Cosmology Day July 2018

Lecture 3: The missing Galaxy problem

Dr Olivia Keenan


Dr Keenan completed her Astrophysics PhD at Cardiff University on the topic of extragalactic astronomy. She focused on dwarf galaxies in the context of the low surface brightness universe and also worked on the AGES project – using Arecibo radio telescope to explore nearby galactic environments. She is now the London and South East regional manager for the Institute of Physics where she focuses on using public engagement to take physics to underrepresented audiences in order to lower barriers to participation.


Lambda CDM (cold dark matter) is the leading cosmological model used to describe the universe. However, state of the art cosmological simulations based on this model produces a number of discrepancies with the real universe. One of these is known as the ‘Dwarf Galaxies problem’; since the models predict a factor of 10 more dwarf galaxies than are observed. We want to understand why this is; whether it is due to problems with observations, simulations or both. Dr Keenan focused on the former – are these missing galaxies out there but eluding us? One potential solution is that these galaxies could be ‘dark galaxies’, galaxies which are dominated by gas but contain few or no stars. There have been many dark galaxy candidates, however, so far none have been confirmed. Another possibility is that the galaxies are star dominated but still extremely faint and hard to detect in crowded, large field surveys. If so these galaxies could be detected using a combination of their physical properties to separate them from background field galaxies.

In the lecture, Dr Keenan discussed her work in this area, and how it has contributed to understanding the Dwarf Galaxies problem.

My notes from the lecture (if they don’t make sense then it is entirely my fault)

The dwarf galaxy problem, also known as the missing satellites problem, arises from numerical cosmological simulations that predict the evolution of the distribution of matter in the universe. Dark matter seems to cluster hierarchically and in ever increasing number counts for smaller-and-smaller-sized halos. However, although there seem to be enough observed normal-sized galaxies to account for this distribution, the number of dwarf galaxies is orders of magnitude lower than expected from simulation. For comparison, there were observed to be around 38 dwarf galaxies in the Local Group, and only around 11 orbiting the Milky Way, yet one dark matter simulation predicted around 500 Milky Way dwarf satellites.

There are two main alternatives to resolving this problem. One is that the smaller halos do exist but only a few of them end up becoming visible because they have not been able to attract enough baryonic matter to create a visible dwarf galaxy. In support of this, Keck observations in 2007 of eight newly discovered ultra-faint Milky Way dwarf satellites showed that six were around 99.9% dark matter (with a mass-to-light ratio of about 1000). Other solutions may be that dwarf galaxies tend to be merged into or tidally stripped apart by larger galaxies due to complex interactions. This tidal stripping has been part of the problem in identifying dwarf galaxies in the first place, which is an extremely difficult task since these objects have low surface brightness and are highly diffused, so much that they are virtually unnoticeable.

A galactic halo is an extended, roughly spherical component of a galaxy which extends beyond the main, visible component. Several distinct components of galaxies comprise the halo:

the stellar halo

the galactic corona (hot gas, i.e. a plasma)

the dark matter halo

A dark matter halo is a theoretical component of a galaxy that envelops the galactic disc and extends well beyond the edge of the visible galaxy. The halo’s mass dominates the total mass. Thought to consist of dark matter, halos have not been observed directly. Their existence is inferred through their effects on the motions of stars and gas in galaxies. Dark matter halos play a key role in current models of galaxy formation and evolution. The dark matter halo is not fully explained by the presence of massive compact halo objects (MACHOs).

Simulated dark matter halo from a cosmological N-body simulation


The problem arises from cosmology. The idea is to look at a real galaxy such as Andromeda, containing 34 dwarf galaxies, run a simulation of it and look for the differences.

The Andromeda Galaxy, also known as Messier 31, M31, or NGC 224, is a spiral galaxy approximately 780 kiloparsecs (2.5 million light-years) from Earth, and the nearest major galaxy to the Milky Way. Its name stems from the area of the sky in which it appears, the constellation of Andromeda.

All galaxies seem to live in dark matter halos and the simulations indicate there are at least 10 times more dwarf galaxies than those seen.


The above graph shows there are many more small dim galaxies than large bright ones.

The Schechter luminosity function provides a parametric description of the space density of galaxies as a function of their luminosity. The form of the function is


where L is galaxy luminosity, and L* is a characteristic galaxy luminosity where the power-law form of the function cuts off. The parameter f* has units of number density and provides the normalization.

According to this formula, the number of galaxies brighter than the luminosity L* drops very rapidly

The dwarf galaxies are extremely dark matter dominated. Missing dwarfs are there but are not detected.

Many simulations are ‘dark matter only’ and don’t take into account normal atomic matter (Baryons).

The lambda CDM model is flawed.

Virgo H121 is a possible dark galaxy located about 50 million light-years from Earth. It doesn’t have stars and can only be seen with radio waves.

Cardiff University located it by tuning in to its radio emissions. The emissions identified it as a swirling cloud of hydrogen gas containing 100 million times the mass of the Sun, making it larger than many dwarf visible galaxies. The speed that it rotates indicates that it is full of some other kind of material mass that emits nothing. At the moment we don’t know what dark matter is but it seems to make up most of what galaxies are made of. This galaxy has an even heavier dark matter than ordinary dark matter.

Some astronomers suggest that this galaxy is at the extreme end of galaxy evolution. Others think that dark galaxies form stars but very slowly. Because of its low density, Virgo H121 may never pull gas clouds together into stars.


The Triangulum Galaxy is a spiral galaxy approximately 3 million light-years (ly) from Earth in the constellation Triangulum. It is catalogued as Messier 33 or NGC 598. The Triangulum Galaxy is the third-largest member of the Local Group of galaxies, behind the Milky Way and the Andromeda Galaxy. It is one of the most distant permanent objects that can be viewed with the naked eye.

The galaxy is the smallest spiral galaxy in the Local Group and it is believed to be a satellite of the Andromeda Galaxy due to their interactions, velocities, and proximity to one another in the night sky. It also has an H-II nucleus.

An HI region or H-I region (read H one) is a cloud in the interstellar medium composed of neutral atomic hydrogen (HI), in addition to the local abundance of helium and other elements. (H is the chemical symbol for hydrogen, and “I” is the Roman numeral. It is customary in astronomy to use the Roman numeral I for neutral atoms, II for singly-ionised—HII is H+ in other sciences—III for doubly-ionised, e.g. OIII is O++, etc.) These regions do not emit detectable visible light (except in spectral lines from elements other than hydrogen) but are observed by the 21-cm (1,420 MHz) region spectral line. This line has a very low transition probability, so requires large amounts of hydrogen gas for it to be seen. At ionization fronts, where HI regions collide with expanding ionized gas (such as an H II region), the latter glows brighter than it otherwise would. The degree of ionization in an HI region is very small at around 10−4 (i.e. one particle in 10,000). At typical interstellar pressures in galaxies like the Milky Way, HI regions are most stable at temperatures of either below 100 K or above several thousand K; gas between these temperatures heats or cools very quickly to reach one of the stable temperature regimes. Within one of these phases, the gas is usually considered isothermal, except near an expanding H II region. Near an expanding H II region is a dense HI region, separated from the undisturbed HI region by a shock front and from the H-II region by an ionization front.

Mapping HI emissions with a radio telescope is a technique used for determining the structure of spiral galaxies. It is also used to map gravitational disruptions between galaxies. When two galaxies collide, the material is pulled out in strands, allowing astronomers to determine which way the galaxies are moving.

HI regions effectively absorb photons that are energetic enough to ionize hydrogen, which requires an energy of 13.6 electron volts. They are ubiquitous in the Milky Way galaxy, and the Lockman Hole is one of the few “windows” for clear observations of distant objects at extreme ultraviolet and soft x-ray wavelengths.

The ALFALFA research project is located at the Arecibo Observatory and started on February 4th 2005. It is still ongoing (April 2013). The name is the abbreviation of Arecibo Legacy Fast ALFA. ALFA is the abbreviation of Arecibo

L-Band Feed Array.

ALFALFA is designed to detect the cool (not hot; not cold) atomic gas in and near galaxies.

ALFALFA is a blind survey; we observe the whole area of sky, whether or not we think/know there is an optical galaxy there.

ALFALFA is a spectroscopic survey; not only do we detect the HI line flux, we also measure its frequency (velocity) and the width of the HI line (a measure of rotational velocity).

ALFALFA is a census of HI in the local universe

ALFALFA: Are there “dark galaxies”?

In agreement with previous results, ALFALFA finds that fewer than 2% of (clearly extragalactic; not ALFALFA UCHVCs) HI sources cannot be identified with an optical counterpart.

ALFALAFA was used to observe HI in M33

Previous observations of M33

Braun, R.; Thilker, D. A.…417..421B


The Arecibo radio telescope in Puerto Rico – a single dish radio telescope with a dish diameter of 305 m

As part of the AGES (Arecibo Galaxy Environment Survey) the Arecibo telescope, in Puerto Rico, was used to observe the neutral hydrogen gas in and around M33. THE data cube showing M33 and the surrounding area is shown in the following diagram. One aim of this work was to search for dark dwarf galaxies: those with hydrogen gas but no detected stars.


M33 data cube. The x and y-axes show two spatial dimensions (right ascension and declination) and the z-axis shows velocity. Gas from the Milky Way is shown in blue, M33 and associated clouds are shown in red with big features labelled. The red dappling shows the noise in the cube.

32 clouds in the area around M33 were detected, eleven of which were previously undetected. Twenty-two of these were discrete clouds meaning they are well separated from the gas disk of M33. The other 10 were all previously discovered clouds that were detected as over-densities in the extended gas disk of M33. You can distinguish this extended gas disk, as the AGES observations of M33 are more sensitive than those of previous surveys.

To ascertain whether the clouds found were dwarf galaxies or perhaps something else, their shapes, masses and sizes were analysed and compared, as were their rotational motion, as it is expected that a gravitationally bound galaxy would show ordered rotation. Stars associated with the clouds were looked for, however, none were found.

The Arecibo galaxy HI survey covers 200 degrees of sky

Numerical simulations:

Where Are the Missing Galactic Satellites?

Authors: Anatoly Klypin; Andrey V Kravtsov; Octavio Valenzuela; Francisco Prada

Atomic Hydrogen Gas in Dark Matter Minihalos and the Compact High-Velocity Clouds

Authors: Amiel Sternberg; Christopher F. McKee; Mark G. Wolfire

‘AGESM33-31’ is an interesting cloud. It is very large and if it is located at the same distance as M33, it is as large as M33 itself (almost 6000 light years across!). AGESM33-31 has a small amount of rotation and it appears to be moving as one cloud.

These features are similar to those that a face-on disk galaxy would have (such as the Milky Way if we were to have a bird’s eye view of it). Additionally, this cloud has a large hole in the middle, which is very perplexing indeed. What could this ring-cloud be?

AGESM33-31 was previously detected by Thilker, Braun and Walterbos in 2002 as an unresolved blob (they couldn’t make out the hole), and it was listed as a possible ‘dark companion’ to M33 – this is certainly a possibility. If this is the case, then AGESM33-31 could be the remnants of a dark galaxy that has undergone an interaction with another galaxy, carving out a hole in the process as the two collided, therefore forming a ring.

This ring-cloud could be an extension of the Magellanic stream. This is a hydrogen stream that crosses the sky, originating at the Magellanic clouds (the small and large Magellanic clouds are actually two dwarf galaxies orbiting near our own Milky Way.)

However, the Magellanic stream is very sparse in this area of the sky and it only consists of many small fragmented blobs of hydrogen. The ring-cloud and is a very large object that contains a lot of hydrogen, so it is unlikely to be part of the stream.

The hole in the ring-cloud could have been formed by a supernova (the death explosion of a massive star). This would have driven gas out of the local area and caused the hole. However, this too seems unlikely as the hole is about 10 times larger than would be expected if it had been caused by a supernova. All we can say for certain about this cloud is that it is a very interesting object indeed!

High-density regions were detected by M. Grossi et al (2008)

The total amount of HI around M 33 detected by the survey is about 107 Sun masses. At least 50% of this mass is made of HI clouds that are related both in space and velocity to the galaxy.

What about looking for the galaxies optically? We are reaching a dwarf galaxy limit of confusion.

Searching through optical images of the Virgo Cluster – a galaxy cluster containing around 2000 galaxies showed the dwarf galaxy problem also applies to it and observations show around 10 times fewer dwarf galaxies than is expected from simulations.

To conduct the search, data was taken from the Canada France Hawaii Telescope in four colour bands. Having multiple bands is useful as some galaxies are brighter in some bands than others. For example, spiral galaxies have a lot of young, blue stars so are brighter in some bands, whereas Elliptical galaxies have older, red stars, so are easier to see in other bands.

Initially, just one band was used to search for new dwarf galaxies. An automatic object detection programme was used to search for new objects by putting limits on their size and brightness. 443 dwarf galaxies were found, of which 303 were new detections.

However, one big challenge when looking at galaxy clusters is determining how to tell apart dwarf galaxies in the cluster from bigger, background galaxies. How do you tell which galaxies appear faint because they really are faint from the ones that appear faint because they’re further away?

Small and far away according to Father Ted

A possible solution to this problem is to look at the colour properties of all of the cluster galaxies to see if this could be used to tell them apart from background galaxies. With accurate colour and brightness measurements for around 1500 galaxies, it is hoped to explore the differences between dwarf galaxies and other galaxies, while correspondingly discovering many new dwarf galaxies.

A short tour of the Virgo cluster

The next generation of the Virgo survey

It will cover 100 square degrees of sky and can reach fainter surfaces.

What does the galaxy colour tell us?

There are a few colours that might be seen in galaxy images. They are usually caused by:

Blue: a region with many young stars. High-mass stars live fast and die young, using fuel at a high rate to maintain high temperatures. This causes them to emit hot radiation, which is blueish (google “blackbody radiation” to find out why).

Red: a region of old stars. The high-mass stars have swollen and cooled, and the low-mass stars were never hot to begin with, so they both emit cool radiation, which is reddish.

Patches of red/pink: a so-called HII (“H-two”) region. This is a cloud of ionized hydrogen (a cloud of free protons and electrons). When a proton captures an electron, it can give off light of various wavelengths as the electron hops down through energy levels. One particular hop, which is pretty common, emits red light, causing the HII region to appear reddish. HII regions are ionized in the first place by ultraviolet radiation from hot stars, so they indicate star-forming regions.

These are just a few features that happen to fall in the visible wavelength range; looking at radiation in the radio, infrared, ultraviolet, x-ray, and gamma-ray wavelength regions can reveal many more galactic characteristics.

There is some alteration of the colour before it is collected. Dust can make the image redder than it would be without dust. This happens because high-frequency (blue) light is more easily scattered by the dust than low-frequency (red) light. The colour of a few galaxies is affected by distance; an extremely distant galaxy has a high recessional velocity due to the expansion of the universe, which causes its light to be shifted toward the red. Most galaxies distant enough for this to have a noticeable effect on the colour of an image are very faint. The exceptions are known as “quasars”, and they produce so much radiation that they can be seen despite their extreme distance. In quasars, the shift is so large that the light that we see wasn’t even in the visible range when it was emitted.

By the way, whenever you look at an astronomical image, you need to check that the colours represent visible colours. Often astronomers will take images in wavelengths that are not visible to the eye, and then use colours to represent various wavelength bands.

NGC 852 is a barred spiral galaxy located in the constellation Eridan. It was discovered by British astronomer John Herschel in 1834.

Measurements not based on the redshift give a distance of 89,411 ± 11,486 Mpc (~292 million of al), which is within the distances calculated by using the value of the offset

Known information on galaxies is used in simulations

Background galaxies could be early dwarf type galaxies.

It is believed that dwarf galaxies played a significant role during the reionization era in transforming the early universe from being dark, neutral and opaque to one that is bright, ionized and transparent.

Despite their importance, distant dwarf galaxies remain elusive, because they are extremely faint and beyond the reach of even the best telescopes.

However, there is a way around this limitation. As predicted by Einstein’s general theory of relativity, a massive object such as a galaxy located along the line of sight to another distant object, can act as a natural lens, magnifying the light coming from that background source.

This phenomenon, known as gravitational lensing, causes the background object to appear brighter and larger. Therefore, these natural telescopes can allow us to discover unseen distant dwarf galaxies.

As a proof of concept, in 2014, the UC Riverside team targeted one cluster of galaxies that produce the gravitational lensing effect and got a glimpse of what appeared to be a large population of distant dwarf galaxies.

The team used the Wide Field Camera 3 on the Hubble Space Telescope to take deep images of three clusters of galaxies. They found the large population of distant dwarf galaxies from when the universe was between two to six billion years old. This cosmic time is critical as it is the most productive time for star formation in the universe.

Cluster membership

The greatest problem concerning the study of the faint cluster population is distinguishing members from the background population. Few clusters have known radial velocities of their faintest members. The 2dF and Flair-II Fornax cluster surveys provide the largest sample of confirmed faint members.

The following graph shows surface brightness–magnitude diagram for cluster members and background galaxies. Cluster members (from FCC): unconfirmed (open circles), confirmed by Flair-II/2dF (filled circles). Background galaxies (from FCCB): unconfirmed (points), confirmed by Flair-II (crosses). Triangles are Flair-II confirmed members originally classified as background galaxies. Squares are 2dF confirmed background galaxies originally classified as cluster members. The dashed line represents the segregation between populations (SB = 1.06V+ 4.16).


Members and background galaxies exhibit strong surface brightness relations, background galaxies having a higher central surface brightness for a given magnitude. The dashed line separates the two populations; member galaxies occupy the region to the left. This separation between the relations has been optimized such that the minimum number of confirmed background galaxies would be classified as cluster members. Of the 224 galaxies to the left of the dashed line, only 6 per cent are confirmed background galaxies.

Three background galaxies were detected from the FCC, which from the Flair-II survey, were found to be cluster members. Their location on the surface brightness–magnitude plot places them clearly within the cluster member population. Similarly, all four of the new background galaxies were detected, confirmed by the FCSS observations. Three of these lie well within the population of background galaxies.

For surface brightnesses <21.5 mag arcsec−2 the separation between populations is obvious, however, as surface brightness increases distinguishing between background and member galaxies becomes more difficult. In this regime, a higher proportion of background galaxies may masquerade as cluster members and vice versa. The inability of the surface brightness relation to characterize the faintest cluster members is entirely due to the resolution and seeing of the data. This effect and its implications will be left to the discussion.

The small sample of galaxies to the right of the background sample are the ultracompact dwarf galaxies. These high surface brightness galaxies occupy the intermediate region between the cluster dwarf population and galactic globular clusters (Ferguson & Binggeli 1994). In this case membership judgements based on surface brightness and luminosity break down and radial velocities are needed to determine membership.

Answers to questions

Are all the missing galaxies dark galaxies? Probably not, but it’s likely that some are.

Are the galaxies all hiding optional data? While many are, whether or not this will solve the problems, remains to be seen.

Dwarf galaxies orbit major galaxies.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s