Quantum Detection of Inertial Frame Dragging

Wan Cong,1, 2, ∗ Jiří Bičák,3, † David Kubizňák,2, 1, ‡ and Robert B. Mann1, 2, § 1Department of and Astronomy, University of Waterloo, Waterloo, Ontario, Canada, N2L 3G1 2Perimeter Institute, 31 Caroline St., Waterloo, Ontario, N2L 2Y5, Canada 3Institute of , Faculty of Mathematics and Physics, V Holešovičkách 2, 180 00 Prague 8, Czech Republic A relativistic of like produces phenomena differing fundamentally from Newton’s theory. An example, analogous to electromagnetic induction, is gravitomagnetism, or the dragging of inertial frames by mass-energy currents. These effects have recently been confirmed by classical observations. Here we show, for the first time, that they can be observed by a quantum detector. We study the response function of Unruh De-Witt detectors placed in a slowly rotating shell. We show that the response function picks up the presence of rotation even though the inside the shell is flat and the detector is locally inertial. The detector can distinguish between the static situation when the shell is non-rotating and the stationary case when the shell rotates and the dragging of inertial frames, i.e. gravitomagnetic effects, arise. Moreover, it can do so when the detector is switched on for a finite time interval within which a signal cannot travel to the shell and back to convey the presence of rotation.

I. INTRODUCTION a rotating shell. They can in principle determine its rota- tion by, for example, sending out a spherical pulse which, Frame-dragging, also known as the Lense–Thirring ef- upon reflection, will experience a differential Doppler ef- fect [1, 2], is a general-relativistic effect that arises due fect, with different shell latitudes Doppler shifting dif- to moving, in particular rotating, [3] and “rotat- ferently. Meanwhile, frame-dragging outside a rotating ing” gravitational waves [4, 5]. If a gyroscope is located body, the Earth, has taken the satellite in the vicinity of a rotating body, it will keep its direc- mission [12, 13] almost a half-century since its inception tion with respect to the axes of a local inertial frame at to detect. the same place but both the inertial axes and the gy- In this paper, we show for the first time that frame- roscope will be rotating with respect to static distant dragging inside a slowly rotating shell can be observed observers (“fixed stars” at asymptotically flat infinity). by an inertial quantum Unruh-DeWitt (UDW) detector Its profound explicit manifestation can be seen for a ro- [14]. Specifically, it can do so in a time shorter than tating , which drags particles into co-rotation, the light time, ts, of the shell, and hence more the dragging becoming so strong inside the ergosphere efficiently than a classical detector. We consider this to that no particle there can remain at rest with respect to be a quantum detection of inertial frame dragging [15– fixed stars [6]. Frame-dragging is also behind various as- 18]. trophysical phenomena such as relativistic jets and the We note that the ability of such detectors to ob- Bardeen-Petterson effect [7], which aligns accretion disks tain non-local information about spacetime structure has perpendicular to the axis of a rotating black hole. been demonstrated in other contexts [19–23]. In particu- In addition, frame-dragging inside a rotating shell was lar, it was shown in [19, 20] that UDW detectors can dis- taken by Einstein to be in support of Mach’s principle. tinguish between global flat Minkowkian and For a nice discussion on Mach’s principle, dragging ef- local flat spacetimes inside massive shells. Our paper is fects and their impact on astrophysics and cosmology, an important step forward from these results, much like see [8] (also [3, 9]). Consider a slowly rotating material the transition from electrotatics to , or arXiv:2009.10584v3 [gr-qc] 14 Jan 2021 shell [10, 11]. Observers inside the shell who are at rest the transition from Schwarzchild to Kerr; it demonstrates with respect to distant fixed stars will find that a particle for the first time that a fundamentally relativistic (non- moving inside the shell experiences a Coriolis acceleration Newtonian) effect of dragging of inertial frames, namely (the centrifugal acceleration is of the second order in the the existence of gravitomagnetism, can in principle be shell’s angular velocity). These observers are not inertial, observed by a quantum detector in settings that are not therefore fictitious arise. classically possible. For inertial observers, without looking at or outside the rotating shell, there is no way of determining, by em- ploying , whether they are surrounded by

II. SLOWLY ROTATING SHELLS ∗ [email protected] † [email protected]ff.cuni.cz ‡ [email protected] Let us begin by describing the spacetime metric of a § [email protected] slowly rotating shell. The metric the shell can be written 2 as III. NORMALIZED MODE SOLUTIONS

2Ma 2 2 2 2 2 In order to compute the response of the UDW detector ds+ = −f(r)dt + r sin θ(dφ − 3 dt) r (1) we need to obtain the normalized mode solutions to the −1 2 2 2 +f(r) dr + r dθ , scalar wave equation. For the scalar field Ψ the wave equation is where f(r) = 1 − 2M/r, M is the mass of the shell and √ µν a = J/M is the angular momentum per unit mass. The ∂µ( −gg ∂ν Ψ) = 0 , (4) r-coordinate ranges from [R, ∞), R being the radius of the shell. To first order in a, the above metric agrees where g is the determinant of the metric. Upon sub- with the and satisfies the vacuum Einstein’s stituting in the metric (1) and (2) , this equation can equations. Inertial frame-dragging is characterized by be solved by separation of variables. To leading or- 3 the function $(r) = gφt/gφφ = 2J/r , where J = Ma der in a, one can employ the usual mode expansion is the fixed total angular momentum as measured at in- Ψ (t, r, θ, φ) = √ 1 e−iωtY (θ, φ)ψ(r) in spherical ω`m 4πω m` finity. The gradients of $(r) determine the precession of harmonics Ym`. This yields a separated radial equation: gyroscopes relative to the orthonormal frame of locally non-rotating observers [6]. On the shell itself, r = R, and  2  α d α 2 dψ  α `(` + 1) 2 3 r − + γ + ω ψ = 0 . (5) $s = 2J/R . β r2 dr β dr r2 For an inertial observer inside the shell (who rotates as seen from infinity) spacelike geodesics (for example, The functions α, β and γ are φ = 0, θ = π/2, r = constant, t ∈ R) connected to fixed ( points at infinity rotate backwards; the shell is rotating pf(R), r ≤ R, forward (the dragging of the inertial frame becomes com- α(r) = pf(r), r > R, plete i.e., inertial observers rotate at the same angular ve- ( locity as the shell, only if the shell is at its Schwarzschild 1, r ≤ R β(r) = (6) radius); the fixed stars are rotating backwards. In [18] 1/pf(r), r > R, these effects are expressed quantitatively1. ( 2 The metric (1) must be joined at r = R to the metric 4Mamω − 2Mam  , r ≤ R, γ(r) = R3 R3 4Mamω 2Mam 2 r3 − r3 , r > R.  2Ma 2 ds2 = −f(R)dt2 + r2 sin2 θ dφ − dt − R3 (2) For r ≤ R, the radial equation reduces to the spherical +dr2 + r2dθ2 Bessel equation, with the solution being

p ω2  2Mam2 inside the shell, which can be seen to be flat using the j`( b(ω)r) , b(ω) = 1 − . (7) coordinate transformation, f(R) R3ω

2Ma However, the solution outside the shell has to be deter- ϕ = φ − t , (3) mined numerically and matched to the solution on the R3 shell. Specifically, we impose continuity of the solution p which transforms the metric (2) to the flat metric in stan- at the shell, ψ(R) = j`( b(ω)R). To find the value of dard coordinates. The stress energy tensor of the shell dψ/dr|R+ , we integrate Eq. (5) across the shell, obtaining giving rise to the above spacetime can be found using the the condition Israel junction condition [24] and has been well-studied α(r) d   ∂ψ  in the literature [25]. ψ = eµ = 0 , (8) β(r) dr ∂xµ r The coordinates used in (1) are (spherical) Lorentzian at infinity and are naturally associated with stationary where eµ is the radial element of the tetrad and the observers at infinity. All observers at fixed (r, θ, ϕ) inside r 3 square brackets represent the difference in the value of the the shell rotate rigidly at the rate dφ/dt = 2Ma/R with term across the shell. Noting from (6) the discontinuity respect to observers at rest at infinity (φ =constant). in β(r), this yields the required initial conditions ψ(R+) This effect is called the dragging of inertial frames, first and ψ0(R+) for numerically solving the radial equation discovered in 1918 by Thirring and Lense [1] and dis- outside the shell. cussed in the introduction. Finally, to normalize the solution, we will follow the scheme presented in [19]. Defining r? such that d/dr? = α β d/dr and ρ = rψ, the radial equation (5) reads

1 In [18] the shell is in general considered to be collapsing but the d2 results can be immediately specialized if it is just rotating. ρ + (ω2 − V (r))ρ = 0 , (9) dr?2 3

π π for r > R, where τ ∈ (− 2k , 2k ) We will denote the total duration of the interaction by ∆τ = π/k. α2`(` + 1) 1 α d α V (r) = + γ + . (10) If the detector starts off in the and inter- r2 r β dr β acts with via the above Hamilto- nian, there may be a non-zero probability of finding the Asymptotically, V (r) → 0 as r → ∞ and hence ψ ∼ detector in its after the interaction. The ? ? sin(ωr )/r . Let the normalized radial solution be de- probability of excitation of the detector can be calcu- ˜ ? ? noted as ψω`m(r ) = Aω`mψ(r ). Given any two wave- lated using and is well-known in the functions Ψ1, Ψ2, their Klein-Gordon inner product is literature. It is given by [26, 27] Z µ ? ? Z ∞ Z ∞ (Ψ1, Ψ2) = i dσn (Ψ1∇µΨ2 − Ψ2∇µΨ1) , (11) 2 −iΩ(τ2−τ1) Σ P = λ dτ1 dτ2χ(τ1)χ(τ2)e −∞ −∞ (14) µ where Σ is a Cauchy surface with normal n . A solu- × W (x(τ1), x(τ2)) tion will be normalized with respect to the Klein-Gordon to second order in λ, where W (x(τ1), x(τ2)) is the Wight- inner product if we choose the normalisation constant man function of the field evaluated along the detector ∗ ? ? ? Aω`m such that Aω`mψ(r ) → 2 sin(ωr )/r as r → ∞ trajectory. [19]. The field can be expanded in terms of the Now that the normalised mode solutions are obtained, normalized field modes Ψω`m of the previous section as we are ready to compute the response of UDW detectors. Z ∞ ˆ X ψ(x(τ)) = dω aˆω`mΨω`m(x(τ)) IV. UDW DETECTOR RESPONSE `,m 0 † † +a ˆω`mΨω`m(x(τ)) , (15) A UDW detector [14, 26] is a 2-level quantum mechan- ical system that interacts locally with a scalar quantum with aˆω`m denoting the mode annihilation operators. Let field as it moves along some trajectory x(τ) in space- |0i denote the field vacuum such that aˆω`m |0i = 0. time. Letting Ω denote the energy gap of the detector This corresponds to the vacuum with respect to an ob- −iΩτ + iΩτ − and µˆ(τ) = e σˆ + e σˆ its monopole moment (in server located at infinity, who is in a non-rotating frame. the ), the Hamiltonian governing the The Wightman function with respect to this vacuum ˆ ˆ detector/field interaction is W (x(τ1), x(τ2)) := h0| ψ(x(τ2))ψ(x(τ1)) |0i is given by

Hˆ (τ) = λχ(τ)ˆµ(τ) ⊗ Ψ(ˆ x(τ)) , (12) Z ∞ X † W (x(τ1), x(τ2)) = dωΨω`m(x(τ1))Ψω`m(x(τ2)) . where σˆ± are the ladder operators, τ is the `,m 0 of the detector, and λ is the dimensionless coupling con- (16) stant. The duration of interaction is controlled by the From the previous section, we have seen that the switching function χ(τ), which we will choose to be normalized mode solutions are given by Ψω`m = √ 1 e−iωtY (θ, φ)ψ˜ (r). Recall that we are inter- ( 4πω `m ω`m cos4(kτ), − π ≤ τ ≤ π ested in studying how the response of the detector differs χ(τ) = 2k 2k (13) 0, otherwise , when placed respectively in a rotating shell and a sta- tionary shell. A simple choice for the trajectory x(τ) of 2Ma which has a shape similar to the Gaussian switching func- the detector is r = rd, θ = π/2, ϕ = 0 i.e., φ = 3 t. p R tion χG [20] used in the static case [19], but ensures for Noting that t = τ/h, where h = f(R), we find the some k > 0 that the interaction takes place only between response function F = P/λ2 of the field in the form

Z ∞ Z ∞ Z ∞ −iΩ(τ2−τ1) X † F = dτ1 dτ2χ(τ1)χ(τ2)e dωΨω`m(x(τ1))Ψω`m(x(τ2)) −∞ −∞ `m 0 Z ∞ Z ∞ Z ∞ X dω −i(Ω+ ω − 2Mam )(τ −τ ) π 2 2 p 2 = dτ dτ χ(τ )χ(τ )e h R3h 2 1 |Y ( , 0)| |A | |j ( b(ω) r )| , 4πω 1 2 1 2 `m 2 ω`m ` d `m 0 −∞ −∞ Z ∞   2 X dω ω 2Mam 2 π 2 p 2 = χˆ Ω + − |Aω`m| |Y`m( , 0)| |j`( b(ω)rd)| , (17) 2ω h R3h 2 `m 0

where we switched the order of integration since the in- tegrand is smooth and integrated over the τ1 and τ2 vari- 4 ables, which amounts to performing Fourier transforms ℱ 0.0322 1 Z ∞ χˆ(y) = √ dτχ(τ)e−iyτ (18) 0.0320 2π −∞ 0.0318 on the switching functions, noting that χˆ(−y) =χ ˆ(y) for 0.0316 a real switching function. 0.0314 We pause to comment that we have computed (17) 0.0312 from the modes Ψω`m assuming (5) is exact. However rd/R the metric (1) is a valid solution of the Einstein equa- 0.2 0.4 0.6 0.8 1.0 tions only to order a while the leading corrections to static ak=0.9 the Wightman function (16) (and thus detector response ℱrot-ℱstat (17)) are of order a2. For sufficiently small Ma/R2, terms of higher order in a will not significantly affect our quan- 0.00004 titative results, and so we shall plot (17) in what follows. 0.00003

0.00002 V. RESULTS 0.00001 We are now ready to look at how rotation of the shell rd/R affects the response of UDW detectors. We do this by 0.2 0.4 0.6 0.8 1.0 computing the expression (17) numerically, terminating ak=0.7 ak=0.8 ak=0.9 the sum over ` at sufficiently large `, chosen to give re- sultant errors not larger than 1%. FIG. 2. Plot of F against rd/R. These plots are obtained for Mk = 1, and Rk = 3. Above: Plot of detector response

ℱrot-ℱstat against rd/R for static and rotating (ak = 0.9) shells, Ω/k = Ω/k 0.5. Below: Frot − Fstat plots for different ak settings with -5 5 Ω/k = 0.5. 1⨯10-7

-1⨯10-5 -1⨯10-7 rd)k = 5, the time needed for a light signal to travel -1 1 2 3 4 5 from the detector to the shell and back. This is in strik- -2⨯10-5 ing contrast to the classical case, where the fastest way a detector inside the shell (with all possible classical fields

-3⨯10-5 in their vacuum states) can detect the presence of rota- tion is by sending and waiting for a light signal to come ak=0.5 ak=0.6 ak=0.7 ak=0.8 ak=0.9 back from the shell. FIG. 1. Detector response against Ω/k. Shown here is the In the top figure of Fig. 2, we plot both Frot(ak = 0.9) plot of the difference Frot − Fstat against Ω for different (di- and Fstat against the detector location rd/R. The re- mensionless) rotation parameters ak with Mk = 1, Rk = sponses peak at some intermediate rd, in agreement with 3 , rdk = 0.5. The inset shows a zoom-in of the plot around the results of ref.[20]. From the bottom figure, we see Ω/k = 0. The difference Frot − Fstat is small but non-zero, that the detector response increases by more than an or- and is more sensitive to the rotation for negative Ω. der of magnitude as compared to Fig. 1 as rd/R → 1. We find that the shape of the curves in Fig 1 remains Fig. 1 shows a plot of Frot −Fstat ≡ Frot −Frot(a = 0) qualitatively the same as rd/R increases, though the in- against Ω for various (dimensionless) rotation parameters teraction duration is eventually no longer less than the ak. The difference between the response of a detector light crossing time. A detector placed at the origin rd = 0 placed in a slowly rotating shell Frot and that placed cannot distinguish between a rotating and a static shell. in a static shell Fstat, though small, is clearly non-zero. We can understand this explicitly by noting that the ro- The difference is more pronounced when the energy gap tation parameter a appears in the radial equation (5) Ω/k < 0, which physically means that the detector starts through the term γ, where it is multiplied with the az- off in the excited state. The rotation parameter a enters imuthal number m. Hence, it has only nontrivial effects the response function F in three positions in eq. (17): when m 6= 0. However since θ = 0 along the axis of rota- in the Fourier transform of the switching function, in tion and Y`m(0, 0) is non-zero only when m = 0, the mode the normalisation constant Aω`m, and in the b(ω) of the solutions and hence the response function are insensitive spherical Bessel function. The net effect of these is an to effects of rotation along this axis. As another illustra- expected increase in |Frot − Fstat| with a. tion of this, we plot in Fig. 3 Frot − Fstat against θ, the We emphasize that the interaction duration ∆τ k = π angle measured from the rotation axis. From this, we see between the field and detector is less than tsk = 2(R − that the sensitivity to rotation of detectors placed at the 5

ℱrot-ℱstat can be observed without the need of quantum detectors, 1.× 10 -7 for example as the Larmor precession of charged parti- 8.× 10 -8 cles. By solving the scalar field equation numerically, we 6.× 10 -8 have obtained the response function of the detector and 4.× 10 -8 seen how it depends on the rotation parameter a. Cor- rections to the metric (1) to higher orders in a will quan- -8 2.× 10 titatively modify (17) but will not qualitatively affect our

θ results. Alternatively, we can regard (1) as a ‘kinematic π/4 π/2 spacetime’ that could be employed in analogue gravity laboratory simulations, in which case our results would FIG. 3. Plot of Frot − Fstat against θ for Mk = 1, Rk = 3, ak = 0.8 and r k = 0.5. hold exactly. Whether or not such effects can be directly d detected remains a challenge for future experiments. same rd increases monotonically as θ increases from 0 to π/2. ACKNOWLEDGMENTS

We would like to thank Keith Ng for helpful discussions VI. CONCLUSIONS on this work and Markus King for a helpful correspon- dence. This work was supported in part by the Natural Classically, the physical effect of a slowly rotating shell Sciences and Engineering Research Council of Canada is the dragging of inertial frames. We have shown that and by the Perimeter Institute. J.B. thanks the kind hos- this effect can be discerned from local measurements of a pitality of the Perimeter Institute, Waterloo, where this quantum particle detector inside the shell, on timescales work started and the Albert-Einstein Institute, Golm, much shorter than the light travel time from the detector where it continued. J.B. also acknowledges the support to the edge of the shell and back. from the Grant Agency of the Czech Republic, Grant No. We note that the gravitational effects inside a rotat- GACR˘ 19-01850S. Research at Perimeter Institute is sup- ing material shell are analogous to the electromagnetic ported in part by the Government of Canada through the effects inside a rotating charged shell; but there are also Department of Innovation, Science and Economic Devel- fundamental differences. For a rotating charged shell, a opment Canada and by the Province of Ontario through dipolar magnetic field will be formed inside. Such a field the Ministry of Colleges and Universities.

[1] J. Lense and H. Thirring, Phys. Z. 19, 156 (1918). [13] C. W. F. Everitt, B. Muhlfelder, D. B. DeBra, et al., [2] H. Thirring, Phys. Z. 19, 33 (1918). Classical and Quantum Gravity 32, 224001 (2015). [3] I. Ciufolini and J. A. Wheeler, and Inertia [14] B. S. DeWitt, in General Relativity: An Einstein Cen- (Princeton University Press, 1995). tenary Survey, edited by S.W. Hawking and W. Is- [4] J. Bi˘cák, J. Katz, T. Ledvinka, and D. Lynden-Bell, rael (Cambridge University Press, Cambridge, England, Phys. Rev. D85, 124003 (2012). 1979) pp. 680–745. [5] W. Barker, T. Ledvinka, D. Lynden-Bell, and J. Bičák, [15] D. Lynden-Bell, J. Katz, and J. Bičák, Monthly Notices Classical and Quantum Gravity 34, 205006 (2017). of the Royal Astronomical Society 272, 150 (1995). [6] C. Misner, K. S. Thorne, and J. A. Wheeler, Gravitation [16] C. Hoefer, Stud. Hist. Phil. Sci. B 48, 128 (2014). (W. H. Freeman, San Francisco, 1973). [17] L. Lindblom and D. R. Brill, Phys. Rev. D 10, 3151 [7] J. M. Bardeen and J. A. Petterson, Astrophys. J. Lett. (1974). 195, L65 (1975). [18] J. Katz, D. Lynden-Bell, and J. Bi˘cák,Classical and [8] J. Barbour and H. Pfister, eds., Mach’s principle: From Quantum Gravity 15, 3177 (1998). Newton’s bucket to quantum gravity. Proceedings, Con- [19] K. K. Ng, R. B. Mann, and E. Martin-Martinez, Phys. ference, Tuebingen, Germany, July 1993 (Birkhauser, Rev. D94, 104041 (2016). Boston, USA, 1995). [20] W. Cong, J. Bičák, D. Kubizňák, and R. B. Mann, Phys. [9] H. Pfister and M. King, Inertia and Gravitation: The Rev. D 101, 104060 (2020). Fundamental and Structure of Space-Time, Lec- [21] E. Martín-Martínez, A. R. H. Smith, and D. R. Terno, ture Notes in Physics, Vol. 897 (Springer International Phys. Rev. D 93, 044001 (2016). Publishing, Cham, 2015). [22] S.-Y. Lin, C.-H. Chou, and B. L. Hu, Journal of High [10] H. Pfister and K. H. Braun, Classical and Quantum Energy Physics 2016, 47 (2016). Gravity 2, 909 (1985). [23] A. R. Smith and R. B. Mann, Class. Quant. Grav. 31, [11] D. R. Brill and J. M. Cohen, Phys. Rev. 143, 1011 (1966). 082001 (2014). [12] C. W. F. Everitt, D. B. DeBra, B. W. Parkinson, et al., [24] W. Israel, Il Nuovo Cimento B (1965-1970) 44, 1 (1966). Phys. Rev. Lett. 106, 221101 (2011). 6

[25] E. Poisson, A Relativist’s Toolkit: The Mathematics [26] N. D. Birrell and P. C. W. Davies, Quantum Fields in of Black-Hole (Cambridge University Press, Curved Space, Cambridge Monographs on Mathematical 2004). Physics (Cambridge University Press, 1982). [27] A. Pozas-Kerstjens and E. Martín-Martínez, Phys. Rev. D 92, 064042 (2015).