LIGO/Virgo status & estimated detection rates
![Jonathan Thornburg [remove -animal to reply]'s profile photo](https://lh3.googleusercontent.com/a/default-user=s40-c)
Jonathan Thornburg [remove -animal to reply]
2009" meeting,
https://nrda2009.aei.mpg.de/
which brings together people working on numerical-relativity
simulations of gravitational-wave sources, and people working on data
analysis from the main gravitational-wave detectors. The conference
presentations are all online at
https://nrda2009.aei.mpg.de/program
Here I'd like to pass on to s.p.r some of the latest news from this
conference on the present and next-5-years prospects for directly
detecting gravitational waves. Ilya Mandel's talk at the meeting
https://nrda2009.aei.mpg.de/program/Mandel_Monday.pdf
is the source for considerable parts of what I'll say here.
[Caveat: My own expertise is primarily in numerical relativity, not in
the building and running of gravitational-wave detectors, so there may
be small technical errors in what I'm saying. But I'm fairly confident
that I have the main points correct...]
Background -- Gravitational Waves
=================================
Gravitational waves are "ripples in the fabric of spacetime". They're
a firm (i.e. no-adjustable-parameters) prediction of general relativity.
They have never been directly detected, but observations of the orbital
decay of various binary pulsars (most famously PSR B1913+16) give
strong indirect evidence that gravitational waves exist and have
roughly the properties predicted by general relativity. (Hulse and
Taylor shared the 1993 Nobel prize in physics for their work on the
PSR B1913+16 binary pulsar.)
General relativity predicts that any mass which is accelerated in a
non-spherically-symmetric manner emits gravitational waves, but in
amounts that turn out to be very tiny unless you're dealing with very
massive objects accelerating quite rapidly. In particular, the only
known sources of gravitational waves which are even close to being
detectable with any detector now built or planned for the next 5 years
are astrophysical -- notably colliding neutron stars or black holes.
Background -- Gravitational-Wave Detectors
==========================================
Today's major gravitational-wave detectors are basically Fabry-Perot
interferometers with long arms, arranged in an "L" shape. The basic
idea is that a passing gravitational wave will slightly change the
relative lengths of the two arms' Fabry-Perot cavities, which in turn
causes a slight change in the interference pattern.
Realistic astrophysical gravitational waves are *very* weak, inducing
arm-length changes on the order of 1 part in 10^21 (i.e., 1e-18 meters
for an arm length of 1km), so the detected signals are *very* small,
and being confused by ground vibrations ("seismic noise") is an
ever-present danger. To counter this, all the major interferometer
components (beamsplitters, mirrors, etc) are suspended on special
vibration-isolating mountings. Moreover, the data analysis is done
looking for events which occur simultaneously (up to light-travel-time)
at different detectors which are many thousands of km apart; this
latter technique is very effective at screening out local seismic &
other disturbances. (And, the light-travel time can be used to
estimate roughly where on the sky the signal is coming from.)
In practice, the sensitivity of such a detector is thus set by its
noise level -- gravitational waves sufficiently stronger than the
noise can be confidently detected (with an acceptably low false-alarm
probability), while too-weak signals can't be seen above the noise.
Much of the work in detector development goes into hardware/software
improvements to lower or better-quantify the noise level.
[In reality, both signals and noise are strongly
frequency-dependent, and there's lots of matched
filtering with expected-signal templates. And even
at a single frequency, the noise spectra aren't
Gaussian -- there are lots of "glitches" that have
to be filtered out (hopefully without discarding
valid astrophysical signals). But for the purposes
of what I'm discussing here, these complications
can mostly be ignored.]
There are a number of such detectors located around the world.
Notably:
* The US's LIGO has 3 interferometers, with 4km- and 2km-long-arms ones
co-located in Hanford, Washington, and another 4km one in Livingston,
Louisiana.
* The French-Italian Virgo interferometer has 3km arms, and is located
near Pisa, Italy.
* The UK-German GEO interferometer has 600 meter arms (but fancier
technology than the bigger interferometers in some other ways), and
is located near Hannover, Germany.
[There are also other smaller projects, notably in Japan
and Australia. The Japanese TAMA-300 project was the first
such large interferometer (it has 300 meter arms), but it's
in a relatively seismically-noisy location. That team is
now working on a next-generation detector where the mirrors
will be cryogenically cooled to reduce thermal noise.
The Australian team has a small prototype running, and is
trying to get funding for a full-sized detector.]
LIGO and Virgo have roughly comparable noise levels (sensitivities);
GEO is roughly a factor of 3-5 worse than LIGO/Virgo. All of these
projects now pool & jointly analyze their data.
These detectors typically alternate between "engineering time", when
the detector people upgrade/debug hardware & make test measurements
(often involving injecting simulated gravitational-wave signals) to
try to make the (very complicated) hardware & software work properly,
and "science runs", when they keep relatively stable hardware and
software configurations and actually try to observe gravitational
waves.
Past Observations and Expected Event Rates
==========================================
The best (most sensitive, i.e. lowest-noise-level) science runs so
far have involved LIGO and Virgo collecting data simultaneously for
about a year, in what's known as S5 (Science run #5). S5 marked the
first major LIGO collection of science data at roughly the original
LIGO design sensitivity (noise level); I think (although I'm less
certain) that this was also true for Virgo. LIGO's term for the
hardware configuration used for S5 is "initial LIGO".
Considering only known & relatively well-understood astrophysical
objects, the most likely things to be detected by LIGO/Virgo/GEO are
the final decay and coalescence of binary neutron star (NS-NS), black
hole (BH-BH) or "mixed" (NS-BH) systems. For a given NS-NS, BH-NS,
or BH-BH system, it's relatively easy to roughly estimate the emitted
gravitational waves
[To calculate them *accurately* has only become
possible since 2005, and remains the subject of
much ongoing research in numerical relativity.]
and thus the distance at which an interferometer with a given noise
level could detect the decay/coalescence signal (at a signal/noise
ratio high enough to be fairly confident that the detection was real).
Turning this calculation around, if we know the space density of
NS-NS, NS-BH, and BH-BH systems,
[This, in turn, can be estimated from an analysis of
how we think these systems form. Very roughly, massive
stars in binary systems live out their lives, explode
in supernovae, and leave behind NS and/or BH remnants.
So we "just" have to figure out how common such massive
binary star systems are, and what fractions of them
leave behind NS or BH remnants, and what fractions of
these binaries stay bound after supernova explosions.
Much of Mandel's talk was about the many complications
involved in model these processes more accurately.]
and the timescales on which these systems decay
[that turns out to be easy to calculate]
then we can fairly easily calculate how often one we can expect such
a decay/coalescence to happen close enough to the Earth for the
gravitational waves to be detected (at a signal/noise ratio high
enough to be fairly confident that the detection was real).
Page 13 of Mandel's slides is a nice table which summarizes the
results of such calculations for the baseline LIGO/Virgo level of
sensitivity (i.e. the hardware status as it was during S5):
Source expected detected-event rate (events/year)
lowest-plausible best-estimate highest-plausible
NS-NS 2e-4 0.02 0.2
NS-BH 9e-5 0.006 0.2
BH-BH 2e-4 0.009 0.7
As you can see, all of the event rates have very large uncertainties,
with a factor of 1000 or more between the lowest-plausible and highest-
-plausible event rates. As discussed in Mandel's talk, this mainly
reflects our rather poor knowledge of the massive-star astrophysics
which goes into forming NS-NS, NS-BH, and BH-BH systems, and thus of
the space density of such systems.
It's also clear that with the best-estimate event rates much less than
one event detected per year of running time, and the lowest-plausible
rates at or below 1 per 5000 years of running, it's not in any way
surprising that S5 (representing about a year of observing) didn't see
anything. (Previous runs used hardware configurations that were
significantly *less* sensitive than S5, so it's even more unsurprising
that they didn't see anything.)
Near-Future Observations and Expected Event Rates
=================================================
Since S5, LIGO and Virgo have both made hardware upgrades which have
cut their noise levels roughly in half. LIGO refers to this new hardware
configuration as "enhanced LIGO"; I think (but am less certain) that
Virgo calls their new configuration "Virgo+". In both cases, the
"hardware upgrades" are things like higher-power lasers and better
(very sophisticated) suspensions to isolate the interferometer
components from seismic noise.
Since these detectors detect gravitational-wave *amplitudes* (not
*energies*), and amplitudes fall off as 1/R (energies fall off as 1/R^2),
a factor of 2 reduction in the noise level means the detectors can now
detect any given event even if it's twice as distant as the previous
outer-limit-detection-distance. This in turn means that the detectors
are now sensitive to events in roughly 8 times the volume they were
before, so the expected event rates are roughly a factor of 8 larger.
Actually, the detector people at the meeting said that if you do the
numbers a bit more carefully, you come out with about a factor of 10
increase in expected event rates.
On 7 July 2009, LIGO, Virgo, and I think also GEO started their next
major science run, S6, using this new hardware. S6 is planned to last
for about a year. Multiplying the event rates I gave above by 10,
it's easy to see that it's now plausible that LIGO/Virgo *might*
detect gravitational waves (at a sufficiently high signal/noise ratio,
in coincidence betwen detectors at widely-separated locations, with
good data quality, etc etc) in S6, we certainly shouldn't be surprised
if they don't.
Farther-Future Observations and Expected Event Rates
====================================================
After S6, LIGO/Virgo plan to shut down for some more major hardware
upgrades. LIGO calls these "advanced LIGO"; I think (but am again
less certain) Virgo calls them "advanced Virgo". The detector people
expect these upgrades to take 2-3 years, i.e., until 2012-2013 or so,
and (once debugged) to yield noise levels roughly a factor of 10 below
the "initial" (S5) configuration. This should give event rates which
are roughly a factor of 1000 higher than the initial configurations.
Page 13 of Mandel's slides gives the results of more careful
calculations of the expected event rates for advanced LIGO/Virgo:
Source expected detected-event rate (events/year)
lowest-plausible best-estimate highest-plausible
NS-NS 0.4 40 400
NS-BH 0.2 10 300
BH-BH 0.5 20 1000
As you can see, these event rates are now (finally!) up into a range
where we can realistically expect to see signals in (much less than!)
a year's running. In fact, even the lowest-plausible event rates
add up to about 1 event/year.
Thus (given the large expected-event-rate error bars) we can say that
advanced LIGO/Virgo is fairly likely to detect gravitational waves
somewhere around 2013-2014, or perhaps 2015 if we allow an extra year
for hardware debugging.
[If advanced LIGO/Virgo reach their expected sensitivity levels, but
don't see any signals in a year or so's running, then I think there
would probably be a big push to further tweak the hardware to gain
another factor of 2 or so, to try to boost the expected event rates
by another factor of 10 or so. At that point (expected event rates
10 x advanced LIGO/Virgo), a continued failure to detect anything in
a year or so's running would definitely imply that some part(s) of
the "theory chain" I described above are wrong.]
--
-- "Jonathan Thornburg [remove -animal to reply]" <jth...@astro.indiana-zebra.edu>
Dept of Astronomy, Indiana University, Bloomington, Indiana, USA
"Washing one's hands of the conflict between the powerful and the
powerless means to side with the powerful, not to be neutral."
-- quote by Freire / poster by Oxfam
eric gisse
[...]
That was an excellent discussion of the subject.
People who harp on the non-detection of gravitational waves need to sit down
for a minute and think about the event frequencies and how small they are
even for best-case scenarios. Patience is the key here, not "we didn't see
anything immediately. Mothball the facilities!"
To add, the Einstein@home people have published the results of their search
through 66 days of S5 data. Unfortunately nothing was found.
clifford wright
Gravitational wave detectors at present. As an engineer however I have
also noted the problems with beam intensity effects etc that they have
already had, I wonder if they can push the sensitivity quite as high in
practice? Although I remain sceptical of the existence of such
"radiation" feeling that other localised effects could produce similar
observations, such as localised electromagnetic forces, I do feel that
the direction of future research should be aimed at space based
measurements. Basically we need wide band detectors, which as has been
shown are simply not practical on the Earth's surface.
In any event we are working "blind" since we still have no coherent
theory of gravitation. Therefore we need to look at such potential
forces over as wide a range as possible.
In any event I am hoping that something likely to detect events on a
reasonable time scale can get going before too long! I'm no longer a
young man and I would like to see some conclusion before I leave this
universe!
Clifford Wright.
================Moderator's note ===============================
I've cut the full quote and set your posting on top! That makes
it easier to read for everybody. HvH.
"Jonathan Thornburg [remove -animal to reply]"
<jth...@astro.indiana-zebra.edu> wrote in
news:alpine.BSO.1.10.0...@nitrogen.astro.indiana.edu:
Oliver Jennrich
> Many thanks for your very clear summary of the position with
> Gravitational wave detectors at present. As an engineer however I have
> also noted the problems with beam intensity effects etc that they have
> already had, I wonder if they can push the sensitivity quite as high in
> practice?
What do you mean with 'quite as high'? LIGO operates at the design
sensitivity (with a few exceptions here and there) so they already did
push the sensitivity to where it was expected to be.
Mind you, that alone is a phantastic achievement.
> Although I remain sceptical of the existence of such "radiation"
> feeling that other localised effects could produce similar
> observations, such as localised electromagnetic forces, I do feel that
> the direction of future research should be aimed at space based
> measurements. Basically we need wide band detectors, which as has been
> shown are simply not practical on the Earth's surface.
Let's take a look: LIGO has about 2 frequency decades sensitivity. In
the electromagnetic world, the equivalent would be a telescope from near
IR (1 micrometer) to soft x-ray (10 nm). I guess any optical
astronomer would call that 'wideband' (well, not quite, they would
probably call it 'overwhelmingly wideband').
LISA's bandwidth will be about 4-5 decades, that would be equivalent of
an optical instrument sensitive from near IR to gamma rays.
> In any event we are working "blind" since we still have no coherent
> theory of gravitation. Therefore we need to look at such potential
> forces over as wide a range as possible.
For all practical purposes of GW detection, GR is good enough.
> In any event I am hoping that something likely to detect events on a
> reasonable time scale can get going before too long! I'm no longer a
> young man and I would like to see some conclusion before I leave this
> universe!
Amen to that!
--
Space - The final frontier
eric gisse
> Many thanks for your very clear summary of the position with
> Gravitational wave detectors at present. As an engineer however I have
> also noted the problems with beam intensity effects etc that they have
> already had, I wonder if they can push the sensitivity quite as high in
> practice?
Ask the people who are doing it. They might know.
> Although I remain sceptical of the existence of such
> "radiation" feeling that other localised effects could produce similar
> observations
http://nobelprize.org/nobel_prizes/physics/laureates/1993/press.html
> , such as localised electromagnetic forces,
Electromagnetic waves are the result of a changing electric dipole
moment whereas gravitational waves are the result of a changing mass
quadrupole moment.
> I do feel that
> the direction of future research should be aimed at space based
> measurements.
LISA is coming along. And is sensitive to frequencies that LIGO, et.al
are not.
> Basically we need wide band detectors, which as has been
> shown are simply not practical on the Earth's surface.
Nothing of the sort has been shown.
>
> In any event we are working "blind" since we still have no coherent
> theory of gravitation. Therefore we need to look at such potential
> forces over as wide a range as possible.
Royal "we", I am assuming. We, as in "the scientific community" we, have
general relativity which have been shown to be quite capable.
[...]
Jonathan Thornburg [remove -animal to reply]
I outlined the expected event rates now and for the next few years
for the big ground-based gravitational-wave (GW) detectors LIGO, Virgo,
and GEO. There's another important point I should add:
The event rates I quoted are all (necessarily) based on *known* GW
sources, i.e., they're predictions of the rates at which already-known
astrophysical objects emit GWs strong enough to be detected.
The history of astrophysics has lots of examples of opening up
"new windows" of observation, e.g., infrared, ultraviolet, radio,
X-ray, gamma-rays, neutrinos, etc. In many of these cases, as well as
seeing the "known" sources, we also discovered completely unexpected
sources (think pulsars, gamma-ray bursts, hot cluster gas, ...).
This might happen again with GW astrophysics, and this possibility adds
to the interest in observing in this "new window".
[This possibility also argues in favor of data-analysis algorithms
(like the "burst searches" I described in another posting) which
presume as little as possible about the precise nature of the GWs.]
clifford wright
Followups should include something substantially new. -P.H.]
"Jonathan Thornburg [remove -animal to reply]"
<jth...@astro.indiana-zebra.edu> wrote in
news:alpine.BSO.1.10.0...@nitrogen.astro.indiana.edu:
> In article
> <alpine.BSO.1.10.0...@nitrogen.astro.indiana.edu> I
> outlined the expected event rates now and for the next few years for
> the big ground-based gravitational-wave (GW) detectors LIGO, Virgo,
> and GEO. There's another important point I should add:
>
> The event rates I quoted are all (necessarily) based on *known* GW
> sources, i.e., they're predictions of the rates at which already-known
> astrophysical objects emit GWs strong enough to be detected.
>
> The history of astrophysics has lots of examples of opening up
> "new windows" of observation, e.g., infrared, ultraviolet, radio,
> X-ray, gamma-rays, neutrinos, etc. In many of these cases, as well as
> seeing the "known" sources, we also discovered completely unexpected
> sources (think pulsars, gamma-ray bursts, hot cluster gas, ...).
> This might happen again with GW astrophysics, and this possibility
> adds to the interest in observing in this "new window".
>
> [This possibility also argues in favor of data-analysis algorithms
> (like the "burst searches" I described in another posting) which
> presume as little as possible about the precise nature of the GWs.]
>
Well many thanks to posters in this group for some very enlightening
information.
As an engineer (electronics) however and after having yet another look at
the LIGO site updates I still feel VERY strongly that here is a system
pushed beyond its practical limits of sensitivity and situated in a
far too "noisy" environment to be really useful as a means of giving some
definitive prooof of theory.
The space based system might well be the only way to test things, and at
great expense (can we raise that much interest?).
I was especially interested in the post which suugested that gravity
might not have all its effects in our particular universe.
It has always seemed to me that something very basic must explain the
enormous difference between electromagnetic and gravitational forces.
Of philosophical concern however are a couple of posts which say
something like "at colloquium X or seminar Y some authority said that Z
must be so" and using this statement as a sort of "proof" of the posters
position.
This is downright alarming!
One may have a lot of respect for expert opinion, but quoting "authority"
like this is more appropriate for a religious seminary than a serious
scientific newsgroup.
It is clear that gravity is still a very fluid and little understood
field, we are far too early to take doctrinaire positions.
The most important attribute is perhaps an open but critical mind>
Clifford Wright.
Chalky
<jth...@astro.indiana-zebra.edu> wrote:
> LISA will be sensitive
> to gravitational waves with frequencies in the milliHertz range
> (LIGO/Virgo/etal are sensitive between roughly 50 and 3000 Hertz).
On Jul 30, 8:59 am, "Jonathan Thornburg [remove -animal to reply]"
<jth...@astro.indiana-zebra.edu> wrote:
> LISA will be sensitive between roughly 0.3 and 30
> milliHertz.
On Jul 30, 8:59 am, Oliver Jennrich <oliver.jennr...@gmx.net> wrote:
> LISA will observe a different frequency range form LIGO (0.03 mHz... 1Hz
> vs 10Hz .. 1 kHz) and signals in that frequency range are very much
> stronger than in the 'acoustic' range.
On Jul 31, 2:06 pm, Oliver Jennrich <oliver.jennr...@gmx.net> wrote:
> LISA's bandwidth will be about 4-5 decades, that would be equivalent of
> an optical instrument sensitive from near IR to gamma rays.
Why such a large discrepancy between these two contributors, for the
expected bandwidth of LISA ?
Jonathan Thornburg [remove -animal to reply]
of LISA's (planned) sensitivity in various postins of mine and of Oliver
Jennrich:
> Why such a large discrepancy between these two contributors, for the
> expected bandwidth of LISA ?
I can't speak for Oliver, but in my case the cause was my imperfect
memory of the actual (planned) LISA frequency response. There's a nice
graph showing both the (planned) LISA and (actual) LIGO frequency
responses at
http://en.wikipedia.org/wiki/File:LIGO-LISA.jpg
http://upload.wikimedia.org/wikipedia/commons/e/eb/LIGO-LISA.jpg
Looking at this graph should clarify the situation a lot.
Roughly speaking, LISA's peak sensitivity is at 0.005 Hertz, and it's
within a factor of 10 of that peak sensitivity from 0.001 Hertz to 0.2
Hertz.
ciao,
Oliver Jennrich
> Chalky <chalk...@bleachboys.co.uk> asked about the differing portrayals
> of LISA's (planned) sensitivity in various postins of mine and of Oliver
> Jennrich:
>> Why such a large discrepancy between these two contributors, for the
>> expected bandwidth of LISA ?
>
> I can't speak for Oliver, but in my case the cause was my imperfect
> memory of the actual (planned) LISA frequency response. There's a nice
> graph showing both the (planned) LISA and (actual) LIGO frequency
> responses at
> http://en.wikipedia.org/wiki/File:LIGO-LISA.jpg
> http://upload.wikimedia.org/wikipedia/commons/e/eb/LIGO-LISA.jpg
> Looking at this graph should clarify the situation a lot.
>
Hm. I thought I wrote something, but it seems that it never made it into
the group. Anyway - I agree that the graphs are nice visualistion of
the sensitivities.