Friday Lunchtime Talks - MOdified Gravity

On Friday we had the latest of the Durham Astronomy groups Friday Lunchtime Talks, these usually consist of two members of the group (which consists of over 50 members including students) giving a half hour talk, either about their own work or a paper that has appeared recently in which they have some interest. Last Fridays was something different however, it was an hour long talk by Professor Emeritus John Moffat of the Perimeter Institute on the subject of his own version of modified gravity, MOG.

I always have a lot of respect to someone that comes to Durham to talk about modified gravity, it really can be like entering the Lions den. Tinkering with gravity is usually done in an attempt to explain astrophysical phenomena without the need to demand that most of the mass in the Universe is invisible and can only be detected through its gravitational influence (the so called Dark Matter). Durham is world renowned as a centre of research into Dark Matter, as such some members of the group could be expected to be particularly partisan on the issue. I think its a great credit to the people involved that things never seem to get too heated, in fact in this Fridays talk, most of the difficult questions were directed from the one Professor in the group who has never been too happy with the current Cosmological Paradigm. He's generally good value for entertainment, and important in keeping everything honest.

Anyway, onto the talk itself. John Moffat has been working on modifying the gravitational laws for many years and has produced several different models. These models differ from most of the work on MOND (MOdified Netwonian Dynamics) in that from the beginning they were specifically chosen to be relativistic, that is that they are modifications to Einsteins General Relativity and not Newtons law of gravity. Essentially in MOdified Gravity (or if you prefer, MOffat Gravity, MOG) the gravitational constant G (also called Newtons Constant) is not in fact a constant, but can vary both in time and in space, so that the G that relates the force between two masses separated by one distance R1 is not the same G that relates the force between the same masses if they are separated by distance R2, similarly for time T1 and T2. These changes in G have to be negligible for objects within the solar system otherwise we should be able to detect deviations from the GR predictions which so far have not been observed, but they can become significant over galactic scales.

The speaker showed several examples of where his proposed changes would allow us to fit astronomical observations without the need to invoke Dark Matter. In particular he mentioned fitting the rotation curves of dwarf galaxies, spiral galaxies and clusters of galaxies (for clusters its technically not a rotation curve but the principle is the same). This in particular interested me as rotation curves are something I have personal experience with, both in my MSci project and the first paper I have published. Below is an example of a spiral galaxy rotation curve, it simply measures the speed at which stars at different radii in the disk orbit the galaxy, the data is the black circles (with error bars) the various dashed lines show the amount of velocity provided by the mass of various components of the galaxy including a DM halo. MOG allows you to explain the observed rotation of the galaxy without this DM component, by assuming the influence of the other two components is stronger than you would naively expect using Newtonian (or GR) gravity.

For me the first and largest problem that appeared during the talk appeared when JM was talking about these fits to the rotation curves. It seems it is possible using his model to fit all spiral galaxies using the same values for two parameters, as far as I could tell these parameters deal with the scale over which G begins to diverge from a constant value and some sort of normalisation of the size of the divergence. The problem was that when he fit the rotation curves of dwarf galaxies, or of clusters of galaxies the values of these two free parameters were different, all dwarfs had the same values, all clusters had the same values but different from the dwarfs or the spirals. This would appear to mean that there has to be yet another effect going on, meaning we need more free parameters to explain everything in terms of one unified gravitational theory. If this is true it would seem to be a problem. I have to go and look at the actual papers and see if in fact this is the case, or if I simply misunderstood something, it could be that the values he was quoting were actually telling you something about the scales involved, i.e that dwarf galaxies are smaller than spirals which are smaller than clusters.

The second major part of the talk dealt with JM's attempts to explain the results from the Bullet Cluster without the need for Dark Matter. This cluster is so far unique and very exiting. Its unique because it is actually two clusters, one of which has passed through the other, during this interaction the hot gas that resides in the clusters hit each other and slowed down, the gas from the smaller cluster having shocked and formed a bullet like shape as seen as the red triangle on the right of the image below. In the image below you can see the hot intracluster gas as the red regions. This is interesting because the hot intracluster gas has been separated from the galaxies of the clusters (seen in the blue regions), normally there is around 10x as much mass in this hot gas as there is in the galaxies of a cluster. By crashing through each other the two clusters have managed to separate the collisional material (the gas) from the collision less material (the galaxies and any Dark Matter).


It is possible to use gravitational lensing of background galaxies by the clusters of galaxies to work out the mass that must be contained within the galaxy clusters. This cluster is so important because it can be convincingly shown that even though the vast majority of the visible mass (the intracluster gas) has been removed the effect of gravity is still very strong, considerably stronger than can possibly be explained by the visible mass contained in the galaxies in the cluster. The traditional explanation is therefore that there must be some invisible mass (DM), which from this cluster we can see must be collision less, otherwise it would have piled up where the intracluster gas is. In MOG the explanation for the extra lensing, above what is predicted by GR for just the normal mass is that the gravity from the galaxies is stronger at larger distances, and also I believe that there is some lensing caused by the gravity from the intracluster gas.

If this is the case then there is happily a way to test which approach is correct, MOG or DM, the intracluster gas is much more massive than the mass contained in the luminous matter of the galaxies, but it is also located on one side of the cluster. Therefore it would seem that if MOG is correct there should be an increase in lensing on the side of the cluster nearest to the gas, of course the size of the effect would depend on many factors, but may in principle be measurable. If MOG is incorrect and DM is really at work, then the lensing should be more symmetric around the cluster, both because the DM is by far the largest mass contribution but also because the gravity of the gas is much less at larger distances than predicted by MOG.

Whatever the result of work such as this, we are really going to learn something fundamental about the Universe. Either most of the mass of the Universe is in some really exotic form, or else the force of gravity is even stranger than we have ever dreamt.

Unfortunately time caught up with us and we didn't get to find out much on the implications of this modified gravity on Cosmology, for example whether it could explain the mysterious Dark Energy at the same time that does away with Dark Matter. Many other interesting questions were raised however, it appears that MOG doesn't allow for singularities, so no black holes, though objects observationally indistinguishable from them probably can exist. This led one prominent member of the group to spend the time to see if the Metric for MOG is compatible with these condensed objects having a "last stable orbit", apparently it doesn't seem like it can, and as these are thought to be observed around BHs this is probably a problem for the theory.

All in all a very interesting talk, a perfect example of the kind of research that is ongoing in Astronomy. I don't think anyone is going to be packing in the DM work anytime soon, but if the DM particles continue to remain so illusive, its good to see that there are concrete alternatives being formulated.

Hubble Heritage

Click for VERY Big.

Thanks to the BadAstronomer I have found out about a cool new HST image from the Hubble Heritage site, what's even better is that two of the people that worked on the project to get the data work along the corridor, hence I am actually better informed than the BadAstronomer for the first time. The picture released by Hubble Heritage shows a cluster of Galaxies called Abell S0740. There are two versions of the image one without annotations and one showing zoomed regions of interest. See them both here.

This data is of particular interest because of what was found in the inner regions of the large elliptical galaxy at the centre of the cluster, when examined carefully it was discovered (by Russell Smith, here's the paper to prove it) that there were 3 gravitational arcs within the galaxy, you can see them in the inset in the top right of the image above. These arcs are the result of the light from a background galaxy being bent by the gravity of the large foreground elliptical galaxy, this has the effect of making several images of the same galaxy which are brighter and larger than they would usually be, making it possible to see galaxies further away. The gravitational arcs in this galaxy are the closest known example of strong lensing by an individual galaxy, as opposed to lensing by a cluster of galaxies.

The reason these lenses are so interesting is that when the position and brightness of the lenses are combined with the distance to the actual background galaxy (which can be worked out from its redshift) it becomes possible to measure precisely the amount of mass in the lensing galaxy within the radius of the arcs, with little or no assumptions. By looking at the way brightness of the lensing galaxy it is then possible (assuming that brightness is related to the amount of stars) to work out fraction of the mass in the lensing galaxy that is from normal matter and the fraction from dark matter. But wait that's not all.

By using the mass determined with the observed brightness of the lensing galaxy and the measured motion of stars in the galaxy (measured from spectroscopy) it becomes possible to work out exactly how the stars in the galaxy move as well, allowing a determination of parameters that can be used to constrain models of how galaxies form. Essentially these lenses could allow determinations of all the physical parameters of the lensing galaxy at accuracies previously unheard of.

Great you say, lets do it, but this is where things get difficult, it turns out that the arcs are so narrow that with current technology is actually impossible to get enough signal from them to determine a redshift, bugger. Basically they are so narrow that you add much more lensing galaxy light then you do lensed galaxy light. If STIS on the HST is ever repaired it may be possible, but if not it could be 10 years before its feasible. Note: if you happen to be someone sitting on a TAC (telescope allocation committee) and I, R. Smith or J. Lucey have a proposal to do what I've just claimed is impossible, I was wrong, we can do it, please give us the time.

Another thing I like about this work is shown in the other inset, a whole bunch of globular clusters. I study globulars, and when observing them from the ground it becomes very difficult to observe them at distances probably around 30-50Mpc away, these ones are located at 142Mpc, makes me even more annoyed that ACS has broken down, there will be no more new data like this for at least two years.

Fun with Gravity

What is gravity? One of the fundamental questions of physics, a force which affects our every move. Any successful theory seeking to explain the Universe must be able to explain gravity. To date there have been two successful theories of gravity (or gravitation).

The first, Newtons theory of gravity postulates that mass gives off a force that attracts other mass, this force decreases in strength as the inverse of the square of the distance between the two masses, such that

F = G*m1*m2 / r^2

This theory was incredibly successful in predicting the motions of planets, moons and other heavenly bodies, but it lacks a cause, no particle has been found that carries the gravitational force, a problem that Newton himself worried about, never quite being able to reconcile himself to this "action at a distance".

The second theory Einsteins General Relativityhas superseded Newton's theory for an important reason, it better reproduces the observations, especially in those regimes where gravity is strong. Einstein explained gravity as being due to mass curving space and time, so gravity is not a force pulling on a mass, it is a result of mass curving space/time such that what appears to be a straight path is actually curved, the moon follows a straight path but to an observer because of the curvature of space/time it appears to orbit the Earth. Of course this merely shifts the question, its now no longer a question of how gravity transmits force from one mass to another, but how does mass itself influence space/time? It is an active field of physics at present to try to answer this question.

The ADherents to Autodynamics believe that they have come up with a theory that explains the inverse square law of gravity naturally and whats more provides a physical explanation for it. This explanation is actually just a re-tread of a theory that was popular until the turn of the 19th century, the Le Sage theory of gravitation. I'll let the Autodynamics people to explain the theory (in which they use the term pico-graviton instead of Le Sage ultramundane corpuscle):

The Pico-Graviton (PG) is very similar to the concept called the "graviton". The concept of gravitons is that gravity is created by the "shadowing" affect. In affect, celestial bodies partially block the the flux of pico gravitons causing a low pressure area between the bodies. This pushes the bodies together.

The main difference between the pico-graviton and gravitons, is that pico-gravitons are somehow absorbed by mass. Most pico-gravitons push on the mass, but a very very tiny fraction are absorbed.

Now, there are many problems with this theory, as listed on this wikipedia page, go ahead its an interesting read. For now I think what would be more interesting that picking the overall theory apart (as the wiki article effectively does) would be to examine the logical inconsistencies in Autodynamics through this theory of gravitation. On this page the Autodynamics people explain their investigations of their version of the le Sage theory:

Using calculations for perihelion advance, he came up with the following numbers: the pico graviton has the mass of around 1x10-81 kilograms, and travels at around 27 times the speed of light. Another person in the SAA has claimed to find approximately the same values for the properties of the pico-graviton through completely different methods but the calculations have not be confirmed.

So the pico-graviton to explain the observations, must be both massive (as in it has some non-zero mass) and travel at superluminal velocities. Now this is a problem, even for Autodynamics, never mind the real world, on this page when talking about faster than light travel they state:

Autodynamics says that if E=mc^2 is correct, than no mass can accelerate beyond the speed of light. Why? Because in Autodynamics, mass moves by expending energy to push itself forward. According to current equations, all mass, no matter what the size could only result in a single photon going at the speed of light.

So far so good, they correctly point out that if E=mc^2 is true, then faster than light travel is impossible for massive objects even in Autodynamics. So for the pico-graviton to be real, and for it to match the observations, E=mc^2 must be false. Now lets just ignore for the moment the mountains of experimental data that show E=mc^2 to be true, lets just look deeper at the Autodynamics for a second. Specifically here. At the bottom of the page you will notice that the very same E=mc^2 has been used to derive the Autodynamics kinetic energy equations. Those very same kinetic energy equations that are supposed to explain amongst other things: muon decay, Compton effect, and nucleus-nucleus interactions.

So the question is, is E=mc^2 correct and hence the pico-graviton non-existent, or is E=mc^2 false and all of the AD "solutions" to atomic processes incorrect? A perfect example of why in Physics its usually impossible to change one "small" thing, everything is connected tweak one parameter to better fit some observations and you may well make the fit to other observations much worse. There are many other objections to the whole pico-graviton idea (especially the superluminal part), which I may get round to posting if I can sort out the maths.