SPONTANEOUS SYMMETRY BREAKING IN CONFORMAL

WEYL GRAVITY

ALEX NELSON

Abstract. By spontaneously breaking conformal symmetry, a scalar field

emerges which we identify as the cosmological constant. The first instance
is in the Schwarzschild solution for the Weyl Gravity Field Equations. The
second is in the homogeneous and isotropic universe with fermionic and bosonic
matter. In the second case, an effective gravitational constant emerges, as well
as a positive cosmological constant, both give theoretical justification within
the conformal gravity framework of an accelerating universe.

1. Introduction
Recall, in General Relativity, a heuristic derivation usually begins with the New-
tonian gravitys Poisson equation
(1.1) 2 N = 4GN
where N is the Newtonian potential, GN is the gravitation constant, and is the
mass-density. It has a solution of
c0
(1.2) N (r) =
r
where c0 is constant. From here, one typically identifies the right hand side of Eq
(1.1) as the time-time component of stress-energy tensor, and the left hand side
is identified as the time-time component of the Ricci tensor. This is how most
approaches to gravity begin.
In conformal gravity, we begin with a different Poisson equation. Instead of a
second order one, we begin with a fourth order one [1]
(1.3) 4 B(r) = f (r)
which has the general solution for a spherical source
r R
Z Z R
1 c1
dxf (x)x2 dxf (x)x4 = e
e
(1.4) B(r) = c0 r + .
2 0 6r 0 r
Observe that when r  1, Eq (1.4) has e c1 /r be the dominant term and ec0 r 0.
Thus for small r, we can recover the Newtonian Poisson equation (1.1).
At first, this may be startling to see Eq (1.4) as being proposed for the gravita-
tional potential. It is counter-intuitive to propose adding an O(r) term, as we dont
observe it at small scales (dropping an apples behaves as being in a O(1/r) po-
tential!). However, at such scales, the potential for a fourth order Poisson equation
behaves as the potential for a second order one. Additionally, there is observational
problems with gravity that departs from a second order Poisson equation at large

Date: 29 November 2008.

1
2 ALEX NELSON

distances. So the fourth order approach modifies only what is expected at large
distances, and agrees with what we expect at small distances.
Here, we must intervene and confess that there is a terribly strong no-go theorem:
Ostrogradskis theorem (for a beautiful introduction, see section 2 of Woodard [2]).
With Lagrangians involving terms of second order (or higher) time derivatives of
the position term is unstable. Woodard notes that in Lagrangians of the form
L(qi , qj , qk ) there is not even any barrier to decay. Adding insult to injury, the
situation does not improve if we add in higher order derivatives!
However, Adler [3] has proposed recovering Einsteins General Relativity from
Spontaneous Symmetry Breaking1 and Zee [4] has proposed using spontaneously
breaking conformal invariance to give rise to masses in Weyl gravity. This is very
much analogous to how spontaneous symmetry breaking in the weak force generates
the Fermi constant. The aim of Adler and Zee is that in breaking symmetry, there
is a sort of macroscopic/low energy limit in which we can recover Newtonian
gravity (or some generalization of it).
If instead we consider conformal gravity as a theory in and of itself, as Mannheim
suggests, we find that from the solution for a static, spherically symmetric body
scale invariance is spontaneously broken to give rise to a nonzero cosmological
constant term. Further, if we consider working with a stress energy tensor involving
a spontaneous symmetry breaking boson (which is to be expected if the standard
model of particle physics is correct), we get a sort of induced conformal cosmology
model which avoids a Big Bang singularity. So the plan is to first examine the
case of the static, spherically symmetric gravitating body which resembles (up to
some negligibly small terms at the solar-system level) the Schwarzschild solution.
Then we will proceed to consider symmetry breaking at the cosmological scale
and review the consequences in conformal cosmology. Included is an appendix
which constitutes an extremely brief (five-minute) introduction to spontaneous
symmetry breaking.

2. Action Principle and Field Equations

One begins with the conformally-invariant fourth order action

Z
(2.1) IW = g d4 x gC C

where g is the coupling constant and C is the Weyl tensor. This action is
invariant under conformal transformations of the metric
g (x) e2(x) g (x)
(Originally Weyl considered in the action (2.1) to be used for the both the con-
formal and the electromagnetic gauge transformations. This way one has for the
covariant derivative g = A g where A is the electromagnetic 4-potential.
The problem with this approach is that conformal invariance implied the particles
are massless, which is observably false.) We can now simplify this Lagrangian a
bit.

1In all fairness, Mannheim credits Adlers work with recovering Einsteins general relativity,
but it appears that Adlers paper is more related to the role of spontaneous symmetry breaking
in the context of avoiding some problems in induced gravity like the logarithmic and quadratic
divergence of the effective Gravitation constant.
 SPONTANEOUS SYMMETRY BREAKING IN CONFORMAL WEYL GRAVITY 3

By plugging in the definition of the Riemann tensor, and recalling that any
contraction of any pair of indices of the Weyl tensor vanishes, we see
1
(2.2) R R = C C + 2R R R2
3
where R is the Ricci scalar. Rearranging terms, we have
1
(2.3) C C = R R 2R R + R2 .
3
Before beginning, note that the quantity [5, 6]

g R R 4R R + R2


(2.4)
is a total divergence. So instead of having our Lagrangian be

(2.5) L = gC C
we can equivalently use the Lagrangian

(2.6) L = 2 g(R R R2 /3)
since we would be working with an extra term (by Stokes theorem a surface inte-
gral), and by demanding the variation vanishes on the boundary the only nonzero
contribution would be this Lagrangian.
Now, De Witt [7] explicitly calculates out the equations of motion for two La-
grangians:

(2.7) L2 = gR2 , and L1 = gR R
Our Lagrangian is a linear combination of these two, so we use a linear combination
of the variation of their respective actions
(2.8)
S2 g g
= (R )+ R R R 2R R + R R
g 2 2
(where is the covariant derivative operator) and
S1
(2.9) = 2g R 2 R 2R R + g R2 /2
g
where S1 and S2 are the actions of the Lagrangians L1 and L2 respectively. In the
(2)
literature, these two quantities are typically referred to as W = S2 /g and
(1)
W = S1 /g . From them, we can construct the quantity
(2) 1 (1)
(2.10) 2g W = 2(W W )
3
which is precisely the variation of the conformal action. So we end up with the field
equations being
(2.11) 4g W = T
where T is the stress-energy tensor we all know and love.
Although this is an intimidating system of coupled fourth order, nonlinear par-
tial differential equations, there are a few solutions calculated out. Mannheim and
Kazanas [8] have computed the exact solution exterior to a static, spherically sym-
metric gravitating source, which is
(2 3)
(2.12) g00 = 1/grr = 1 3 + r kr2
r
 4 ALEX NELSON

where the parameters , and k are three dimensionful integrations constants which
appear in the solution but not in the equations of motion. They spontaneously break
the scale symmetry. This should look familiar as it resembles the Schwarzschild
solution with a cosmological constant
2m
(2.13) g00 = 1/grr = 1 + r2
3 r
in units where GN = 1 and c = 1. The only difference is a constant term and a
term that linearly depends on r.
Breaking symmetry Here we need to reiterate so one appreciates the beauty of the situation. In a
gives information about Lagrangian of the form (2.1) which has no boundary term or constant term added
cosmological constant in by hand, makes no assumptions about the cosmological constant, one can solve
for the spherically symmetry, static gravitating body and one naturally gets a term
which yields information about the cosmological constant and a term which breaks
symmetry to give masses to the massless particles.

3. Conformal Cosmology
We begin by thinking about breaking symmetry differently (read: in the naive
way) by considering2 the Lagrangian of matter conformally coupled to gravity [10,
11]
(3.1)
1
Z
1 2
IM = d4 x g 4

2 12 R + + i| (x)[{z+ (x)]} g
|

{z }
| {z } interaction
fermion
scalar

where (x) is the fermion spin connection, and g are the dimensionless coupling
constants, (x) is the (symmetry breaking) scalar field and is a fermionic field.
We will demonstrate that the scalar spontaneously breaks symmetry. Observe
that the potential term for the scalar field is
2 R
(3.2) V () = 4
12
we take its derivative
R
(3.3) V 0 () = 43
6
then set it to zero and solve for . The resulting value is
r
R
(3.4) v=
24
then we plug it back into the potential to find
r !  2
R2
 
R R R R
(3.5) V = =
24 24 12 24 576
which is nonzero, which implies that symmetry is spontaneously broken. Note
that if we included the fermion-scalar interaction term, the results would not have
changed as it would have been equivalent to adding a term linear in into the
2Note that this is for De-Sitter spacetime, to make this anti-de-Sitter spacetime we need to
change the sign of the 4 term. This has been calculated in [9].
 SPONTANEOUS SYMMETRY BREAKING IN CONFORMAL WEYL GRAVITY 5

potential (for explicit calculations refer to appendix B). (Observe the dependence
on R is directly proportional too.)
When the scalar field (x) in IM obtains a nonzero mass (which we are free to
rotate to some spacetime constant 0 due to conformal invariance), the fermion
then obeys the curved space Dirac equations
(3.6) i~ (x)( + (x)) = ~g0
and acquires a mass ~g0 . The scalar fields equation of motion is
R = 0.
(3.7) + 43 + g
6
The corresponding stress-energy tensor to (3.1) is

( + ) + 2 g
h
T = ~ i
3 6 3
g 2 g i
(3.8) + (R R) g 4
3 6 2
which can be rewritten grouping terms in a more elegant manner. If we think of

(3.9) u u = i~ ( + ) + ~
2
where is the pressure of an ideal fluid, u is thought of as the worldline (so it
satisfies g u u = 1), and
~ ~
(3.10) pu u = +
3 6
where p is the pressure of an ideal fluid, then the stress energy tensor may be
written as
 
1 ~
(3.11) T = ( + p)U U + pg 20 R g R g ~40 .
6 2
This is the right hand side of our fourth order field equations.
By working in an isotropic and homogeneous geometry, the left hand side of
(3.11) necessarily vanishes, giving us the equation
 
1 2 1
(3.12) R g R = ( + p)U U + pg g 4
6 2
Thus conformal cosmology looks like the standard cosmology with a perfect mat-
ter fluid and a nonzero cosmological constant with the important exception that Replace Newtons GN
Newtons constant has been replaced by an effective constant of the form with an effective one

3
(3.13) Geff = .
420
This is not Newtons constant as Cavendish measured, but instead a term which we
identify to be analagous to the Newton gravitational constant! Observe that as we
change scales, Geff changes in inverse proportion.
We can also identify the = 40 term as being a cosmological constant. Note Cosmological Constant
that this term really is effectively a cosmological constant since it is a homogeneous Emerges
and isotropic global scalar field. Observe that this cosmological constant is (1) scale
dependent (that is, quartic in ), (2) always positive (that is, we have de Sitter
spacetime, so gravity is repulsive but at this scale). The notion that the cosmological
constant is scale dependent is novel, but the important change is that the sign
6 ALEX NELSON

explains how gravity is repulsive instead of attractive. Due to the sign, there is
no initial singularity in this model. Instead the universe expands from a finite
minimum radius, and is not subject to the same problems that one encounters in
the standard cosmological model.
Despite finding gravity being globally repulsive, it is locally attractive. This
reconciles the use of the fourth order Poisson equation (1.3) which merely adds an
extra term linear in r (radial distance) to the gravitational potential that would be
negligible at local scales. It turns out that Mannheim et al [12] demonstrate the
empirical strength of such a proposition at the galactic level, but that is beyond
the scope of this article to review it too.

4. Conclusion
We introduced a different action which is based off of Weyls attempt to unify
gravity and electromagnetism. Instead of attempting such a unified field theory, we
observed that it has interesting gravitational properties.
The vacuum satisfies the Schwarzschild solution for general relativity with a
nonzero cosmological constant, plus some nonzero term and a term linear in r
negligibly small at the local scale. Due to these extra terms, the scale invariance
was spontaneously broken. This was purely accidental.
We also observed that when we solve the fourth order field equations for the
isotropic and homogeneous case, we end up breaking symmetry again. But in doing
so, we recover the standard cosmological model, and we explained why gravity is
accelerating within the framework of the Conformal gravity model. Further, we
have an effective gravitational constant that is scale dependent which allows gravity
to be repulsive globally but (due to inhomogeneities in the scalar field) is locally
attractive. This is consistent with the first investigation of spontaneous symmetry
breaking in solving the static, spherically symmetric bodys gravitational field as
locally (for small enough r) resembling Schwarzschilds solution.
Observe that this is really nothing surprising, since this is just another version
of the Brans-Dicke theory. The Brans-Dicke action is
4
Z  
1
(4.1) I= d x g R + Lmatter
16
one can rearrange it by introducing 2 = to look like

Z
1
d4 x g 2 R 4 + Lmatter


(4.2) I=
16
which resembles the action in Eq (3.1). What the Brans-Dicke theory effectively
does is replace k = 16G/c4 with a scalar field . We did something similar, except
our scalar field spontaneously broke the scale invariance (so, analogously, we had a
bare minimum value for k) which gave rise to a cosmological constant in addition to
recovering the standard cosmological model. Further, we used covariant derivatives
instead of partial derivatives, so we would need to include in the Lmatter the extra
terms, the 4 term, and the coupling to matter. Nonetheless, the cosmological
constant naturally emerges when we break symmetry.

Appendix A. Spontaneous Symmetry Breaking

Spontaneous symmetry breaking occurs whenever a given field in a given La-
grangian has a nonzero vacuum expectation value. Why exactly is this breaking
 SPONTANEOUS SYMMETRY BREAKING IN CONFORMAL WEYL GRAVITY 7

the symmetry? Well, the Lagrangian appears symmetric under a symmetry group,
but its vacuum state fails to be symmetric. The system no longer behaves sym-
metrically. So we went from symmetry to no symmetry due to a nonzero vacuum
expectation value. It came about from condensed matter physics (see [13] for ap-
plications of it in condensed matter physics) but has since been applied to quantum
field theory and particle physics (see [14] for examples in particle physics).
Consider the scalar Lagrangian given by

1 1
(A.1) L= ( )2 + 2 2 4
|2 {z } |2 {z 4! }
kinetic term potential term

where is the scalar field, is a sort of mass parameter, and is the coupling.
Observe that there is a symmetry of (a discrete symmetry). We can think
of the potential as being

1
(A.2) V () = 2 2 + 4
2 4!
which has extrema when its derivative is zero. There are two, given by
r
6
(A.3) 0 = v =

where the constant v is the vacuum expectation value.

We can then write

(A.4) (x) = v + (x)

and then rewrite the Lagrangian as

r
1 1 3 4
(A.5) L = ( )2 (22 ) 2
2 2 6 4!

where we dropped the constant terms. We see that the symmetry is no

longer identifiable.

Appendix B. Proof of Spontaneous Symmetry Breaking with extra

Linear Term
The potential we are investigating has the form

2 R
(B.1) V () = c + 4
12
where c is some constant term, so its first derivative would be

R
(B.2) V 0 () = c + 43 .
6
8 ALEX NELSON

By the fundamental theorem of algebra, it has exactly three roots at the values

 p 2/3
3
2 3 (5832c2 R3 ) 108c2
2R +
v1 = q p
622/3 3 2 3 (5832c2 R3 ) 108c2

 p 2/3
2 3 2R + 1 i 3 2 3 (5832c2 R3 ) 108c2
v2 = q p
1222/3 3 2 3 (5832c2 R3 ) 108c2

 p 2/3
1+i 3 2 3 (5832c2 R3 ) 108c2 2(1)2/3 3 2R
v3 = q p .
1222/3 3 2 3 (5832c2 R3 ) 108c2

We can plug each of these into the function with the extra linear term and find


 p 2/3 4
3 3 2 3 2
2R + 2 (5832c R ) 108c
V (v1 ) =  p 4/3
518422/3 3 2 3 (5832c2 R3 ) 108c2

 p 2/3 2
3 3 2 3 2
R 2R + 2 (5832c R ) 108c
+  p 2/3
864 3 22 2 3 (5832c2 R3 ) 108c2
 p
 2/3 
3 3 2 3 2
c 2R + 2 (5832c R ) 108c
q p
622/3 3 2 3 (5832c2 R3 ) 108c2

4

 p
 2/3
2 3 2R + 1 i 3 2 3 (5832c2 R3 ) 108c2
V (v2 ) =  p 4/3
8294422/3 3 2 3 (5832c2 R3 ) 108c2


 p 2/3 2
R 2 3 2R + 1 i 3 2 3 (5832c2 R3 ) 108c2
+  p 2/3
3456 3 22 2 3 (5832c2 R3 ) 108c2

 p
 2/3 
c 2 3 2R + 1 i 3 2 3 (5832c2 R3 ) 108c2
+ q p
1222/3 3 2 3 (5832c2 R3 ) 108c2
 SPONTANEOUS SYMMETRY BREAKING IN CONFORMAL WEYL GRAVITY 9

4

 p
 2/3
3 2 3 2 2/3 3
1+i 3 2 (5832c R ) 108c 2(1) 2R
V (v3 ) =  p 4/3
8294422/3 3 2 3 (5832c2 R3 ) 108c2
2

 p
 2/3
3 2 3 2 2/3 3
R 1+i 3 2 (5832c R ) 108c 2(1) 2R
+  p 2/3
3456 3 22 2 3 (5832c2 R3 ) 108c2

 p
 2/3 
3 2 3 2 2/3 3
c 1+i 3 2 (5832c R ) 108c 2(1) 2R
+ q p .
1222/3 3 2 3 (5832c2 R3 ) 108c2
Observe that these are nonzero quantities, just a whole lot messier than
neglecting that linear term.

References
[1] P. D. Mannheim, Four-dimensional conformal gravity, confinement, and galactic rotation
curves, arXiv:gr-qc/9407010.
[2] R. P. Woodard, Avoiding dark energy with 1/R modifications of gravity, Lect. Notes
Phys. 720 (2007) 403433, arXiv:astro-ph/0601672.
[3] S. L. Adler, Einstein gravity as a symmetry-breaking effect in quantum field theory,
Reviews of Modern Physics 54 (1982) 729766.
http://link.aps.org/abstract/RMP/v54/p729.
[4] A. Zee, Einstein gravity emerging from quantum Weyl gravity, Annals of Physics 151
(1983) 431443. http://www.sciencedirect.com/science/article/B6WB1-4DF54GS-24B/2/
ed008688cccdcdc295fd25144002d614.
[5] D. Kazanas and P. D. Mannheim, General Structure of the Gravitational Equations of
Motion in Conformal Weyl Gravity, Astrophys. J. Suppl. 76 (1991) 431453.
[6] C. Lanczos, A Remarkable property of the Riemann-Christoffel tensor in four dimensions,
Annals Math. 39 (1938) 842850.
[7] B. D. Witt, Dynamical theory of groups and fields, in Relativity, Groups and Topology,
C. DeWitt and B. DeWitt, eds., pp. 587822. 1964.
[8] P. D. Mannheim and D. Kazanas, Exact Vacuum Solution to Conformal Weyl Gravity and
Galactic Rotation Curves, Astrophys. J. 342 (1989) 635638.
[9] A. Edery, L. Fabbri, and M. B. Paranjape, Spontaneous breaking of conformal invariance
in theories of conformally coupled matter and Weyl gravity, Class. Quant. Grav. 23 (2006)
64096423, arXiv:hep-th/0603131.
[10] P. D. Mannheim, Cosmic acceleration as the solution to the cosmological constant
problem, Astrophys. J. 561 (2001) 112, arXiv:astro-ph/9910093.
[11] P. D. Mannheim, Conformal Gravity Challenges String Theory, arXiv:0707.2283
[hep-th].
[12] P. D. Mannheim and J. Kmetko, Linear potentials and galactic rotation curves - detailed
fitting, arXiv:astro-ph/9602094.
[13] X.-G. Wen, Quantum Field Theory of Many-Body Systems. Oxford Graduate Texts, 2008.
[14] M. E. Peskin and D. V. Schroeder, An Introduction to Quantum Field Theory. Westview
Publisher, 1995.
E-mail address: [email protected]