Uniformly accelerated traveler in an FLRW universe

Morteza Kerachian1, ∗ 1Institute of Theoretical Physics, Faculty of Mathematics and Physics, Charles University, CZ-180 00 Prague, Czech Republic This paper introduces an analytical treatment of accelerated and geodesic motion within the framework of the Friedmann -Lemaˆıtre-Robertson-Walker (FLRW) spacetime. By employing conformal time transformations we manage to convert second order differential equations of motion in FLRW spacetime to first order equations in the conformally transformed spacetime. This allows us to derive a general analytical solution in closed-form for accelerated motion in spatially curved FLRW spacetime. We provide few examples of this general solution. The last part of our work focuses on the return journey for a traveler exploring a FLRW universe. We derive certain conditions for a de Sitter universe that have to be satisfied in order to achieve a return journey.

PACS numbers: Keywords: Gravitation, Cosmology; Dynamical systems

1. INTRODUCTION issue from a different point of view, i.e. our rocketeer travels with uniform deceleration in order to achieve a return trip. The paradigm of a homogeneous expanding isotropic uni- The formalism presented in this work reduces to the verse in the framework of General Relativity is realised via the geodesic motion in a spatially curved FLRW spacetime in the Friedmann-Lemaˆıtre-Robertson-Walker (FLRW) model [1– limit of zero proper acceleration. Since geodesic in an expand- 4]. In this work we are going to investigate the accelerated ing universe has vast applications in cosmology, astrophysics motion of a test particle in FLRW. Such a test particle cor- and quantum gravity, many attempts have been undertaken to responds to a rocketeer traveling in an FRLW universe. The solve the geodesic equations of motion (for more details see derivation and the interpretation of accelerated motion have references [12 − 39] of [12]). The first attempt to tackle this suffered from ambiguous treatments which will be discussed issue was initiated by Whiting [13]. Whiting derived the equa- later on1. tions of motion for a free particle with Newtonian background and its relativistic generalization. In [14] geodesic in low- The motion of a uniformly accelerated traveller in an ex- velocity regime has been studied. These efforts by [13, 14] for panding universe is described by a set of differential equations solving geodesic motion was not sufficient due to the number which are in general non-trivial coupled. This set of equations of shortcomings in calculation and interpretation. Latter on, does not become less complicated even if a specific cosmo- Grøn & Elgarøy [15] derived a general solution for geodesics logical model is applied. Thus, an analytical derivation of the in the full general relativity framework. Moreover, Ref. [13] path of a rocketeer is highly challenging. Actually a goal of claimed that particle moving uniformly in an expanding uni- this work is to present a general formulation which allows an verse will join the Hubble flow. This claim has been refuted analytical treatment. In particular, this work has been inspired in [16], in which it has been formally proven that particles by [10] in which W. Rindler proposed a generalization of the following the geodesic motion in an eternally expanding uni- hyperbolic motion in Minkowski spacetime to solve the cor- verse do not asymptotically rejoin the Hubble flow. Recently, responding set of equations. However, Rindler solved it only a method for deriving both timelike and spacelike geodesic for the de Sitter spacetime [10]. distances in spatially flat FLRW spacetime with given initial- Studying the accelerated motion of a rocketeer is useful for value or boundary-value constraints was presented in [12]. the future accelerated space probe. For our universe ( with In this work, we use conformal time transformations in or- Ωm ≈ 0.27, ΩΛ ≈ 0.73 and nearly spatially flat) a space trav- der to get a general analytical formulation. Thus, it is useful arXiv:1911.05436v2 [gr-qc] 24 Apr 2020 eler could visit a galaxy which is observed today at a red- to provide a brief overview of what has been already done in shift of 1.7 on a one-way journey with proper acceleration FLRW with conformal transformations. Conformal transfor- equal to the terrestrial gravitational acceleration, in almost mation and its symmetries help us to grasp the notion of the 100 years [11]. However, for galaxies at redshift less than causal structure of spacetime [17]. FLRW metric has vanish- 1.7, e.g. 0.65, it is not clear whether the traveller would suc- ing Weyl tensor, therefore, all Friedmann cosmological mod- ceed to return back home. Therefore, it might be appropriate els are conformally flat and their systematic description has to consider a traveller of intermittent accelerations to explore been studied in detail in Ref. [18–20]. The nature of FLRW the universe [9]. In this study, we are going to address this models in conformal coordinates has been studied in [21]. It has been demonstrated in [22] that transformation into confor- mal coordinates do not eliminate superluminal recession ve- locities for open or flat matter dominated FLRW cosmologies, ∗ Electronic address: [email protected] and all of them possess superluminal expansion. Ref. [23] de- 1 For example, it has been debated whether analysing uniformly accelerated motion in an expanding universe could clarify the physics behind the ex- rived the scalar field and the electromagnetic field of a moving pansion. Namely, the debate has been about if the expansion is a trick of charged particle in de Sitter spacetime, when the particle is coordinates or a physical phenomenon [5–9]. following geodesic trajectories or it is uniformly accelerated. 2

In order to achieve this, conformal transformation between de where dot means derivation with respect to t and C is a con- Sitter and Minkowski spacetime was applied. stant proportional to the matter density (see e.g. [24]). The layout of this work is as follows; Sec. 2 provides the es- The four-acceleration of a particle is given by sential mathematical background in which the conformal time µ ν σ transformation is applied. In Sec. 3 a novel general formalism µ µ ν du µ dx dx a = u;ν u = + Γνσ , (4) is presented. Namely, using the transformed FLRW space- dλ dλ dλ time enables us to solve the equations of motion of acceler- where uµ is the four-velocity and λ is the proper time. aµ and ated particle. In this way, second order differential equations uµ satisfy the following constraints reduce to first order differential equations which allow us to µ solve the trajectories for accelerated particle and free motion. u uµ = −1, (5) In addition, this formalism specifies the four-velocities of par- ticles. This extends previous results [12, 15] covering only µ 2 geodesic motion. We prove that accelerated and geodesic mo- a aµ = A , (6) tions in FLRW universe depend on the expansion factor and µ its integral for any specific FLRW model. In Sec. 4 we give a uµ = 0, (7) some examples for known FLRW models that have an ana- lytic solution. In cases where there is no analytical solutions, where A is the norm of the acceleration. Having uniform ac- we use numerical integration to solve them. Furthermore, in celeration means that A =const. Sec. 4.4 we provide solutions for both the uniformly acceler- Solving Eq. (4) for a given acceleration (say for A =const) ated and the geodesic motion in the global anti-de Sitter space- is almost analytically intractable (see e.g [10]). Here we in- time by implementing similar prescription as we did in Sec. 3. troduce the conformal time transformation in order to tackle This accelerated trajectory is indeed the generalized form of this problem. In particular, the conformal time η is such that the known uniformly accelerated observer in the anti-de Sit- Z dt ter spacetime [24–26]. In Sec. 5 we discuss the return jour- η = . (8) ney. We show that, in order to achieve a return journey hav- R(t) ing uniformly deceleration is not sufficient condition for every Additionally, by putting χ˜ = χ, the FLRW metric reads spacetime. For a de Sitter spacetime we derive the boundary 2 ˜2 2 2 2 2 2 2 condition that must be satisfied to be able to fulfill the return ds = R (η)[−dη + dχ˜ + Sk (χ˜ )(dθ + sin θdφ )], (9) journey. Concluding remarks are driven in Sec. 6. Notice that, R˜(η) = R(t). When we have cosmological models with Λ = 0 or Λ 6= 0 2. MATHEMATICAL BACKGROUND but without matter, it holds that (see e.g. [24])

 R˜ b( η ), k = + , We begin by introducing the line element of the FLRW  c sin b 1 ˜ ˜ b spacetime, which describes the metric of an expanding, ho- R(η) = Rcη , k = 0,  ˜ b η mogeneous and isotropic universe Rc sinh ( b ), k = −1, 2 2 2 2 2 2 2 2 2 ds = −c dt + R (t)[dχ + Sk (χ)(dθ + sin θdφ )], (1) where R˜c is a constant length which determines the scale of the universe. The power coefficient b for Λ = 0 is b = 2 . where c is light speed (hereafter c = 1), 3γ−2 The value of b distinguishes between different cosmological  sin χ, k = +1, closed, models. For example, if b = 2 then the universe is filled with  1 Sk(χ) = χ, k = 0, flat, dust;for stiff matter b = 2 ; while b = 1 describes the radia-  sinh χ, k = −1, open, tion case. Moreover, for non-negative curvature when Λ 6= 0 and without matter, which is actually a de-Sitter cosmological expresses the space curvature and R(t) is the scale factor model, then b = −1. which describes the expansion of the universe. t is the co- ordinate time t ∈ [0,∞); χ lies in the range χ ∈ [0,∞) for k = 0,−1 and χ ∈ [0,π] for k = 1; while the angles θ ∈ [0,π] 3. PATH OF PARTICLES IN FLRW SPACETIME and φ ∈ [0,2π) independently of the curvature. Let us assume a cosmological model with a cosmological We would like first to present the general formulation for constant Λ and a fluid with equation of state given by the motion of particles in the transformed FLRW spacetime P = P(ρ) = (γ − 1)ρ (2) (9) by considering only the radial motion. To do that, we shall define the four-velocity as follows [11] where P is the pressure, ρ is the energy density and we as- dη coshζ(λ) dχ˜ sinhζ(λ) sume that constant γ can take any values. Then, the Friedmann uη = = , uχ˜ = = , (10) equation reads dλ R˜(η) dλ R˜(η) R˙2(t) Λ k C where ζ(λ) is the rapidity, which will be determined later. = − + , (3) R2(t) 3 R2(t) R3γ (t) Note that equations (10) automatically satisfy constraint (5). 3

The only needed non-vanishing Christoffel symbols for this 3.1. Accelerated Radial Motion case are It is convenient to express the equation of motion of the η η χ˜ 1 dR˜(η) Γηη = Γ = Γ = . (11) rocketeer in terms of η. Therefore, from the four-velocity χ˜ χ˜ ηχ˜ R˜(η) dη (10), we get The four-acceleration in the set of coordinates (9) can be dχ˜ dχ˜ /dλ written in the following way = = tanhζ(λ). (22) dη dη/dλ η η du η η 2 η χ˜ 2 To find the unknown rapidity ζ(λ) we need to use Eq. (20) a = + Γηη (u ) + Γ ˜ ˜ (u ) , (12) dλ χχ and reparametrize it in terms of η ˜ χ˜ η dζ(η) χ˜ ˜ du χ˜ ˜ A − u − Γu = 0, (23) aχ = + 2Γ uη uχ . (13) dη dλ ηχ˜ ˜ ˜ From now on, since all used Christoffel symbols have equal where ζ(η) = ζ(η(λ)) = ζ(λ) and A =const.. value, we shall denote them by Γ. Integrating Eq. (23) with respect to η, we obtain By differentiating the first term in the right-hand side of υ ζ˜(η) = arcsinh(AR˜(η) + ), (24) Eq. (12) and by using Eq. (10) we obtain R˜(η) duη sinhζ(λ) dζ(λ) coshζ(λ) dR˜(η) where = − . (14) R η d ˜ d ˜ 2 d R˜(η˜ )2dη˜ λ R(η) λ R(η) λ ˜( ) = , (25) R η ˜ ˜ R(η) Since the dR(η) can be written in terms of η dλ and υ is an integration constant which is related to the initial dR˜(η) dη dR˜(η) dR˜(η) velocity of particle. Consequently, = = uη , (15) dλ dλ dη dη Z R˜(η) χ˜ = A q dη by substituting it into Eq. (14) together with Eqs. (10) and (11) (A ˜( ) + υ )2 + R η R˜( ) 1 we arrive to η Z ˜ η 1/R(η) du χ˜ dζ(λ) η 2 + υ q dη. (26) = u − (u ) Γ. (16) (A ˜( ) + υ )2 + 1 dλ dλ R η R˜(η) Thus Now, we go back to the coordinates of the original FLRW dζ(λ) dζ(λ) metric (1). This is achieved by using the inverse transfor- aη = uχ˜ + (uχ˜ )2Γ = uχ˜ ( + Γuχ˜ ). (17) mation of Eq. (8), i.e. R˜(η)dη = dt, and by recalling that dλ dλ R(t) = R˜(η). Thus, we obtain for Eq. (24) ˜ Similar calculation can be undertaken for aχ where υ ζˆ(t) = arcsinh(AR(t) + ), (27) duχ˜ dζ(λ) R(t) = uη − uη uχ˜ Γ, (18) dλ dλ where t which finally gives R R(t0)dt0 R(t) = . (28) R(t) dζ(λ) dζ(λ) aχ˜ = uη + uη uχ˜ Γ = uη ( + Γuχ˜ ). (19) dλ dλ Using Eq. (27) enables us to derive the four-velocity in stan- dard FLRW spacetime as follows We denote dt dχ sinhζˆ(t) dζ(λ) ut = = coshζˆ(t), uχ = = . (29) A = + Γuχ˜ , (20) dλ dλ R(t) dλ Finally, the trajectory of uniform acceleration motion is where A, is the norm of acceleration as mentioned earlier in given by Eq. (6). As a result, the four-acceleration becomes Z 1 R(t) aη = Auχ˜ , aχ˜ = Auη . (21) χ = A dt R(t) q υ 2 (AR(t) + R(t) ) + 1 Note that Eq. (21) satisfies also the constraints (6) and (7). Z In the transformed FLRW metric (9) the coordinates η and 1 1 + υ 2 q dt, (30) χ˜ share a common coefficient, i.e. the R(˜η). If we constraint R(t) υ 2 (AR(t) + R(t) ) + 1 the motion only on the radial direction through this transfor- mation we get a solvable set of equations from Eq. (4). This By specifying the evolution of the scale factor R(t) Eq. (30) allows us to analyze the radial motion of the rocketeer in the provides the accelerated radial path of the rocketeer in the FLRW spacetime. standard FLRW coordinate. 4

3.2. Some characteristic types of motion and

2+k  2 2 2+2k  a. Purely accelerated motion. When one ignores the in- Aκ η 1 2 + k 4 + 3k A κ η χ˜a = 2F1 , ; ;− , tegration constant υ (i.e. υ = 0) in trajectories (26) and (30) (2 + k)(1 + 2k) 2 2 + 2k 2 + 2k (1 + 2k)2 the motion is called purely accelerated. In the conformally (39) transformed coordinates the trajectory is given by where 2F1(a,b; c; z) is the Gauss hypergeometric function. These two trajectories become equivalent if Z R˜(η) χ˜a = A q dη, (31)  A2 ˜( )2 + 1 + k = −m, R η 1  Aκ υ 1+2k = µ , (40) and in the original FLRW coordinates we get  1 + 2k 6= 0. Z 1 R(t) χa = A p dt, (32) For instance, the uniformly accelerated trajectories in the de R(t) A2R(t)2 + 1 Sitter spacetime get transformed to the geodesic motion in where index a in both Eqs. (31) and (32) refers to the purely the Minkowski spacetime and vice versa ( See Sec. 4.3 for accelerated motion. more details). Moreover, one can show that an observer with b. Geodesic motion. To get the trajectory for the a suitable acceleration moving in a decelerating Friedmann geodesic motion one has to substitute A = 0 into the Eqs. (24), universe, i.e. a dust field universe, has the same cosmological redshift as the observer in the ΛCDM model [27]. (26) and (30). Thus, the rapidity function ζ˜(η) becomes  υ  ζ˜(η) = arcsinh . (33) R˜(η) 4. SOME EXAMPLES Consequently, Eq. (26) will be Z υ In this section of section’s 3 formalism is applied to specific χ˜v = p dη, (34) FLRW universe models. The motion of a particle both in the R˜(η)2 + υ2 original and in the transformed coordinates depends only on and for Eq. (30) we obtain the scale factor and its integral (Eq. (28) and Eq. (25) respec- tively). Thus, specifying the expansion factor for each cos- Z υ 1 mological model enables us to determine the particles world- χv = p dt. (35) R(t) R(t)2 + υ2 lines. In this section the behavior of the trajectories presented in paragraphs a and b of Sec. 3.2 is studied. Here index v in Eqs. (34) and (35) denotes geodesic motion. Recently, the solution of Friedmann Eq. (3) was presented For all υ values, υ2 > 0, which guarantees that Eq. (35) is a for various FLRW models with k = 0 [28]. Namely, Chavanis timelike geodesic [12]. has derived an analytical solution for R(t) in a universe un- Eqs. (34) and (35) are geodesics in any conformal time dergoing a various combination of eras, e.g. stiff matter era, FLRW spacetime and FLRW spacetime respectively. Eq. (35) dark matter era, and dark energy era due to the cosmological is the same as equation derived in [15] and recently in [12]. constant. From this study we use the form of the scale factor c. Null geodesics. We can see from Eq. (30) that for in the cosmological examples of Sec. 4 and Sec. 5. large acceleration A the particle’s trajectory asymptotically reaches the null geodesic, that means Note that, although the transformation (8) is not, in general, conformally flat transformation (CFT) for spatially curved Z 1 FLRW models, it is CFT for all the flat FLRW models. Thus, lim χa = ± dt. (36) A>> R(t) since the cosmological models appearing in Secs. 4.1 and 4.3 have zero spatial curvature, the transformation (8) is a CFT, Moreover, this statement holds for large υ value, i.e. i.e. it holds that Z 1 lim χa = ± dt. (37) 2 2 2 υ>> R(t) dsk=0 = Ω ds f lat , (41) d. Transformation conditions. For the flat spatial cur- where Ω = R˜(η). vature FLRW spacetime the accelerated motion (31) can be transformed into geodesic motion (34) under certain condi- It is clear that this formalism is able to reproduce the known tions. In order to investigate this statement, we consider two hyperbolic motion in the Minkowski spacetime [29]. In a sim- m ilar manner as in the Minkowski spacetime, the uniformly ac- different spacetimes having scale factors R˜m(η) = µη and k celerated motion can be derived in the Einstein static universe, R˜k(η) = κη . By substituting R˜m(η) and R˜k(η) into the Eqs. (34) and (31) respectively we get since for both spacetimes the scale factor R(t) = 1. To provide visualization for our examples we are going to υ η1−m 1 m − 1 3m − 1 υ2 η−2m  plot some trajectories in Penrose diagrams with coordinates η χ˜v = F , ; ;− , (38) µ (m − 1) 2 1 2 2m 2m µ2 and χ given by the metric (9). 5

4.1. Flat FLRW spacetime without Λ +

A=0.1 In this section we consider spatially flat FLRW spacetime with vanishing cosmological constant, i.e. Λ = 0, and a single A=1 fluid content provided by the EoS (2). From the Friedmann υ=1 A=5 equation (3) one can obtain the scale factor υ=0.1 ℐ+ 2 χ = 0 3γ R(t) = Rct , (42) Null √ 2 A=-4,υ=0.34 3
3γ where Rc = 2 γ C . Substituting this scale facor into the η=Const. equations Eqs. (32) and (35) we obtain

χ = Const. 0 2−2/3 2  9At γ γ η=0 χa = η=0 Rc (3γ + 2)(6γ − 2)  2! 1 1 1 3Aγ t FIG. 1: Penrose diagram for the dust-filled universe. The thick black F ,1 − ;2 − ;− , (43) 2 1 2 3γ 3γ 2 + 3γ solid line denotes the photon trajectory. Solid lines have constant acceleration A and υ = 0, while the dotted dashed lines have constant and acceleration A and υ 6= 0. Dashed curves are geodesics, i.e. A = 0. 3t1−2/3γ χv = Rc (3 − 2γ) 4.2. Milne Universe  2 2! 1 3γ 1 3γ 1 Rc t /3γ F , − ; + ;− , (44) Vacuum FLRW model with Λ = 0 and k = −1 is known 2 1 2 4 2 4 2 υ as Milne universe [30], where R(t) = t. For this spacetime where a and v denote uniform acceleration and geodesic mo- particle’s paths are R˜( ) =   tion respectively. In conformal representation, where η p 2 2 ˜ b 2 χa = ln At + A t + 4 + ln(C5), (47) Rcη where b = 3γ−2 , we have

2+b where C5 is an integration constant. For a particle starting AR˜c η 1 χ˜ = from χa = t = 0 (C5 = 2 ), χa reduces to a (2 + b)(1 + 2b) A 2 2 2+2b 1 2 + b 4 + 3b A R˜ η  χa = arcsinh( t), (48) F , ; ;− c (45) 2 2 1 2 2 + 2b 2 + 2b (1 + 2b)2 and χv becomes and υ υ χv = −arctanh(√ ) + arctanh(−√ ). (49) υ η1−b 1 b − 1 3b − 1 υ2 η−2b  t2 + υ2 1 + υ2 χ˜ = F , ; ;− . (46) v ˜ 2 1 ˜2 Rc (b − 1) 2 2b 2b Rc For Milne universe we cannot use Eqs. (26) and (34) since The dynamical features of the above trajectories on a Pen- transformation (8) is not CFT. In order to plot the above case rose diagram are very similar for all the usual barotropic flu- in a Penrose diagram we have to use the transformation ids, i.e. for fluids with 1 ≤ γ ≤ 2. Thus, in Fig. 1 we plot p R t = T 2 − R2, χ = arctanh( ), (50) just the case γ = 1, which shows different trajectories in a T spatially flat FLRW universe with dust. Namely, Fig. 1 shows between a Milne Universe and the Minkowski spacetime [29]. accelerated trajectories with zero υ and non zero υ (solid blue Using this transformation we get and dotted dashed red lines respectively), the geodesic trajec- tories (orange dashed lines) together with the null geodesic 1 1 (R + )2 − T 2 = , (51) (black thick solid line). In the case of constant acceleration, a A A2 the greater the acceleration the faster the rocketeer approaches and the null geodesic behavior. We plot also one decelerating υ T trajectory with non-zero υ, which exhibits a return journey: Rv = √ , (52) 2 such trajectories will be discussed in Sec. 5. Regarding the 1 + υ geodesic motion the greater the initial velocity, the further the for the trajectories (48) and (49) respectively in the traveller can reach. Note that even if initially geodesic trav- Minkowski spacetime. ellers overtake accelerated ones, eventually as expected the It is known that Eqs. (51) and (52) describe hyperbolic and accelerated ones prevail. Additionally, this figure provides the geodesic motion respectively in Minkowski spacetime. In asymptotic behavior of the trajectories as t → ∞. One can see Fig. 2 examples of these types of motion are depicted in the that, accelerated trajectories reach the future null infinity ,i.e. same manner as Fig. 1. The shaded region in this figure indi- I +, whereas the geodesics motion ends up to timelike infin- cates the part of the Penrose diagram that does not belong to ity, i.e. i+. the Milne universe. 6

+ in de Sitter spacetime (see Fig. 3). This result confirms previ- ous works of Rindler2 [10] and Bicˇak´ & Krtousˇ [23]. A=0.3 Fig. 3 shows the trajectories in the de Sitter spacetime. In A=-0.5,υ=2 υ=1 this spacetime, all trajectories have the same description as A=1 in Fig 1, but some of them have different initial conditions. A=2 Namely, some of them do not pass through the origin t = 0. υ=0.5 ℐ+ Another difference is that the de Sitter spacetime covers only the lower part of the Penrose diagram, since all trajectories Null end up at the I +. We continue plotting the trajectories even

η=Const. to the upper shaded region in order to provide a global view of the behavior of these trajectories. χ = Const. 0

4.4. Anti-de Sitter spacetime FIG. 2: Penrose diagram for the Milne universe. Curves have the same coloring as described in Fig. 1. Note that, the shaded part of the plot does not belong to the Milne universe. In this section we consider a vacuum FLRW universe with a negative cosmological constant Λ and negative spatial cur- vature k = −1 namely Anti-de Sitter universe. This particular case of Anti-de Sitter universe has the scale factor 4.3. de-Sitter Universe, γ = 0 t R(t) = α cos( ) (58) Considering the dark energy dominated universe in the ab- α sence of any matter the scale factor is p where α = 3/ | Λ |. Thus, the accelerated and geodesic tra- q Λ jectories become 3 t R(t) = R0e , (53) q  2  where Λ is a cosmological constant [28]. This solution is A2 α2 + Aα A2 α2 sin (t/α) + cos2(t/α) χa = ln2 , known as the de Sitter solution. cos(t/α) Accelerated and geodesic motion in this particular space- time are described by (59) and A ! χa = − q +C1, (54) υ sin(t/α) q Λ t q Λ 3 2 Λ χv = arctanh p . (60) 3 R0e A + 3 α2 cos2(t/α) + υ2

In Fiq. 4 we illustrate these trajectories and we denote the r q Λ different types of trajectories as we did in Fig. 1. r 3 t 2 2 3 (R0e ) + υ Note that this coordinate system does not cover the whole χv = − q . (55) Λ Λ anti-de Sitter spacetime. In order to study the accelerated mo- 3 t υR0e tion in the whole anti-de Sitter spacetime we use the following line element In conformal coordinates, where scale factor is R˜(η) = q 3 1 ds2 = −cosh2(r)dt2 + α2 dr2 + sinh2(r)(dθ 2 + sin2 θ dφ 2)
, − Λ η , we get (61) A χ˜a = q η, (56) A2 + Λ where r is dimentionless. The accelerated and geodesic mo- 3 tions of this metric can not be studied from the formulation presented in Sec. 3. However, we can introduce a similar pre- r scription to obtain those trajectories. Namely, we introduce 3 2 χ˜v = − + η +C . (57) Λυ2 3

It is clear from Eqs. (56) and (57) that the geodesic equa- 2 Note that, W.Rindler obtained only one special case of accelerated motion tion in conformally flat coordinates ( Minkowski spacetime) in de Sitter spacetime. Namely, he studied the case when a particle leaves get transformed to uniformly accelerated worldline in de Sitter the origin (t = χ = 0) from rest, i.e. ut = 1 and uχ = 0. One can rederive q 3 spacetime, whereas the trajectory of a uniformly accelerated Rindelr’s trajectory by putting υ = − Λ A into the Eq. (30) together with particle in Minkowski spacetime get transformed to geodesic χ(t=0) = 0. 7

η=∞

A=-1 A=1

A=-1,υ=0.25 υ=1 A=3

A=-2,υ=-1 A=4,υ=-2 η=0 ℐ+ t=+∞

η=Const.

χ = Const. t=0

Null Null υ=-1 χ=-∞ χ=∞

η=-∞ t=-∞

FIG. 3: Penrose diagram for de Sitter universe.The shaded part is not covered by the flat de Sitter universe. the conformal coordinate by setting sinh(r) = tan( ) to- 1 dF(χ) χ χ and Γ = F(χ) dχ . Therefore, for the rapidity function gether with t = α η . Then, the metric (61) takes the form ξ(χ) = ξ(χ(λ)) = ξ(λ), which can be determined from Eq. (65), we get α2 ds2 = [−dη2 + dχ2 + sin2(χ)(dθ 2 + sin2 θ dφ 2)]. cos2(χ)  υ  ξ(χ) = arcCosh AF (χ) + , (66) (62) F(χ) It is clear that the anti-de Sitter spacetime time covers only half of the Einstein static universe, namely in the range χ ∈ where [0,π/2). R Now, similarly to the formulation presented in Sec. 3, we F(χˆ )2dχˆ F (χ) = , (67) introduce the radial four-velocity F(χ)

dη cosh(ξ(λ)) dχ sinh(ξ(λ)) and υ is an initial velocity of the accelerated particle. Thus, uη = = , uχ = = , (63) dλ F(χ) dλ F(χ) from the four-velocity (63) and Eq. (66) we get α Z where F(χ) = . This radial motion has the four- η = coth(ξ(χ))dχ. (68) cos(χ) acceleration given by By substituting υ = 0 into the Eq. (68), namely for the purely aη = Auχ , aχ = Auη , (64) accelerated motion, after some manipulation we get the fol- lowing trajectory where √ ! A2 α2 − 1 tan(∆η) dξ(λ) η χa = π − arccos , (69) A = + Γu (65) Aα p 2 dλ 1 + tan (∆η) 8

χ = ∞ , stationary. Similar to the accelerated motion, maximum dura- η = π 2 α tion of this motion is π/2. Setting ξ(χ0) = 0 for any fixed χ = χ0 reduces the acceler- ated radial motion to the family of the timelike worldlines rep-

υ=0.2 resenting uniformly accelerated observers studied in previous υ=1 works [24–26]. Thus, the newly found radially moving ac- ℐ celerated observers have as a limiting case the already known stationary ones. Null Fig. 5 shows these trajectories denoted in the same manner

η=Const. as in Fig. 1. As we discussed previously, worldlines of fixed χ represent uniformly accelerated observers. In this figure,

χ = Const. trajectory number (1) shows that the stationary observer from η=0 η ∈ (−∞,0]. Then, at η = 0 it starts to accelerated radially χ = 0 with an acceleration A = 3 and goes toward I ( trajectory (4) has the similar behavior). Observer number (6) has a decel- eration A = −1.8 from η ∈ (−π/2,0). Then, at η = 0 its uχ vanishes and becomes stable. On the other hand, trajectories (5) and (7) which have the υ > 0 start from η = χ = 0 travel-

A=4 ing with for ∆η = π/2 with constant υ. Its radial component χ A=1 of four-velocity, i.e. u is decreasing until at η = 0 it become zero. After this point, the observer becomes stationary. Ob- A=0.3 servers (2) and (3) are at rest from η ∈ (−∞,−π/2] and then they move with negative υ towards O.

π α - = 2 η , ∞ = χ 5. RETURN JOURNEY η=0

FIG. 4: Penrose diagram of the part of the anti-de Sitter spacetime as In this section we are going to focus our study on analyz- a particular case of the FLRW universe. Trajectories are collored in ing the return journey of the rocketeer in spatially flat FLRW the same manner as Fig. 1. universe. In particular, we are going to study the behavior of the Eq. (30) or Eq. (26) in the spacetimes studied in the previ- ous section. Actually, in order to fulfill the return journey, our A2 2 > = − where α 1 has to be satisfied, ∆η η η0 and the rocketeer must begin to decelerate, i.e. Ad < 0, long enough constant η0 is an integration constant from integral appears in time before it reaches the designated proper distance. Eq. (68). Observers with an acceleration A travel radially in Assuming that, the spaceship is travelling with non-zero the range positive value υ, at t = λ = χ = 0 the rocketeer applies a de- celeration A . Thus, the rocketeer reaches the maximum co-  1  π d arcsin < χa < . moving distance from the origin at the return point with coor- Aα 2 ˆ χ dinates {t1,x1} > 0 when ζ(t1) = 0 or equivalently u (t1) = 0. At χa = arcsin(1/(Aα)) the χ component of the four-velocity Therefore, depending on the form of scale factor, one can vanishes and the radial moving observer turns to the station- analyze the return journey of the rocketeer. ary observer. Furthermore, Eq. (69) show that the maximum duration of the radial accelerated traveler in the anti-de sitter n spacetime is ∆η = π/2, since the trajectories reach the I . 5.1. R(t) = t spacetimes Moreover, by putting A = 0 into the Eq. (68) we derive the trajectories for geodesic motion In this section we analyze the return journey in the space- √ ! time studied in sections 4.1- 4.2 , namely spacetimes having υ2 − α2 tan(∆η) the scale factor like R(t) = tn. In these particular spacetimes χv = arcsin p , (70) υ 1 + tan2(∆η) the rockeeter reaches the maximum comoving distance from the origin at which holds under the condition that υ2 > α2 (for υ = α the   1 geodesic trajectory vanishes). Eq. (70) shows that, the ob- υ(n + 1) n+1 t1 = − . (71) server moving with the constant υ moves in the range Ad α  Afterwards, the rocketeer returns towards the origin. As t → ∞ 0 < χv < arccos . υ the trajectory of the rocketeer asymptotically becomes χ When the observer reaches at χv = arccos(α/υ) the u be- Z 1 lim χ = − dt, (72) comes zero and therefore the radial moving observer becomes t→∞ tn 9 χ = Const. 5 6 η=Const. 7 χ = π / 2

η=π/2

Null υ=2 A=1.1 υ=1.1

ℐ O A=3 η=0 χ = 0

υ=-1.1 A=-1.8 υ=-2 Null

η=-π/2

1 2 3 4

η=0

FIG. 5: The Penrose diagram for the global anti-de Sitter spacetime. See the text for more details.

which means that there is a finite t2 > t1, when the rocke- Actually, to attain an actual return journey, the following rela- teer arrives back to the origin, χ = 0. In Figs. (1)- (2), the tion dotted dash trajectories represent the return journeys in each r Λ spacetimes. 2A < − υ < A , (75) d 3 d has to be satisfied. In Fig. 6 we show several cases of return 5.2. de-Sitter case journeys in de Sitter spacetime for Λ = 3 and A = −2. The negative values of χ represents the opposite direction from the one that the rocketeer is supposed to explore. The return journey in the de Sitter spacetime has a differ- ent behavior with respect to the previous examples, since the q Line 1. For − Λ υ ≥ A , there isn’t any return point for the scale factor (53) is given by a different function of time. In 3 d this specific spacetime, the rocketeer reaches the maximum particle and rocketeer will move toward the −χ direc- comoving distance from the origin at tion. q A < − Λ < A r r ! Line 2. For 2 d 3 υ d, there is a return point and 3 Λ υ t1 = ln − . (73) the rocketeer is able to come back to the origin. Λ 3 Ad q Λ Line 3. For − 3 υ = 2Ad, there is a return point but the Moreover, the total cosmic time t2 needed to cover the return rocketeer will return back to origin in a infinite cosmic journey for a rocketeer that leaves the origin at t = λ = 0, is time. derived from Eq. (30) and it is given by q Λ Line 4. For − 3 υ > 2Ad, there is a return point but the r √ ! 3 3Λυ rocketeer will never go back to the origin. t2 = ln −√ . (74) Λ 3Λυ + 6Ad Thus, we have seen that in de Sitter spacetime, having the uniform deceleration motion is not sufficient for the rocketeer Thus, from Eqs. (73) and (74) one can see that the return jour- to come back to the origin. One has to apply the deceleration ney does not happen for all values of Ad and υ (see Fig. 6). which satisfies Eq. (75). 10

η=0 1 2 3 4 ℐ+ t=+∞ leads us to derive the trajectory (30) of an accelerated traveler. We have shown that the accelerated and the geodesic mo- tion in an expanding universe are solely determined by the expansion factor and its integral (28). The scale factor is the solution of the Friedmann equation (3), which depends on the spatial curvature k, the cosmological constant Λ, and the equa-

χ = 0 tion of state P = P(ρ). Although, we have chosen a specific

Null Null form of the equation of state in Sec. 2, this formulation is in- dependent of the choice of an equation of state. It depends only on whether the scale factor can be expressed analytically χ = Const. η=Const. as a function of time or not. Additionally, we have provided a similar formulation in the t=0η=-1 η=0 case of the anti-de Sitter spacetime for the uniformly accel- erated and geodesic radial motion. The newly found radially FIG. 6: Penrose diagram for the worldlines of return journeys in de accelerated trajectories are generalizations of the known uni- Sitter spacetime. Here Λ = 3 and A = −2 and the spacetime is de- formly accelerated stationary observers in the anti-de Sitter picted only for t ≥ 0. universe. In the last part of our work we have focused on the return journey of the rocketeer. It had been suggested that having 6. CONCLUSIONS uniform deceleration would be enough in order to have an ac- tual return journey [10]. Here we have proved that even if this In 1960, W.Rindler [10] proposed the problem of a ”Hyper- condition is necessary, it is not sufficient for all spacetimes. bolic Motion in Curved Spacetime” to study the accelerated In particular, among the cosmological models analyzed here, motion in curved spacetime. in the de Sitter case Eq. (75) must be satisfied for a return In particular, it was suggested that the accelerated motion journey to be possible. is the best way of exploring our universe in a reasonably short time [9]. Taking the above suggestion into account, the mo- tion of an accelerated traveller in an expanding universe has been studied in this work. This involves solving the non-trivial Acknowledgments Eq. (4) for a given acceleration. To achieve this, we applied a conformal time transformation (8) to the generic FLRW uni- I am grateful to Georgios Loukes-Gerakopoulos, Giovanni verse. Using the method introduced in Sec. 3 has helped us to Acquaviva and Jirˇ´ı Bicˇak´ for valuable discussions on this determine a generalized form of rapidity function (27), which work and useful comments on the manuscript.

[1] A. Friedmann. Uber¨ die krummung¨ des raumes. Z. Physik, [9] J. Kwan, G. F. Lewis, and J. B. James. The adventures of the 10:377–386, 1922. rocketeer: accelerated motion under the influence of expanding [2] G. Lemaˆıtre. Expansion of the universe, a homogeneous uni- space. Publ. Astron. Soc. Aust., 27(1):15–22, 2010. verse of constant mass and increasing radius accounting for the [10] W Rindler. Hyperbolic motion in curved space time. Phys. Rev., radial velocity of extra-galactic nebulae. Mon. Notices Royal 119(6):2082, 1960. Astron. Soc., 91:483–490, 1931. [11] J. S. Heyl. The long-term future of space travel. Phys. Rev. D, [3] H. P. Robertson. Kinematics and world-structure. Astrophys. J., 72(10):107302, 2005. 82:284, 1935. [12] W. J. Cunningham, D. Rideout, J. Halverson, and D. Krioukov. [4] A.G. Walker. On milne’s theory of worldstructure. Proc. Lon- Exact geodesic distances in flrw spacetimes. Phys. Rev. D, don Math. Soc., 42:90, 1936. 96(10):103538, 2017. [5] Marek A Abramowicz, Stanislaw Bajtlik, Jean-Pierre Lasota, [13] A. B. Whiting. The expansion of space: Free particle motion and Audrey Moudens. Eppur si espande. arXiv preprint astro- and the cosmological redshift. Observatory, 124:174, 2004. ph/0612155, 2006. [14] T. M. Davis, C. H. Lineweaver, and J. K. Webb. Solutions [6] Marek A Abramowicz. Spacetime is not just space and time. to the tethered galaxy problem in an expanding universe and New Astronomy Reviews, 51(10-12):799–802, 2008. the observation of receding blueshifted objects. Am. J. Phys., [7] Marek A Abramowicz, Stanislaw Bajtlik, Jean-Pierre Lasota, 71(4):358–364, 2003. and Audrey Moudens. A short answer to critics of our article” [15] Ø. Grøn and Ø. Elgarøy. Is space expanding in the friedmann eppur si espande”[arxiv: astro-ph/0612155]. arXiv preprint universe models? Am. J. Phys., 75(2):151–157, 2007. arXiv:0812.3266, 2008. [16] L. A. Barnes, M. J. Francis, J. B. James, and G. F. Lewis. Join- [8] Geraint F Lewis, Matthew J Francis, Luke A Barnes, Juliana ing the hubble flow: implications for expanding space. Mon. Kwan, and J Berian James. Cosmological radar ranging in an Notices Royal Astron. Soc., 373(1):382–390, 2006. expanding universe. Monthly Notices of the Royal Astronomical [17] S. W. Hawking and G. F. R. Ellis. The large scale structure of Society, 388(3):960–964, 2008. space-time. Cambridge university press, 1973. 11

[18] Ø. Grøn and S. Johannesen. Frw universe models in confor- time and the insufficiency of retarded fields. Phys. Rev. D, mally flat-spacetime coordinates. i: General formalism. Eur. 64:124020, 2001. Phys. J. C, 126(3):28, 2011. [24] J. B. Griffiths and Jirˇ´ı Podolsky.´ Exact space-times in Einstein’s [19] Ø. Grøn and S. Johannesen. Frw universe models in confor- general relativity. Cambridge University Press, 2009. mally flat-spacetime coordinates ii: Universe models with neg- [25] J Podolsky.` Accelerating black holes in anti-de sitter universe. ative and vanishing spatial curvature. Eur. Phys. J. C, 126(3):29, Czechoslovak journal of physics, 52(1):1–10, 2002. 2011. [26] Pavel Krtous.ˇ Accelerated black holes in an anti-de sitter uni- [20] Ø. Grøn and S. Johannesen. Frw universe models in confor- verse. Physical Review D, 72(12):124019, 2005. mally flat-spacetime coordinates iii: Universe models with pos- [27] Antonio Feoli and Elmo Benedetto. Dark energy or local accel- itive spatial curvature. Eur. Phys. J. C, 126(3):30, 2011. eration? Gravitation and Cosmology, 23(3):240–244, 2017. [21] M. J. Chodorowski. A direct consequence of the expansion [28] P. Chavanis. Cosmology with a stiff matter era. Phys. Rev. D, of space? Mon. Notices Royal Astron. Soc., 378(1):239–244, 92(10):103004, 2015. 2007. [29] W. Rindler. Essential relativity: special, general, and cosmo- [22] G. F. Lewis, M. J. Francis, L. A. Barnes, and J. B. James. Coor- logical. Springer Science & Business Media, 2012. dinate confusion in conformal cosmology. Mon. Notices Royal [30] E. A Milne. World structure and the expansion of the universe. Astron. Soc., 381(1):L50–L54, 2007. Nature, 130:9, 1932. [23] J. Bicˇak´ and P. Krtous.ˇ Accelerated sources in de sitter space-