) H ( x , m ) is locally invertible with respect to m m m , this parametrization describes a general spherically symmetric geometry in the diagonal areal gauge. From the metric, we can read
N = β F , N x = 0 , E 1 = x 2 , ( E 2 ) 2 μ = x 2 H . (82) N=\beta\sqrt{F},\qquad N^{x}=0,\qquad E^{1}=x^{2},\qquad\frac{(E^{2})^{2}}{\mu}=\frac{x^{2}}{H}. \tag*{(82)} N = β F , N x = 0 , E 1 = x 2 , μ ( E 2 ) 2 = H x 2 . (82)
To calculate G ˙ μ ν \dot{\mathcal{G}}_{\mu\nu} G ˙ μν , we first use (60) to determine M e f f M_{\mathrm{eff}} M eff in terms of the metric components and their derivatives, as discussed above. Substituting (82) into (60) , we obtain
0 = x 2 4 [ 4 H ( x , m ) + μ ( x 2 , M e f f ) Z ( x 2 , M e f f ) ] . (83) 0=\frac{x^{2}}{4}\left[4H(x,m)+\mu(x^{2},M_{\mathrm{eff}})\mathcal{Z}(x^{2},M_{\mathrm{eff}})\right]. \tag*{(83)} 0 = 4 x 2 [ 4 H ( x , m ) + μ ( x 2 , M eff ) Z ( x 2 , M eff ) ] . (83)
To solve this equation for M e f f M_{\mathrm{eff}} M eff , we need to know the functions μ ( s 1 , M e f f ) \mu(s_{1},M_{\mathrm{eff}}) μ ( s 1 , M eff ) and Z ( s 1 , M e f f ) \mathcal{Z}(s_{1},M_{\mathrm{eff}}) Z ( s 1 , M eff ) . Recall that these functions were determined by prescribing M e f f = M ( x ; m ) M_{\mathrm{eff}}=M(x;m) M eff = M ( x ; m ) in the vacuum sector and requiring the static metric (10) to be a vacuum solution of the theory. These requirements led to the reconstruction equations (17). In the vacuum case, (60) still holds. Substituting the static vacuum solution, for which N = F ( x , m ) N=\sqrt{F(x,m)} N = F ( x , m ) , E 1 = x 2 E^{1}=x^{2} E 1 = x 2 , and ( E 2 ) 2 / μ = x 2 / H ( x , m ) (E^{2})^{2}/\mu=x^{2}/H(x,m) ( E 2 ) 2 / μ = x 2 / H ( x , m ) , into this equation, we must therefore recover M e f f = M ( x ; m ) M_{\mathrm{eff}}=M(x;m) M eff = M ( x ; m ) . In other words, the reconstruction procedure guarantees that
0 = x 2 4 [ 4 H ( x , m ) + μ ( x 2 , M ( x ; m ) ) Z ( x 2 , M ( x ; m ) ) ] , (84) 0=\frac{x^{2}}{4}\left[4H(x,m)+\mu(x^{2},M(x;m))\mathcal{Z}(x^{2},M(x;m))\right], \tag*{(84)} 0 = 4 x 2 [ 4 H ( x , m ) + μ ( x 2 , M ( x ; m )) Z ( x 2 , M ( x ; m )) ] , (84)
which holds for arbitrary m m m . Comparing this equation with (83) , we conclude that the latter admits the solution
M e f f = M ( x , m ) . (85) M_{\mathrm{eff}}=M(x,m). \tag*{(85)} M eff = M ( x , m ) . (85)
Consequently, all quantities depending on s 1 s_{1} s 1 and M e f f M_{\mathrm{eff}} M eff , including μ \mu μ , Z \mathcal{Z} Z , R \mathcal{R} R , and their derivatives with respect to M e f f M_{\mathrm{eff}} M eff , retain the same functional forms as in the static vacuum case, with the constant mass parameter m m m replaced by the spacetime-dependent function m m m . In particular, we get from the reconstruction equation (17) that
μ = H F ( ∂ m M ) 2 . (86) \mu=\frac{H}{F(\partial_{m}M)^{2}}. \tag*{(86)} μ = F ( ∂ m M ) 2 H . (86)
which, together with (82) , leads to
E 2 = x F ∂ m M . (87) E^{2}=\frac{x}{\sqrt{F}\,\partial_{m}M}. \tag*{(87)} E 2 = F ∂ m M x . (87)
Furthermore, the above arguments also apply to ∂ s 2 2 M e f f \partial_{s_{2}}^{2}M_{\mathrm{eff}} ∂ s 2 2 M eff , which is determined by (59) . With this result, we can now calculate G ˙ μ ν \dot{\mathcal{G}}_{\mu\nu} G ˙ μν . We first calculate G ˙ t t \dot{\mathcal{G}}_{tt} G ˙ tt . Since ∂ t E 1 = 0 \partial_{t}E^{1}=0 ∂ t E 1 = 0 , (58) simplifies to
D = − 8 μ N 2 ∂ s 2 2 M e f f . (88) D=-8\mu N^{2}\partial_{s_{2}}^{2}M_{\mathrm{eff}}. \tag*{(88)} D = − 8 μ N 2 ∂ s 2 2 M eff . (88)
Using this result, we could get from (68) for G ˙ t t \dot{\mathcal{G}}_{tt} G ˙ tt as
G ˙ t t = N 2 μ E 1 ∂ x E 1 ( R ∂ x E 1 + ∂ x M e f f ) . (89) \dot{\mathcal{G}}_{tt}=\frac{N^{2}\sqrt{\mu E^{1}}}{\partial_{x}E^{1}}\left(\mathcal{R}\partial_{x}E^{1}+\partial_{x}M_{\mathrm{eff}}\right). \tag*{(89)} G ˙ tt = ∂ x E 1 N 2 μ E 1 ( R ∂ x E 1 + ∂ x M eff ) . (89)
where we have assumed regular matter, such that the scalar E ( μ ) \mathcal{E}^{(\mu)} E ( μ ) , which is proportional to the trace of the two-dimensional energy-momentum tensor, remains finite. Substituting E 1 = x 2 E^{1}=x^{2} E 1 = x 2 and M e f f = M ( x ; m ) M_{\mathrm{eff}}=M(x;m) M eff = M ( x ; m ) , we obtain
∂ x M e f f = ∂ x M ( x ; m ) ∣ m = m + ∂ m M ( x ; m ) ∂ x m . (90) \partial_x M_{\mathrm{eff}}=\left.\partial_x M(x;m)\right|_{m=m}+\partial_m M(x;m)\partial_x m. \tag*{(90)} ∂ x M eff = ∂ x M ( x ; m ) ∣ m = m + ∂ m M ( x ; m ) ∂ x m . (90)
Then, we get
G ˚ t t = N 2 μ E 1 ∂ x E 1 ( R ∂ x E 1 + ∂ x M ( x ; m ) ∣ m = m ) + N 2 μ E 1 ∂ x E 1 ∂ m M ∂ x m . (91) \begin{aligned}
\mathring{G}_{tt}=\frac{N^{2}\sqrt{\mu E^{1}}}{\partial_x E^{1}}\left(R\partial_x E^{1}+\left.\partial_x M(x;m)\right|_{m=m}\right) \\
&+\frac{N^{2}\sqrt{\mu E^{1}}}{\partial_x E^{1}}\partial_m M\partial_x m. \tag*{(91)}
\end{aligned} G ˚ tt = ∂ x E 1 N 2 μ E 1 ( R ∂ x E 1 + ∂ x M ( x ; m ) ∣ m = m ) + ∂ x E 1 N 2 μ E 1 ∂ m M ∂ x m . (91)
The first term is exactly G ˚ t t \mathring{G}_{tt} G ˚ tt evaluated on the static vacuum solution, with m m m replaced by m m m , and therefore vanishes. To simplify the second term, we note that from the second relation of Eqs. Eqs. (17) and (82)
N 2 μ ( ∂ m M ) = β 2 F H , (92) N^{2}\sqrt{\mu}(\partial_m M)=\beta^{2}\sqrt{FH}, \tag*{(92)} N 2 μ ( ∂ m M ) = β 2 F H , (92)
where we assume that ∂ m M > 0 \partial_m M>0 ∂ m M > 0 consistent with the convention adopted in Eq. (16) . These results thus lead to
G ˚ t t = β 2 F H 2 ∂ x m . (93) \mathring{G}_{tt}=\frac{\beta^{2}\sqrt{FH}}{2}\partial_x m. \tag*{(93)} G ˚ tt = 2 β 2 F H ∂ x m . (93)
For G ˚ t x \mathring{G}_{tx} G ˚ t x , from (69) , we could get
G ˚ t x = μ E 1 ∂ x E 1 4 ∂ s 2 2 M e f f ∂ t E 2 E 2 . (94) \mathring{G}_{tx}=\frac{\sqrt{\mu E^{1}}\partial_x E^{1}}{4\partial_{s_2}^{2}M_{\mathrm{eff}}}\frac{\partial_t E^{2}}{E^{2}}. \tag*{(94)} G ˚ t x = 4 ∂ s 2 2 M eff μ E 1 ∂ x E 1 E 2 ∂ t E 2 . (94)
According to Eq. (87) , we get
∂ t E 2 E 2 = ∂ M e f f Z 8 ∂ m M F ∂ t m (95) \frac{\partial_t E^{2}}{E^{2}}=\frac{\partial_{M_{\mathrm{eff}}}\mathcal{Z}}{8\partial_m M F}\partial_t m \tag*{(95)} E 2 ∂ t E 2 = 8 ∂ m M F ∂ M eff Z ∂ t m (95)
where we used that, according to the last relation in Eq. (17) ,
∂ M e f f Z = − 4 ( ∂ m F ∂ m M + 2 F ∂ m 2 M ) (96) \partial_{M_{\mathrm{eff}}}\mathcal{Z}=-4\left(\partial_m F\partial_m M+2F\partial_m^{2}M\right) \tag*{(96)} ∂ M eff Z = − 4 ( ∂ m F ∂ m M + 2 F ∂ m 2 M ) (96)
Moreover, according to Eq. (59) , we get
∂ s 2 2 M e f f = 1 8 E 1 μ ∂ M e f f Z . (97) \partial_{s_2}^{2}M_{\mathrm{eff}}=\frac{1}{8}E^{1}\mu\partial_{M_{\mathrm{eff}}}\mathcal{Z}. \tag*{(97)} ∂ s 2 2 M eff = 8 1 E 1 μ ∂ M eff Z . (97)
Substituting Eqs. (95) and (97) , as well as E 1 = x 2 E^{1}=x^{2} E 1 = x 2 and the second relation in Eq. (17) , into Eq. (94) , we finally get
G ˚ t x = 1 2 F H ∂ t m (98) \mathring{G}_{tx}=\frac{1}{2\sqrt{FH}}\partial_t m \tag*{(98)} G ˚ t x = 2 F H 1 ∂ t m (98)
Finally, we evaluate G ˚ x x \mathring{G}_{xx} G ˚ xx using Eq. (70) . To simplify the calculation, we first establish the following useful identity:
μ E 1 ∂ x E 1 4 E 2 ∂ s 2 2 M e f f + ∂ s 4 M e f f = 0. (99) \frac{\mu E^{1}\partial_x E^{1}}{4E^{2}\partial_{s_2}^{2}M_{\mathrm{eff}}}+\partial_{s_4}M_{\mathrm{eff}}=0. \tag*{(99)} 4 E 2 ∂ s 2 2 M eff μ E 1 ∂ x E 1 + ∂ s 4 M eff = 0. (99)
To begin with, the first equation in Eq. (66) yields
μ E 1 ∂ x E 1 4 E 2 ∂ s 2 2 M e f f + ∂ s 4 M e f f = ∂ s 2 M e f f ∂ s 2 ∂ s 4 M e f f ∂ s 2 2 M e f f (100) \frac{\mu E^{1}\partial_x E^{1}}{4E^{2}\partial_{s_2}^{2}M_{\mathrm{eff}}}+\partial_{s_4}M_{\mathrm{eff}}=\frac{\partial_{s_2}M_{\mathrm{eff}}\partial_{s_2}\partial_{s_4}M_{\mathrm{eff}}}{\partial_{s_2}^{2}M_{\mathrm{eff}}} \tag*{(100)} 4 E 2 ∂ s 2 2 M eff μ E 1 ∂ x E 1 + ∂ s 4 M eff = ∂ s 2 2 M eff ∂ s 2 M eff ∂ s 2 ∂ s 4 M eff (100)
On the other hand, evaluating the second equation in Eq. (51) , which is equivalent to Eq. (53) , on the metric (81), and noting that ∂ t E 1 = 0 \partial_t E^{1}=0 ∂ t E 1 = 0 , we obtain
∂ s 2 M e f f = 0. (101) \partial_{s_2}M_{\mathrm{eff}}=0. \tag*{(101)} ∂ s 2 M eff = 0. (101)
Then, Eq. (99) is just a consequence of substituting this result into (100).
Using Eq. (99) , we can simplify Eq. (70) to
G ˚ x x = ( E 2 ) 2 16 N μ E 1 ∂ s 2 2 M e f f [ 4 μ E 1 ∂ x E 1 ∂ x N ( E 2 ) 2 + 2 μ N E 1 ∂ s 1 Z − 4 μ N E 1 ( ∂ x E 1 ) 2 ( E 2 ) 2 ∂ M e f f R ] − ( E 2 ) 2 μ E 1 R + ( ∂ x E 1 ) 2 ∂ M e f f μ 8 μ N 2 ∂ s 2 2 M e f f G ˚ t t . (102) \begin{aligned}
\mathring{G}_{xx}={}&\frac{(E^{2})^{2}}{16N\sqrt{\mu E^{1}}\partial_{s_2}^{2}M_{\mathrm{eff}}}\left[\frac{4\mu E^{1}\partial_x E^{1}\partial_x N}{(E^{2})^{2}}+2\mu N E^{1}\partial_{s_1}\mathcal{Z}-\frac{4\mu N E^{1}(\partial_x E^{1})^{2}}{(E^{2})^{2}}\partial_{M_{\mathrm{eff}}}\mathcal{R}\right]\\
&-\frac{(E^{2})^{2}}{\sqrt{\mu E^{1}}}\mathcal{R}+\frac{(\partial_x E^{1})^{2}\partial_{M_{\mathrm{eff}}}\mu}{8\mu N^{2}\partial_{s_2}^{2}M_{\mathrm{eff}}}\mathring{G}_{tt}. \tag*{(102)}
\end{aligned} G ˚ xx = 16 N μ E 1 ∂ s 2 2 M eff ( E 2 ) 2 [ ( E 2 ) 2 4 μ E 1 ∂ x E 1 ∂ x N + 2 μ N E 1 ∂ s 1 Z − ( E 2 ) 2 4 μ N E 1 ( ∂ x E 1 ) 2 ∂ M eff R ] − μ E 1 ( E 2 ) 2 R + 8 μ N 2 ∂ s 2 2 M eff ( ∂ x E 1 ) 2 ∂ M eff μ G ˚ tt . (102)
To simplify this expression, let us first consider ∂ s 1 Z \partial_{s_1}\mathcal{Z} ∂ s 1 Z . According to the last equation in Eq. (17) , we get
2 x ∂ s 1 Z + ∂ M e f f Z ∂ x M ∣ m = m = − 4 ( ( ∂ m M ) 2 ∂ x F ∣ m = m + 2 F ∂ m M ∂ x ∂ m M ∣ m = m ) . (103) 2x\partial_{s_1}\mathcal{Z}+\left.\partial_{M_{\mathrm{eff}}}\mathcal{Z}\partial_x M\right|_{m=m}=-4\left((\partial_m M)^{2}\left.\partial_x F\right|_{m=m}+2F\partial_m M\left.\partial_x\partial_m M\right|_{m=m}\right). \tag*{(103)} 2 x ∂ s 1 Z + ∂ M eff Z ∂ x M ∣ m = m = − 4 ( ( ∂ m M ) 2 ∂ x F ∣ m = m + 2 F ∂ m M ∂ x ∂ m M ∣ m = m ) . (103)
Here, ∂ x O ∣ m = m \left.\partial_x O\right|_{m=m} ∂ x O ∣ m = m indicates that the derivative ∂ x \partial_x ∂ x is first evaluated with m m m held fixed, after which m m m is replaced by the spacetime-dependent function m ( t , x ) m(t,x) m ( t , x ) . Thus, ∂ x \partial_x ∂ x acts only on the explicit x x x dependence and not on m m m . Moreover, by the first equation in Eq. (17) , we have
∂ x ∂ m M ∣ m = m = − 2 x ∂ M e f f R ∂ m M (104) \left.\partial_x\partial_m M\right|_{m=m}=-2x\partial_{M_{\mathrm{eff}}}\mathcal{R}\partial_m M \tag*{(104)} ∂ x ∂ m M ∣ m = m = − 2 x ∂ M eff R ∂ m M (104)
Then, substituting Eqs. (97) and (104) into Eq. (103) yields
μ E 1 ( ∂ s 1 Z − 2 ( ∂ x E 1 ) 2 ( E 2 ) 2 ∂ M e f f R ) = − 2 x H ∂ x F ∣ m = m F + 8 ∂ s 2 2 M e f f R . (105) \mu E^{1}\left(\partial_{s_1}\mathcal{Z}-\frac{2(\partial_x E^{1})^{2}}{(E^{2})^{2}}\partial_{M_{\mathrm{eff}}}\mathcal{R}\right)=-\frac{2xH\partial_x F|_{m=m}}{F}+8\partial_{s_2}^{2}M_{\mathrm{eff}}\mathcal{R}. \tag*{(105)} μ E 1 ( ∂ s 1 Z − ( E 2 ) 2 2 ( ∂ x E 1 ) 2 ∂ M eff R ) = − F 2 x H ∂ x F ∣ m = m + 8 ∂ s 2 2 M eff R . (105)
where we have applied Eqs. (82) and (87) , and the first and second relations in Eq. (17) . With this result, we have
G ˚ x x = ( E 2 ) 2 16 N μ E 1 ∂ s 2 2 M e f f [ 4 μ E 1 ∂ x E 1 ∂ x N ( E 2 ) 2 − 4 N x H ∂ x F ∣ m = m F ] + ( ∂ x E 1 ) 2 ∂ M e f f μ 8 μ N 2 ∂ s 2 2 M e f f G ˚ t t . (106) \mathring{G}_{xx}=\frac{(E^{2})^{2}}{16N\sqrt{\mu E^{1}}\partial_{s_2}^{2}M_{\mathrm{eff}}}\left[\frac{4\mu E^{1}\partial_x E^{1}\partial_x N}{(E^{2})^{2}}-\frac{4NxH\partial_x F|_{m=m}}{F}\right]+\frac{(\partial_x E^{1})^{2}\partial_{M_{\mathrm{eff}}}\mu}{8\mu N^{2}\partial_{s_2}^{2}M_{\mathrm{eff}}}\mathring{G}_{tt}. \tag*{(106)} G ˚ xx = 16 N μ E 1 ∂ s 2 2 M eff ( E 2 ) 2 [ ( E 2 ) 2 4 μ E 1 ∂ x E 1 ∂ x N − F 4 N x H ∂ x F ∣ m = m ] + 8 μ N 2 ∂ s 2 2 M eff ( ∂ x E 1 ) 2 ∂ M eff μ G ˚ tt . (106)
To simplify the term involving ∂ M e f f μ \partial_{M_{\mathrm{eff}}}\mu ∂ M eff μ , we again use the relation given in the reconstruction equation (17) to get
∂ M e f f μ ∂ m M = μ [ ∂ m H H − ∂ m F F − 2 ∂ m 2 M ∂ m M ] . (107) \partial_{M_{\mathrm{eff}}}\mu\partial_m M=\mu\left[\frac{\partial_m H}{H}-\frac{\partial_m F}{F}-2\frac{\partial_m^{2}M}{\partial_m M}\right]. \tag*{(107)} ∂ M eff μ ∂ m M = μ [ H ∂ m H − F ∂ m F − 2 ∂ m M ∂ m 2 M ] . (107)
Moreover, the combination of Eqs. (82) and (87) , we get
1 ∂ s 2 2 M e f f = − 1 μ E 1 F ∂ m ( F ∂ m M ) (108) \frac{1}{\partial_{s_2}^{2}M_{\mathrm{eff}}}=-\frac{1}{\mu E^{1}\sqrt{F}\partial_m\left(\sqrt{F}\partial_m M\right)} \tag*{(108)} ∂ s 2 2 M eff 1 = − μ E 1 F ∂ m ( F ∂ m M ) 1 (108)
Substituting Eqs. (107) , (108) , (82) , (87) and (93) into Eq. (106) , we get
G ˚ x x = ∂ x m 2 F H − ∂ m M Ξ 2 H ∂ m ( F ∂ m M ) (109) \mathring{\mathcal{G}}_{xx}=\frac{\partial_x m}{2\sqrt{F H}}-\frac{\partial_m M\Xi}{2\sqrt{H}\partial_m\left(\sqrt{F}\partial_m M\right)} \tag*{(109)} G ˚ xx = 2 F H ∂ x m − 2 H ∂ m ( F ∂ m M ) ∂ m M Ξ (109)
where we introduced
Ξ = β x β + ∂ m ln F H ∂ x m (110) \Xi=\frac{\beta_x}{\beta}+\partial_m\ln\sqrt{F H}\partial_xm \tag*{(110)} Ξ = β β x + ∂ m ln F H ∂ x m (110)
for abbreviation.
To calculate G ˚ θ θ \mathring{\mathcal{G}}_{\theta\theta} G ˚ θ θ , we note that μ 2 = ∂ M e f f μ ∂ s 2 M e f f = 0 \mu_2=\partial_{M_{\mathrm{eff}}}\mu\partial_{s_2}M_{\mathrm{eff}}=0 μ 2 = ∂ M eff μ ∂ s 2 M eff = 0 due to Eq. (101) . Thus, we have
G ˚ θ θ = 1 2 x ∇ ˚ a G ˚ x a (111) \mathring{\mathcal{G}}_{\theta\theta}=\frac{1}{2x}\mathring{\nabla}_a\mathring{\mathcal{G}}^a_x \tag*{(111)} G ˚ θ θ = 2 x 1 ∇ ˚ a G ˚ x a (111)
leading to
G ˚ θ θ = ( ∂ m M ) 2 4 x β F H [ ∂ x ( β 2 F H ∂ x m β F H ) − ∂ t ( ∂ t m β F H ) ] − ∂ m M 4 β x H ∂ x [ β F ( ∂ m M ) 2 ∂ m ( F ∂ m M ) Ξ ] . (112) \mathring{\mathcal{G}}_{\theta\theta}=\frac{(\partial_m M)^2}{4x\beta}\sqrt{\frac{F}{H}}\left[\partial_x\left(\frac{\beta^2\sqrt{F H}\partial_xm}{\beta\sqrt{F H}}\right)-\partial_t\left(\frac{\partial_t m}{\beta F H}\right)\right]-\frac{\partial_m M}{4\beta x\sqrt{H}}\partial_x\left[\beta\frac{F(\partial_m M)^2}{\partial_m(\sqrt{F}\partial_m M)}\Xi\right]. \tag*{(112)} G ˚ θ θ = 4 x β ( ∂ m M ) 2 H F [ ∂ x ( β F H β 2 F H ∂ x m ) − ∂ t ( β F H ∂ t m ) ] − 4 β x H ∂ m M ∂ x [ β ∂ m ( F ∂ m M ) F ( ∂ m M ) 2 Ξ ] . (112)
Generalized Vaidya solution
In this section, we construct generalized Vaidya solutions for the case E ( μ ) = 0 \mathcal{E}^{(\mu)}=0 E ( μ ) = 0 . We emphasize that this condition concerns the matter sector rather than imposing an additional restriction on the gravitational theory. According to Eq. (26) , this condition implies that the trace of the four-dimensional energy-momentum tensor is entirely determined by the tangential pressure. A simple example satisfying this condition is null dust, whose energy-momentum tensor has a vanishing trace and no tangential pressure. Nevertheless, the following construction does not assume null dust as the matter source.
In general relativity, the Vaidya solution can be obtained by promoting the constant parameter m m m of the Schwarzschild metric to a function of a null coordinate. In a modified theory of gravity, however, such a promotion does not automatically yield a solution sourced by standard null dust. The resulting matter source depends on the modified gravitational dynamics and must therefore be determined from the corresponding field equations. Our aim in this section is to show that, within the reconstructed covariant effective theory, this procedure remarkably leads to precisely the standard Vaidya energy-momentum tensor.
Starting from a general spherically symmetric static metric (10), we define
d v ± = d t ± 1 F H d x . (113) dv_{\pm}=dt\pm\frac{1}{\sqrt{F H}}dx. \tag*{(113)} d v ± = d t ± F H 1 d x . (113)
In each coordinate patch, we may choose either ( v + , x ) (v_+,x) ( v + , x ) or ( v − , x ) (v_-,x) ( v − , x ) as the coordinates. The metric then takes the form
d s 2 = − F ( x , m ) d v ± 2 ± 2 α d v ± d x + x 2 d Ω 2 , (114) ds^2=-F(x,m)dv_{\pm}^2\pm2\sqrt{\alpha}dv_{\pm}dx+x^2d\Omega^2, \tag*{(114)} d s 2 = − F ( x , m ) d v ± 2 ± 2 α d v ± d x + x 2 d Ω 2 , (114)
where we define
α = F H . (115) \alpha=\frac{F}{H}. \tag*{(115)} α = H F . (115)
The Vaidya extension of this metric is obtained by promoting the constant mass parameter m m m to a function of the null coordinate v ± v_{\pm} v ± . The resulting metric takes the form
d s 2 = − F ( x , m ( v ± ) ) d v ± 2 ± 2 α ( x , m ( v ± ) ) d v ± d x + x 2 d Ω 2 . (116) ds^2=-F(x,m(v_{\pm}))dv_{\pm}^2\pm2\sqrt{\alpha(x,m(v_{\pm}))}dv_{\pm}dx+x^2d\Omega^2. \tag*{(116)} d s 2 = − F ( x , m ( v ± )) d v ± 2 ± 2 α ( x , m ( v ± )) d v ± d x + x 2 d Ω 2 . (116)
To apply the results derived in Sec. IV, we need to come back the coordinates corresponding to the diagonal areal gauge. The time coordinate t t t in this gauge is related to v ± v_{\pm} v ± by
d v ± = β d t ± 1 F H d x (117) dv_{\pm}=\beta dt\pm\frac{1}{\sqrt{F H}}dx \tag*{(117)} d v ± = β d t ± F H 1 d x (117)
where, unlike the static case with constant m m m , the factor β \beta β is necessary to ensure the right hand side is an exact differential. β \beta β thus satisfies
∂ x β = ± ∂ t ( 1 F H ) (118) \partial_x\beta=\pm\partial_t\left(\frac{1}{\sqrt{F H}}\right) \tag*{(118)} ∂ x β = ± ∂ t ( F H 1 ) (118)
With this equation, the metric in ( t , x ) (t,x) ( t , x ) coordinates returns to the form given in Eq. (81) explicitly with m = m ( v ± ) m=m(v_{\pm}) m = m ( v ± ) . Noting that F F F and H H H depends on m m m , we get from Eq. (118) that
Ξ = ∂ x β β + ∂ m ln ( F H ) ∂ x m = 0 (119) \Xi=\frac{\partial_x\beta}{\beta}+\partial_m\ln\left(\sqrt{F H}\right)\partial_xm=0 \tag*{(119)} Ξ = β ∂ x β + ∂ m ln ( F H ) ∂ x m = 0 (119)
Substituting this results into Eq. (109) , we get
G ˚ x x = ∂ x m 2 F H . (120) \mathring{\mathcal{G}}_{xx}=\frac{\partial_xm}{2\sqrt{F H}}. \tag*{(120)} G ˚ xx = 2 F H ∂ x m . (120)
To calculate G ˚ θ θ \mathring{\mathcal{G}}_{\theta\theta} G ˚ θ θ , we note that, in this case,
± β F H ∂ x m = ∂ t m . (121) \pm\beta\sqrt{F H}\partial_xm=\partial_t m. \tag*{(121)} ± β F H ∂ x m = ∂ t m . (121)
Thus, G ˚ θ θ \mathring{\mathcal{G}}_{\theta\theta} G ˚ θ θ in this case can be simplified to, with considering Ξ = 0 \Xi=0 Ξ = 0 ,
G ˚ θ θ = ± ( ∂ m M ) 2 4 x β F H [ ∂ x ( β ∂ t m β F H ) − ∂ t ( ∂ x m F H ) ] = ( ∂ m M ) 2 ∂ x m 4 x β F H [ ± ∂ x β − ∂ t ( 1 F H ) ] = 0. (122) \begin{aligned}
\mathring{\mathcal{G}}_{\theta\theta}
&=\frac{\pm(\partial_m M)^2}{4x\beta}\sqrt{\frac{F}{H}}\left[\partial_x\left(\frac{\beta\partial_t m}{\beta\sqrt{F H}}\right)-\partial_t\left(\frac{\partial_xm}{\sqrt{F H}}\right)\right]\\
&=\frac{(\partial_m M)^2\partial_xm}{4x\beta}\sqrt{\frac{F}{H}}\left[\pm\partial_x\beta-\partial_t\left(\frac{1}{\sqrt{F H}}\right)\right]\\
&=0. \tag*{(122)}
\end{aligned} G ˚ θ θ = 4 x β ± ( ∂ m M ) 2 H F [ ∂ x ( β F H β ∂ t m ) − ∂ t ( F H ∂ x m ) ] = 4 x β ( ∂ m M ) 2 ∂ x m H F [ ± ∂ x β − ∂ t ( F H 1 ) ] = 0. (122)
Combining all of the above results, including the expression of G ˚ t t \mathring{\mathcal{G}}_{tt} G ˚ tt and G ˚ t x \mathring{\mathcal{G}}_{tx} G ˚ t x given in (93) and (98) respectively, we finally get
T μ ν d x μ d x ν = G ˚ μ ν d x μ d x ν = ± 1 2 d m d v ± d v ± 2 . (123) \mathcal{T}_{\mu\nu}dx^\mu dx^\nu=\mathring{\mathcal{G}}_{\mu\nu}dx^\mu dx^\nu=\pm\frac{1}{2}\frac{dm}{dv_\pm}dv_\pm^2. \tag*{(123)} T μν d x μ d x ν = G ˚ μν d x μ d x ν = ± 2 1 d v ± d m d v ± 2 . (123)
Since E ( μ ) = 0 \mathcal{E}^{(\mu)}=0 E ( μ ) = 0 , the definition of T μ ν \mathcal{T}_{\mu\nu} T μν in Eq. (50) implies that T a b = T a b \mathcal{T}_{ab}=T_{ab} T ab = T ab and T ( θ ) = − 2 T θ θ = 0 \mathcal{T}^{(\theta)}=-2T_{\theta\theta}=0 T ( θ ) = − 2 T θ θ = 0 . Using Eq. (26) , we therefore obtain the four-dimensional energy-momentum tensor
T μ ν d x μ d x ν = ± 1 4 π x 2 d m d v ± d v ± 2 . (124) T_{\mu\nu}dx^\mu dx^\nu=\pm\frac{1}{4\pi x^2}\frac{dm}{dv_\pm}dv_\pm^2. \tag*{(124)} T μν d x μ d x ν = ± 4 π x 2 1 d v ± d m d v ± 2 . (124)
This is precisely the standard energy-momentum tensor of null dust, demonstrating that the generalized Vaidya metric constructed above provides a solution to the coupled system of gravity and null dust within our effective framework. Thus, standard null-dust sources can support dynamical black-hole geometries beyond general relativity, including regular black holes. Moreover, since the energy-momentum tensor of ingoing null dust is proportional to m ′ ( v + ) / x 2 m'(v_+)/x^2 m ′ ( v + ) / x 2 , Hawking-like mass loss, characterized by m ′ ( v + ) < 0 m'(v_+)<0 m ′ ( v + ) < 0 , necessarily entails a violation of the null energy condition. This suggests that quantum gravity corrections to the gravitational dynamics alone may be insufficient to account for Hawking-like evaporation.
Conclusion
In this work, building upon the covariant Hamiltonian gravity in spherical symmetry, we have derived the effective Einstein tensor G ˚ μ ν \mathring{\mathcal{G}}_{\mu\nu} G ˚ μν associated with a prescribed one-parameter family of static, spherically symmetric metrics. This allows the field equations of the framework to be expressed in an Einstein-like form, G ˚ μ ν = T μ ν \mathring{\mathcal{G}}_{\mu\nu}=\mathcal{T}_{\mu\nu} G ˚ μν = T μν , where T μ ν \mathcal{T}_{\mu\nu} T μν is the tensor related to the energy-momentum tensor of the matter source. The association is understood in the sense that G ˚ μ ν \mathring{\mathcal{G}}_{\mu\nu} G ˚ μν vanishes when evaluated on any metric in the prescribed family, demonstrating that these metrics are vacuum solutions of the reconstructed effective theory.
As an application, we construct generalized Vaidya solutions of the reconstructed effective theory. More precisely, starting from a given one-parameter family of static, spherically symmetric metrics characterized by a mass parameter m m m , we first transform the metrics into Eddington–Finkelstein coordinates and then promote m m m to a function m ( v ) m(v) m ( v ) of the null coordinate v v v . Remarkably, evaluating the effective Einstein tensor on the resulting metric yields a matter source whose energy-momentum tensor takes precisely the standard Vaidya form. This demonstrates that null dust can serve as the matter source for the dynamical extensions of the prescribed static geometries, including regular black holes. An immediate consequence of the above result is that Hawking-like mass loss in the ingoing Vaidya solution, characterized by m ′ ( v ) < 0 m'(v)<0 m ′ ( v ) < 0 , necessarily requires a violation of the null energy condition. This suggests that quantum gravity corrections to the gravitational dynamics alone may be insufficient to account for Hawking-like evaporation within our effective framework.
Acknowledgments
This work was supported in part by the National Key Research and Development Program of China Grant No. 2021YFC2203001 and in part by the NSFC (Grants No. 12275022, No. 12475046, and No. 12505055) and “the Fundamental Research Funds for the Central Universities”.