A New Independent Limit on the Cosmological Constant/Dark Energy from the

Relativistic Bending of Light by Galaxies and Clusters of Galaxies

Mustapha Ishak∗ , Wolfgang Rindler, Jason Dossett, Jacob Moldenhauer, Chris Allison
Department of Physics, The University of Texas at Dallas, Richardson, TX 75083, USA
(Dated: December 6, 2007)
We derive new limits on the value of the cosmological constant, Λ, based on the Einstein bending
of light by systems where the lens is a distant galaxy or a cluster of galaxies. We use an amended
lens equation in which the contribution of Λ to the Einstein deflection angle is taken into account
and use observations of Einstein radii around several lens systems. We use in our calculations a
Schwarzschild-de Sitter vacuole exactly matched into a Friedmann-Robertson-Walker background
and show that a Λ-contribution term appears in the deflection angle within the lens equation. We
find that the contribution of the Λ-term to the bending angle is larger than the second-order term
arXiv:0710.4726v2 [astro-ph] 6 Dec 2007

for many lens systems. Using these observations of bending angles, we derive new limits on the
value of Λ. These limits constitute the best observational upper bound on Λ after cosmological con-
straints and are only two orders of magnitude away from the value determined by those cosmological
constraints.

PACS numbers: 95.30.Sf,98.80.Es,98.62.Sb

I. INTRODUCTION II. THE BENDING ANGLE IN THE PRESENCE

OF A COSMOLOGICAL CONSTANT

Cosmic acceleration and the dark energy associated We outline here the main steps of the calculation of
with it constitute one of the most important and chal- [3] and expand it using the second-order terms for the
lenging current problems in cosmology and all physics, solution of the null geodesic equation. Let us consider
see for example the reviews [1] and references therein. the Schwarzschild-de Sitter (SdS) metric [12]
The cosmological constant, Λ, is among the favored can-
didates responsible for this acceleration. Current con- ds2 = f (r)dt2 − f (r)−1 dr2 − r2 (dθ2 + sin2 (θ)dφ2 ) (1)
straints on Λ are coming from cosmology, see e.g. [2],
and it is important to obtain constraints or limits from where
other astrophysical observations.
2m Λr2
Very recently, the authors of reference [3] demon- f (r) ≡ 1 − − , (2)
strated that, contrarily to previous claims (e.g. [4, 5, r 3
6, 7, 8, 9]), when the geometry of the Schwarzschild- and where we use relativistic units (c = G = 1), m being
de Sitter spacetime is taken into account, the cosmolog- the mass of the central object.
ical constant does contribute to the light-bending and As shown in many text books, e.g. [13, 15], the null
the Einstein deflection angle. See also the discussions in geodesic equation in SdS spacetime is given exactly by
[10, 11].
In this Letter, we incorporate that result into the d2 u
+ u = 3mu2 , (u ≡ 1/r). (3)
broadly used lens equation and then apply it to current dφ2
observations of Einstein radii around distant galaxies and
clusters of galaxies. Using observational data of a se- In the usual way, the null orbit is described as a pertur-
lected list of Einstein radii around clusters and galaxies, bation of the undeflected line (i.e. the solution of equa-
we show that the contribution of the cosmological con- tion (3) without the RHS)
stant to the bending angle can be larger than the second-
r sin(φ) = R. (4)
order term of the Einstein bending angle. These new re-
sults allow us to put new independent limits on the value
After substitution of (4) into (3), one obtains the follow-
of the cosmological constant based on the observations
ing equation for u (Eq. (11.64) in [13])
of the bending angle by galaxies and clusters of galaxies.
These limits provide the best observational upper bound 1 sin(φ) 3m  1 
on Λ after cosmological constraints and are only two or- =u= + 1 + cos(2φ) , (5)
r R 2R2 3
ders of magnitude away from the value determined by
those cosmological constraints. where R is a constant related to the physically meaningful
area distance r0 of closest approach (when φ = π/2) by

1 1 m
∗ Electronic address: mishak@utdallas.edu
= + 2. (6)
r0 R R
 2

Apparent Position

αDLS α
y
θDOS h t Ra Lig
ht R
Lig ay
S R
θ O
βDOS β
φ
Coordinate Baseline α/2
Baseline L
SE Observer ’s
Source Position DLS DOL
for Einstein Ring
( β=0) DOS

FIG. 1: The lens equation geometry. Observer, lens, and source are at O, L, and S, respectively. The position of the unlensed
source is at an angle β, the apparent position is at the angle θ and the deflection angle is α. The distance from the observer
to the source is DOS , from the observer to the lens is DOL , and from the lens to the source is DLS . The angle φ is as shown
on the figure. As usual, the lens equation follows from the geometry as θDOS = βDOS + αDLS .

Other authors, see for example [14, 15], use the im- where u ≡ r1 and u0 ≡ R1 . Substituting this into equation
pact parameter b to discuss the bending of light in (3) and collecting terms of equal powers of M u0 gives the
Schwarzschild spacetime, but SdS spacetime is not following two equations:
asymptotically flat and one needs to define another pa-
d2 δu1
rameter such as R. As shown in [3], the contribution of + δu1 = 3 sin2 φ (11)
Λ to the bending angle comes from the spacetime metric dφ2
itself, independently of the parameterization of the null
d2 δu2
geodesic equation. + δu2 = 6δu1 sin φ. (12)
It was shown in [3] that the angle θ of our Figure 1 dφ2
(denoted by ψ in [3]) is given by Solving (11) and (12) for δu1 and δu2 and substituting
them into (10) gives the solution
f (r)1/2 r
tan(θ) = . (7) 1 sin φ 3m  cos 2φ  3m2 
|dr/dφ| = + 1 + + 10π cos φ −
r R 2R2 3 16R3
with f (r) as in Eq.(2) above (f (r) is α(r) in [3]) and 20φ cos φ − sin 3φ . (13)

dr mr2 r2 Now, we differentiate (13) and multiply by r2 to obtain

= 2 sin(2φ) − cos(φ) (8)
dφ R R dr r2 mr2 15m2 r2 
= − cos φ + 2 sin 2φ + cos φ +
dφ R R 4R3
to lowest order. The total bending angle α (at coordinate
φ = 0) was found in [3] to be 3 π 
cos 3φ + ( − φ) sin φ . (14)
20 2
m ΛR3 After some manipulation, it follows from (7) and (14)
α≈4 − . (9)
R 6m that the total bending angle (at φ = 0) to the third-order
is given by
to first order in m/R. This result shows that a positive
Λ diminishes α, as might well be expected from the re- m 15π m2 305 m3 ΛR3
α≈4 + + − . (15)
pulsive effect of Λ. The first term in (9) is simply the R 4 R2 12 R3 6m
classical Einstein bending angle to first order. The coefficients for the first and second-order terms in
Now, since we plan to compare to observations, it is this expansion are the same as the ones in the expansion
useful to expand the calculation to higher orders includ- in terms of the impact parameter b, see e.g. [17], that is
ing the second-order solution to the null geodesic equa- used for the asymptotically flat Schwarzschild spacetime.
tion. In the usual way, see for example [16], we write In the next section, we put our results into an observa-
tional context using systems where the lens is a galaxy
u = u0 [sin(φ) + (mu0 )δu1 + (mu0 )2 δu2 ] (10) or a cluster of galaxies.
 3

III.OBSERVATIONS OF EINSTEIN-RADII Now, equation (18) yields, to the smallest order, φb =

AND THE CONTRIBUTION OF THE R/rb , so we can finally write from (21)
COSMOLOGICAL CONSTANT TO THE
DEFLECTION 4m ΛRrb
α≈ − (22)
R 3
As one might expect, while the cosmological constant
where R is related to the closest approach by equation (6)
has a very negligible effect on small scales this is not
and rb is the boundary radius between SdS and FLRW,
the case at the level of distant galaxies and clusters of
and is given by equation (17). Using equations (13) and
galaxies. In this section, we evaluate the contribution of
(14), we expand the result immediately to
the cosmological constant to the bending of light using
observations of large Einstein radii where the lens is a m 15π m2 305 m3 ΛRrb
distant galaxy or cluster. α≈4 + 2
+ 3
− (23)
R 4 R 12 R 3
Equations (9) and (15) above were derived based on a
source and an observer located in a Schwarzschild-de Sit- Next, the second method that we employ to calculate
ter background. We will derive here the equivalent equa- α is based on an approximate construction that is fre-
tion in a Friedmann-Lemaitre-Robertson-Walker back- quently used in gravitational lensing literature and where
ground (FLRW). For that, we consider a Schwarzschild- the lens (inhomogeneity) in an FLRW is represented by
de Sitter vacuole exactly embedded into an FLRW space- a Newtonian potential embedded in a post-minkowskian
time using the Isreal-Darmois formalism [18, 19]. The line element or a post-FLRW line element, see for exam-
results for the boundaries are simple and well-known in ple [22]. Then, following for example references [23, 24],
the literature, see for example [20], and are given by the the deflection angle is given by
following two conditions: Z +xb
rb in SdS = a(t) rb in F LRW (16) α=2 ∇⊥ Φ(x, y, z)dx (24)
−xb
where rb stands for the coordinate radius of the vacuole, where ∇⊥ ≡ ∇ − ∇k is the gradient transverse to the
and path, see for example [24], and
4π 3
m SdS = r ρmatter in F LRW . (17) m Λr̃2
3 b in SdS Φ(r̃) = − − . (25)
r̃ 6
Thus, for a given cluster mass, equation (17) provides
a boundary radius where the spacetime transitions from
p
At the boundary, xb = rb2 − r̃02 and a straightforward
a SdS spacetime to an FLRW background. We specifi- integration yields the smallest-order terms in m/r̃0 and
cally assume that all the light-bending occurs in the SdS Λ as
vacuole according to our previous formulae, and that
once the light transitions out of the vacuole and into 4m Λr̃0 rb
α≈ −4 (26)
FLRW spacetime, all bending stops. We use two meth- r̃0 3
ods in order to calculate the deflection angle and the
where the factor 4 in the second term is due to the dif-
Λ-contribution term.
ference between the Schwarzschild coordinates and the
In the first method, we follow the same calculation as
coordinates used in the approximate post-minkowskian
in Rindler and Ishak [3]. For a small angle φb at the
construction, see the discussion in [25].
boundary, equation (5) gives
As expected, independently of the method used, the
1 φb 2m contribution of Λ to the deflection angle is established.
u= = + 2 (18) Finally, following the usual procedure, see e.g. [21, 22,
rb R R
24], we put our results within the lens equation which is
and equation (8) gives given from the geometry (see Figure 1) and small-angle
relations as follows
rb2  2φb m 
|A| = 1− . (19)
R R θDOS = βDOS + αDLS (27)
Next, inserting (18) and (19) into equation (7) yields after
or in the familiar form
a few steps
DLS
2m Λφb rb2 θ =β+α (28)
θ ≈ tan θ ≈ φb + − + higher-order terms. (20) DOS
R 6
The bending angle, α, is given, to the smallest order in where all the quantities are as defined in Figure 1 and
m/R and Λ, by the angular-diameter distance is given by
Z z
α 2m Λφb rb2 c dz ′
≈ θ − φb ≈ − . (21) D(z) = (29)
H0 (1 + z) 0
p
2 R 6 Ωm (1 + z ′ )3 + ΩΛ
 4

Cluster or galaxy Einstein Mass in 1st Order 2nd Order Λ-term Ratio-1 Ratio-2 Upper Limit
name and references Radius (Kpc) Msun h−1 term (rads) term (rads) (rads) 1st/Λ-term Λ-term/2nd on Λ (cm−2 )
Abell 2744 [28, 33] 96.4 1.97 × 1013 5.53E-05 2.25E-09 2.20E-08 2.51E+03 9.78 3.24E-54
SDSS J1004+4112 [30] 110.0 4.26 × 1013 1.05E-04 8.06E-09 3.25E-08 3.22E+03 4.03 4.16E-54
3C 295 [26] 127.7 7.1 × 1013 1.50E-04 1.66E-08 4.47E-08 3.36E+03 2.69 4.33E-54
Abell 1689 [28, 29] 138.2 9.36 × 1013 1.88E-04 2.61E-08 5.35E-08 3.52E+03 2.05 4.54E-54
Abell 2219L [28, 31] 86.3 3.22 × 1013 1.01E-04 7.47E-09 2.32E-08 4.34E+03 3.10 5.60E-54
AC 114 [28, 32] 54.6 9.23 × 1012 4.57E-05 1.54E-09 9.68E-09 4.72E+03 6.30 6.09E-54

TABLE I: Contributions of the cosmological constant to the Einstein bending angle by distant clusters of galaxies. Column-8
shows that the Λ-term contribution is larger than the second-order term in the Einstein bending angle for these lens systems.
The last column shows limits on the cosmological constant based on observations of the bending angle. These limits provide the
best upper bound on Λ after cosmological constraints and are only two orders of magnitude away from the value determined for
Λ by those cosmological constraints, i.e. 1.29 10−56 cm−2 . Previously, the best upper bound after cosmology was determined
from planetary or stellar systems and is Λ ≤ 10−46 cm−2 , see [6, 8] and references therein.

where, for a spatially flat cosmology, Ωm = 0.27, ΩΛ = follows that

0.73, and H0 = 71km/s/M pc.
Thanks to the advancement of observational tech- 3 ∆α
Λ≤ . (30)
niques, one can find in the literature a number of distant R rb
galaxies and clusters of galaxies that are lenses with large For example, with ∆α = 10%, we find from the system
Einstein radii, making them very interesting for apply- Abell 2744 [28, 33] that
ing our results. The selected systems are shown in Table
1 along with our evaluation of the deflection first-order Λ ≤ 3.24 10−54 cm−2 . (31)
term, the second-order term, and the Λ-term, and some
of their ratios. Despite the smallness of the cosmological
The other limits are in Table 1. Remarkably, these limits
constant, Λ, we find that the Einstein first-order term in
are the best observational upper bound on the value of Λ
the bending angle due to these systems is only by some
after cosmological constraints and are only two orders of
103 bigger than the Λ-term. Interestingly, we find that for
magnitude away from the value determined from cosmol-
the lens systems in Table 1, the contribution of the cos-
ogy. Previously, the best upper bound after cosmology
mological constant term is larger than the second-order
was provided from planetary or stellar systems and is
term of the Einstein bending angle.
Λ ≤ 10−46 cm−2 , see for example [6, 8] and references
therein.
IV. A NEW LIMIT ON THE COSMOLOGICAL In conclusion, we showed that a Λ-contribution term
CONSTANT FROM LIGHT-BENDING appears in the deflection angle within the lens equation.
This contribution is larger than the second-order term in
the Einstein bending angle for many cluster lens systems.
From cosmology (e.g. using supernova magnitude-
These results allow us to put new limits on the cosmolog-
redshift relation and the Cosmic Microwave Background
ical constant based on observations of the bending angle
Radiation), the value of the cosmological constant, Λ,
by galaxies and clusters of galaxies. These limits provide
is found to be about 1.29 10−56 cm−2 (using H0 = 71
the best upper bound on Λ after cosmological constraints
km/s/Mpc and ΩΛ = 0.73, see e.g. [2, 34]). It is very de-
and are only two orders of magnitude away from the value
sirable to obtain other limits on Λ that come from other
determined for Λ from those cosmological constraints.
astrophysical constraints. As we show, when we consider
the uncertainty in the measurements of the bending angle
(which is around ∆α ∼ 5-10% for several of the systems
considered in Table 1), we find that the bending angle Acknowledgments
due to distant galaxies and clusters can provide inter-
esting limits on the value of the cosmological constant. The authors thank C. Kochanek and I. Browne for use-
Indeed, if the contribution of Λ cannot exceed the un- ful comments about some lens systems. M.I. acknowl-
certainty in the bending angle for these system, then it edges partial support from the Hoblitzelle Foundation.

[1] S. Weinberg Rev. Mod. Phys., 61, 1 (1989); M.S. Turner, sky Int.J.Mod.Phys. D9, 373 (2000); S.M. Carroll, Liv-
Phys. Rep., 333, 619 (2000); V. Sahni, A. Starobin- ing Reviews in Relativity, 4, 1 (2001); T. Padmanab-

han, Phys. Rep., 380, 235 (2003); P.J.E. Peebles and
B. Ratra, Rev. Mod. Phys. 75, 559 (2003); A. Upadhye,
M. Ishak, P. J. Steinhardt , Phys. Rev. D 72, 063501
(2005); A. Albrecht et al, Report of the Dark Energy Task
Force astro-ph/0609591 (2006). M. Ishak, Foundations of
Physics Journal, Vol.37, No 10, 1470 (2007).
[2] A. G. Riess, et al., Astron. J. 116, 1009-1038 (1998);
S. Perlmutter, et al., Astrophys. J. 517, 565-586 (1999);
R. A. Knop, et al., Astrophys. J. 598, 102-137 (2003);
A. G. Riess, et al., Astrophys. J. 607, 665-687 (2004);
C. L. Bennett, et al., Astrophys. J. Suppl. Ser. 148, 1
(2003); D. N. Spergel, et al., Astrophys. J. Suppl. Ser. ,
175 (2003); L. Page et al., Astrophys. J. Suppl. Ser. 148,
2333 (2003). Seljak et al., Phys.Rev. D71, 103515 (2005);
M. Tegmark, et al., Astrophys. J. 606, 702-740 (2004).
D.N. Spergel , et al., Astrophys.J.Suppl. 170, 377 (2007).
[3] W. Rindler, M. Ishak, Phys. Rev. D 76 043006 (2007).
[4] N.J. Islam, Phys. Lett. A 97, 239 (1983).
[5] W. H. C. Freire, V. B. Bezerra, J. A. S. Lima, Gen. Rel-
ativ. Gravit. 33, 1407 (2001).
[6] V. Kagramanova, J. Kunz, C. Lammerzahl, Phys.Lett. B
634 465-470 (2006).
[7] F. Finelli, M. Galaverni, A. Gruppuso, Phys.Rev. D 75,
043003 (2007).
[8] M. Sereno, Ph. Jetzer, Phys. Rev. D 73, 063004 (2006).
[9] We clarify here that the work by (A. W. Kerr, J. C.
Hauck, B. Mashhoon, Class. Quant. Grav., 20, 2727
(2003)), cited in our previous paper [3], studied the in-
fluence of the cosmological constant on the time delay.
[10] K. Lake, arXiv:0711.0673.
[11] M. Sereno, arXiv:0711.1802.
[12] F. Kottler, Ann. Phys. 361, 401 (1918).
[13] W. Rindler, Relativity: Special, General, and Cosmolog-
ical, Second Edition (Oxford University Press, 2006).
[14] R. Wald General Relativity (University of Chicago Press,
1984).
[15] C.W. Misner, K.S. Thorne and J.A. Wheeler, Gravitation
(Freeman, San Francisco 1973).
[16] J. Bodenner, C.M. Will, Am. J. Phys., Vol. 71, No. 8,
770 (2003).
5
[17] C. R. Keeton, A. O. Petters Phys.Rev. D72, 104006
(2005).
[18] Darmois G (1927) Mémorial de Sciences Mathématiques,
Fascicule XXV, “Les equations de la gravitation ein-
steinienne”, Chapitre V.
[19] Israel W (1966) Nuovo Cim B 44, 1. Erratum 48, 463.
[20] Einstein A. and Strauss E.G., Rev. Mod. Phys. 17,
120.(1945), erratum: ibid 18, 148 (1946). Schucking, E.,
Z. Phys. 137, 595 (1954); and a long list of references
in references Krasinski A., Inhomogeneous Cosmological
Models (Cambridge University Press, Cambridge, 1997).
[21] Y. Mellier, Ann. Rev. Astron. Astrophys. 37, 127 (1999).
[22] M. Bartelmann, P. Schneider Phys.Rept. 340 (2001) 291-
472.
[23] T. Pyne, M. Birkinshaw Astrophys.J. 458 (1996) 46.
[24] S. Carroll, Spacetime and Geometry: An Introduction
to General Relativity (Addison Wesley, San Fransisco,
2004).
[25] In [16], it is was shown how differences in coordinate sys-
tems lead to differences between the coefficients of sec-
ond and higher-order terms. As it was argued there, the
Schwarzschild coordinate r is the physically meaningful
because it represents the observer area distance.
[26] M. Wold, M. Lacy, H. Dahle, P.B. Lilje, S. E. Ridgway,
Mon.Not.Roy.Astron.Soc. 335, 1017 (2002).
[27] J. Miralda-Escude, A. Babul, Astrophys. Jour., Vol. 449,
18, (1995).
[28] S. Allen, M.N.R.A.S, 296, 392, (1998).
[29] M. Limousin et al. Astrophys. Jour., 668:643666, (2007).
[30] K. Sharon, et al. Astrophys. Jour. 629, L73 (2005); N.
Ota, et al., Astrophys. Jour., 647, 215 (2006).
[31] I. Smail, D. Hogg, R. Blanford, J. Cohen, A. Edge, S.
Djorgovski, MNRAS, 277, 1 (1995).
[32] I. Smail, W. Couch, R. Ellis, R. Sharples, Astrophys.
Jour., 440, 501 (1995).
[33] I. Smail,R. Ellis, M. Fitchett, H. Norgaard-Nielsen,
L.Hansen, H. Jorgensen, M.N.R.A.S, 252, 19 (1991).
[34] W. Rindler, Astrophys. J. 157, L147 (1969).