Formation of Closed Timelike Curves with Morris-Thorne wormholes
|
|
Thread rating:  |
Adam Getchell - 14 Feb 2005 20:36 GMT Hello all,
I'm reviewing some basic results from the Morris, Thorne, and Yurtsever paper "Wormholes, Time Machines, and the Weak Energy Condition" (Phys. Rev. Lett., Volume 61, Number 13), and I would like to check my conclusions.
First, I'll assume that a traversable Lorentzian wormhole is constructed and transported at relativistic velocity, and that the mouth is opened only upon arrival to the destination. Let mouth A be the stationary mouth (left in the paper) and mouth B be the moving mouth. Both reference frames start at x = t = 0.
In the first case, for definiteness assume a 100 light year voyage at ..7c. If I haven't made a mistake, with respect to (= wrt) inertial frame A the voyage requires 142.8 years (100 ly / .7 c), so the wormhole is opened at +142.8 years.
WRT to B, the voyage requires t' = t/gamma = 142.8 / 1.4 = 102 years, where gamma = 1/Sqrt[1-v^2/c^2] and gamma(.7) ~ 1.4. Information from A->B propagates instantly through the wormhole, but requires 100 years to propagate outside the wormhole. Since there are no inertial frames in GR, and in particular, A and B are different frames of reference, am I correct in concluding that we don't (yet) have a closed timelike curve?
In case 2, assume that B travels 100 ly and back to be brought within 1 light second of A. Then, wrt A the wormhole mouth opens at Y285.6, but wrt B the wormhole mouth opens at Y204. Because A and B are now in the same reference frame, there is a delta of 81.6 years. Specifically, WRT A the time frame is 286.5, WRT B it is 204, so traveling from A to B goes back in time by 81 years?
Unfortunately, this is the opposite conclusion from the MTY paper, so where have I gone wrong? (In MTY B->A yields a CTC)
For case 3, assume additionally a 3 wormholes: A<->B, B<->C, and C<->, all spaced as in case 1. If I represent this as a directed graph with arrows pointing from "older" to "younger", is this correct?
A --> B --> C --> A
If so, then this cyclic graph produces a time machine, also of some 80 years?
Final query: a directed acyclic graph of wormholes should not produce any CTCs?
Thanks for your comments, clarifications, and corrections.
--Adam Getchell
George Jones - 16 Feb 2005 17:37 GMT > Hello all, > [quoted text clipped - 29 lines] > Unfortunately, this is the opposite conclusion from the MTY paper, so > where have I gone wrong? (In MTY B->A yields a CTC) Suppose we have a spacetime that has a wormhole, and that spacetime is (nearly) flat outside of the wormhole. Suppose further that the throat of the wormhole is infinitesimally short.
The wormhole is a time machine, i.e., spacetime has a closed timelike curve (CTC), if Han Solo can do the following: jump into mouth 2, move through the wormhole's throat, emerge mouth 1, move from mouth 1 to mouth 2 through spacetime external to the thoat, and bump into himself as he jumps into mouth 2.
Call jumping into 2 event C' and emerging from 1 event C. The above is possible if event C' is in the future of C in the the external spacetime. In other words, there has to be a (future-directed) timelike worldline joining C to C' for Han to traverse. The CTC C'CC' consists of a segment from C' to C through the throat of the wormhole together with a timelike segment from C to C' external to the wormhole.
Because the throat of the wormhole is short (and thin), Every event on the worldline of mouth 1 is identified with a corresponding event on the worldline of mouth 2. A and A' are one such pair of identified points.
Suppose the mouths of the wormhole are initially close together and at rest with respect to a particular inertial reference frame for the external flat spacetime. Now let the mouths play the roles of the twins in the twin paradox, i.e., mouth 1 stays at rest with respect to the original reference frame, while mouth 2 moves out and back.
Which pairs of events are identified in this scenario? At the instant that the mouths start to move apart, place a (zeroed) clock in mouth 1. The clock is also in mouth 2, again because of the infinitesimal throat length. Therefore, the clock measures proper time for both "twins", and an event A on the worldline of mouth 1 is identified with an event A' on the worldline of mouth 2 iff the proper time elapsed for event A according to mouth 1 equals the proper time elapsed for A' according to mouth 2.
The following spacetime diagram shows the worldlines of the 2 mouths and 3 pairs of identified events. If Han goes into the wormhole at one event in the pair, he comes out of the wormhole at the other event of the pair. As explained above, there has to be a timelike relationship between events in a pair in order to have a CTC.
t
| |\ | \ | \ | \ | \C' | \ | \ | \ C| \ | \B' | / B| / | / | / | /A' | / A| / | / | / |/ O----------------------------- x
The proper time of any event on the worldline of mouth 1 is just the t coordinate of that event. The proper time tau of any event on the worldline of mouth 2 is related to the t coordinate of that event by
tau = t*sqrt(1 - v^2). (1)
Consider the A and A' events. The identification condition is
t_A = tau_A'. (2)
To find the causal relationship between A and A', consider
(x_A' - x_A)/(t_A' - t_A)
= v*t_A'/(t_A' - t_A'*sqrt(1 - v^2)) (from (1) and (2))
= v/(1 - sqrt(1 - v^2)). (3)
Note: 1) (3) is independent of A and A' , and so is valid for any part of the outgoing part of mouth 2's worldline;
2) (3) > 1, so the relationship between A and A' is spacelike, and thus is part of any worldline for Han.
Like you said, no time machine forms as mouth 2 moves away from mouth 1.
Consider C and C'.
x_C' = x_B' - v*(t_C' - t_B')
= 2*x_B' - v*t_C'
gives
t_C' = (2*x_B' - x_C')/v (4)
Suppose a time machine forms at C', i.e., C and C' lightlike related. Then,
1 = (x_C' - x_C)/(t_C' - t_C)
= x_C'/(t_C' - t_C'*sqrt(1 - v^2))
= v*x_C'/[ (2*x_B' - x_C')*(1 - sqrt(1 - v^2))]. (5)
After playing around with (5), I get (maybe incorrectly)
x_C' = x_B'*[1 - sqrt((1 - v)/(1 + v))]. (6)
In your example, v = 7/10 and x_B' = lightyears.
I get x_C' = 58 lightyears. A time machine forms when mouth 2 get within 58 lightyears of mouth 1 on the return trip.
I haven't checked the above carefully, so there may be mistakes.
Note the Doppler shift factor in (6). In terms of of twins, a time machine forms when twin 2 sees via light, maybe through a telescope, the time on twin 1's watch to be the same as the time on his watch.
Regards, George
Adam Getchell - 18 Feb 2005 17:42 GMT > Suppose we have a spacetime that has a wormhole, and that spacetime is > (nearly) flat outside of the wormhole. Suppose further that the throat > of the wormhole is infinitesimally short. Hmmm ... I don't think you can make this assumption.
Using the Morris-Thorne wormhole metric:
ds^2=-exp(2*phi(l))*dt^2+dl^2+r(l)^2*(d(theta)^2+sin^2(theta)d(phi)^2))
where l is the proper radial distance (-infinity to +infinity), then a traveler of length L through the wormhole throat experiences a tidal acceleration g equal to:
g/L = Abs((1-b/r)(-phi''-(phi')^2)+ (1/(2r^2))*(b'r-b)phi')
where primes denote derivatives wrt l.
(Visser, eq. 13.9)
Since the beings would like to traverse the wormhole without getting shredded, this effectively sets a minimum "size" (shape and lapse/redshift function) for the wormhole.
> Because the throat of the wormhole is short (and thin), Every event on > the worldline of mouth 1 is identified with a corresponding event on the > worldline of mouth 2. A and A' are one such pair of identified points. I don't think you can identify events from mouth 1 to events on mouth 2: the metric is definitely not flat.
In particular, traversing the wormhole from distance -l1 to +l2 takes:
Proper time for traveler: Integral(from -l1 to +l2)*dl*(v*gamma)^-1
Time wrt to static observer: Integral(from -l1 to +l2)*dl*(v*exp(phi))^-1
For the solution with exotic matter limited to the throat in the original 1988 Morris-Thorne paper, the transit time is about 200 days.
> Which pairs of events are identified in this scenario? At the instant > that the mouths start to move apart, place a (zeroed) clock in mouth 1. > The clock is also in mouth 2, again because of the infinitesimal throat > length. Therefore, the clock measures proper time for both "twins", and I think in this case the clocks have to be placed at the entrance/exit to the wormhole, rather than at the throat, so the time should be given by the second equation wrt to a static observer, i.e. involving exp(phi). Then (I think) the moving mouth will have the additional Lorentz factor described.
Although I'm not yet sure that relativistic movement of the wormhole won't collapse the metric. As a very crude approximation (I don't know how accurate it is):
For perturbation theory, g ~ g0+ delta(g), where g = wormhole metric, g0 is the unperturbed Morris-Thorne metric, and delta(g) is the perturbation.
Roughly speaking, delta(g) < g0, else perturbation theory doesn't apply. To first order, we can approximate delta(g) by the Lorentz contraction of the length in the z direction. This gives us the condition that:
gamma(dz) * z < 2 z
where dz=v, and gamma(dz) is 1/Sqrt(1-v^2/c^2). This gives us dz = ..866 c, so in order for our perturbation to die off, maximum velocity in the z direction is .86 c.
Another way of saying this is that we want a convergent geometric series:
Sum(n=0 to infinity) a*r^n converges for r < 1
Now, as a SWAG, I'm going to aim for quadratic damping. This means we want the gamma factor to be no more than 1.5, so that the perturbation terms go as 1/2^n (remembering the factor of 1 goes to the g0 term). This arises from integrating the geometric series term by term: Integral of 1/x^2 --> 1/x, but the series 1/x diverges.
So limiting gamma(dz) < 1.5 yields dz < .74c.
This is why I limited the wormhole speed to ~.7c in my example.
> The following spacetime diagram shows the worldlines of the 2 mouths and > 3 pairs of identified events. If Han goes into the wormhole at one event [quoted text clipped - 32 lines] > > tau = t*sqrt(1 - v^2). (1) Where t is the integral I gave above, and I think that properly we should consider the length contraction of the shape function b for the moving mouth.
> Suppose a time machine forms at C', i.e., C and C' lightlike related. > Then, [quoted text clipped - 15 lines] > > I haven't checked the above carefully, so there may be mistakes. I'll have to work this out with the caveats I've mentioned, thanks! Unless I'm completely barking up the wrong tree.
> Regards, > George Adam
George Jones - 20 Feb 2005 23:04 GMT >>Suppose we have a spacetime that has a wormhole, and that spacetime is >>(nearly) flat outside of the wormhole. Suppose further that the throat [quoted text clipped - 5 lines] > > ds^2=-exp(2*phi(l))*dt^2+dl^2+r(l)^2*(d(theta)^2+sin^2(theta)d(phi)^2)) I definitely did not use the metric from the original Morris and Thorne paper, Am. J. Phys. v56, pp395-412 (988).
> where l is the proper radial distance (-infinity to +infinity), then a > traveler of length L through the wormhole throat experiences a tidal [quoted text clipped - 9 lines] > shredded, this effectively sets a minimum "size" (shape and > lapse/redshift function) for the wormhole. Agreed.
I sacrificed the traversable aspect of the wormhole in order to focus on the time machine aspect in the most simplistic way possible.
>>Because the throat of the wormhole is short (and thin), Every event on >>the worldline of mouth 1 is identified with a corresponding event on the >>worldline of mouth 2. A and A' are one such pair of identified points. > > I don't think you can identify events from mouth 1 to events on mouth 2: > the metric is definitely not flat. See section 18.1 of Visser's Lorentzian Wormholes, which, in part, says "It is sufficient for the purposes of this chapter, to take an extremely simple idealized model for a traversable wormhole. Start by considering a short-throat wormhole embedded in flat Minkowski spacetime - one of the 'cut and paste' wormholes described by the thin-shell formalism. .... The wormhole is now mathematically modeled by Minkowski space with two timelike worldlines identified."
The thin-shell formalism, described in chapter 14 of Visser, uses curvature tensors that have delta-function contributions, so, as you say, the tidal forces on Han are a bit much :-) in this model.
In the model I gave, the timelike worldlines are the worldlines of the twins in the twin paradox and the identification is (t,0) ~ (gamma*t , v*gamma*t) for t <= x_B'/(v*gamma) and (t,0) ~ (gamma*t , 2*x_B' - v*gamma*t) for t > x_B'/(v*gamma).
The incoming identification is analogous to (part of) the Misner space identification given in the first full paragraph on page 1448 of (MTY), Morris, Thorne, and Yurtsever, Phys. Rev. Lett., v61, pp1146-1149 (1988). This paragraph and the next also point out the interesting counteracting roles of the Doppler shift and defocusing in the closed lightlike curve C'CC' on my diagram.
Finally, since the caustic in Fig. 2 of MTY joins nearly identical clock readings, I think my model does approximate the more realistic model drawn in Fig.2.
Regards, George
Nick Maclaren - 18 Feb 2005 17:43 GMT |> Adam Getchell wrote: |> > [quoted text clipped - 7 lines] |> in the twin paradox, i.e., mouth 1 stays at rest with respect to the |> original reference frame, while mouth 2 moves out and back. Is this really the best way to think of this problem? I may be confused, but my take on this is rather different.
What I think that that paper is saying is that, if spacetime is multiply connected, more-or-less arbitrarily manipulable AND Einstein's formulae apply over the whole of its surface, then there is a breach of causality. But all three conditions are needed to prove this for a single wormhole.
It is well-known that a breach of causality is an inconsistency in temporal ordering, and it is also well-known that that any inconsistency in a basic model can lead to two equivalent analyses giving different answers, so it is not surprising that two people get different results.
Regards, Nick Maclaren.
George Jones - 20 Feb 2005 23:04 GMT > |> Adam Getchell wrote: > |> > [quoted text clipped - 10 lines] > Is this really the best way to think of this problem? I may be > confused, but my take on this is rather different. Maybe not, but I do think it is the best way to get at the time machine aspects of the situation. However, my presentation maybe wasn't as clear as it might have been.
I give a few more details in my response to Adam's response to my response to Adam's original post.
Regards, George
> What I think that that paper is saying is that, if spacetime is > multiply connected, more-or-less arbitrarily manipulable AND [quoted text clipped - 7 lines] > giving different answers, so it is not surprising that two people > get different results.
|
|
|