Multi-wavelength observations of NGC 4666
One of my "pet" objects, i.e. one that is suitable for the kind of research that I pursue, is NGC 4666, a galaxy viewed by us almost edge-on. It forms stars at a very high rate, which is the reason for its classification as a "starburst galaxy". Taking a special interest in starburst galaxies, in particular in their interstellar medium, I have been involved in many observations of this object that are available to me to present a case here for the use of many datasets to gain a deep knowledge of the workings of a single object.
First I will list which observations are available (i.e., have been obtained by us or retrieved from publicly accessible data archives), before describing which new information we gained from these.
We started off with one essential piece of information:
- IRAS far-infrared total flux measurements
The 60 micron and 100 micron total fluxes of NGC 4666 are both high, as is the 60-to-100 micron flux ratio, of 0.45, indicating the presence of large amounts of warm dust. In the absence of an active galactic nucleus (AGN), this implied that NGC 4666 produces a lot of massive stars that heat the dust. Galaxies with lots of massive stars are the ones most likely to have gaseous halos. Since we were trying to find galaxies with halos, we applied for observations using this selection criterion and subsequently obtained the following data (sorted by ascending photon energy; radio data first, X-ray last).
- HI line imaging spectroscopy
- 1.4 GHz radio continuum image
- 4.9 GHz radio continuum image
- 1.4 and 4.9 GHz radio continuum polarisation imaging
- CO line imaging spectroscopy
- CO line ratios (see text below)
- Hα emission line image
- Optical R-band image
- Optical image from the Digital Sky Survey (DSS; external link)
- Optical spectroscopy (see text below)
- UV image
- X-ray imagery (see also others in the picture gallery)
- X-ray spectroscopy (see text below)
Also available from the online 2MASS data archive:
Together, this adds up to quite a considerable database. In most cases one does not have this many complementary pieces of information.
Now let us consider what these observations have told us about NGC 4666 that was previously not known (or at least more uncertain because of inferior quality of earlier observations). Because our primary research goal was a study of the gas halo of NGC 4666, a lot of what follows is directly related to the nature of gaseous halos of spiral galaxies (especially the descriptions of the various phases of the interstellar medium given there might help).
Note that NGC 4666 is at a distance of roughly 26 Megaparsec from us (almost 80 million lightyears). This is, even with current astronomical technology, quite far away for detailed investigations of its properties. What this means is that each of the observations listed above has taken up a significant amount of time - a huge investment into the study of a single object; let us try to judge at the end whether it was worth it...
HI line data
From our HI line observations we learned that, unexpectedly, NGC 4666 is interacting with at least one partner galaxy, NGC 4668. However, our data also show that only the gas outside the optically visible galaxy is affected by the resulting gravitational disturbance and that thus our assumption is correct that gas motions and properties in and near the galaxy disk itself are unaffected by external influences.
The HI data also give essential information about the internal gas kinematics, out to far beyond the optically visible disk. Knowledge of this is important in determining what has initially caused the starburst phenomenon, which is not found in all galaxies. In the case of NGC 4666 the most likely cause for the onset of a starburst is a disturbance of gas by a far-field interaction with NGC 4668, which led to gas accretion in some parts of the disk with subsequent gas cloud collapse, thus forming a large number of massive stars.
The HI observations provide no evidence for the existence of an HI gas halo around NGC 4666, which is what we were originally searching for. All visible features beyond the optically visible galaxy can be explained as either outer spiral arms or tidal tails.
Radio continuum imaging
The two total intensity images of NGC 4666 show clearly the presence of lots of supernovae (which produce large amounts of synchrotron radio continuum emission) and thus active star formation, with lots of massive stars (because only these end up as supernovae). Radio continuum imaging is therefore a good tracer of the energy input into the interstellar medium of a galaxy by supernovae. Furthermore, our data prove beyond any doubt that there is emission not only from the galaxy disk of NGC 4666, but also from a halo around it.
There are several interesting things about such a radio halo. The particles emitting the observed radiation originally come from the galaxy disk; they must therefore travel long distances (several kiloparsecs [kpc], a kpc corresponding to about 3000 light years) to reach their current locations.
This implies that one would expect a dependence between the level and distribution of massive star formation in the galaxy disk and the existence and extent of the radio halo. And indeed such a dependence is found! Note how the halo is less extended than the disk, because in its outer part the disk does not produce enough energy to expel gaseous particles to form a halo.
And there is more to come. Where there is synchrotron radio continuum emission, there must also be a magnetic field (B-field). If not, no radiation would be emitted by the cosmic ray particles (relativistic electrons).
Having two radio continuum images at different wavelengths we can also produce a spectral index image, which is the equivalent to an optical "colour image" (cf. the page on optical astronomy). While an optical colour image contains information on the stellar population of a galaxy, a radio spectral index image provides information on the properties of the cosmic ray particles emitting the synchrotron radiation, such as their mean lifetime, processes leading to energy losses they undergo, etc.
Polarised radio continuum
From the radio total intensity images we know already that there are magnetic fields. Polarisation data also tell us which directions the magnetic fields have. In the case of NGC 4666 there is not only the usual plane-parallel magnetic field, but in addition a field component outwards into the halo, perpendicular to the disk plane, is found. This is unusual and also important, because all ionised gas can move fastest following magnetic field lines.
CO imaging spectroscopy
Interferometric imaging of the lowest rotational transition emission line of the CO molecule provides us with valuable information about both the distribution and the kinematics of molecular gas in NGC 4666.
Knowing the distribution of molecular gas is important, because molecular clouds are the birthsites of stars, meaning that by studying NGC 4666's molecular gas we can find clues about why there is an ongoing starburst.
CO data also yields the best kinematical information about the central part of the gaseous disk (where HI observations, see above, are for various reasons less suitable).
Our CO image of NGC 4666 exhibits a "hole" in the gas distribution (see arrow). Using the kinematical information contained in the dataset, we could verify that this hole is probably an expanding gas shell. Its size and expansion velocity give us a measure of the local energy input into the gas by multiple supernova explosions. In the area concerned there must have been about 10000 supernova explosions to drive such a shell.
CO line ratios
Depending on the amount of energy absorbed by molecular gas, different numbers of individual molecules are found in various allowed energy states. There are thus several allowed energy states also for CO molecules. We have made use of this fact and observed CO line emission arising from various transitions to deduce from these the level of energy input into the gas. In NGC 4666 widespread star formation (i.e. a high level of energy input) over large parts of the disk causes the gas to have a relatively high mean temperature, with little variation over several kpc radius (which is unusual). This way one can trace the effects of massive star formation on its ambient medium (from which the stars had initially formed). Such investigations of the feedback of star formation processes into the interstellar medium are important to assess whether a galaxy is going to continue forming (massive) stars at a high rate in the future.
As mentioned in the section on infrared astronomy, near-infrared images give a representation of the distribution of relatively old stars in galaxies and are a good tracer of the galaxies' gravitational potential; the one of NGC 4666 shows that the stellar distribution is apparently undisturbed, as opposed to the outermost HI gas (see above). In this waveband the influence of dust is basically negligible. This is the reason why the stellar distribution looks so homogeneous. Colour differences indicate the presence of stars with different temperatures.
Hα line imaging
Hα imaging is one of the best tools to study the distribution of warm ionised gas in galaxies in both HII regions (star formation regions in which hydrogen exists in its ionised form) and in the diffuse medium outside HII regions. In NGC 4666, there is obviously an enormous number of bright HII regions in the disk. Whisps of very faint emission are also detected in the halo. These delineate the boundaries of ionised halo gas, as we will see below when discussing the distribution of hot ionised gas in NGC 4666 based on X-ray imagery. They have their footpoints in the brightest part of the disk.
Optical R-band image
An image with an optical R-band filter samples relatively cool, mostly low-mass stars. It therefore shows the distribution of stars, similar to the near-infrared image discussed above. However, in the optical dust absorption in the galaxy disk plays a much more dominant role, chaning the galaxy's appearance; a cut along NGC 4666's major axis shows that different stellar populations exist: One in the central, most actively star-forming part of the disk, one out to the boundary of the visible disk in the R-band image, and a third one that becomes apparent only after averaging many datapoints, indicating the presence of stars also in the areas of tidally disturbed HI gas (see above).
The hydrogen Hα recombination line mentioned above is only one out of a large number of emission lines in the optical regime, many of which can tell us a lot about the state of the warm ionised gas emitting them. In the case of NGC 4666 we used optical spectra to determine the excitation conditions of gas in disk and halo. A comparison with models predicting absolute and relative line strengths for different excitation conditions, it was found that the spectra of HII regions in the disk are consistent with pure photo-ionisation (by UV photons from massive stars). In the halo there is clear evidence for a second, different heating source, most probably shocks. This is an important result, because it implies that, under the conditions found in NGC 4666 (in particular its massive starburst), stellar UV continuum is not the only source of energy input heating the warm ionised gas. In other galaxies, such as e.g. NGC 2188 (with a lower star formation rate, but a halo of warm ionised gas nevertheless), pure photo-ionisation can explain all optical line ratios observed in the disk and the halo.
UV images trace directly the distribution of massive stars; this usually matches the distribution of star formation regions traced by emission such as radio continuum or optical Hα (also FIR dust continuum; which is not available here with sufficient angular resolution), because that is where the massive stars reside. The measured UV flux, together with a model of how much UV radiation is absorbed within the galaxy, can tell us how many massive stars a galaxy contains and how much energy they deposit into the interstellar medium.
X-ray imaging observations show the distribution of hot ionised gas and of compact X-ray sources in external galaxies (mostly high-mass X-ray binaries). In NGC 4666 hot gas, with temperatures well above a million Kelvin, was found both in the disk and in its halo. The halo emission is contained within a volume delineated by the faint Hα wisps described above. As described on the page about X-ray astronomy an absorption band caused by neutral gas in the disk of NGC 4666 is seen to cut through the diffuse emission distribution. In principle (although, due to a low signal-to-noise ratio, not in this case), such an absorption feature can also be used to determine some properties of the absorbing gas.
As in other wavebands, imaging normally provides mostly qualitative information, but one cannot derive many quantitative details describing the emitting gas. A much more powerful tool for in-depth investigations is spectroscopy. X-ray spectroscopy of NGC 4666 led to the measurement of several its hot ionised medium's key properties. In particular, the gas mass, temperature(s) and metallicity could be determined. Since this hot gas was produced by massive stars, it is enriched with elements created in the stellar fusion processes. It is expelled by the subsequent supernovae and their remnants and therefore plays a crucial role in the redistribution of heavy elements and thus the chemical evolution of galaxies over time.
Phew! I am sure that the above account of our findings is not even complete yet. But it should be good enough to answer the question: Was it worth it or not?! The observations described above have drawn a "panchromatic picture" of NGC 4666. From these we have learned to understand this particular galaxy, and also in general the whole class of similar objects, a lot better than we did before, by establishing general rules of behaviour. The importance of this field of research is to establish how gas enriched with heavy elements produced in massive stars can be redistributed through galaxies and eventually, if outflows occur at or above the systems' escape velocity, how such material is released into intergalactic space, thereby contributing to the chemical evolution of the Universe as a whole.
Most datasets used in this process of deduction provide complementary bits of information, but there are also some overlaps, which render possible consistency checks to ensure that our findings are not flawed by previously undetected instrumental effects. In this respect astronomy is like a lawsuit: Don't believe the evidence, if it is not independently corroborated!