This paper was converted on www.awesomepapers.org from LaTeX by an anonymous user.
Want to know more? Visit the Converter page.

Nonlogarithmic divergence of a deflection angle by a marginally unstable photon sphere of the Damour-Solodukhin wormhole in a strong deflection limit

Naoki Tsukamoto1 tsukamoto@rikkyo.ac.jp 1Department of General Science and Education, National Institute of Technology, Hachinohe College, Aomori 039-1192, Japan
Abstract

Static, spherically symmetric black holes and compact objects without an event horizon have unstable (stable) circular orbits of a light called photon (antiphoton) sphere. A Damour-Solodukhin wormhole has been suggested as a simple black hole mimicker and the difference of its metric tensors from a black hole is described by a dimensionless parameter λ\lambda. The wormhole with two flat regions has two photon spheres and an antiphoton sphere for λ<2/2\lambda<\sqrt{2}/2 and a photon sphere for λ2/2\lambda\geq\sqrt{2}/2. When the parameter λ\lambda is 2/2\sqrt{2}/2, the photon sphere is marginally unstable because of degeneration of the photon spheres and antiphoton sphere. We investigate gravitational lensing by the wormhole in weak and strong gravitational fields. We find that the deflection angle of a light ray reflected by the marginally unstable photon sphere diverges nonlogarithmically in a strong deflection limit for λ=2/2\lambda=\sqrt{2}/2, while the deflection angle reflected by the photon sphere diverges logarithmically for λ2/2\lambda\neq\sqrt{2}/2. We extend a strong deflection limit analysis for the nonlogarithmic divergence case. We expect that our method can be applied for gravitational lenses by marginally unstable photon spheres of various compact objects.

I Introduction

Recently, LIGO and VIRGO Collaborations have reported the direct detection of gravitational waves from black holes Abbott:2016blz ; LIGOScientific:2018mvr and Event Horizon Telescope Collaboration has reported the ring image of supermassive black hole candidates at the center of a giant elliptical galaxy M87 Akiyama:2019cqa . The black holes and the other compact objects with a strong gravitational field described by general relativity will be more important to understand our universe.

It is well known that static, spherically symmetric compact objects have unstable (stable) circular photon orbit called photon (antiphoton) sphere Claudel:2000yi ; Perlick_2004_Living_Rev . The upper bound of the radius of the (anti)photon sphere of the static, spherically symmetric black hole under the weak energy condition is given by r=3Mr=3M, where MM is the mass of the black hole Hod:2017xkz . The (anti)photon sphere has important roles in several phenomena in a strong gravitational field: Light rays emitted by a source and reflected by the photon sphere make infinite number of dim images Hagihara_1931 ; Darwin_1959 ; Atkinson_1965 ; Luminet_1979 ; Ohanian_1987 ; Nemiroff_1993 ; Frittelli_Kling_Newman_2000 ; Virbhadra_Ellis_2000 ; Bozza_Capozziello_Iovane_Scarpetta_2001 ; Bozza:2002zj ; Perlick_2004_Phys_Rev_D ; Bozza_2010 , which are named relativistic images in Ref. Virbhadra_Ellis_2000 , on the both sides of the photon sphere and we can survey the compact object with the photon sphere by the images even if the compact objects themselves do not emit light rays. The photon sphere can be observed during a collapsing star to be a black hole Ames_1968 ; Synge:1966okc ; Yoshino:2019qsh . The photon sphere has strong influence on the high-frequency behavior of the photon absorption cross section Sanchez:1977si ; Decanini:2010fz and the high-frequency spectrum of quasinormal modes of compact objects Press:1971wr ; Goebel_1972 ; Raffaelli:2014ola . An observer moving on the photon sphere feels no centrifugal force and no gyroscopic precession Abramowicz_Prasanna_1990 ; Abramowicz:1990cb ; Allen:1990ci ; Hasse_Perlick_2002 and it is the fastest way to circle a static, spherically symmetric black hole for massless particles Hod:2012nk . The photon sphere is correspond to Bondi’s sonic horizon of a radial fluid Mach:2013gia ; Chaverra:2015bya ; Cvetic:2016bxi ; Koga:2016jjq ; Koga:2018ybs ; Koga:2019teu .

Stability of light rings, i.e., circular photon orbits Koga:2019uqd , of compact objects are an important property of the spacetime. Instability of the compact objects with the stable light rings has been concerned since they cause the slow decay of linear waves Keir:2014oka ; Cardoso:2014sna ; Cunha:2017qtt . The numbers of light rings of stationary, axisymmetic compact objects without an event horizon under energy conditions are two at least, they are even number in general, and the inner light ring is stable Cunha:2017eoe . Hod has shown that spherically symmetric compact objects without an event horizon have odd number light rings because of degeneration Hod:2017zpi . We note that we cannot apply the theorem of the number of the light rings for wormholes since a trivial topology has been assumed in Refs. Cunha:2017eoe ; Hod:2017zpi .

General relativity permits the wormholes which can have nontrivial topological structures Visser_1995 ; Morris_Thorne_1988 . The wormholes have a throat which connects two regions of one universe or two universes. Gravitational lensing Schneider_Ehlers_Falco_1992 ; Schneider_Kochanek_Wambsganss_2006 ; Perlick_2004_Living_Rev is used to find dark gravitating objects like wormholes. However, we cannot distinguish the wormhole with a positive mass from other massive objects such as a black hole under a weak-field approximation. Therefore, we must rely on the observation near the throat or the photon sphere in a strong gravitational field such as gravitational lensing of light rays scatter by the throat or the photon sphere Chetouani_Clement_1984 ; Perlick_2004_Phys_Rev_D ; Nandi:2006ds ; Muller:2008zza ; Tsukamoto_Harada_Yajima_2012 ; Perlick:2014zwa ; Tsukamoto:2016qro ; Tsukamoto:2016zdu ; Nandi:2016uzg ; Tsukamoto:2017edq ; Shaikh:2018oul ; Shaikh:2019jfr ; Shaikh:2019itn , visualizations Muller_2004 ; James:2015ima , shadows in an accretion gas Ohgami:2015nra ; Ohgami:2016iqm ; Kuniyasu:2018cgv , wave optics Nambu:2019sqn , and gravitational waves Cardoso:2016rao to distinguish wormhole from the other compact objects.

A Damour-Solodukhin wormhole Damour:2007ap has been suggested as a black hole mimicker. Its metric was created by making a slight modification to the Schwarzschild metric and it can be the simplest metric among black hole mimickers. The difference of the metric from the Schwarzschild spacetime is described by a positive dimensionless parameter λ\lambda. Lemos and Zaslavskii have pointed out that we can distinguish the wormhole from the black hole by a tidal force acting on a body near the throat even if the difference of the metric tensors of the wormhole from the black hole is small, i.e., λ1\lambda\ll 1 Lemos:2008cv . Emissions from its accretion disk Karimov:2019qco , quasinormal modes and gravitational waves Bueno:2017hyj ; Volkel:2018hwb , images of its accretion disks Paul:2019trt , shadow Amir:2018pcu , gravitational lensing Nandi:2018mzm ; Ovgun:2018fnk ; Bhattacharya:2018leh ; Ovgun:2018swe ; Ovgun:2018oxk , and particle collision Tsukamoto:2019ihj in the Damour-Solodukhin wormhole spacetime have been investigated.

Gravitational lensing in a strong deflection limit, which is a semianalytic formalism for gravitational lensing of a light ray reflected by the photon sphere, has been investigated by Bozza Bozza:2002zj . The deflection angle αdef\alpha_{\mathrm{def}} of the light ray in the strong deflection limit has the following form

αdef(b)\displaystyle\alpha_{\mathrm{def}}(b) =\displaystyle= a¯log(bbm1)+b¯\displaystyle-\bar{a}\log\left(\frac{b}{b_{\mathrm{m}}}-1\right)+\bar{b} (1)
+O((bbm1)log(bbm1))\displaystyle+O\left(\left(\frac{b}{b_{\mathrm{m}}}-1\right)\log\left(\frac{b}{b_{\mathrm{m}}}-1\right)\right)

or

αdef(θ)\displaystyle\alpha_{\mathrm{def}}(\theta) =\displaystyle= a¯log(θθ1)+b¯\displaystyle-\bar{a}\log\left(\frac{\theta}{\theta_{\infty}}-1\right)+\bar{b} (2)
+O((θθ1)log(θθ1)),\displaystyle+O\left(\left(\frac{\theta}{\theta_{\infty}}-1\right)\log\left(\frac{\theta}{\theta_{\infty}}-1\right)\right),

where bb and bmb_{\mathrm{m}} are the impact parameter and the critical impact parameter of the light ray, respectively, θ\theta is an image angle, θbm/DOL\theta_{\infty}\equiv b_{\mathrm{m}}/D_{OL} is the image angle of the photon sphere, where DOLD_{OL} is a distance between an observer and a lens object, and a¯\bar{a} is a positive parameter and b¯\bar{b} is a parameter. We can assume that the impact parameter bb is non-negative without loss of generality when we treat one light ray in a spherically symmetric spacetime. The analysis in the strong deflection limit has been improved Eiroa:2002mk ; Bozza:2005tg ; Bozza:2006nm ; Bozza:2007gt ; Tsukamoto:2016qro ; Tsukamoto:2016jzh ; Ishihara:2016sfv ; Shaikh:2019itn ; Shaikh:2019jfr and its relations to high-energy absorption cross section Wei:2011zw and to quasinormal modes Stefanov:2010xz have been investigated. Recently, Shaikh et al. have considered the deflection angle of light rays scattered by a photon sphere at the throat of a wormhole Shaikh:2018oul ; Shaikh:2019jfr .

The deflection angle of a light ray scattered by a photon sphere of the Damour-Solodukhin wormhole in the strong deflection limit was obtained by Nandi et al. Nandi:2018mzm , Ovgun Ovgun:2018fnk , and Bhattacharya and Karimov Bhattacharya:2018leh when the wormhole metric is similar to the black hole, i.e., for λ<2/2\lambda<\sqrt{2}/2. In Ref. Bhattacharya:2018leh , Bhattacharya and Karimov have pointed out that b¯\bar{b} by Ovgun Ovgun:2018fnk is in error.

In this paper, firstly we reexamine the deflection angle of the light ray scattered by the photon spheres of the Damour-Solodukhin wormhole for λ<2/2\lambda<\sqrt{2}/2. We recover Bhattacharya and Karimov’s result Bhattacharya:2018leh and we give a small modification for calculation in Nandi:2018mzm . Secondly we extend the analysis for λ2/2\lambda\geq\sqrt{2}/2. The deflection angles diverge logarithmically in the strong deflection limit if the dimensionless parameter λ2/2\lambda\neq\sqrt{2}/2. Interestingly, we have found that the deflection angle of a light ray scattered by a marginally unstable photon sphere at the throat diverges nonlogarithmically for λ=2/2\lambda=\sqrt{2}/2. We construct the strong deflection limit analysis for the deflection angle with a nonlogarithmic divergence in the Damour-Solodukhin wormhole spacetime.

This paper is organized as follows. In Sec. II, we review the Damour-Solodukhin wormhole spacetime. In Secs. III and IV, we investigate the deflection angle and observables, respectively, in the strong deflection limit. We review the gravitational lensing under a weak-field approximation in Sec. V and we summarize our result in Sec. VI. In Appendixes A and B, the Arnowitt-Deser-Misner (ADM) masses and the violation of energy conditions of the Damour-Solodukhin wormhole are shown, respectively. In this paper we use the units in which a light speed and Newton’s constant are unity.

II Damour-Solodukhin wormhole spacetime

In this section, we review the trajectory of a light ray and its deflection angle αdef\alpha_{\mathrm{def}} in a Damour-Solodukhin wormhole spacetime Damour:2007ap with a line element given by, in coordinates (t~,r,ϑ,φ)(\tilde{t},r,\vartheta,\varphi),

ds2\displaystyle ds^{2} =\displaystyle= (12M~r+λ2)dt~2+dr212M~r\displaystyle-\left(1-\frac{2\tilde{M}}{r}+\lambda^{2}\right)d\tilde{t}^{2}+\frac{dr^{2}}{1-\frac{2\tilde{M}}{r}} (3)
+r2(dϑ2+sin2ϑdφ2),\displaystyle+r^{2}\left(d\vartheta^{2}+\sin^{2}\vartheta d\varphi^{2}\right),

where λ\lambda and M~\tilde{M} are positive parameters and the radial coordinate rr is defined in a range 2M~r<2\tilde{M}\leq r<\infty. A throat exists at r=rth2M~r=r_{\mathrm{th}}\equiv 2\tilde{M}. The wormhole spacetime can take different values of the parameters M~\tilde{M} and λ\lambda in two asymptotically flat regions. For simplicity, we assume the equal parameters in the both regions. The line element is the same the one in the Schwarzschild spacetime in a limit λ0\lambda\rightarrow 0 and it is the same as the static case of a Kerr-like wormhole Bueno:2017hyj . Bozza has considered the deflection angle of a light ray in the strong deflection limit in a general asymptotically flat, static, and spherically symmetric spacetime with its metric tensor behaving

limrgtt(1+2Mr),\displaystyle\lim_{r\rightarrow\infty}g_{tt}\rightarrow-\left(1+\frac{2M}{r}\right), (4)
limrgrr12Mr,\displaystyle\lim_{r\rightarrow\infty}g_{rr}\rightarrow 1-\frac{2M}{r}, (5)
limrgϑϑ=limrgφφsin2ϑr2,\displaystyle\lim_{r\rightarrow\infty}g_{\vartheta\vartheta}=\lim_{r\rightarrow\infty}\frac{g_{\varphi\varphi}}{\sin^{2}\vartheta}\rightarrow r^{2}, (6)

where tt is a time coordinate and MM is a positive parameter Bozza:2002zj . The Damour-Solodukhin wormhole spacetime is an asymptotically flat, static, and spherically symmetric spacetime but the (t~,t~)(\tilde{t},\tilde{t})-component of the metric tensor gt~t~g_{\tilde{t}\tilde{t}} (3) does not satisfy the assumption (4). We notice that (t~,t~)(\tilde{t},\tilde{t})-component of the metric tensor gt~t~g_{\tilde{t}\tilde{t}} shown in Eq. (3) behaving asymptotically

limrgt~t~(1+λ2)\displaystyle\lim_{r\rightarrow\infty}g_{\tilde{t}\tilde{t}}\rightarrow-(1+\lambda^{2}) (7)

is not suitable for a variable zz defined in Sec. III. Thus, we introduce a new time coordinate tt and a positive parameter MM which are defined by

tt~1+λ2\displaystyle t\equiv\frac{\tilde{t}}{\sqrt{1+\lambda^{2}}} (8)

and

MM~1+λ2,\displaystyle M\equiv\frac{\tilde{M}}{1+\lambda^{2}}, (9)

respectively, to satisfy the assumption (4). By using tt and MM, the line element (3) is rewritten in

ds2=A(r)dt2+B(r)dr2+C(r)(dϑ2+sin2ϑdφ2),\displaystyle ds^{2}=-A(r)dt^{2}+B(r)dr^{2}+C(r)\left(d\vartheta^{2}+\sin^{2}\vartheta d\varphi^{2}\right),

where A(r)A(r), B(r)B(r), and C(r)C(r) are given by

A(r)12Mr,\displaystyle A(r)\equiv 1-\frac{2M}{r}, (11)
B(r)[12M(1+λ2)r]1,\displaystyle B(r)\equiv\left[1-\frac{2M(1+\lambda^{2})}{r}\right]^{-1}, (12)

and

C(r)r2,\displaystyle C(r)\equiv r^{2}, (13)

respectively. 111The metric tensor (II)(\ref{eq:line_element2}) does not satisfy the condition (5) but it does not give us troubles to define the variable zz. Notice that rthr_{\mathrm{th}} is rewritten as rth=2M(1+λ2)r_{\mathrm{th}}=2M(1+\lambda^{2}). Since the spacetime is a static, spherically symmetric spacetime, there are time-translational and axial Killing vectors tμμ=tt^{\mu}\partial_{\mu}=\partial_{t} and φμμ=φ\varphi^{\mu}\partial_{\mu}=\partial_{\varphi}, respectively.

From kμkμ=0k^{\mu}k_{\mu}=0, where kμx˙μk^{\mu}\equiv\dot{x}^{\mu} is the wave number of a light ray and where the dot denotes the differentiation with respect to an affine parameter, the trajectory of the light ray is obtained as

A(r)t˙2+B(r)r˙2+C(r)φ˙2=0.\displaystyle-A(r)\dot{t}^{2}+B(r)\dot{r}^{2}+C(r)\dot{\varphi}^{2}=0. (14)

Here we have set ϑ=π/2\vartheta=\pi/2 without loss of generality. Equation (14) can be expressed as

r˙2+V~(r)=0,\displaystyle\dot{r}^{2}+\tilde{V}(r)=0, (15)

where the effective potential V~(r)\tilde{V}(r) is defined by

V~(r)E2B(r)(1A(r)b2C(r)),\displaystyle\tilde{V}(r)\equiv-\frac{E^{2}}{B(r)}\left(\frac{1}{A(r)}-\frac{b^{2}}{C(r)}\right), (16)

where bL/Eb\equiv L/E is the impact parameter of the light ray and the conserved energy EgμνtμkνE\equiv-g_{\mu\nu}t^{\mu}k^{\nu} and the conserved angular momentum LgμνφμkνL\equiv g_{\mu\nu}\varphi^{\mu}k^{\nu} of the light ray are constant along the trajectory. Since V~(r)E2>0\tilde{V}(r)\rightarrow E^{2}>0 in spatial infinity rr\rightarrow\infty, the light ray exists there.

By introducing a proper radial distance ll from the throat given by

l\displaystyle l \displaystyle\equiv rthrB(r)𝑑r\displaystyle\int^{r}_{r_{\mathrm{th}}}\sqrt{B(r)}dr
=\displaystyle= r22M(1+λ2)r\displaystyle\sqrt{r^{2}-2M\left(1+\lambda^{2}\right)r}
+M(1+λ2)logrM(1+λ2)+r22M(1+λ2)rM(1+λ2),\displaystyle+M\left(1+\lambda^{2}\right)\log\frac{r-M(1+\lambda^{2})+\sqrt{r^{2}-2M(1+\lambda^{2})r}}{M(1+\lambda^{2})},

the line element, the equation of trajectory of the light ray, and the effective potential are rewritten in

ds2=(12Mr(l))dt2+dl2+r2(l)(dϑ2+sin2ϑdφ2),\displaystyle ds^{2}=-\left(1-\frac{2M}{r(l)}\right)dt^{2}+dl^{2}+r^{2}(l)\left(d\vartheta^{2}+\sin^{2}\vartheta d\varphi^{2}\right),
l˙2+v~(l)=0,\displaystyle\dot{l}^{2}+\tilde{v}(l)=0, (19)

and

v~(l)E2(1A(l)b2C(l)),\displaystyle\tilde{v}(l)\equiv-E^{2}\left(\frac{1}{A(l)}-\frac{b^{2}}{C(l)}\right), (20)

respectively. Note that the proper radial distance ll is defined in a range <l<-\infty<l<\infty and the throat is at l=0l=0.

From v~=dv~dl=0\tilde{v}=\frac{d\tilde{v}}{dl}=0, we get the circular light orbits with b=2M(1+λ2)32/λb=2M(1+\lambda^{2})^{\frac{3}{2}}/\lambda and b=33Mb=3\sqrt{3}M at r=rth=2M(1+λ2)r=r_{\mathrm{th}}=2M(1+\lambda^{2}) and r=3Mr=3M, respectively. From straightforward calculations, we obtain d2v~dl2|r=rth>0\left.\frac{d^{2}\tilde{v}}{dl^{2}}\right|_{r=r_{\mathrm{th}}}>0 for λ<2/2\lambda<\sqrt{2}/2 and d2v~dl2|r=rth0\left.\frac{d^{2}\tilde{v}}{dl^{2}}\right|_{r=r_{\mathrm{th}}}\leq 0 for λ2/2\lambda\geq\sqrt{2}/2. Therefore, the throat is a photon (antiphoton) sphere for λ>2/2\lambda>\sqrt{2}/2 (λ<2/2\lambda<\sqrt{2}/2) and the photon sphere is marginally unstable for λ=2/2\lambda=\sqrt{2}/2. Figure 1 shows the dimensionless effective potential vv~/E2v\equiv\tilde{v}/E^{2} for the (marginally unstable) photon sphere and antiphoton sphere.

Refer to caption
Figure 1: A dimensionless effective potential v(l)v(l) as a function of a proper distance ll from the throat. It shows a (anti)photon sphere at the throat. The solid (red), broken (green), and dotted (magenta) curves denote the effective potential v(l)v(l) when λ=0.4\lambda=0.4, 2/2\sqrt{2}/2, and 11, respectively. We have set M=1M=1 and b=2M(1+λ2)32/λb=2M(1+\lambda^{2})^{\frac{3}{2}}/\lambda.

The circular orbit of a light ray with b=33b=3\sqrt{3} at r=3Mr=3M is unstable for λ2/2\lambda\leq\sqrt{2}/2, i.e., it is a photon sphere. Notice that the wormhole has two asymptotically flat regions and it has two photon spheres at r=3Mr=3M for λ<2/2\lambda<\sqrt{2}/2. The photon spheres at r=3Mr=3M and the antiphoton sphere at the throat r=rth=2M(1+λ2)r=r_{\mathrm{th}}=2M(1+\lambda^{2}) degenerate to be a marginally unstable photon sphere at the throat just for λ=2/2\lambda=\sqrt{2}/2. The wormhole has only one photon sphere at the throat for λ>2/2\lambda>\sqrt{2}/2.

As shown Fig. 2, we can classify the light ray which coming from a spatial infinity into a falling case with b<bmb<b_{\mathrm{m}}, a critical case with b=bmb=b_{\mathrm{m}}, and a scattered case with b<bmb<b_{\mathrm{m}}. Here the critical impact parameter bmb_{\mathrm{m}} is defined by

bmlimr0rmb(r0)=CmAm,\displaystyle b_{\mathrm{m}}\equiv\lim_{r_{0}\rightarrow r_{\mathrm{m}}}b(r_{0})=\sqrt{\frac{C_{\mathrm{m}}}{A_{\mathrm{m}}}}, (21)

where r0r_{0} is the closest distance of the light ray and hereafter subscript mm denotes quantities of the photon sphere at r=rmr=r_{\mathrm{m}}.

Refer to caption
Refer to caption
Figure 2: A dimensionless effective potential v(l)v(l) as a function of a proper distance ll from the throat. The solid (red), broken (green), and dotted (magenta) curves denote the effective potential v(l)v(l) in scattered (b=8M)(b=8M), critical (b=33M)(b=3\sqrt{3}M), and falling (b=2M)(b=2M) cases, respectively. The upper and lower panels show cases with λ=0.1\lambda=0.1 and λ=2/2\lambda=\sqrt{2}/2, respectively. All the cases, we have set M=1M=1. The effective potential v(l)v(l) for λ>2/2\lambda>\sqrt{2}/2 has a quietly similar shape to the one for λ=2/2\lambda=\sqrt{2}/2.

We concentrate on the scattered case. In this case, a light ray comes from a spatial infinity, it is deflected by the wormhole at a reflection point r=r0r=r_{0}, and it goes back to the same spatial infinity. The reflection point is obtained as the largest positive solution of the equation v(l)=0v(l)=0 or V(r)=0V(r)=0. At the reflection point r=r0r=r_{0}, the equation of the trajectory (14) gives

A0t˙02=C0φ˙02,\displaystyle A_{0}\dot{t}^{2}_{0}=C_{0}\dot{\varphi}^{2}_{0}, (22)

where the subscript 0 denotes the quantity at r=r0r=r_{0}. The impact parameter b=b(r0)b=b(r_{0}) is expressed by

b(r0)=LE=C0φ˙0A0t˙0=C0A0.\displaystyle b(r_{0})=\frac{L}{E}=\frac{C_{0}\dot{\varphi}_{0}}{A_{0}\dot{t}_{0}}=\sqrt{\frac{C_{0}}{A_{0}}}. (23)

Here we have used Eq. (22).

From Eq. (14), the deflection angle αdef(r0)\alpha_{\mathrm{def}}(r_{0}) of the light ray as a function of the reflection point r0r_{0} is obtained as

αdef(r0)I(r0)π,\displaystyle\alpha_{\mathrm{def}}(r_{0})\equiv I(r_{0})-\pi, (24)

where I(r0)I(r_{0}) is defined by

I(r0)\displaystyle I(r_{0}) \displaystyle\equiv 2r0B(r)drC(r)C(r)A0A(r)C01\displaystyle 2\int^{\infty}_{r_{0}}\frac{\sqrt{B(r)}dr}{\sqrt{C(r)}\sqrt{\frac{C(r)A_{0}}{A(r)C_{0}}-1}} (25)
=\displaystyle= 2r0bdrC(r)V(r)\displaystyle 2\int^{\infty}_{r_{0}}\frac{bdr}{C(r)\sqrt{-V(r)}}
=\displaystyle= 2l0bdlC(l)v(l),\displaystyle 2\int^{\infty}_{l_{0}}\frac{bdl}{C(l)\sqrt{-v(l)}},

where V(r)V~(r)/E2V(r)\equiv\tilde{V}(r)/E^{2} is a dimensionless effective potential in the radial coordinate rr and where l0l(r)|r=r0l_{0}\equiv\left.l(r)\right|_{r=r_{0}} is the position of the reflection point in the radial coordinate ll.

III Deflection angle in a strong deflection limit

In this section, we investigate the deflection angle in the strong deflection limit r0rmr_{0}\rightarrow r_{\mathrm{m}} or bbmb\rightarrow b_{\mathrm{m}}. We treat it in the cases for λ<2/2\lambda<\sqrt{2}/2, λ>2/2\lambda>\sqrt{2}/2, and λ=2/2\lambda=\sqrt{2}/2 in this order. In the strong deflection limit bbmb\rightarrow b_{\mathrm{m}}, the deflection angle of the light ray in a strong deflection is expressed by a following form Bozza:2002zj 222The subleading term is O(bbm)O(b-b_{\mathrm{m}}) in Ref. Bozza:2002zj but we should read it as O((bbm1)log(bbm1))O\left(\left(\frac{b}{b_{\mathrm{m}}}-1\right)\log\left(\frac{b}{b_{\mathrm{m}}}-1\right)\right). See Refs. Tsukamoto:2016jzh ; Tsukamoto:2016qro ; Iyer:2006cn .

αdef(b)\displaystyle\alpha_{\mathrm{def}}(b) =\displaystyle= a¯log(bbm1)+b¯\displaystyle-\bar{a}\log\left(\frac{b}{b_{\mathrm{m}}}-1\right)+\bar{b} (26)
+O((bbm1)log(bbm1)).\displaystyle+O\left(\left(\frac{b}{b_{\mathrm{m}}}-1\right)\log\left(\frac{b}{b_{\mathrm{m}}}-1\right)\right).

We introduce a variable zz Bozza:2002zj defined by

zA(r)A01A0=1r0r.\displaystyle z\equiv\frac{A(r)-A_{0}}{1-A_{0}}=1-\frac{r_{0}}{r}. (27)

III.1 λ<2/2\lambda<\sqrt{2}/2

In the case for λ<2/2\lambda<\sqrt{2}/2, the photon sphere is at r=rm=3Mr=r_{\mathrm{m}}=3M and the critical impact parameter bmb_{\mathrm{m}} is given by bm=33Mb_{\mathrm{m}}=3\sqrt{3}M. By using zz, we rewrite I(r0)I(r_{0}) as

I(r0)=01R(z,r0)f(z,r0)𝑑z,\displaystyle I(r_{0})=\int^{1}_{0}R(z,r_{0})f(z,r_{0})dz, (28)

where R(z,r0)R(z,r_{0}) is given by

R(z,r0)\displaystyle R(z,r_{0}) \displaystyle\equiv 2ABAC(1A0)C0\displaystyle\frac{2\sqrt{AB}}{A^{\prime}C}(1-A_{0})\sqrt{C_{0}} (29)
=\displaystyle= 2r02M(1z)r02M(1+λ2)(1z),\displaystyle 2\sqrt{\frac{r_{0}-2M(1-z)}{r_{0}-2M(1+\lambda^{2})(1-z)}},

where is the differentiation with respect to rr and f(z,r0)f(z,r_{0}) is given by

f(z,r0)\displaystyle f(z,r_{0}) \displaystyle\equiv 1A0[(1A0)z+A0]C0C\displaystyle\frac{1}{\sqrt{A_{0}-\left[(1-A_{0})z+A_{0}\right]\frac{C_{0}}{C}}}
=\displaystyle= 1α0z+β0z22Mr0z3,\displaystyle\frac{1}{\sqrt{\alpha_{0}z+\beta_{0}z^{2}-\frac{2M}{r_{0}}z^{3}}},

and α0\alpha_{0} and β0\beta_{0} are defined by

α0=α(r0)26Mr0,\displaystyle\alpha_{0}=\alpha(r_{0})\equiv 2-\frac{6M}{r_{0}}, (31)
β0=β(r0)1+6Mr0.\displaystyle\beta_{0}=\beta(r_{0})\equiv-1+\frac{6M}{r_{0}}. (32)

R(z,r0)R(z,r_{0}) is regular but f(z,r0)f(z,r_{0}) diverges in a limit z0z\rightarrow 0. We define f0(z,r0)f_{0}(z,r_{0}) as

f0(z,r0)1α0z+β0z2.\displaystyle f_{0}(z,r_{0})\equiv\frac{1}{\sqrt{\alpha_{0}z+\beta_{0}z^{2}}}. (33)

Since αm=0\alpha_{\mathrm{m}}=0 and βm=1\beta_{\mathrm{m}}=1, the integral of f0(z,r0)f_{0}(z,r_{0}) diverges in the strong deflection limit r0rmr_{0}\rightarrow r_{\mathrm{m}}. By using f0(z,r0)f_{0}(z,r_{0}), we separate I(r0)I(r_{0}) into a divergent part ID(r0)I_{\mathrm{D}}(r_{0}) and a regular part IR(r0)I_{\mathrm{R}}(r_{0}):

I(r0)=ID(r0)+IR(r0),\displaystyle I(r_{0})=I_{\mathrm{D}}(r_{0})+I_{\mathrm{R}}(r_{0}), (34)

where

ID(r0)=01R(0,rm)f0(z,r0)𝑑z,\displaystyle I_{\mathrm{D}}(r_{0})=\int^{1}_{0}R(0,r_{\mathrm{m}})f_{0}(z,r_{0})dz, (35)
IR(r0)=01g(z,r0)𝑑z,\displaystyle I_{\mathrm{R}}(r_{0})=\int^{1}_{0}g(z,r_{0})dz, (36)

where g(z,r0)g(z,r_{0}) is defined by

g(z,r0)R(z,r0)f(z,r0)R(0,rm)f0(z,r0).\displaystyle g(z,r_{0})\equiv R(z,r_{0})f(z,r_{0})-R(0,r_{\mathrm{m}})f_{0}(z,r_{0}). (37)

The divergent part ID(r0)I_{\mathrm{D}}(r_{0}) can be integrated and it becomes

ID(r0)=2R(0,rm)β0logβ0+α0+β0α0.\displaystyle I_{\mathrm{D}}(r_{0})=\frac{2R(0,r_{\mathrm{m}})}{\sqrt{\beta_{0}}}\log\frac{\sqrt{\beta_{0}}+\sqrt{\alpha_{0}+\beta_{0}}}{\sqrt{\alpha_{0}}}. (38)

We expand α0\alpha_{0} and β0\beta_{0} in powers of (r0rm)(r_{0}-r_{\mathrm{m}}):

α0=23M(r0rm)+O((r0rm)2),\displaystyle\alpha_{0}=\frac{2}{3M}(r_{0}-r_{\mathrm{m}})+O\left((r_{0}-r_{\mathrm{m}})^{2}\right), (39)
β0=1+O(r0rm).\displaystyle\beta_{0}=1+O\left(r_{0}-r_{\mathrm{m}}\right). (40)

By substituting them into Eq. (38), we obtain

ID(r0)\displaystyle I_{\mathrm{D}}(r_{0}) =\displaystyle= 212λ2log(r0rm1)+2log212λ2\displaystyle-\frac{2}{\sqrt{1-2\lambda^{2}}}\log\left(\frac{r_{0}}{r_{\mathrm{m}}}-1\right)+\frac{2\log 2}{\sqrt{1-2\lambda^{2}}} (41)
+O((r0rm1)log(r0rm1)).\displaystyle+O\left(\left(\frac{r_{0}}{r_{\mathrm{m}}}-1\right)\log\left(\frac{r_{0}}{r_{\mathrm{m}}}-1\right)\right).

By using the impact parameter bb expanded in powers of (r0rm)(r_{0}-r_{\mathrm{m}}) obtained as

b=bm+32M(r0rm)2+O((r0rm)3),\displaystyle b=b_{\mathrm{m}}+\frac{\sqrt{3}}{2M}(r_{0}-r_{\mathrm{m}})^{2}+O\left((r_{0}-r_{\mathrm{m}})^{3}\right), (42)

we can rewrite ID=ID(b)I_{\mathrm{D}}=I_{\mathrm{D}}(b) as

ID(b)\displaystyle I_{\mathrm{D}}(b) =\displaystyle= 112λ2log(bbm1)+log612λ2\displaystyle-\frac{1}{\sqrt{1-2\lambda^{2}}}\log\left(\frac{b}{b_{\mathrm{m}}}-1\right)+\frac{\log 6}{\sqrt{1-2\lambda^{2}}} (43)
+O((bbm1)log(bbm1)).\displaystyle+O\left(\left(\frac{b}{b_{\mathrm{m}}}-1\right)\log\left(\frac{b}{b_{\mathrm{m}}}-1\right)\right).

The regular part IR(r0)I_{\mathrm{R}}(r_{0}) can be expanded in powers of (r0rm)(r_{0}-r_{\mathrm{m}}) and it is expressed by

IR(r0)=j=01j!(r0rm)j01jgr0j|r0=rmdz.\displaystyle I_{\mathrm{R}}(r_{0})=\sum^{\infty}_{j=0}\frac{1}{j!}(r_{0}-r_{\mathrm{m}})^{j}\int^{1}_{0}\left.\frac{\partial^{j}g}{\partial r^{j}_{0}}\right|_{r_{0}=r_{\mathrm{m}}}dz. (44)

We are interested in j=0j=0 term:

IR(r0)\displaystyle I_{\mathrm{R}}(r_{0}) =\displaystyle= 01g(z,rm)𝑑z\displaystyle\int^{1}_{0}g(z,r_{\mathrm{m}})dz (45)
+O((r0rm1)log(r0rm1)),\displaystyle+O\left(\left(\frac{r_{0}}{r_{\mathrm{m}}}-1\right)\log\left(\frac{r_{0}}{r_{\mathrm{m}}}-1\right)\right),

where g(z,rm)g(z,r_{\mathrm{m}}) is given by

g(z,rm)\displaystyle g(z,r_{\mathrm{m}}) =\displaystyle= 23+6zz12λ2+2(1+λ2)z32z\displaystyle\frac{2\sqrt{3+6z}}{z\sqrt{1-2\lambda^{2}+2(1+\lambda^{2})z}\sqrt{3-2z}} (46)
2z12λ2.\displaystyle-\frac{2}{z\sqrt{1-2\lambda^{2}}}.

Thus, we have obtained the deflection angle αdef(b)\alpha_{\mathrm{def}}(b) of the light ray in the strong deflection limit bbmb\rightarrow b_{\mathrm{m}} in the form of Eq. (26) with

a¯\displaystyle\bar{a} =\displaystyle= 112λ2,\displaystyle\frac{1}{\sqrt{1-2\lambda^{2}}}, (47)
b¯\displaystyle\bar{b} =\displaystyle= log612λ2+IRπ.\displaystyle\frac{\log 6}{\sqrt{1-2\lambda^{2}}}+I_{\mathrm{R}}-\pi. (48)

This is the same as the deflection angle obtained by Bhattacharya and Karimov Bhattacharya:2018leh . We plot a¯\bar{a} and b¯\bar{b} for λ2/2\lambda\neq\sqrt{2}/2 in Fig. 3.

Refer to caption
Figure 3: a¯\bar{a} and b¯\bar{b}. The solid (red) and dashed (green) curves denote a¯\bar{a} and b¯\bar{b}, respectively. In a limit λ2/2\lambda\rightarrow\sqrt{2}/2, we obtain a¯\bar{a}\rightarrow\infty and b¯\bar{b}\rightarrow-\infty.

We comment on the details of earlier work on it. First, the deflection angle in the strong deflection limit was calculated by Nandi et al. Nandi:2018mzm . They considered the metric tensor (3) in the coordinates (t~,r,ϑ,φ)(\tilde{t},r,\vartheta,\varphi) and they defined the variable 333Note that we use (,+,+,+)(-,+,+,+), while Nandi et al. have used (+,,,)(+,-,-,-) Nandi:2018mzm .

z[74]gt~t~(r)gt~t~(r0)1gt~t~(r0)=2M~2M~λ2r0(1r0r).\displaystyle z_{[74]}\equiv\frac{g_{\tilde{t}\tilde{t}}(r)-g_{\tilde{t}\tilde{t}}(r_{0})}{1-g_{\tilde{t}\tilde{t}}(r_{0})}=\frac{2\tilde{M}}{2\tilde{M}-\lambda^{2}r_{0}}\left(1-\frac{r_{0}}{r}\right). (49)

Note that the integration range of Eq. (2.10) in Ref. Nandi:2018mzm , 0z[74]10\leq z_{[74]}\leq 1 should be modified to be

0z[74]2M~2M~λ2r0.\displaystyle 0\leq z_{[74]}\leq\frac{2\tilde{M}}{2\tilde{M}-\lambda^{2}r_{0}}. (50)

The error affects to b¯\bar{b} and the difference is small when λ\lambda is small as shown Table I. Their results in Ref. Nandi:2018mzm will be valid since they discussed in the case of λ1\lambda\ll 1.

Table 1: Parameters a¯\bar{a} and b¯\bar{b} as a function with respect to λ\lambda. We have obtained the same a¯\bar{a} as Ref. Nandi:2018mzm . We compare b¯[74]\bar{b}_{[74]} obtained in Ref. Nandi:2018mzm and b¯\bar{b} obtained from Eq. (48).
λ\lambda a¯\bar{a} b¯[74]\bar{b}_{[74]} b¯\bar{b}
0 1.00001.0000 0.4002-0.4002 0.4002-0.4002
0.0010.001 1.00001.0000 0.4002-0.4002 0.4002-0.4002
0.010.01 1.00011.0001 0.4008-0.4008 0.4004-0.4004
0.020.02 1.00041.0004 0.4028-0.4028 0.4007-0.4007
0.030.03 1.00091.0009 0.4046-0.4046 0.4014-0.4014
0.040.04 1.00161.0016 0.4105-0.4105 0.4022-0.4022
0.050.05 1.00251.0025 0.4163-0.4163 0.4034-0.4034

The deflection angle was calculated also by Ovgun Ovgun:2018fnk , in the coordinates (t,r,φ,ϑ)(t,r,\varphi,\vartheta),

b¯=log(6)2λ2+1log(10)2λ21+rmπ.\displaystyle\bar{b}=\frac{\log(6)\sqrt{-2\lambda^{2}+1}}{\log(10)2\lambda^{2}-1}+r_{\mathrm{m}}-\pi. (51)

See Eq. (2.32) in Ref. Ovgun:2018fnk . Bhattacharya and Karimov pointed out that b¯\bar{b} obtained in Ref. Ovgun:2018fnk , i.e., Eq. (51), is in error Bhattacharya:2018leh .

III.2 λ>2/2\lambda>\sqrt{2}/2

We consider the case of λ>2/2\lambda>\sqrt{2}/2. In this case, the throat is the photon sphere at r=rm=rth=2M(1+λ2)r=r_{\mathrm{m}}=r_{\mathrm{th}}=2M(1+\lambda^{2}) and the critical impact parameter is bm=2M(1+λ2)32/λb_{\mathrm{m}}=2M(1+\lambda^{2})^{\frac{3}{2}}/\lambda. We notice that the regular factor R(z,rm)R(z,r_{\mathrm{m}}) is given by

R(z,rm)=2λ2+z(1+λ2)z\displaystyle R(z,r_{\mathrm{m}})=2\sqrt{\frac{\lambda^{2}+z}{(1+\lambda^{2})z}} (52)

and the form of z12z^{-\frac{1}{2}} is not suitable for analysis of the strong deflection limit. Thus, we express I(r0)I(r_{0}) as

I(r0)=01S(z,r0)h(z,r0)𝑑z,\displaystyle I(r_{0})=\int^{1}_{0}S(z,r_{0})h(z,r_{0})dz, (53)

where a new regular factor S(z,r0)S(z,r_{0}) and a new divergent factor h(z,r0)h(z,r_{0}) are given by

S(z,r0)2r02M(1z)\displaystyle S(z,r_{0})\equiv 2\sqrt{r_{0}-2M(1-z)} (54)

and

h(z,r0)\displaystyle h(z,r_{0})\equiv
1γ0z+η0z2+[2M(r0rm)r0+rmβ0]z32Mrmr0z4,\displaystyle\sqrt{\frac{1}{\gamma_{0}z+\eta_{0}z^{2}+\left[-\frac{2M(r_{0}-r_{\mathrm{m}})}{r_{0}}+r_{\mathrm{m}}\beta_{0}\right]z^{3}-\frac{2Mr_{\mathrm{m}}}{r_{0}}z^{4}}},

where γ0\gamma_{0} and η0\eta_{0} are defined as

γ0=γ(r0)(r0rm)α0,\displaystyle\gamma_{0}=\gamma(r_{0})\equiv(r_{0}-r_{\mathrm{m}})\alpha_{0}, (56)
η0=η(r0)(r0rm)β0+rmα0.\displaystyle\eta_{0}=\eta(r_{0})\equiv(r_{0}-r_{\mathrm{m}})\beta_{0}+r_{\mathrm{m}}\alpha_{0}. (57)

Notice that αm=(1+2λ2)/(1+λ2)\alpha_{\mathrm{m}}=(-1+2\lambda^{2})/(1+\lambda^{2}) and βm=(2λ2)/(1+λ2)\beta_{\mathrm{m}}=(2-\lambda^{2})/(1+\lambda^{2}) in the case of λ>2/2\lambda>\sqrt{2}/2. Since we obtain γm=0\gamma_{\mathrm{m}}=0 and ηm=2M(2λ21)\eta_{\mathrm{m}}=2M(2\lambda^{2}-1), the integral of h0(z,r0)h_{0}(z,r_{0}) defined as

h0(z,r0)1γ0z+η0z2\displaystyle h_{0}(z,r_{0})\equiv\sqrt{\frac{1}{\gamma_{0}z+\eta_{0}z^{2}}} (58)

gives the divergent part of I(r0)I(r_{0}) in the strong deflection limit r0rmr_{0}\rightarrow r_{\mathrm{m}}. We separate I(r0)I(r_{0}) into the divergent part Id(r0)I_{\mathrm{d}}(r_{0}) and a regular part Ir(r0)I_{\mathrm{r}}(r_{0}), i.e.,

I(r0)=Id(r0)+Ir(r0),\displaystyle I(r_{0})=I_{\mathrm{d}}(r_{0})+I_{\mathrm{r}}(r_{0}), (59)

where

Id(r0)=01S(0,rm)h0(z,r0)𝑑z,\displaystyle I_{\mathrm{d}}(r_{0})=\int^{1}_{0}S(0,r_{\mathrm{m}})h_{0}(z,r_{0})dz, (60)
Ir(r0)=01k(z,r0)𝑑z.\displaystyle I_{\mathrm{r}}(r_{0})=\int^{1}_{0}k(z,r_{0})dz. (61)

Here, k(z,r0)k(z,r_{0}) is

k(z,r0)S(z,r0)h(z,r0)S(0,rm)h0(z,r0).\displaystyle k(z,r_{0})\equiv S(z,r_{0})h(z,r_{0})-S(0,r_{\mathrm{m}})h_{0}(z,r_{0}). (62)

The divergent part Id(r0)I_{\mathrm{d}}(r_{0}) can be integrated and we obtain

Id(r0)=2S(0,rm)η0logη0+γ0+η0γ0.\displaystyle I_{\mathrm{d}}(r_{0})=\frac{2S(0,r_{\mathrm{m}})}{\sqrt{\eta_{0}}}\log\frac{\sqrt{\eta_{0}}+\sqrt{\gamma_{0}+\eta_{0}}}{\sqrt{\gamma_{0}}}. (63)

By substituting γ0\gamma_{0} and η0\eta_{0} expanded in powers of (r0rm)(r_{0}-r_{\mathrm{m}}) as

γ0=αm(r0rm)+O((r0rm)2),\displaystyle\gamma_{0}=\alpha_{\mathrm{m}}(r_{0}-r_{\mathrm{m}})+O\left((r_{0}-r_{\mathrm{m}})^{2}\right), (64)
η0=2M(2λ21)+O(r0rm)\displaystyle\eta_{0}=2M(2\lambda^{2}-1)+O\left(r_{0}-r_{\mathrm{m}}\right) (65)

into Eq. (63), we obtain

Id(r0)\displaystyle I_{\mathrm{d}}(r_{0}) =\displaystyle= 2λ2λ21log(r0rm1)+4λlog22λ21\displaystyle-\frac{2\lambda}{\sqrt{2\lambda^{2}-1}}\log\left(\frac{r_{0}}{r_{\mathrm{m}}}-1\right)+\frac{4\lambda\log 2}{\sqrt{2\lambda^{2}-1}} (66)
+O((r0rm1)log(r0rm1)).\displaystyle+O\left(\left(\frac{r_{0}}{r_{\mathrm{m}}}-1\right)\log\left(\frac{r_{0}}{r_{\mathrm{m}}}-1\right)\right).

From the impact parameter bb which is expanded in powers of (r0rm)(r_{0}-r_{\mathrm{m}}) as

b=bm+(2λ21)1+λ22λ3(r0rm)+O((r0rm)2),\displaystyle b=b_{\mathrm{m}}+\frac{(2\lambda^{2}-1)\sqrt{1+\lambda^{2}}}{2\lambda^{3}}(r_{0}-r_{\mathrm{m}})+O\left((r_{0}-r_{\mathrm{m}})^{2}\right),

we obtain Id=Id(b)I_{\mathrm{d}}=I_{\mathrm{d}}(b) as

Id(b)\displaystyle I_{\mathrm{d}}(b) =\displaystyle= 2λ2λ21log(bbm1)\displaystyle-\frac{2\lambda}{\sqrt{2\lambda^{2}-1}}\log\left(\frac{b}{b_{\mathrm{m}}}-1\right) (68)
+2λ2λ21log2(2λ21)λ2\displaystyle+\frac{2\lambda}{\sqrt{2\lambda^{2}-1}}\log\frac{2(2\lambda^{2}-1)}{\lambda^{2}}
+O((bbm1)log(bbm1)).\displaystyle+O\left(\left(\frac{b}{b_{\mathrm{m}}}-1\right)\log\left(\frac{b}{b_{\mathrm{m}}}-1\right)\right).

The regular part Ir(r0)I_{\mathrm{r}}(r_{0}) can be expanded in powers of (r0rm)(r_{0}-r_{\mathrm{m}}) as

Ir(r0)=j=01j!(r0rm)j01jkr0j|r0=rmdz\displaystyle I_{\mathrm{r}}(r_{0})=\sum^{\infty}_{j=0}\frac{1}{j!}(r_{0}-r_{\mathrm{m}})^{j}\int^{1}_{0}\left.\frac{\partial^{j}k}{\partial r^{j}_{0}}\right|_{r_{0}=r_{\mathrm{m}}}dz (69)

and the term of j=0j=0 gives

Ir(r0)\displaystyle I_{\mathrm{r}}(r_{0}) =\displaystyle= 01k(z,rm)𝑑z\displaystyle\int^{1}_{0}k(z,r_{\mathrm{m}})dz (70)
+O((r0rm1)log(r0rm1)),\displaystyle+O\left(\left(\frac{r_{0}}{r_{\mathrm{m}}}-1\right)\log\left(\frac{r_{0}}{r_{\mathrm{m}}}-1\right)\right),

where k(z,rm)k(z,r_{\mathrm{m}}) is given by

k(z,rm)=2λ2+z(2λ21)z2+(2λ2)z3z42λz2λ21.\displaystyle k(z,r_{\mathrm{m}})=2\sqrt{\frac{\lambda^{2}+z}{(2\lambda^{2}-1)z^{2}+(2-\lambda^{2})z^{3}-z^{4}}}-\frac{2\lambda}{z\sqrt{2\lambda^{2}-1}}.

Therefore, we have obtained

a¯\displaystyle\bar{a} =\displaystyle= 2λ2λ21,\displaystyle\frac{2\lambda}{\sqrt{2\lambda^{2}-1}}, (72)
b¯\displaystyle\bar{b} =\displaystyle= 2λ2λ21log2(2λ21)λ2+Irπ.\displaystyle\frac{2\lambda}{\sqrt{2\lambda^{2}-1}}\log\frac{2(2\lambda^{2}-1)}{\lambda^{2}}+I_{\mathrm{r}}-\pi. (73)

III.3 λ=2/2\lambda=\sqrt{2}/2

In the case of λ=2/2\lambda=\sqrt{2}/2, the throat corresponds with the photon sphere, i.e., rm=rth=3Mr_{\mathrm{m}}=r_{\mathrm{th}}=3M. The critical impact parameter bmb_{\mathrm{m}} is given by bm=33Mb_{\mathrm{m}}=3\sqrt{3}M. From αm=γm=ηm=0\alpha_{\mathrm{m}}=\gamma_{\mathrm{m}}=\eta_{\mathrm{m}}=0 and βm=1\beta_{\mathrm{m}}=1, when r0=rmr_{0}=r_{\mathrm{m}}, the divergent factor h(z,r0)h(z,r_{0}) gives

h(z,rm)=13Mz32Mz4\displaystyle h(z,r_{\mathrm{m}})=\sqrt{\frac{1}{3Mz^{3}-2Mz^{4}}} (74)

and it causes the integral II to diverge as

Iz12|z=0.\displaystyle I\sim\left.z^{-\frac{1}{2}}\right|_{z=0}. (75)

This implies that I(r0)I(r_{0}) has the following form, in the strong deflection limit r0rm=3Mr_{0}\rightarrow r_{\mathrm{m}}=3M,

I(r0)=𝒜r0rm1++O(r0rm1),\displaystyle I(r_{0})=\frac{\mathcal{A}}{\sqrt{\frac{r_{0}}{r_{\mathrm{m}}}-1}}+\mathcal{B}+O\left(\sqrt{\frac{r_{0}}{r_{\mathrm{m}}}-1}\right), (76)

where 𝒜\mathcal{A} and \mathcal{B} are constant.

We separate the integral II as

I=I𝒟+I,\displaystyle I=I_{\mathcal{D}}+I_{\mathcal{R}}, (77)

where a divergent part I𝒟I_{\mathcal{D}} and a regular part II_{\mathcal{R}} are defined by

I𝒟01S(0,rm)h(z,rm)𝑑z,\displaystyle I_{\mathcal{D}}\equiv\int^{1}_{0}S(0,r_{\mathrm{m}})h(z,r_{\mathrm{m}})dz, (78)
I01q(z,r0)𝑑z,\displaystyle I_{\mathcal{R}}\equiv\int^{1}_{0}q(z,r_{0})dz, (79)

respectively, and where

q(z,r0)S(z,r0)h(z,r0)S(0,rm)h(z,rm).\displaystyle q(z,r_{0})\equiv S(z,r_{0})h(z,r_{0})-S(0,r_{\mathrm{m}})h(z,r_{\mathrm{m}}). (80)

The divergent part I𝒟I_{\mathcal{D}} can be integrated as

I𝒟\displaystyle I_{\mathcal{D}} \displaystyle\sim 43z|z=043\displaystyle\left.\frac{4}{\sqrt{3z}}\right|_{z=0}-\frac{4}{3} (81)
=\displaystyle= 433r0rm143+O(r0rm1).\displaystyle\frac{4\sqrt{3}}{3\sqrt{\frac{r_{0}}{r_{\mathrm{m}}}-1}}-\frac{4}{3}+O\left(\sqrt{\frac{r_{0}}{r_{\mathrm{m}}}-1}\right).

From Eq. (42), I𝒟=I𝒟(b)I_{\mathcal{D}}=I_{\mathcal{D}}(b) is given by

I𝒟(b)=274314(bbm1)1443+O((bbm1)34).\displaystyle I_{\mathcal{D}}(b)=\frac{2^{\frac{7}{4}}3^{-\frac{1}{4}}}{\left(\frac{b}{b_{\mathrm{m}}}-1\right)^{\frac{1}{4}}}-\frac{4}{3}+O\left(\left(\frac{b}{b_{\mathrm{m}}}-1\right)^{\frac{3}{4}}\right). (82)

The regular part I(r0)I_{\mathcal{R}}(r_{0}) can be expanded in powers of (r0rm)(r_{0}-r_{\mathrm{m}}) as

I(r0)=j=01j!(r0rm)j01jqr0j|r0=rmdz\displaystyle I_{\mathcal{R}}(r_{0})=\sum^{\infty}_{j=0}\frac{1}{j!}(r_{0}-r_{\mathrm{m}})^{j}\int^{1}_{0}\left.\frac{\partial^{j}q}{\partial r^{j}_{0}}\right|_{r_{0}=r_{\mathrm{m}}}dz (83)

and the term of j=0j=0 gives

I(r0)=01q(z,rm)𝑑z+O(r0rm1),\displaystyle I_{\mathcal{R}}(r_{0})=\int^{1}_{0}q(z,r_{\mathrm{m}})dz+O\left(\sqrt{\frac{r_{0}}{r_{\mathrm{m}}}-1}\right), (84)

where q(z,rm)q(z,r_{\mathrm{m}}) is given by

q(z,rm)=2(1+2z1)z3z2z2.\displaystyle q(z,r_{\mathrm{m}})=\frac{2\left(\sqrt{1+2z}-1\right)}{z\sqrt{3z-2z^{2}}}. (85)

We obtain I(rm)=2.3671I_{\mathcal{R}}(r_{\mathrm{m}})=2.3671 in a numerical calculation. Therefore, the deflection angle of the light in the strong deflection limit is obtained as

αdef(b)=c¯(bbm1)14+d¯+O((bbm1)34),\displaystyle\alpha_{\mathrm{def}}(b)=\frac{\bar{c}}{\left(\frac{b}{b_{\mathrm{m}}}-1\right)^{\frac{1}{4}}}+\bar{d}+O\left(\left(\frac{b}{b_{\mathrm{m}}}-1\right)^{\frac{3}{4}}\right), (86)

where c¯\bar{c} and d¯\bar{d} are given by

c¯274314=2.5558\displaystyle\bar{c}\equiv 2^{\frac{7}{4}}3^{-\frac{1}{4}}=2.5558
d¯43+I(rm)π=2.1078.\displaystyle\bar{d}\equiv-\frac{4}{3}+I_{\mathcal{R}}(r_{\mathrm{m}})-\pi=-2.1078. (87)

IV Observables in the strong deflection limit

We consider a small angle lens equation Bozza:2008ev

DLSα¯=DOS(θϕ),\displaystyle D_{\mathrm{LS}}\bar{\alpha}=D_{\mathrm{OS}}(\theta-\phi), (88)

where DLSD_{\mathrm{LS}} and DOSD_{\mathrm{OS}} are angular distances between a lens object and a source object and between the observer and the source object, respectively, α¯\bar{\alpha} is an effective deflection angle defined by

α¯αdefmod 2π,\displaystyle\bar{\alpha}\equiv\alpha_{\mathrm{def}}\;\mathrm{mod}\;2\pi, (89)

θ\theta is an image angle, and ϕ\phi is a source angle as shown Fig. 4.

Refer to caption
Figure 4: Lens configuration. A light ray, which is emitted by a source S with a source angle ϕ\phi and reflected by a lens L with a deflection angle αdef\alpha_{\mathrm{def}} and an impact parameter bb, reaches an observer O. The observer sees an image I with an image angle θ\theta. Here, α¯=αdef2πn\bar{\alpha}=\alpha_{\mathrm{def}}-2\pi n is the effective deflection angle of the light with a winding number nn. DOSD_{\mathrm{OS}}, DLSD_{\mathrm{LS}}, and DOL=DOSDLSD_{\mathrm{OL}}=D_{\mathrm{OS}}-D_{\mathrm{LS}} are distances between the observer and the source, the lens and the source, and the observer and the lens, respectively.

Here, we have assumed that all the angles are small, i.e., α¯\bar{\alpha}, θ\theta, ϕ1\phi\ll 1. Under the assumption, an impact parameter bb is expressed by b=θDOLb=\theta D_{\mathrm{OL}}, where DOL=DOSDLSD_{\mathrm{OL}}=D_{\mathrm{OS}}-D_{\mathrm{LS}} is an angular distance between the observer and the lens.

We can express the deflection angle αdef\alpha_{\mathrm{def}} of a light ray which rotates around the photon sphere nn times as

αdef=α¯+2πn,\displaystyle\alpha_{\mathrm{def}}=\bar{\alpha}+2\pi n, (90)

where the winding number nn is a positive integer in this section and n=0n=0 in Sec. V. We define θn0\theta^{0}_{n} as

αdef(θn0)=2πn.\displaystyle\alpha_{\mathrm{def}}(\theta^{0}_{n})=2\pi n. (91)

We expand the deflection angle αdef(θ)\alpha_{\mathrm{def}}(\theta) around θ=θn0\theta=\theta^{0}_{n} as

αdef(θ)\displaystyle\alpha_{\mathrm{def}}(\theta) =\displaystyle= αdef(θn0)+dαdefdθ|θ=θn0(θθn0)\displaystyle\alpha_{\mathrm{def}}(\theta^{0}_{n})+\left.\frac{d\alpha_{\mathrm{def}}}{d\theta}\right|_{\theta=\theta^{0}_{n}}\left(\theta-\theta^{0}_{n}\right) (92)
+O((θθn0)2).\displaystyle+O\left(\left(\theta-\theta^{0}_{n}\right)^{2}\right).

IV.1 λ2/2\lambda\neq\sqrt{2}/2

For λ2/2\lambda\neq\sqrt{2}/2, we can rewrite the deflection angle in the strong deflection limit as

αdef(θ)\displaystyle\alpha_{\mathrm{def}}(\theta) =\displaystyle= a¯log(θθ1)+b¯\displaystyle-\bar{a}\log\left(\frac{\theta}{\theta_{\infty}}-1\right)+\bar{b} (93)
+O((θθ1)log(θθ1)),\displaystyle+O\left(\left(\frac{\theta}{\theta_{\infty}}-1\right)\log\left(\frac{\theta}{\theta_{\infty}}-1\right)\right),

where θbm/DOL\theta_{\infty}\equiv b_{\mathrm{m}}/D_{\mathrm{OL}} is the image angle of the photon sphere 444 It is known that Eq. (93) is a good approximation in some examples by comparing the exact deflection angle even if the winding number n=1n=1. See Ref. Tsukamoto:2014dta as an example.. In this case, we obtain

dαdefdθ|θ=θn0=a¯θn0θ\displaystyle\left.\frac{d\alpha_{\mathrm{def}}}{d\theta}\right|_{\theta=\theta^{0}_{n}}=-\frac{\bar{a}}{\theta^{0}_{n}-\theta_{\infty}} (94)

and, from Eqs. (91) and (93),

θn0=[1+exp(b¯2πna¯)]θ.\displaystyle\theta^{0}_{n}=\left[1+\exp\left(\frac{\bar{b}-2\pi n}{\bar{a}}\right)\right]\theta_{\infty}. (95)

We consider the solution θ=θn\theta=\theta_{n} of the lens equation (88) for the light ray with the winding number nn. From Eqs. (90)-(92), (94), and (95), we get

α¯(θn)=a¯eb¯+2πna¯θ(θn0θn).\displaystyle\bar{\alpha}(\theta_{n})=\frac{\bar{a}e^{\frac{-\bar{b}+2\pi n}{\bar{a}}}}{\theta_{\infty}}\left(\theta^{0}_{n}-\theta_{n}\right). (96)

From Eqs. (88) and (96), we obtain the image angle

θn(ϕ)θn0+θeb¯2πna¯DOS(ϕθn0)a¯DLS\displaystyle\theta_{n}(\phi)\sim\theta^{0}_{n}+\frac{\theta_{\infty}e^{\frac{\bar{b}-2\pi n}{\bar{a}}}D_{\mathrm{OS}}\left(\phi-\theta^{0}_{n}\right)}{\bar{a}D_{\mathrm{LS}}} (97)

and its Einstein ring angle θEn\theta_{\mathrm{E}n} as

θEnθn(0)=θn0[1θeb¯2πna¯DOSa¯DLS].\displaystyle\theta_{\mathrm{E}n}\equiv\theta_{n}(0)=\theta^{0}_{n}\left[1-\frac{\theta_{\infty}e^{\frac{\bar{b}-2\pi n}{\bar{a}}}D_{\mathrm{OS}}}{\bar{a}D_{\mathrm{LS}}}\right]. (98)

We get the magnification of the image

μn\displaystyle\mu_{n} \displaystyle\equiv θnϕdθndϕ\displaystyle\frac{\theta_{n}}{\phi}\frac{d\theta_{n}}{d\phi} (99)
\displaystyle\sim θ2DOS(1+eb¯2πna¯)eb¯2πna¯a¯DLSϕ,\displaystyle\frac{\theta_{\infty}^{2}D_{\mathrm{OS}}\left(1+e^{\frac{\bar{b}-2\pi n}{\bar{a}}}\right)e^{\frac{\bar{b}-2\pi n}{\bar{a}}}}{\bar{a}D_{\mathrm{LS}}\phi},

the sum of the magnifications of the infinite numbers of images

n=1μnθ2DOS(1+e2πa¯+eb¯a¯)eb¯a¯a¯DLSϕ(e4πa¯1),\displaystyle\sum^{\infty}_{n=1}\mu_{n}\sim\frac{\theta_{\infty}^{2}D_{\mathrm{OS}}\left(1+e^{\frac{2\pi}{\bar{a}}}+e^{\frac{\bar{b}}{\bar{a}}}\right)e^{\frac{\bar{b}}{\bar{a}}}}{\bar{a}D_{\mathrm{LS}}\phi\left(e^{\frac{4\pi}{\bar{a}}}-1\right)}, (100)

the magnification without the outermost image

n=2μnθ2DOS(e4πa¯+e2πa¯+eb¯a¯)eb¯4πa¯a¯DLSϕ(e4πa¯1),\displaystyle\sum^{\infty}_{n=2}\mu_{n}\sim\frac{\theta_{\infty}^{2}D_{\mathrm{OS}}\left(e^{\frac{4\pi}{\bar{a}}}+e^{\frac{2\pi}{\bar{a}}}+e^{\frac{\bar{b}}{\bar{a}}}\right)e^{\frac{\bar{b}-4\pi}{\bar{a}}}}{\bar{a}D_{\mathrm{LS}}\phi\left(e^{\frac{4\pi}{\bar{a}}}-1\right)}, (101)

and the ratio of the magnification of the outermost image to the others

rμ1n=2μn(e4πa¯1)(e2πa¯+eb¯a¯)e4πa¯+e2πa¯+eb¯a¯.\displaystyle\mathrm{r}\equiv\frac{\mu_{1}}{\sum^{\infty}_{n=2}\mu_{n}}\sim\frac{\left(e^{\frac{4\pi}{\bar{a}}}-1\right)\left(e^{\frac{2\pi}{\bar{a}}}+e^{\frac{\bar{b}}{\bar{a}}}\right)}{e^{\frac{4\pi}{\bar{a}}}+e^{\frac{2\pi}{\bar{a}}}+e^{\frac{\bar{b}}{\bar{a}}}}. (102)

The difference of the image angles between the outermost images and the photon sphere is obtained as

sθ1θθ10θ0=θeb¯2πa¯.\displaystyle\mathrm{s}\equiv\theta_{1}-\theta_{\infty}\sim\theta^{0}_{1}-\theta^{0}_{\infty}=\theta_{\infty}e^{\frac{\bar{b}-2\pi}{\bar{a}}}. (103)

IV.2 λ=2/2\lambda=\sqrt{2}/2

For λ=2/2\lambda=\sqrt{2}/2, the deflection angle in the strong deflection limit can be expressed by

αdef(θ)=c¯(θθ1)14+d¯+O((θθ1)34).\displaystyle\alpha_{\mathrm{def}}(\theta)=\frac{\bar{c}}{\left(\frac{\theta}{\theta_{\infty}}-1\right)^{\frac{1}{4}}}+\bar{d}+O\left(\left(\frac{\theta}{\theta_{\infty}}-1\right)^{\frac{3}{4}}\right). (104)

In this case, we obtain

dαdefdθ|θ=θn0=c¯4θ(θn0θ1)54\displaystyle\left.\frac{d\alpha_{\mathrm{def}}}{d\theta}\right|_{\theta=\theta^{0}_{n}}=-\frac{\bar{c}}{4\theta_{\infty}}\left(\frac{\theta^{0}_{n}}{\theta_{\infty}}-1\right)^{-\frac{5}{4}} (105)

and, from Eqs. (91) and (104),

θn0=[1+(c¯2πnd¯)4]θ.\displaystyle\theta^{0}_{n}=\left[1+\left(\frac{\bar{c}}{2\pi n-\bar{d}}\right)^{4}\right]\theta_{\infty}. (106)

Thus, from Eqs. (90)-(92), (105), and (106), the effective deflection angle is obtained by

α¯(θn)=(2πnd¯)54θc¯4(θn0θn).\displaystyle\bar{\alpha}(\theta_{n})=\frac{\left(2\pi n-\bar{d}\right)^{5}}{4\theta_{\infty}\bar{c}^{4}}\left(\theta^{0}_{n}-\theta_{n}\right). (107)

From Eqs. (88) and (107), we get the image angle with the winding number nn

θn(ϕ)θn0+4c¯4DOSθ(ϕθn0)(2πnd¯)5DLS\displaystyle\theta_{n}(\phi)\sim\theta^{0}_{n}+\frac{4\bar{c}^{4}D_{\mathrm{OS}}\theta_{\infty}(\phi-\theta^{0}_{n})}{(2\pi n-\bar{d})^{5}D_{\mathrm{LS}}} (108)

and the Einstein ring angle

θEnθn(0)θn0[14c¯4DOSθ(2πnd¯)5DLS].\displaystyle\theta_{\mathrm{E}n}\equiv\theta_{n}(0)\sim\theta^{0}_{n}\left[1-\frac{4\bar{c}^{4}D_{\mathrm{OS}}\theta_{\infty}}{(2\pi n-\bar{d})^{5}D_{\mathrm{LS}}}\right]. (109)

Therefore, we obtain the magnification of the image

μnθnϕdθndϕ4c¯4DOSθ2F(n)DLSϕ,\displaystyle\mu_{n}\equiv\frac{\theta_{n}}{\phi}\frac{d\theta_{n}}{d\phi}\sim\frac{4\bar{c}^{4}D_{\mathrm{OS}}\theta_{\infty}^{2}F(n)}{D_{\mathrm{LS}}\phi}, (110)

where F(n)F(n) defined by

F(n)1+(c¯2πnd¯)4(2πnd¯)5\displaystyle F(n)\equiv\frac{1+\left(\frac{\bar{c}}{2\pi n-\bar{d}}\right)^{4}}{(2\pi n-\bar{d})^{5}} (111)

can be calculated numerically. We can calculate the magnification of the infinite number of images

n=1μn4c¯4DOSθ2DLSϕn=1F(n)\displaystyle\sum^{\infty}_{n=1}\mu_{n}\sim\frac{4\bar{c}^{4}D_{\mathrm{OS}}\theta_{\infty}^{2}}{D_{\mathrm{LS}}\phi}\sum^{\infty}_{n=1}F(n) (112)

and the sum of the magnifications of the images without the outermost image

n=2μn4c¯4DOSθ2DLSϕn=2F(n),\displaystyle\sum^{\infty}_{n=2}\mu_{n}\sim\frac{4\bar{c}^{4}D_{\mathrm{OS}}\theta_{\infty}^{2}}{D_{\mathrm{LS}}\phi}\sum^{\infty}_{n=2}F(n), (113)

where

n=1F(n)2.6077×105,\displaystyle\sum^{\infty}_{n=1}F(n)\sim 2.6077\times 10^{-5}, (114)
n=2F(n)1.8303×106.\displaystyle\sum^{\infty}_{n=2}F(n)\sim 1.8303\times 10^{-6}. (115)

Therefore, the ratio of the magnifications of the outermost image to the other images is obtained as

r=μ1n=2μnF(1)n=2F(n)=13.248,\displaystyle\mathrm{r}=\frac{\mu_{1}}{\sum^{\infty}_{n=2}\mu_{n}}\sim\frac{F(1)}{\sum^{\infty}_{n=2}F(n)}=13.248, (116)

where we have used

F(1)2.4247×105.\displaystyle F(1)\sim 2.4247\times 10^{-5}. (117)

The difference of the image angles between the outermost images and the photon sphere is

s=θ1θθ10θ0=(c¯2πd¯)4θ.\displaystyle\mathrm{s}=\theta_{1}-\theta_{\infty}\sim\theta^{0}_{1}-\theta^{0}_{\infty}=\left(\frac{\bar{c}}{2\pi-\bar{d}}\right)^{4}\theta_{\infty}. (118)

V Gravitational lens under the weak-field approximation

Under a weak-field approximation M/r01M/r_{0}\ll 1, the line element is given by

ds2\displaystyle ds^{2} =\displaystyle= (12Mr)dt2+[1+2M(1+λ2)r]dr2\displaystyle-\left(1-\frac{2M}{r}\right)dt^{2}+\left[1+\frac{2M\left(1+\lambda^{2}\right)}{r}\right]dr^{2} (119)
+r2(dϑ2+sin2ϑdφ2).\displaystyle+r^{2}\left(d\vartheta^{2}+\sin^{2}\vartheta d\varphi^{2}\right).

From Eq. (8.5.8)(8.5.8) in Ref. Weinberg:1972kfs , the deflection angle αdef\alpha_{\mathrm{def}} of a light ray is obtained as

αdef=4Mr0+O((Mr0)2)\displaystyle\alpha_{\mathrm{def}}=\frac{4M_{*}}{r_{0}}+O\left(\left(\frac{M_{*}}{r_{0}}\right)^{2}\right) (120)

where MM_{*} is

MM(1+λ22).\displaystyle M_{*}\equiv M\left(1+\frac{\lambda^{2}}{2}\right). (121)

From Eq. (23), b/r0=1+O(M/r0)b/r_{0}=1+O(M/r_{0}) is satisfied. Thus, the deflection angle is rewritten as

αdef=4Mb+O((Mb)2).\displaystyle\alpha_{\mathrm{def}}=\frac{4M_{*}}{b}+O\left(\left(\frac{M_{*}}{b}\right)^{2}\right). (122)

This is the same as Eq. (2.21)(2.21) in Ref. Ovgun:2018fnk .

From Eqs. (90), (122)(\ref{eq:deflection_angle_weak2}), n=0n=0, and b=θDOLb=\theta D_{\mathrm{OL}}, the solution θ=θ±0(ϕ)\theta=\theta_{\pm 0}(\phi) of the lens equation (88) is given by

θ±0(ϕ)=12(ϕ±ϕ2+4θE02),\displaystyle\theta_{\pm 0}(\phi)=\frac{1}{2}\left(\phi\pm\sqrt{\phi^{2}+4\theta_{\mathrm{E}0}^{2}}\right), (123)

where θE0\theta_{\mathrm{E}0} is the angle of an Einstein ring defined as

θE0θ+0(0)=4MDLSDOSDOL.\displaystyle\theta_{\mathrm{E}0}\equiv\theta_{+0}(0)=\sqrt{\frac{4M_{*}D_{\mathrm{LS}}}{D_{\mathrm{OS}}D_{\mathrm{OL}}}}. (124)

Notice that θ0(ϕ)\theta_{-0}(\phi) has a negative value and its impact parameter is also negative. We obtain the magnifications of the images

μ±0\displaystyle\mu_{\pm 0} \displaystyle\equiv θ±0ϕdθ±0dϕ\displaystyle\frac{\theta_{\pm 0}}{\phi}\frac{d\theta_{\pm 0}}{d\phi} (125)
=\displaystyle= 14(2±ϕϕ2+4θE02±ϕ2+4θE02ϕ)\displaystyle\frac{1}{4}\left(2\pm\frac{\phi}{\sqrt{\phi^{2}+4\theta_{\mathrm{E}0}^{2}}}\pm\frac{\sqrt{\phi^{2}+4\theta_{\mathrm{E}0}^{2}}}{\phi}\right)

and its total magnification

μ0tot\displaystyle\mu_{0\mathrm{tot}} \displaystyle\equiv |μ+0|+|μ0|\displaystyle\left|\mu_{+0}\right|+\left|\mu_{-0}\right| (126)
=\displaystyle= 12(ϕϕ2+4θE02+ϕ2+4θE02ϕ).\displaystyle\frac{1}{2}\left(\frac{\phi}{\sqrt{\phi^{2}+4\theta_{\mathrm{E}0}^{2}}}+\frac{\sqrt{\phi^{2}+4\theta_{\mathrm{E}0}^{2}}}{\phi}\right).

VI Discussion and conclusion

From Secs. I to IV, we have concentrated on an infinite number of images with positive impact parameters bb or positive image angles θn\theta_{n}. The each image has a partner with a negative impact parameter. The image angle θn\theta_{-n} and the magnification μn\mu_{-n} of the partner of the image with θn\theta_{n} are given by θnθn\theta_{-n}\sim-\theta_{n} and μnμn\mu_{-n}\sim-\mu_{n}, respectively. Thus, the diameter of the pair images on a sky is obtained as θnθn2θn\theta_{n}-\theta_{-n}\sim 2\theta_{n} and the total magnification of the pair images is given by μntot|μn|+|μn|2|μn|\mu_{n\mathrm{tot}}\equiv\left|\mu_{n}\right|+\left|\mu_{-n}\right|\sim 2\left|\mu_{n}\right|. The observables and the parameters a¯\bar{a}, b¯\bar{b}, c¯\bar{c}, and d¯\bar{d} of the deflection angle in the strong deflection limit are summarized in Table II.

Table 2: Parameters a¯\bar{a}, b¯\bar{b}, c¯\bar{c}, and d¯\bar{d} in the deflection angle in the strong deflection limit, the diameters of the innermost ring 2θ2\theta_{\infty}, and the outermost ring 2θE12\theta_{\mathrm{E}1} among rings scattered by the photon sphere, the difference of the radii of the outermost ring and the innermost ring s=θE1θ\mathrm{s}=\theta_{\mathrm{E}1}-\theta_{\infty}, the magnification of the pair images of the outermost ring μ1tot2|μ1|\mu_{1\mathrm{tot}}\sim 2\left|\mu_{1}\right|, the ratio of the magnification of the outermost ring to the other rings r=μ1/n=2μn\mathrm{r}=\mu_{1}/\sum^{\infty}_{n=2}\mu_{n} for given λ\lambda. Here we have set DOS=16D_{\mathrm{OS}}=16kpc, DOL=8D_{\mathrm{OL}}=8kpc, DLS=DOSDOL=8D_{\mathrm{LS}}=D_{\mathrm{OS}}-D_{\mathrm{OL}}=8kpc, M=M(1+λ2/2)=4×106MM_{*}=M\left(1+\lambda^{2}/2\right)=4\times 10^{6}M_{\odot}, and the source angle ϕ=1\phi=1 arcsecond for μ1tot\mu_{1\mathrm{tot}}. Notice that the diameter of the Einstein ring 2θE0=2.86182\theta_{\mathrm{E}0}=2.8618 arcsecond and the magnification of a pair of images μ0tot=1.6807\mu_{0\mathrm{tot}}=1.6807 for the source angle ϕ=1\phi=1 arcsecond do not depend on λ\lambda under the weak-field approximation since we make MM_{*} constant. We find that 2θ2\theta_{\infty} monotonically decreases as λ\lambda increases from 0 to 2\sqrt{2} and it monotonically increases as λ\lambda increases from 2\sqrt{2} to \infty if MM_{*} is constant.
λ\lambda 0 0.40.4 0.60.6 0.70.7 2/2\sqrt{2}/2 0.710.71 0.80.8 11 2\sqrt{2} 55
a¯\bar{a} 1.00001.0000 1.21271.2127 1.88981.8898 7.07117.0711 -- 15.68115.681 3.02373.0237 2.0002.000 1.63301.6330 1.42861.4286
b¯\bar{b} 0.4002-0.4002 0.7591-0.7591 2.8784-2.8784 40.837-40.837 -- 116.02-116.02 4.8720-4.8720 1.3632-1.3632 0.77941-0.77941 0.7264-0.7264
c¯\bar{c} -- -- -- -- 2.55582.5558 -- -- -- -- --
d¯\bar{d} -- -- -- -- 2.1078-2.1078 -- -- -- -- --
2θ2\theta_{\infty} [μ\muas] 51.58051.580 47.75947.759 43.71243.712 41.43041.430 41.26441.264 41.19741.197 39.48539.485 37.43637.436 36.47336.473 38.99338.993
2θE12\theta_{\mathrm{E}1} [μ\muas] 51.64551.645 47.90347.903 44.05544.055 41.48341.483 41.61941.619 41.21441.214 40.47240.472 38.25438.254 36.95536.955 39.28239.282
s\mathrm{s} [μ\muas] 0.0322770.032277 0.0717850.071785 0.171450.17145 0.0264390.026439 0.177580.17758 0.00844460.0084446 0.493380.49338 0.409130.40913 0.241350.24135 0.144230.14423
μ1tot×1017\mu_{1\mathrm{tot}}\times 10^{17} 1.61631.6163 2.74952.7495 3.87553.8755 0.150390.15039 3.41623.4162 0.0215210.021521 6.40336.4033 7.58787.5878 5.29605.2960 3.84533.8453
r\mathrm{r} 535.16535.16 177.42177.42 26.99626.996 1.43301.4330 13.24813.248 0.492970.49297 7.14317.1431 22.60422.604 46.47646.476 80.88680.886

As shown in Sec V, the gravitational lensing under the weak-field approximation is not characterized by MM but MM_{*}. Under an assumption that MM_{*} is constant, the size of the photon sphere θ\theta_{\infty} monotonically decreases as λ\lambda increases from 0 to 2\sqrt{2} and θ\theta_{\infty} monotonically increases as λ\lambda increases from 2\sqrt{2} to \infty. The minimal value of θ\theta_{\infty} is given by 36M/(4DOL)3\sqrt{6}M_{*}/(4D_{\mathrm{OL}}) for λ=2\lambda=\sqrt{2}555When MM is constant, the size of the photon sphere θ\theta_{\infty} monotonically increases as λ\lambda increases from 0 to \infty and it takes a constant and minimum value 33M/DOL3\sqrt{3}M/D_{\mathrm{OL}} for λ2/2\lambda\leq\sqrt{2}/2.

We summarize our result. We have shown that the Damour-Solodukhin wormhole with two flat regions has two photon spheres and an antiphoton sphere for λ<2/2\lambda<\sqrt{2}/2 and only one photon sphere for λ2/2\lambda\geq\sqrt{2}/2 and the photon sphere is marginally unstable when λ=2/2\lambda=\sqrt{2}/2. We have reexamined that deflection angle in the strong deflection limit for λ<2/2\lambda<\sqrt{2}/2 and we have extent the analysis for λ=2/2\lambda=\sqrt{2}/2 and λ>2/2\lambda>\sqrt{2}/2. We have found that the deflection angle of a light ray reflected by the marginally unstable photon sphere diverges nonlogarithmically in the strong deflection limit for λ=2/2\lambda=\sqrt{2}/2, while the deflection angle of the light reflected by the photon sphere diverges logarithmically for λ2/2\lambda\neq\sqrt{2}/2. We expect that our method can be applied for gravitational lenses by marginally unstable photon spheres of various compact objects.

Acknowledgements

The author thanks R. Izmailov, K. K. Nandi, A. Övgün, and an anonymous referee for their useful comments.

Appendix A Arnowitt-Deser-Misner masses

Wormholes have two ADM masses since the ADM mass is defined in every asymptotically flat region Poisson . For simplicity, we have assumed the equal ADM masses of the Damour-Solodukhin wormhole. See Visser Visser_1995 for the details of the mass of the wormholes.

We show that MADM=M(1+λ2)M_{\mathrm{ADM}}=M\left(1+\lambda^{2}\right) is the ADM mass of the wormhole. By using a radial coordinate rr_{*}, which is given by

r12[rM(1+λ2)+r22M(1+λ2)r]\displaystyle r_{*}\equiv\frac{1}{2}\left[r-M\left(1+\lambda^{2}\right)+\sqrt{r^{2}-2M\left(1+\lambda^{2}\right)r}\right] (127)

or

r=r[1+M(1+λ2)2r]2,\displaystyle r=r_{*}\left[1+\frac{M\left(1+\lambda^{2}\right)}{2r_{*}}\right]^{2}, (128)

the line element (119)(\ref{eq:line_element0}) under the weak field approximation is rewritten as

ds2\displaystyle ds^{2} =\displaystyle= (12Mr(r))dt2\displaystyle-\left(1-\frac{2M}{r(r_{*})}\right)dt^{2}
+[1+2M(1+λ2)r][dr2+r2(dϑ2+sin2ϑdφ2)].\displaystyle+\left[1+\frac{2M\left(1+\lambda^{2}\right)}{r_{*}}\right]\left[dr_{*}^{2}+r_{*}^{2}\left(d\vartheta^{2}+\sin^{2}\vartheta d\varphi^{2}\right)\right].

We consider the hypersurfaces Σt\Sigma_{\mathrm{t}}, which are surfaces of constant tt with a unit normal nα=(1M/r)αtn_{\alpha}=-(1-M/r)\partial_{\alpha}t. The induced metric on Σt\Sigma_{\mathrm{t}} is obtained as

habdyadyb=[1+2M(1+λ2)r][dr2+r2(dϑ2+sin2ϑdφ2)].\displaystyle h_{ab}dy^{a}dy^{b}=\left[1+\frac{2M\left(1+\lambda^{2}\right)}{r_{*}}\right]\left[dr_{*}^{2}+r_{*}^{2}\left(d\vartheta^{2}+\sin^{2}\vartheta d\varphi^{2}\right)\right].

The induced metric on a two-sphere StS_{\mathrm{t}} at r=Rtr_{*}=R_{\mathrm{t}} with a unit normal ra=[1+M(1+λ2)/r]arr_{a}=\left[1+M(1+\lambda^{2})/r_{*}\right]\partial_{a}r_{*} is

σABdθAdθB=[1+2M(1+λ2)Rt]Rt2(dϑ2+sin2ϑdφ2).\sigma_{AB}d\theta^{A}d\theta^{B}=\left[1+\frac{2M\left(1+\lambda^{2}\right)}{R_{\mathrm{t}}}\right]R_{\mathrm{t}}^{2}\left(d\vartheta^{2}+\sin^{2}\vartheta d\varphi^{2}\right). (131)

The extrinsic curvature of StS_{\mathrm{t}} embedded in Σt\Sigma_{\mathrm{t}} is obtained as k=σABkAB=r|aa=2[Rt2M(1+λ2)]/Rt2k=\sigma^{AB}k_{AB}=r^{a}_{\;\;\left|a\right.}=2\left[R_{\mathrm{t}}-2M\left(1+\lambda^{2}\right)\right]/R_{\mathrm{t}}^{2}, where |a\,{}_{\left|a\right.} is the covariant differentiation on Σt\Sigma_{\mathrm{t}}. The extrinsic curvature of StS_{\mathrm{t}} embedded in flat space is given by k0=2[RtM(1+λ2)]/Rt2k_{0}=2\left[R_{\mathrm{t}}-M\left(1+\lambda^{2}\right)\right]/R_{\mathrm{t}}^{2}. The ADM mass is obtained as

MADM\displaystyle M_{\mathrm{ADM}} \displaystyle\equiv 18πlimSt±St(kk0)σd2θ\displaystyle-\frac{1}{8\pi}\lim_{S_{\mathrm{t}}\rightarrow\pm\infty}\oint_{S_{\mathrm{t}}}(k-k_{0})\sqrt{\sigma}d^{2}\theta (132)
=\displaystyle= M(1+λ2),\displaystyle M(1+\lambda^{2}),

where σ=Rt4sin2ϑ\sigma=R_{\mathrm{t}}^{4}\sin^{2}\vartheta. Therefore, the ADM mass is not MM but M(1+λ2)M(1+\lambda^{2}).

Appendix B Violation of energy conditions

The Ricci tensor and Ricci scalar are given by

tt=M2λ2r3(2Mr),\displaystyle\mathcal{R}_{tt}=-\frac{M^{2}\lambda^{2}}{r^{3}(2M-r)}, (133)
rr=(2r3M)Mλ2r(r2M)2[r2M(1+λ2)],\displaystyle\mathcal{R}_{rr}=-\frac{(2r-3M)M\lambda^{2}}{r(r-2M)^{2}\left[r-2M\left(1+\lambda^{2}\right)\right]}, (134)
ϑϑ=λ2M2Mr,\displaystyle\mathcal{R}_{\vartheta\vartheta}=-\frac{\lambda^{2}M}{2M-r}, (135)
φφ=sin2ϑϑϑ.\displaystyle\mathcal{R}_{\varphi\varphi}=\sin^{2}\vartheta\mathcal{R}_{\vartheta\vartheta}. (136)

and

=2M2λ2r2(r2M)2,\displaystyle\mathcal{R}=-\frac{2M^{2}\lambda^{2}}{r^{2}(r-2M)^{2}}, (137)

respectively. The Einstein tensor is given by

Gtt=0,\displaystyle G^{t}_{\;\>t}=0, (138)
Grr=2Mλ2r2(r2M),\displaystyle G^{r}_{\;\>r}=-\frac{2M\lambda^{2}}{r^{2}(r-2M)}, (139)
Gϑϑ=Gφφ=M(rM)λ2r2(r2M)2.\displaystyle G^{\vartheta}_{\;\>\vartheta}=G^{\varphi}_{\;\>\varphi}=\frac{M(r-M)\lambda^{2}}{r^{2}(r-2M)^{2}}. (140)

The nonzero components of the stress energy tensor 𝒯νμ\mathcal{T}^{\mu}_{\;\>\nu} are 𝒯tt=ρ\mathcal{T}^{t}_{\;\>t}=-\rho, 𝒯rr=p\mathcal{T}^{r}_{\;\>r}=p, and 𝒯ϑϑ=𝒯φφ=pT\mathcal{T}^{\vartheta}_{\;\>\vartheta}=\mathcal{T}^{\varphi}_{\;\>\varphi}=p_{\mathrm{T}}, where ρ\rho is the energy density, pp is the radial pressure, and pTp_{\mathrm{T}} is the tangential pressure.

From the Einstein equation Gνμ=8π𝒯νμG^{\mu}_{\;\>\nu}=8\pi\mathcal{T}^{\mu}_{\;\>\nu}, we obtain

ρ=0,\displaystyle\rho=0, (141)
p=Mλ24πr2(r2M),\displaystyle p=-\frac{M\lambda^{2}}{4\pi r^{2}(r-2M)}, (142)
pT=M(rM)λ28πr2(r2M)2.\displaystyle p_{\mathrm{T}}=\frac{M(r-M)\lambda^{2}}{8\pi r^{2}(r-2M)^{2}}. (143)

The weak, null, and strong energy conditions Visser_1995 are violated everywhere because of ρ+p<0\rho+p<0.

References

  • (1) B. P. Abbott et al. [LIGO Scientific and Virgo Collaborations], Phys. Rev. Lett.  116, 061102 (2016).
  • (2) B. P. Abbott et al. [LIGO Scientific and Virgo Collaborations], Phys. Rev. X 9, 031040 (2019).
  • (3) K. Akiyama et al. [Event Horizon Telescope Collaboration], Astrophys. J.  875, L1 (2019).
  • (4) C. M. Claudel, K. S. Virbhadra and G. F. R. Ellis, J. Math. Phys.  42, 818 (2001).
  • (5) V. Perlick, Living Rev. Relativity 7, 9 (2004).
  • (6) S. Hod, Phys. Lett. B 727, 345 (2013).
  • (7) Y. Hagihara, Jpn. J. Astron. Geophys., 8, 67 (1931).
  • (8) C. Darwin, Proc. R. Soc. Lond. A 249, 180 (1959).
  • (9) R. d’ E. Atkinson, Astron. J. 70, 517 (1965).
  • (10) J.-P. Luminet, Astron. Astrophys. 75, 228 (1979).
  • (11) H. C. Ohanian, Am. J. Phys. 55, 428 (1987).
  • (12) R. J. Nemiroff, Am. J. Phys. 61, 619 (1993).
  • (13) S. Frittelli, T. P. Kling, and E. T. Newman, Phys. Rev. D 61, 064021 (2000).
  • (14) K. S. Virbhadra and G. F. R. Ellis, Phys. Rev. D 62, 084003 (2000).
  • (15) V. Bozza, S. Capozziello, G. Iovane, and G. Scarpetta, Gen. Relativ. Gravit. 33, 1535 (2001).
  • (16) V. Bozza, Phys. Rev. D 66, 103001 (2002).
  • (17) V. Perlick, Phys. Rev. D 69, 064017 (2004).
  • (18) V. Bozza, Gen. Relativ. Gravit. 42, 2269 (2010).
  • (19) W. L. Ames and K. S. Thorne, Astrophys. J. 151, 659 (1968).
  • (20) J. L. Synge, Mon. Not. Roy. Astron. Soc.  131, no. 3, 463 (1966).
  • (21) H. Yoshino, K. Takahashi, and K. i. Nakao, Phys. Rev. D 100, 084062 (2019).
  • (22) N. G. Sanchez, Phys. Rev. D 18, 1030 (1978).
  • (23) Y. Decanini, A. Folacci, and B. Raffaelli, Phys. Rev. D 81, 104039 (2010).
  • (24) W. H. Press, Astrophys. J.  170, L105 (1971).
  • (25) C. J. Goebel, Astrophys. J. 172, L95 (1972).
  • (26) B. Raffaelli, Gen. Rel. Grav.  48, 16 (2016).
  • (27) M. A. Abramowicz and A. R. Prasanna, Mon. Not. Roy. Astr. Soc. 245, 720 (1990).
  • (28) M. A. Abramowicz, Mon. Not. Roy. Astr. Soc. 245, 733 (1990).
  • (29) B. Allen, Nature 347, 615 (1990).
  • (30) W. Hasse and V. Perlick, Gen. Relativ. Gravit. 34, 415 (2002).
  • (31) S. Hod, Phys. Rev. D 84, 104024 (2011).
  • (32) P. Mach, E. Malec, and J. Karkowski, Phys. Rev. D 88, 084056 (2013)
  • (33) E. Chaverra and O. Sarbach, Class. Quant. Grav.  32, 155006 (2015)
  • (34) M. Cvetic, G. W. Gibbons, and C. N. Pope, Phys. Rev. D 94, 106005 (2016).
  • (35) Y. Koga and T. Harada, Phys. Rev. D 94, 044053 (2016).
  • (36) Y. Koga and T. Harada, Phys. Rev. D 98, 024018 (2018).
  • (37) Y. Koga, Phys. Rev. D 99, 064034 (2019).
  • (38) Y. Koga and T. Harada, Phys. Rev. D 100, 064040 (2019).
  • (39) J. Keir, Class. Quant. Grav.  33, 135009 (2016).
  • (40) V. Cardoso, L. C. B. Crispino, C. F. B. Macedo, H. Okawa, and P. Pani, Phys. Rev. D 90, 044069 (2014).
  • (41) P. V. P. Cunha, E. Berti, and C. A. R. Herdeiro, Phys. Rev. Lett.  119, 251102 (2017).
  • (42) P. V. P. Cunha, C. A. R. Herdeiro, and E. Radu, Phys. Rev. D 96, 024039 (2017).
  • (43) S. Hod, Phys. Lett. B 776, 1 (2018).
  • (44) M. Visser, Lorentzian Wormholes: From Einstein to Hawking (American Institute of Physics, Woodbury, NY, 1995).
  • (45) M. S. Morris and K. S. Thorne, Am. J. Phys. 56, 395 (1988).
  • (46) P. Schneider, J. Ehlers, and E. E. Falco, Gravitational Lenses (Springer-Verlag, Berlin, 1992).
  • (47) P. Schneider, C. S. Kochanek, and J. Wambsganss, Gravitational Lensing: Strong, Weak and Micro, Lecture Notes of the 33rd Saas-Fee Advanced Course, edited by G. Meylan, P. Jetzer, and P. North (Springer-Verlag, Berlin, 2006).
  • (48) L. Chetouani and G. Clément, Gen. Relativ. Gravit. 16, 111 (1984).
  • (49) K. K. Nandi, Y. Z. Zhang, and A. V. Zakharov, Phys. Rev. D 74, 024020 (2006).
  • (50) T. Muller, Phys. Rev. D 77, 044043 (2008).
  • (51) N. Tsukamoto and T. Harada, Phys. Rev. D 95, 024030 (2017).
  • (52) N. Tsukamoto, T. Harada, and K. Yajima, Phys. Rev. D 86, 104062 (2012).
  • (53) V. Perlick, AIP Conf. Proc.  1577, 94 (2015).
  • (54) N. Tsukamoto, Phys. Rev. D 94, 124001 (2016).
  • (55) K. K. Nandi, R. N. Izmailov, A. A. Yanbekov, and A. A. Shayakhmetov, Phys. Rev. D 95, 104011 (2017).
  • (56) N. Tsukamoto, Phys. Rev. D 95, 084021 (2017).
  • (57) R. Shaikh, P. Banerjee, S. Paul, and T. Sarkar, Phys. Rev. D 99, 104040 (2019).
  • (58) R. Shaikh, P. Banerjee, S. Paul, and T. Sarkar, JCAP 1907, 028 (2019).
  • (59) R. Shaikh, P. Banerjee, S. Paul, and T. Sarkar, Phys. Lett. B 789, 270 (2019) Erratum: [Phys. Lett. B 791, 422 (2019)].
  • (60) T. Muller, Am. J. Phys. 72, 1045,(2004).
  • (61) O. James, E. von Tunzelmann, P. Franklin, and K. S. Thorne, Am. J. Phys.  83, 486 (2015).
  • (62) T. Ohgami and N. Sakai, Phys. Rev. D 91, 124020 (2015).
  • (63) T. Ohgami and N. Sakai, Phys. Rev. D 94, 064071 (2016).
  • (64) M. Kuniyasu, K. Nanri, N. Sakai, T. Ohgami, R. Fukushige, and S. Komura, Phys. Rev. D 97, 104063 (2018).
  • (65) Y. Nambu, S. Noda, and Y. Sakai, Phys. Rev. D 100, 064037 (2019).
  • (66) V. Cardoso, E. Franzin and P. Pani, Phys. Rev. Lett.  116, no. 17, 171101 (2016) Erratum: [Phys. Rev. Lett. 117, 089902 (2016)].
  • (67) T. Damour and S. N. Solodukhin, Phys. Rev. D 76, 024016 (2007)
  • (68) J. P. S. Lemos and O. B. Zaslavskii, Phys. Rev. D 78, 024040 (2008).
  • (69) R. K. Karimov, R. N. Izmailov, and K. K. Nandi, Eur. Phys. J. C 79, 952 (2019).
  • (70) P. Bueno, P. A. Cano, F. Goelen, T. Hertog, and B. Vercnocke, Phys. Rev. D 97, 024040 (2018).
  • (71) S. H. Volkel and K. D. Kokkotas, Class. Quant. Grav.  35, 105018 (2018).
  • (72) S. Paul, R. Shaikh, P. Banerjee, and T. Sarkar, JCAP 03, 055 (2020).
  • (73) M. Amir, K. Jusufi, A. Banerjee, and S. Hansraj, Class. Quant. Grav.  36, no. 21, 215007 (2019).
  • (74) K. K. Nandi, R. N. Izmailov, E. R. Zhdanov, and A. Bhattacharya, JCAP 1807, 027 (2018).
  • (75) A. Ovgun, Phys. Rev. D 98, 044033 (2018).
  • (76) A. Bhattacharya and R. K. Karimov, arXiv:1811.00768 [gr-qc].
  • (77) A. Ovgun, arXiv:1811.06870 [gr-qc].
  • (78) A. Ovgun, Universe 5, 115 (2019).
  • (79) N. Tsukamoto and T. Kokubu, Phys. Rev. D 101, 044030 (2020).
  • (80) N. Tsukamoto, Phys. Rev. D 95, 064035 (2017).
  • (81) E. F. Eiroa, G. E. Romero, and D. F. Torres, Phys. Rev. D 66, 024010 (2002).
  • (82) V. Bozza, F. De Luca, G. Scarpetta, and M. Sereno, Phys. Rev. D 72, 083003 (2005). [gr-qc/0507137].
  • (83) V. Bozza, F. De Luca, and G. Scarpetta, Phys. Rev. D 74, 063001 (2006).
  • (84) V. Bozza and G. Scarpetta, Phys. Rev. D 76, 083008 (2007).
  • (85) A. Ishihara, Y. Suzuki, T. Ono, and H. Asada, Phys. Rev. D 95, 044017 (2017).
  • (86) S. W. Wei, Y. X. Liu, and H. Guo, Phys. Rev. D 84, 041501 (2011).
  • (87) I. Z. Stefanov, S. S. Yazadjiev, and G. G. Gyulchev, Phys. Rev. Lett.  104, 251103 (2010).
  • (88) S. V. Iyer and A. O. Petters, Gen. Rel. Grav. 39, 1563 (2007).
  • (89) V. Bozza, Phys. Rev. D 78, 103005 (2008).
  • (90) N. Tsukamoto, T. Kitamura, K. Nakajima, and H. Asada, Phys. Rev. D 90, 064043 (2014).
  • (91) S. Weinberg, “Gravitation and Cosmology : Principles and Applications of the General Theory of Relativity,” John Wiley & Sons, New York, USA, 1972.
  • (92) E. Poisson, A Relativist’s Toolkit: The Mathematics of Black-Hole Mechanics, (Cambridge University Press, Cambridge, 2004).