1

Metamaterial-based model of the Alcubierre warp drive

Igor I. Smolyaninov

Department of Electrical and Computer Engineering, University of Maryland, College

Park, MD 20742, USA

Electromagnetic metamaterials are capable of emulating many exotic space-time

geometries, such as black holes, rotating cosmic strings, and the big bang

singularity. Here we present a metamaterial-based model of the Alcubierre warp

drive, and study its limitations due to available range of material parameters. It

appears that the material parameter range introduces strong limitations on the

achievable “warp speed”, so that ordinary magnetoelectric materials cannot be

used. On the other hand, newly developed “perfect” bi-anisotropic non-reciprocal

magnetoelectric metamaterials should be capable of emulating the physics of warp

drive gradually accelerating up to 1/4c.

PACS numbers: 78.20.Ci, 42.25.Bs.

Metamaterial optics [1,2] greatly benefited from the field theoretical ideas developed to

describe physics in curvilinear space-times [3]. Unprecedented degree of control of the

local dielectric permittivity εik and magnetic permeability μik tensors in electromagnetic

metamaterials has enabled numerous recent attempts to engineer highly unusual “optical

spaces”, such as electromagnetic black holes [4-8], wormholes [9], and rotating cosmic

strings [10]. Phase transitions in metamaterials are also capable of emulating physical
 2

processes which took place during and immediately after the big bang [11,12]. These

models can be very informative for phenomena where researchers have no direct

experience and therefore limited intuition.

Since its original introduction by Alcubierre [13], the warp drive spacetime has

become one of the most studied geometries in general relativity. In the simplest form it

can be described by the metric

ds 2 = c 2 dt 2 − (dx − v(r )dt ) − dy 2 − dz 2

2
(1)

(
where r = ( x − v0 t ) + y 2 + z 2
2
)
1/ 2
is the distance from the center of the “warp bubble”,

v0 is the warp drive velocity, and v=v0f(r). The function f(r) is a smooth function

satisfying f(0)=1 and f (r ) → 0 for r → ∞ . This metric describes an almost flat

spheroidal “warp bubble” which is moving with respect to asymptotically flat external

spacetime with an arbitrary speed v0. Such a metric bypasses the speed limitation due to

special relativity: while nothing can move with speeds greater than the speed of light

with respect to the flat background, spacetime itself has no restriction on the speed with

which it can be stretched. One example of fast stretching of the spacetime is given by

the inflation theories, which demonstrate that immediately after the big bang our

Universe expanded exponentially during an extremely short period of time.

Unfortunately, when the spacetime metric (1) is plugged into the Einstein’s

equations, it is apparent that exotic matter with negative energy density is required to

build the warp drive. In addition, it was demonstrated that the eternal superluminal warp

drive becomes unstable when quantum mechanical effects are introduced [14]. Another

line of research deals with a situation in which a warp drive would be created at a very

low velocity, and gradually accelerated to large speeds. Physics of such a process is also

quite interesting [15]. We should point out that the warp drive space-time cannot be
 3

reduced to a simple combination of a white-hole and a black-hole event horizons. Such

a combination would be non-controversial and “easy” to realize. The difference between

the warp drive space-time and such a white-hole/black-hole combination is that the flat

space-time region inside the warp bubble is moving as a whole with respect to the flat

space-time outside the warp bubble (see metric in eq.(1)) . This non-trivial property of

the warp drive space-time has led to conclusion that it cannot be realized even at sub-

luminal speeds. Very recently it was demonstrated that even low speed sub-luminal

warp drives generically require energy-conditions-violating matter [16]: the T00

component of the energy-momentum tensor (the energy density distribution) appears to

be negative even at sub-luminal speeds. Therefore, even subluminal warp drives appear

to be prohibited by the laws of physics.

In this paper we demonstrate that electromagnetic metamaterials are capable of

emulating the warp drive metric (1). Since energy conditions violations do not appear to

be a problem in this case, metamaterial realization of the warp drive is possible. Our

result is interesting because the body of evidence collected so far seemed to indicate that

the warp drives operating at any speed (even sub-luminal) were strictly prohibited by

the laws of Nature

Below we will find out what kind of metamaterial geometry is needed to

emulate a laboratory model of the warp drive, so that we can build more understanding

of the physics involved. It appears that the available range of material parameters

introduces strong limitations on the possible “warp speed”. Nevertheless, our results

demonstrate that physics of a gradually accelerating warp drive can be modeled based

on newly developed “perfect” magnetoelectric metamaterials [17]. Since even low

 4

velocity physics of warp drives is quite interesting [15,16], such a lab model deserves

further study.

To avoid unnecessary mathematical complications, let us consider a 1+1

dimensional warp drive metric of the form

ds 2 = (c / n∞ ) dt 2 − (dx − v0 f ( ~
x )dt ) − dy 2 − dz 2
2 2
(2)

x = ( x − v0 t ) and n∞ is a scaling constant. In the rest frame of the warp bubble it

where ~

can be re-written as

2
(
ds 2 = (c / n∞ ) dt 2 − d~
~
x + v0 f ( ~
2
)
x )dt − dy 2 − dz 2 (3)

~ ~
where f (0) = 0 , and f ( ~
x ) → 1 for ~
x → ±∞ . The resulting metric is

⎛ 1 v02 ~ 2 ~ ⎞ 2 2 ~ 2 ~
ds = ⎜⎜ 2 − 2 f ( x ) ⎟⎟c dt − dx − 2v0 f ( ~
2
x )d~
x dt − dy 2 − dz 2 (4)
⎝ n∞ c ⎠

Following ref.[18], Maxwell equations in this gravitational field can be written in the

three-dimensional form as
r r
r E
D=
rr
+ Hg , [ ] r H
B=
rr
+ gE , [ ] (5)
h h

where h=g00, and gα=-g0α/g00. These equations coincide with the macroscopic Maxwell

equations in a magneto-electric material [19]. In the equivalent material

1
ε = μ = h −1 / 2 = (6)
1 v02 ~ 2 ~
− f (x )
n∞2 c 2

and the only non-zero component of the magneto-electric coupling vector is

v0 ~ ~
f (x )
gx = c
(7)
1 v02 ~ 2 ~
− f (x )
n∞2 c 2
 5

In the subluminal v0<<c limit eqs.(6) and (7) become

⎛ v02 n∞2 ~ 2 ~ ⎞ v ~
ε = μ ≈ n∞ ⎜⎜1 + 2 f ( x ) ⎟⎟ , g x ≈ n∞2 0 f ( ~x ) (8)
⎝ 2c ⎠ c

The magneto-electric coupling coefficients in thermodynamically stable materials are

limited by the inequality [20]:

g x2 ≤ (ε − 1)(μ − 1) , (9)

which means that a subluminal warp drive model based on the magnetoelectric effect

must satisfy inequality

v0 ~ ~ n −1
f (x ) ≤ ∞ 2 (10)
c n∞

This inequality demonstrates that while “the true warp drive” in vacuum ( n∞ = 1 ) is

prohibited, n∞ > 1 values in a material medium make a warp drive model

thermodynamically stable at least at subluminal speeds. This is an important result

because the body of evidence collected so far seems to indicate that the warp drives

operating at any speed are strictly prohibited by laws of Nature. Equation (10) also

provides an upper bound on the largest possible “warp speed”, which is achievable

within the described metamaterial model. This upper bound is reached at n∞ = 2 , and

equals to v0=1/4c. Therefore, at the very least, we can build a toy model of a warp drive

“operating” at v0~1/4c. Coordinate dependence of the metamaterial parameters in such a

x ) = (1 + a 2 / ~
x2) .
~ −1
model is shown in Fig.1 assuming f ( ~

However, in classical magnetoelectric materials, such as Cr2O3 and

multiferroics, actual values of magnetoelectric susceptibilities are two orders of

magnitude smaller than the limiting value described by eq. (9) [21], so that the warp

drive model is impossible to make with ordinary materials. On the other hand, recently
 6

developed “perfect” magnetoelectric metamaterials [17], which can be built based on

such designs as split ring resonators, fishnet structures [22], etc. allow experimentalists

to reach the limiting values described by eq.(9), and make a lab model of the warp drive

possible. Following ref.[17], the effective susceptibilities of the split ring metamaterial

can be written in the RLC-circuit model as

nCd 2ω02 nCS 2ω 2ω02

ε = 1+ 2 , μ = 1+ 2 2
(
ω0 − ω 2 − iωγ ) (
c ω0 − ω 2 − iωγ
, and
) (11)

inCdSωω02
g=
(
c ω02 − ω 2 − iωγ), (12)

where n is the split ring density, d is the gap in the ring, S is the ring area, and C is the

gap capacitance. These expressions explicitly demonstrate that the split ring

metamaterial considered in ref. [17] satisfies the upper bound given by eq.(9), and

therefore can be used as one of the building blocks in the metamaterial warp drive

design. On the other hand, this particular split-ring design cannot be used without

modification, since this metamaterial is reciprocal.

An actual laboratory demonstration of a metamaterial warp drive space time

would require a non-reciprocal bi-anisotropic metamaterial, in which both spatial and

time reversal symmetries are broken. In addition, the metamaterial loss issue has to be

overcome. Since the issue of metamaterial loss compensation using gain media is well

studied (see for example recent experimental demonstration of loss compensation in a

negative index metamaterial [23]) let us concentrate on the experimental ways of

breaking spatial and time reversal symmetries. Breaking the mirror x ↔ -x symmetry is

most easily achieved by deformation of the metamaterial, which can be easily done in

either one of the most popular split ring [17] or fishnet [22] metamaterial designs. As far

as breaking the time-reversal t ↔ -t symmetry is concerned, there are two most natural
 7

ways to break this symmetry in solids: application of external magnetic field [24,25], or

spin-orbit interaction in a nonsymmorfic lattice [26]. In addition, such chiral

superconductors as Sr2RuO4 [27] may be used in superconducting metamaterial designs

[28]. Application of external electric and magnetic fields is known to break both spatial

and time symmetries of such materials as methyl-cyclopentadienyl-Mn-tricarbonil

molecular liquids [24], thus creating an illusion of a moving (non-reciprocal bi-

anisotropic) medium [25]. Experimental results of ref. [24] do indeed demonstrate this

behavior (however, the emulated “medium velocity” is very low, of the order of 50 nm/s

[25]). Since utilization of magnetized particles, such as ferrites is easily applicable in

the metamaterial design, all the ingredients necessary for experimental realization of the

Alcubierre metric have been demonstrated in the experiment. Moreover, we should

point out that very recently it was also asserted [29,30] that material parameters, which

are necessary to achieve a warp drive imitation in a nanostructured metamaterial are

indeed possible.

Let us come back to the split ring-based “perfect magnetoelectric metamaterial”

design implemented in [17], and demonstrate how the time-reversal symmetry may be

broken in this metamaterial. One of the possible metamaterial geometries is shown

schematically in Fig.2. An elementary unit of the split ring-based “perfect

magnetoelectric metamaterial” design of ref.[17] is supplemented with a magnetized

ferrite particle. The particle is magnetized and shifted along the x-direction with respect

to the center of the split ring. Thus, this geometry explicitly violates spatial and time

reversal symmetries, resulting in a non-reciprocal bi-anisotropic metamaterial. As

demonstrated in [31], near the ferromagnetic resonance frequency in such a

metamaterial
 8

ω mω 0
g~ , (13)
ω02 − ω 2

where ω0 is the ferromagnetic resonance frequency, and ωm=γM0. Thus, in the design

presented in Fig.2 gx is proportional to the particle magnetization M0 in a given location,

which explicitly demonstrate the non-reciprocal nature of this metamaterial design.

Time reversal t ↔ -t leads to change of sign of gx. We should also note that very

recently a somewhat related metamaterial design has been proposed in ref.[32] , which

emulates medium motion at an arbitrary speed. While demonstrating the proof of

principle, the designs presented in Fig.2 and ref.[32] may only be considered as a first

step. Such complicated metamaterial designs typically contain many unwanted terms in

ε, μ and g, which must be carefully eliminated by iteration, so that the ideal form of

eqs.(8) may be achieved.

Light rays propagation inside the metamaterial model of the warp drive

“operating” at 1/4c is illustrated in Fig.3. The metamaterial medium (1) outside the

warp bubble is engineered to have properties of a medium moving towards the warp

bubble with the designed warp speed, while medium (3) is “moving” away from the

bubble. In the reference frame moving with the warp speed these media look exactly the
~
same as medium (2) at rest. Ray trajectories were calculated assuming a step-like f ( ~
x)

profile. Rays are emitted by a point source located at the origin point (0,0,0) of the

coordinate frame inside the “warp bubble” (marked as medium (2)). Boundaries of the

warp bubble are located at x=±5 (they are marked by the dashed lines). At large enough

incidence angles light rays originating inside the “warp bubble” cannot penetrate into

medium (3) (in the hypothetical superluminal warp drive this would be true for any

incidence angle: this boundary would look like a white hole event horizon). On the
 9

other hand, all the light rays propagating towards the other boundary of the “warp

bubble” can propagate into medium (1). Note that the metamaterial medium (1) is

identical to medium (3).

In conclusion, we have presented a metamaterial-based model of the Alcubierre

warp drive metric. It appears that the material parameter range introduces strong

limitations on the achievable “warp speed”, so that ordinary magnetoelectric materials

cannot emulate the warp drive. On the other hand, newly developed “perfect”

magnetoelectric bi-anisotropic non-reciprocal metamaterials should be capable of

emulating the physics of gradually accelerating warp drive, which can reach “warp

speeds” up to 1/4c.

References

[1] J. B. Pendry, D. Schurig, D.R. Smith, Science 312, 1780-1782 (2006).

[2] U. Leonhardt, Science 312, 1777-1780 (2006).

[3] U. Leonhardt and T. G. Philbin, New J. Phys. 8, 247 (2006).

[4] I.I. Smolyaninov, New Journal of Physics 5, 147 (2003).

[5] T. G. Philbin, C. Kuklewicz, S. Robertson, S. Hill, F. König, and U. Leonhardt,

Science 319, 1367 (2008).

[6] D.A. Genov, S. Zhang, and X. Zhang, Nature Physics, 5, 687-692 (2009).

[7] E.E. Narimanov and A.V. Kildishev, Appl. Phys. Lett. 95, 041106 (2009).

[8] Q. Cheng and T. J. Cui, arXiv:0910.2159v3

[9] A. Greenleaf, Y. Kurylev, M. Lassas, and G. Uhlmann, Phys. Rev. Lett. 99, 183901

(2007).

[10] T. G. Mackay and A. Lakhtakia, Phys. Lett. A 374, 2305-2308 (2010).

 10

[11] I.I. Smolyaninov and E.E. Narimanov, Phys. Rev. Letters 105, 067402 (2010)

[12] I.I. Smolyaninov, Journal of Optics 12, 024004 (2010).

[13] M. Alcubierre, Classical and Quantum Gravity 11, L73 (1994).

[14] S. Finazzi, S. Liberati and C. Barcelo, Phys. Rev. D 79, 124017 (2009).

[15] C. Clark, W.A. Hiscock, and S.L. Larson, Classical and Quantum Gravity 16,

3965-3972 (1999).

[16] F.S.N. Lobo and M. Visser, Classical and Quantum Gravity 21, 5871 (2004).

[17] A.M. Shuvaev, S. Engelbrecht, M. Wunderlich, A. Schneider, and A. Pimenov,

arXiv:1004.4524v1

[18] L.D. Landau, E.M. Lifshitz, The Classical Theory of Fields (Elsevier, Oxford,

2000).

[19] A.N. Serdyukov, I.V. Semchenko, S.A. Tretyakov, and A. Sihvola,

Electromagnetics of Bi-Anisotropic Materials: Theory and Applications (Gordon and

Breach, Amsterdam 2001).

[20] W.F. Brown, R.M. Hornreich, and S. Shtrikman, Phys. Rev. 168, 574-577 (1968).

[21] M. Fiebig, Journal of Physics D: Applied Physics 38, R123 (2005).

[22] M.S. Rill, C.E. Kriegler, M. Thiel, G. von Freymann, S. Linden, and M. Wegener,

Optics Letters 34, 19-21 (2009).

[23] S. Xiao, V.P. Drachev, A.V. Kildishev, X. Ni, U.K. Chettiar, H.-K. Yuan, and

V.M. Shalaev, Nature 466, 735-740 (2010).

[24] T. Roth, G.L.J.A. Rikken, Phys. Rev. Letters 88, 063001 (2002).

[25] A. Feigel, Phys. Rev. Letters 92, 020404 (2004).

[26] Y. T. Rebane, Phys. Rev. B 52, 1590–1595 (1995).

 11

[27] H. Nobukane, A. Tokuno, T. Matsuyama, and S. Tanda, Phys. Rev. B 83, 144502

(2011).

[28] S.M. Anlage, J. Opt. 13, 024001 (2011).

[29] I.I. Smolyaninov, arXiv:1009.5663v1.

[30] T.H. Anderson, T.G. Mackay, and A. Lakhtakia, arXiv:1101.0044v1.

[31] S.A. Tretyakov, A.H. Sihvola, A.A. Sochava, and C.R. Simovski, J. of

Electromagnetic Waves and Applications 12, 481-497 (1998).

[32] S.A. Tretyakov and I.S. Nefedov, in Proc. of the 3rd Int. Conf. on Advanced

Electromagnetic Materials in Microwaves and Optics, p. 114-116 (2009).

 12

Figure Captions

Figure 1. Spatial distributions of ε, μ, and gx in the metamaterial model of a warp drive

gradually accelerated up to 1/4c.

Figure 2. Example of a non-reciprocal bi-anisotropic metamaterial geometry, which

explicitly violates spatial and time reversal symmetries. An elementary unit of the split

ring-based “perfect magnetoelectric metamaterial” design of ref.[17] is supplemented

with a magnetized ferrite particle. The particle is magnetized and shifted along the x-

direction with respect to the center of the split ring. The particle magnetization is

proportional to the required gx in a given location.

Figure 3. Light rays propagation inside the metamaterial model of the warp drive

operating at 1/4c. Rays are emitted by a point source located at the origin point (0,0,0)

of the coordinate frame inside the “warp bubble”. Boundaries of the warp bubble are

located at x=±5. Note that the metamaterial media (1) and (3) are identical.
 13

v=1/4c
ε=μ
2.0 v=1/8c

ε, μ, gx
1.5

1.0
v=1/4c gx
0.5
v=1/8c

0.0
-15 -10 -5 0 5 10 15
4x/a

Fig.1
 14

Fig.2
 15

(1) (2) (3)

50
40
30
20
10
0
y

-10
-20
-30
-40
-50
-50 -40 -30 -20 -10 0 10 20 30 40 50
x

Fig.3