Friday, December 17, 2004

Astronomers prove string theory?

David Goss has pointed out the following article in the future issue of New Scientist to me.
The organization of this entry is slightly chaotic because I started to write it before I've seen the full version of that text.

As far as I understand the article, a quirky quasar combined with double images of galaxies that look like if they originate from a cosmic string actually lead some astronomers to believe that
  • a huge cosmic string - possibly a macroscopic heterotic or type II string - is stretched across our galaxy
That's of course too cool and one is naturally skeptical. I am doubly skeptical because New Scientist has not been a terribly serious journal in my eyes in the last 5 years. Nevertheless I think it's fun to bring your attention to this article because observing a string in the telescope is the favorite scenario of many leaders of our field how string theory should eventually be proved. ;-)

OK, let me now pretend that I believe this stuff. The cosmic superstrings were recently studied - and "predicted" :-) - by Joe Polchinski et al., see his following papers
and their citations and references. The paper explains how the cosmic superstrings can be distinguished from other, more ordinary types of cosmic strings. (The article in New Scientist only dedicates the last paragraph to the difference between the fundamental strings and other types of cosmic strings.)

I also recommend you a recent review of the cosmic string issues by T. Kibble who is not a string theorist but who has a lot to say about the cosmic strings especially because he is the #1 founding father of "cosmic string theory":
Those who have read The Elegant Universe by Brian Greene also know that a string observed in the telescope is also Edward Witten's favorite scenario how string theory will eventually be proved - because nothing would settle the question whether the strings exist so cleanly.

News - gravitational lensing of CSL-1:

Before I read the whole article, Joe Polchinski wrote me that he had been interviewed, and he believes that the discovery may have been related to the following Russian-Italian observation by Mikhail Sazhin et al.:
When I saw the full article, this rumor was confirmed. The authors keep the coordinates of the object in secret and they want to get a telescope time before someone else studies it. Joe Polchinski is skeptical whether the object is a cosmic string, but Henry Tye is more confident because the tension seems to be much bigger than the upper bound allowed for pulsars. However Eanna Flanagan claims that the pulsar bound is wrong.

The team has observed a pair of galaxies 10 billion light years away and gravitational lensing is supposed to be the origin. The angular separation of the pair is roughly 2 arc-seconds so that the two images of the galaxy (or two galaxies?) almost touch. OK, now the unusual part:

Gravitational lensing, as Andy Neitzke explains me, usually produces an odd number of images. This is a consequence of the Morse theory. Moreover, these images have typically very different intensities if the source of the lensing is a point-like object. The Russian-Italian pair is special because
  • no candidate galaxy whose gravity would be the source of the lensing has been seen
  • exactly two (even number) of images have been observed
  • their intensity is equal, at least in three different intervals of wavelength (within the experimental error) - in fact, a similar pair of radio sources of equal strength may already have been found by Winn et al. in 2000 (the separation is 1.13 arc seconds there)
  • the images are not distorted; it's very unusual for gravitational lensing as you can see if you play with the simulation here:
  • ...
Tom Kibble from London believes that these things together indicate a presence of a cosmic string. In fact, I have not told you the main interesting observation yet:
  • the team of Sazhin has now found 11 pairs of double images in the 16-arc-second square around their original object CSL-1
A galaxy as the source of gravitational lensing is statistically expected to create 2 images (pairs) in average, while the cosmic strings would give you something in between 9 and 200; note that 11 belongs to this interval. ;-) The main skeptics' explanation is that the double images are just random pairs of galaxies etc. that happen to be similar to each other. Incidentally, I also recommend you an Italian page about CSL-1 and why they believe it is a double image caused by a cosmic string:
Moreover, there seems to be one more rumor from Robert Helling, Cambridge, England: the WMAP people may have already obtained the coordinates of CSL-1, and they may have looked in that direction. It is "not ruled out" that they've seen a repetition of the image of the cosmic microwave background, something that you would expect the hypothetical cosmic string to do.

The oldest double image - another evidence

The first gravitational lensing (by an ordinary point-like object, and therefore distorted) was observed in 1979 by the Jodrell Bank telescope, the UK. It's a double quasar Q0957+561A,B. Normally, the oscillations of one image A are mimicked by the other image B roughly 417 days later, because of the different lengths of the two paths. But recently our Harvard astro-colleagues have observed some additional, 100-day-long oscillations in intensity (plus minus 4% or so) which were repeated by both images roughly four times without any time delay. Well, it's the right moment to insert a link to the paper about these strange oscillations:
OK, something is moving in between us and the lensing object - and it must be very close to us because the lag is zero. Moreover, because the images are separated by 6 arc seconds which is a rather large angle, the nearby object(s) causing the oscillations must be pretty large. Because a hypothetical binary star or a similar object would have to be unacceptably heavy to give you the speedy oscillations, the source of the additional variations should be an oscillating cosmic string, they say. A string is able to oscillate very quickly.

However, what seems more comprehensible is that they can measure how this object (string?) is moving, and it is moving by the velocity 0.7c across our line of sight. OK, at any rate, they believe that this string is oscillating, and the smooth period of 100 days of the oscillation is translated to the radius comparable to 1000 astronomical units - and the string should be in our galaxy, just 10,000 light years from the Sun!

Note that this old double-imaged quasar is at a different location than CSL-1, so it would probably be an artifact of a different cosmic string, if I understand it well. It would be interesting to know whether the estimated tensions of these two cases are close.

More tests to be done include the spectral analysis of CSL-1 to determine whether they're really identical images, as well as the attempts to find more examples in the skies. Joe Polchinski explains why the emission of gravitational waves is an interesting signal - a signal that can be seen by LIGO and/or VIRGO and/or LISA, as Damour and Vilenkin pointed out. Only the last paragraph of that article in New Scientist is dedicated to the fact that we still have no idea whether these cosmic strings are generic field-theoretical cosmic strings, or fundamental strings (or D-strings) from string theory.

Finally, let me say that the article looks much more serious and potentially exciting than what I expected. And don't get me wrong: the observation of cosmic strings, even if they're not the known strings from string theory, is a Nobel-prize-scale insight, I think.

Some trivial quantitative statements

The gravitational field of a cosmic string in 3+1 dimensions is simple - the spacetime is flat everywhere except for the locus of the string, but it has a deficit angle, much like a usual cone (multiplied by two more flat dimensions along the stringy worldsheet). The deficit angle is something like "8.pi.G_{Newton}.T" where T is the tension of the string - roughly the energy density per unit length of the string. The gravitational lensing by cosmic strings is special - the cosmic strings create two identical images, and they are visually separated roughly by the deficit angle itself. There are various upper bounds and lower bounds what the tension of the cosmic strings should be - i.e. how large the deficit angle of potential realistic cosmic strings can be if the cosmic strings exist. These upper and lower bounds marginally contradict each other (the uncertainty seems to be large enough so that the picture may still be consistent), but all these bounds are close to the deficit angle slightly below 10^-6. Because the tension has units of "squared mass" and because the deficit angle is roughly the tension in Planck units as we've said, we see that the mass scale associated with the string tension is roughly 10^-3 of the Planck scale - which is nothing else than the GUT scale. This fits the estimate for the tension that you expect from some GUT cosmic strings or the fundamental strings in string theory themselves - in the nearly old-fashioned models where the string scale is close to the GUT scale.

Let me also say some elementary comments about the popularity of cosmic string models that decreased rapidly around 2000 (and may be revived now): the cosmic strings have also been proposed as alternatives against inflation to explain structure formation. And because their equation of state is roughly "pressure equals -1/3 of energy density" in average, you may imagine that they cause the acceleration of the Universe instead of the cosmological constant. Both of these applications now seem highly unlikely with WMAP: WMAP says "pressure is below -0.78 of energy density at 95% confidence level". Also, the inflationary predictions of non-uniformities of the cosmic microwave background (CMB) agree with WMAP while the cosmic-string-dominated models are more or less falsified - the latter would lead to a much smoother profile of the temperature variations.

One more trivial calculation of mine: if they claim that the length of the stringy loop lensing CSL-1 is about 10^14 meters (10^49 Planck lengths) and the tension is 10^-6 squared Planck masses, the total mass is comparable to 10^43 Planck masses which is roughly 10^35 kilograms. So this loop of string would have something like 50,000 solar masses, unless I made an error. That would be a huge chunk of mass.

In this particular case, I decided not to include any links to the articles inspired by this topic on other blogs simply because their quality does not seem sufficient for you to learn something new out of them - their authors have not tried to learn the basic questions about the phenomenon of gravitational lensing.


  1. Indeed, as you mention, Lubos: The question is not so much if there are vortex defects out there, but if they are fundamental strings. I guess no one with an elementary knowledge of gauge theory would doubt that there are non-trivial cycles in the vacuum manifold of spontaneously broken gauge theories. Clearly some solutions are going to wrap such cycles. Polchinski et. al. spend most of there efforts on finding a good mechanism to distiguish such gauge theory strings from the fundamental ones. The "whip mechanism" is kind of cool: A little loop travels down the string at 0.9999999c and emits very characteristic radiation -- something gauge strings supposedly can't do. Also the combinatorics and rate of reconnection/breaking should be markedly different, but this is highly model dependent.

    I really like Polchinski's et. al. ideas and approach. The one thing that I can't get my head around is the following. They argue that the monopoles don't mess the arguments up, because their number scales less than exponential. However, unless I am missing something (and I assume I do), when I count I find that the monopole scale over-exponentially, something like a nasty factorial. If I'm right their argument is ready for the trash bin. I wish I had a chance to ask Joe. Next best thing: Ask Lubos. ;)) Do you happen to know?

    All the best,

  2. Hi Mark! Excellent comments. Unfortunately, I am not Joe - but my guess is that you should not be afraid to send a mail to Joe, Joe Polchinski has e-mail If there's a factorial monopole problem, he should know - and if there's no one, you will probably be told very quickly what you're doing wrong. ;-)


  3. (Sorry this is off-topic)
    I'd be interested in hearing the Morse theory explanation of why gravitational lensing produces an odd number of images. There's a very nice intuitive reason presented by Thorne and Blandford: (section 6.6)

    - Richard

  4. You know that saying that, "If a picture could paint a thousand words," then why can't I paint the universe:)

    Well imagery sometimes helps greatly to imprint a great consolidation of ideas into a hopefuly coherent nutshell. My view may seem all over the place, but hopefully I have succeded in dispelling that view? If not, please inject what is missing, if you have time.

    Warp Field Creates Gravitational Lensing

  5. Regarding the "cosmic string" Lubos meantioned. It merely shows how desperate super string theoriticians are in trying to find some observational evidence, any thing, that may remotely justify their stuff. When you have that kind of desperation, you tend to OVER-INTERPRET your data.

    Lubos said: "The team has observed a pair of galaxies 10 billion light years away and gravitational lensing is supposed to be the origin. The angular separation of the pair is roughly 2 arc-seconds."

    Note the two key numbers, 10 billion (10^10) light years distance and 2 arcseconds (9.7x10^-6 radian angle) angular separation. At that distance and that angular separation, if these are two galaxies their center barely separated by 9.7x10^4 light years.

    A typical galaxy like our galaxy, has a diameter of 2x10^5 light years. If what they observed are two images of a typical size galaxy, they barely separate half of their diameters.

    i.e, what they observe is instead one concrete image from the two half of the same galaxy, but be mis-interpretted as two images. It's that simple.

    The distance from the center of one half of the galaxy to the other half happen to be about 1x10^5 light years, which is the angular "separation" they reported.

    How desperate they have become? They reported that with an area of the sky merely 16 square arc seconds (4 arcseconds x 4 arcseconds) they found 11 pairs of such identical galaxy images.

    They must have counted each individual photons received as individual images :-)


  6. Dear Quantoken,

    you totally misunderstood the whole issue. Sazhin et al. are not string theorists. They are astronomers, and they think that they have observed something amazing (I repeat: amazing, not "desperate"), namely a huge cosmic string 100 astronomical units in size.

    Regardless of the type of that string, one would need an explanation if the observation is confirmed. String theory obviously offers several explanations - from the ordinary ones (cosmic strings from spontaneously broken gauge group) to the super-exciting ones, namely the fundamental superstrings grown to a macroscopic size.

    To some extent, various particular stringy models PREDICT(ed) the existence of cosmic strings created after inflation, and this could very well be an experimental confirmation of these predictions. Polchinski estimate the a priori probability that such cosmic superstrings may exist to be 10 percent.

    Yes, if you open the papers by Sazhin et al., you will see the specific pictures of the pair of galaxies - they are like two circles in the digit "8" and almost touch each other. You can still distinguish that these are 2 objects. The theory that there is one image only - without lensing - is safely ruled out.

    If you looked at the papers with the explicit pictures, you could have avoided writing complete stupidities. You know, I am putting the links to the papers at because I expect the average reader to open these papers. By this sentence I want to say that you are much much worse than what I expect to be the average reader.


  7. This comment has been removed by a blog administrator.

  8. This comment has been removed by a blog administrator.

  9. This comment has been removed by a blog administrator.

  10. Thanks for all of your comments Lumo. I had a chance to read the last of them this morning. Great comments. I should have known that you a personal connection with Checkoslovakia. The average person didn't know that stuff. I will be checking back often. I hope you visit my sight again in the future. Thanks again.

  11. Hey Lubos. Thanks for drawing this to our attention.

    The best competing explanation seems to be that it is an isolated Schwartzschild black hole or neutron star? Note that a perfect Schwartzchild metric will produce just two images (see, e.g., ). However, two arcseconds seems like a very large distance for a neutron star (which should give a splitting on the order of mas at 10 kpc.)

    Now, it could be a black hole; for two arcseconds, it would have to be of mass 10^6 M_sol or so, i.e., to be one of those intermediate mass black holes that everyone is talking about. Of course, you could move the NS/BH closer, but I don't think it is possible to get an isolated neutron star that close to us (it would have to be right on top of us to have an effect.

    The excess # of lenses nearby seems interesting.

  12. I should add that the non-distorted isophotes, and the comparable image intensities argue against a Schwartzschild lens; if the two images are close in intensity, their distortion should be noticeable.

    Very curious!

  13. Thanks for the link. I am going to add your site to my blogroll the next time I change my template. I try to make all of my changes at one time because I have the unique ability to crash my site almost every time I play with it.

  14. Sorry for the unnecessary comment. Is it Lumo or Lubos. I just added you to my blogrolls.

  15. I just read an article in the LA Times where the String Theory was called fanciful. I know it is a theory. But, is it fanciful. To me the tooth fairy is fanciful. Doesn't the String Theory have more validity than the tooth fairy.

  16. I've reviewed the articles by the Russian astronomers and find the evidence that the CSL-1 object is a result of a gravitational lens (much less a cosmic string) to be far from convincing. Much more likely, given the lack of distortion and lack of an obvious lens, is the possibility that these are two separate galaxies with similar brightness and similar spectra.

    The Russian group claims these objects are "identical", but the low angular resolution of their images and the fairly noisy spectra do not establish that. It only establishes them as very similar. And it is not at all unusual for similar galaxies to be found in such close apparent proximity.

    The Russians do admit that they have not conclusively demonstrated this is a gravitational lens:
    "Whether the two components of CSL-1 are two identi-
    cal ellipticals which happen to lay almost along the same line of sight, or rather two images of the same back-ground object produced by a gravitational lens is much more difficult to disentangle."
    And later: "... on the basis of present observational evidence we cannot yet rule out the only other (also unlikely) possible explanation in terms of projection effects ....

    Unfortunately, they do not explain why they consider "projection effects" to be an unlikely explanation.

    As for the other "lens candidates" in the immediate area, I can't say much, since very little evidence is presented in the papers you linked to. But they are clearly still merely "candidates," and so nothing has been verified one way or the other.

    Since there may well not even be a lensing phenomenon at all with this object, the evidence for any cosmic string lensing seems extremely thin at the moment. Fortunately, further, more detailed, observations should be able to definitively answer this question.

  17. the strings there are still in universe,or just there are "ghost actions" in our spacetime frame,"today". the that
    received are informations earliest past informations.but these vibrations are clusters of all the connected and compacted dimensions .

  18. in smallest bubbles.the spacetime are only one frequency of one vibrating string and having its infinity harmonics to each primary vibration-the interactions between those multiples strings create the extradimensions and the interne shapes of each dimension.the extradimensions with its supersymetrics energies define the freedom degrees of each dimensions,and what was the degrees of energy in the
    initial of the cosmo

  19. for what reasons the photons suffer not interactions with the higg'fields,or better why the photon appear has mass equal to zero,if its has energy? or its appear not has mass to justify the equations of STR?