NGC 7632 – with higher redshift double galaxy

NGC 7632 (object 1 in Figure 1) has been reported to have a higher redshift double galaxy (objects 8 and 9) nearby by Arp (1981). Arp (1982) noted further that:

The central galaxy is disturbed on such a large scale that this large interacting pair seems to be the only possibility to account for such a disturbance.

That should of course be impossible due to discordant redshift of the galaxy pair.

I haven’t found any other papers discussing this system. Figure 1 shows all nearby objects that have redshifts available. The system is not particularly interesting in discordant redshift sense. There’s a pair alignment with objects 6 and 10 aligned across NGC 7632. Objects 2, 3, 5, 6, and 7 have similar redshifts, so they form a probable galaxy group, making the pair alignment of objects 6 and 10 less significant because having a background galaxy group near a low redshift galaxy increases possibility for an alignment with some other background object, or in other words, changing the position of object 10 somewhat would still result of an alignment with some object in the group (the position angle for object 10 that would still give good alignments could change dozens of degrees).

Figure 1. The objects with measured redshifts near NGC 7632. Size of the image is 15 x 15 arcmin. Image is from Digitized Sky Survey (POSS2/UKSTU Red), and it has been adjusted for brightness and contrast to bring out the faint objects in the field.

Objects and their data

1 NGC 7632 SB0 0.005107 12.95 0
2 LEDA 195478 galaxy 0.095920 18.77 2.179
3 LEDA 123659 galaxy 0.095820 17.49 3.018
4 LCRS B231901.0-424300 galaxy 0.047423 18.07 3.665
5 LCRS B231942.3-424324 galaxy 0.096330 18.10 5.088
6 LCRS B231932.1-424038 galaxy 0.095152 17.45 5.466
7 LCRS B231940.7-424032 galaxy 0.094919 18.05 6.500
8 ESO 291-023 S 0.034754 15.76 6.799
9 ESO 291-022 S 0.034897 15.38 7.182
10 APMUKS(BJ) B231849.62-425237.5 galaxy 0.065002 17.76 8.847


Arp, 1981, ApJS, 46, 75, “Spectroscopic measures of galaxies, their companions, and peculiar galaxies in the southern hemisphere”

Arp, 1982, ApJ, 256, 54, “Characteristics of companion galaxies”

SDSS J153636.22+044127.0 – Another high redshift NGC 1275 clone

Boroson & Lauer (2009) report about another high redshift (z ~ 0.4) system that resembles the situation in NGC 1275. Recently, I wrote about SDSSJ092712.65+294344.0, which is very similar to this new one. Both have two emission line sets at different redshifts. SDSS J153636.22+044127.0 has redshift difference of about 3500 km/s. There is one big difference between these two systems, however, the SDSSJ092712.65+294344.0 has third emission line set at intermediate redshift, but SDSS J153636.22+044127.0 has an absorption line set at intermediate redshift.

Boroson & Lauer (2009) first point out how extremely exceptional systems these two are:

Of the 17,500 objects in the entire sample, only two objects have multiple redshift systems, SDSS J092712.65+294344.0, described above, and SDSS J153636.22+044127.0 (J1536+0441).

They then go through the basic properties of the system (magnitudes etc.), and they describe the emission line systems:

The spectrum of J1536+0441 shows two broad-line emission systems and one system of narrow absorption lines. The higher redshift “r-system” at z = 0.3889 shows broad Balmer lines (Ha, Hß, and H?) and the usual narrow lines seen in low redshift quasars. The lower redshift “b-system” at z = 0.3727, shows broad Balmer lines (Ha 2 through Hd) and broad Fe II emission, seen most strongly around 3000 Å in the rest frame. A strong narrow absorption-line “a-system” is also present, including 6 unresolved resonance lines, at z =0.38783, which, in the quasar rest frame, is 240 km s-1 less than that of the r-system and 3300 km s-1 greater than that of the b-system.

The “red system” has spectral features like a low redshift quasar typically has, but the “blue system” is nothing like a quasar. “Absorption system” is also not typical for a low redshift quasar. They suggest that one possibility to explain the absorption system is that it is an unrelated gas cloud in front of the object.

They analyse the system as a binary black hole candidate, and calculate system parameters based on that. They also suggest as alternative hypothesis, that the two emission line systems could arise as a chance projection of two quasars. They find a probability of 0.0032 for the chance projection, but they also note that the blue system spectral features argue against chance projection hypothesis. They conclude:

The most obvious interpretation of the presence of absorption at a redshift close to the r-system is that the b-system is background to the r-system. This would be inconsistent with an explanation involving ejection of one of the black holes, though it would be consistent with an infall interpretation, analogous with NGC 1275. In this case, however, the two broad-line systems must represent two active nuclei.

Gaskell (2009) presents an argument against the binary black hole hypothesis:

I argue here, however, that just as the Balmer line profiles in previous SMB candidates have been shown to be due to disc emission, so too the Balmer line profile of J1536+0441 is probably due to disc emission.

Gaskell then compares the spectra of SDSS J153636.22+044127.0 and ARP 102B which is an example of the disk emitter, and finds that they are similar.

Chornock et al. (2009) confirmed the existence of the two emission line sets, but they also suggest a presence of a third emission redshift that is blueshifted by 3800 km/s compared to the red system. They also perform a test to the binary black hole hypothesis. They argue that the system should show measurable redshift evolution if it would be a binary black hole. They compare the two spectra now available of the system, and don’t find evidence for the expected redshift evolution. They also suggest that the system is disk emitter.

Wrobel & Laor (2009) found from VLA images that there is two radio sources separated by 0.97″ at the position of SDSS J153636.22+044127.0. They propose two alternatives:

These could be two related radio sources, both energized by the candidate 0.1-pc binary system. Alternatively, VLA-A and VLA-B could be independent radio sources originating from a binary quasar system, with a projected separation of 5.1 kpc.

They find some similar features from this system than the features of known systems matching the first alternative. For the binary quasar alternative they note:

Boroson & Lauer (2009) noted that the probability for such a random projection in their sample is 0.0032. Therefore, the two quasars are most likely not due to a random projection, but are likely physically related, i.e. a binary quasar system.

They discuss further that it is likely to have one binary quasar in Boroson & Lauer (2009) sample. They also say that if this is a binary quasar, then it is most likely within two strongly interacting galaxies. However:

The velocity separation of 3500 km s-1 (Boroson & Lauer 2009) is rather large, but not implausible in a cluster of galaxies. About half the clusters studied by Carlberg et al. (1996) show such an extent of velocities. The maximum velocity differences in the quasar binaries studied by Hennawi et al. (2006) is 1870 km s-1, but that study imposed a cap of 2000 km s-1 on the binary velocity separation. At a projected relative velocity of 3500 km s-1, two galaxies cannot form a bound system, so the term binary quasar here does not refer to a physically bound binary system.

Decarli et al. (2009) also found two separate objects in the system from VLT-images. Here is a link to their image. Based on that, they argue that the system is a quasar pair.

In a new study based on HST images of the system, Lauer & Boroson (2009) mention another possibility for this system:

A more intriguing possibility is that the b-system is a black hole ejected from the nucleus of a galaxy now just hosting the r-system; however, the large implied ejection velocity and the light source for the absorption system are problematic.

However, they say:

These new observations appear to rule out the “chance superposition” or “ejected black hole” hypotheses, but leave the choice between the “double-peaked” or “binary black hole” hypotheses unresolved.

From the HST images, they find an optical counterpart for the second radio source that Wrobel & Laor (2009) found. Lauer & Boroson don’t find any signs of interaction between the two objects. From the available spectra (they took a new one) they determine that the different emission line systems are not separated spatially, so they must come from the same object. This means that the newly found companion doesn’t contribute to the different redshifts in the system. This also means that the chance projection hypothesis has less evidence; when there was another candidate object for different redshift, the possibility of chance projection seemed better, but now we are again back to the situation where we have apparently one object emitting different redshifts.

They don’t favor the ejected black hole hypothesis because velocity differences are so high. They seem to lean toward the double-peaked disk emitter hypothesis also. At least there seems to be little evidence against that hypothesis. They also mention that the binary black hole hypothesis does not gain any new evidence from their observations, but it can’t be ruled out either. As a binary black hole, the system is so close together that there are expected changes in the two emission velocities to be seen. That we should know within few years.

In a follow-up to their previous paper, Chornock et al. (2009b) report a new optical Keck spectrum. They confirm the third emission line system which they reported in their first paper. They still don’t find the redshift evolution expected in a binary black hole hypothesis, and now the situation is quite clearly such that the evolution should be measurable. They also show, like Lauer & Boroson, that the two main emission line sets are not separated spatially so the double-quasar hypothesis seems to be ruled out. They suggest that the system is double peaked disk emitter.

UPDATE (August 6, 2009):
Decarli et al. (2009b) published new imaging of the system. They verify that the system has two optical components by perfoming an decomposite analysis on the system, see their Figure 1. They checked the neighborhood of the system to see if there is a galaxy cluster that would help to explain the large velocity difference implied by the redshift difference between the emission sets, and the closeness of the two objects. They said:

We find a significant excess of the galaxy density within < 200 kpc from SDSS J1536+0441 with respect to average field. Thus the quasar is found in a moderately rich cluster of galaxies.

However, they only estimated amount of galaxies photometrically consistent to a z = 0.4 galaxy, so the estimate is very uncertain. At any case, the presence of the companion helps to explain the absorption in the spectrum. In the end, they also can’t make a difference between the binary black hole hypothesis and the double peaked emitter hypothesis.

UPDATE (September 3, 2009):
There is a new paper by Tang & Grindlay (2009) in arXiv. They have reanalysed the existing data of the system by fitting double peaked disk emitter model to the spectrum of the system. They found that even though the model doesn’t fully explain the spectrum, it still provides evidence that there is some disk emission in the system. They end up proposing that the system is both a binary black hole and a double peaked emitter. They provide several arguments to support that proposal. First, as mentioned above, the spectral fit with disk emission model suggests that there’s some disk emission involved. Second, any model with a single explanation doesn’t seem to provide adequate explanation. Third, they say:

The existence of a minor black hole is physically more natural than an extremely asymmetric accretion disk. In fact, we do expect SMBH binaries as natural consequences of galaxy mergers (Begelman et al. 1980; Civano et al. 2009; Comerford et al. 2009).

Fourth, the fact that double peaked disk emitters are rare objects can be neatly explained by a presence of a another black hole which helps produce disk emission by disturbing the accretion disk of the central black hole. They also discussed the question why other double peaked disk emitters don’t show evidence of binary black hole and possible answers they came up is that the other black hole might be so small that the central black hole is too dominant for the other to show up in observations and that long-term observations are generally needed (but not available currently) to detect binary black holes anyway. I think their explanation of the system is quite sensible.

Figure 1. The SDSS J153636.22+044127.0 field. Size of the image is 7 x 7 arcmin. There are no objects with measured redshifts within this field, except for SDSS J153636.22+044127.0 which is in the center of the image. Image is from Digitized Sky Survey (POSS2/UKSTU Red).

Objects and their data

1 SDSS J153636.22+044127.0 “blue system” QSO em. line system 0.3727 17.24 (g) 0
2 SDSS J153636.22+044127.0 “red system” QSO em. line system 0.3889 0
3 SDSS J153636.22+044127.0 “absorption system” QSO abs. line system 0.38783 0
4 SDSS J153636.22+044127.0 “even bluer system” QSO em. line system cz of red system – 3800 km/s 0

Boroson & Lauer, 2009, Nature, 458, 53, “A candidate sub-parsec supermassive binary black hole system”

Chornock et al., 2009, ATel, 1955, 1, “SDSS J1536+0441: An Extreme “Double-peaked Emitter,” Not a Binary Black Hole”

Chornock et al., 2009b, arXiv, 0906.0849, “The Quasar SDSS J1536+0441: An Unusual Double-Peaked Emitter”

Decarli et al., 2009, aTel, 2061, 1, “SDSS J1536+0441: A quasar pair, not a binary black hole”

Decarli et al., 2009b, arXiv, 0907.5414, “Probing the nature of the massive black hole binary candidate SDSS J1536+0441”

Gaskell, 2009, arXiv, 0903.4447, “J1536+0441 and the lack of evidence for close supermassive binary black holes”

Lauer & Boroson, 2009, arXiv, 0906.0020, “HST Images and KPNO Spectroscopy of the Binary Black Hole Candidate SDSS J153636.22+044127.0”

Tang & Grindlay, 2009, arXiv, 0909.0258, “The Quasar SDSS J153636.22+044127.0: A Double-Peaked Emitter in a Candidate Binary Black-Hole System”

Wrobel & Laor, 2009, ApJ, 699, 22, “Discovery of Radio Emission from the Quasar SDSS J1536+0441, a Candidate Binary Black-Hole System”

Redshift components

In an earlier post, I briefly described the redshift components of an extragalactic object. It turns out that generally an extragalactic object has two redshift components; kinematical and cosmological. In this post I’m going to describe how the different components are handled, when calculating things relating to redshifts of galaxies and their associated objects. When combining different redshift components, they are not simply added together. Redshift is a curious quantity in that sense. Let us consider a simple example of one galaxy having a cosmological and kinematical redshift components.

The galaxy we observe has a peculiar velocity with respect to a reference frame that is non-moving frame local to the galaxy. For example, the Earth moves around the Sun, Sun is in motion around the Milky Way center, and the Milky Way is in motion with respect to Local Group, and so on. All this makes the Earth to have a peculiar velocity with respect to non-moving space at location of the Earth. So, when the galaxy we are observing emits light, the first thing the light experiences is the kinematical redshift that occurs between the galaxy’s rest frame and the galaxy’s local non-moving reference frame. After that, the light starts moving through space towards us. In the space it encounters something that causes redshift also. That is the cosmological redshift component. In Big Bang cosmology, cosmological redshift component is caused by expanding space. Many alternative cosmologies have different explanations.

In the receiving end, there is the Earth’s peculiar velocity that also causes redshifting, but that is usually not kept in these analyses. Instead, the measured redshift is corrected for the Earth’s motion. Usually redshifts have been given in heliocentric reference frame, but galactocentric and Local Group -centric redshifts have also been used. Lately, the redshifts have been increasingly given in CMBR reference frame. However, it is quite common to just use the heliocentric reference frame. That contains some error in the redshift, but when dealing with objects close to each other, such as a galaxy and an object near it, then the correction due to Earth’s complex motion is approximately the same for both objects, so in practice there’s no effect for the Earth’s motion to the results in that kind of situations. Therefore it is OK for us to use heliocentric redshifts here, which is still most common frame where redshifts are given.

Now, let us derive the equation that can be used to calculate the redshift components. This derivation is given in Davis et al. (2003). We start with the basic redshift equation:

1 + z = Lobs/Lem [eq. 1]

where z is the measured redshift (the heliocentric redshift), Lobs is the wavelength of the light we observe, and Lem is the wavelength of the light when it was emitted. The peculiar velocity of the galaxy causes the wavelength to change. We will call this wavelength Lnm, the wavelength of the light in non-moving reference frame local to the galaxy. We will do a mathematical trick next. We will multiply the basic redshift equation by one (which doesn’t change the result so it’s a valid thing to do), but we just write 1 = Lnm/Lnm. So we have:

1 + z = Lobs/Lem * Lnm/Lnm

Then we move Lnm’s a bit:

1 + z = Lobs/Lnm * Lnm/Lem [eq. 2]

This equation now contains both redshift components. I’ll demonstrate that next. If we would be in the non-moving reference frame local to the galaxy, we would observe the wavelength of the galaxy to be Lobs = Lnm, and putting that to equation 2 we would get kinematical redshift zK:

1 + zK = Lnm/Lnm * Lnm/Lem = Lnm/Lem

=> 1 + zK = Lnm/Lem [eq. 3]

If the galaxy wouldn’t have any peculiar velocity, there wouldn’t be any kinematical redshift component, so wavelength in the non-moving reference frame would be the same as emitted wavelength, i.e. Lnm = Lem. Again from equation 2 we would then get cosmological redshift component zC:

1 + zC = Lobs/Lnm * Lnm/Lnm

1 + zC = Lobs/Lnm [eq. 4]

Now we have seen that equation 2 contains both redshift components, but it still has wavelengths in it. We would want redshifts there explicitly. We can do it by replacing wavelengths in equation 2 with redshift components from equations 3 and 4, and it becomes:

1 + zM = (1 + zC) * (1 + zK) [eq. 5]

Where zM is the measured redshift of an object. So, as you can see, when adding up redshift components, it is not done simply by zM = zC + zK. Let’s try the equation 5 with an example galaxy. Let us assume that we have measured a redshift of zM = 0.01 to a galaxy, but let’s also assume that the galaxy has a peculiar velocity of 500 km/s away from us. First thing we need to do is to convert the peculiar velocity to redshift. There is a simple equation for that: z = v/c. Speed of light, c, is 299792.458 km/s, so our kinematical redshift component is:

zK = 500 km/s / (299792.458 km/s) = 0.001668

Now we know the kinematical redshift component and the measured redshift, so only unknown in equation 5 is the cosmological redshift component. Let’s calculate it:

1 + zM = (1 + zC) * (1 + zK)
=> zC = (1 + zM)/(1 + zK) – 1 = (1 + 0.01)/(1 + 0.001668) – 1 = 0.008318

In practice it is very difficult to determine the redshift components. The kinematical and cosmological redshift components look the same in the spectra of objects. If we are only observing a single galaxy, it is practically impossible to determine the components. Situation gets better if we have two objects physically associated with each other, pair of interacting galaxies for example. Easiest situation is when there is a large galaxy interacting with a clearly smaller galaxy, and we wish to calculate the redshift components of the smaller galaxy. In that situation, we can assume that the smaller galaxy follows the motion of the large galaxy similarily as the Earth follows the motion of the Sun around the Milky Way nucleus, even if the Earth also has its own motion around the Sun. So, the smaller galaxy has its own motion around the large galaxy, but it still follows the path of the large galaxy. This means we can calculate the system so that we ignore any peculiar velocity the large galaxy has because the smaller galaxy also has the same peculiar velocity in addition to its own motion around the large galaxy. Or, to put in other words, we can calculate the redshift component that the smaller galaxy has in addition to large galaxy redshift components. So, the thorough equation for the redshift components of the smaller galaxy would be:

1 + zS = (1 + zC) * (1 + zKlarge) * (1 + zKsmall)

Where zS is the measured redshift of the smaller galaxy, zKlarge is the kinematical redshift component, or the peculiar velocity of the large galaxy, and zKsmall is the additional peculiar velocity of the smaller galaxy. zKlarge is included here because we assume that the smaller galaxy follows the motion of the large galaxy, as discussed above. Also following the discussion above, especially the equation 5, we can now just put all large galaxy components under one heading to derive an equation that we can use to calculate the kinematical redshift component of the smaller galaxy:

1 + zS = (1 + zL) * (1 + zKsmall) [eq. 6]

Where zL is the combined redshift components of the large galaxy, i.e. the measured redshift of the large galaxy. Note that equation 6 is exactly the same as equation 5, only the names of the components have changed. How can we use this equation then? In practice, we should have redshifts measured both for the large and the smaller galaxy. The measured redshift of the large galaxy is the zL, and the measured redshift of the smaller galaxy is the zM. So, let’s assume that we have a large galaxy with measured redshift of zL = 0.010, and we have a smaller galaxy right beside the large one with measured redshift of zS = 0.011. We assume that these galaxies are interacting (sometimes there are some visible signs suggesting that, but not always), and we therefore decide we can use equation 6. First, we’ll solve the equation 6 for zKsmall, because that is what we want to calculate:

zKsmall = (1 + zS)/(1 + zL) – 1 [eq. 7]

And then we will just put the numbers in:

zKsmall = (1 + 0.011)/(1 + 0.010) – 1 = 0.00099

Peculiar velocity of the smaller galaxy with reference to the large galaxy is then:

v = cz = 299792.458 * 0.00099 = 297 km/s

The kinematical redshift component of the smaller galaxy is the only redshift component that we can calculate from the system when we only know redshifts of both galaxies. In order to determine other components, we need additionally a redshift independent distance measurement for the galaxy pair. If we know the distance between us and the galaxy pair, we can calculate how big kinematical redshift component the large galaxy has, and the size of the cosmological redshift component. However, problem here is that distance measurements to galaxies are still quite inaccurate, so this calculation will have lot of uncertainty.

Let’s take NGC 289 as an example. NGC 289 has a redshift of c * zL = 1629 km/s (zL = 0.005434) in NED, and a distance of 23.4 Mpc. NGC 289 has a companion, LSBG F411-024, which has a redshift of cz = 1510 km/s (zS = 0.005037). If you look at the distance measurement section for NGC 289 in NED, you will notice that there’s two measurements, from which the NED value is derived; 19.4 Mpc and 27.4 Mpc, quite a difference, which reflects the above mentioned inaccuracy in distance measurements.

To convert the measured distance to redshift, we will use Hubble law (there are more distinguished methods to do that, but it will serve the purpose for us in this example). From Hubble law we get (using Hubble constant of 72 (km/s)/Mpc):

vDIST = H0 * d = 72 (km/s)/Mpc * 23.4 Mpc = 1684.8 km/s

If we assume that this is accurate value, it would mean that NGC 289 has a peculiar velocity of:

vKlarge = c * zL – vDIST = 1629 km/s – 1684.8 km/s = -55.8 km/s

Minus sign indicates that the velocity is towards us (redshift is positive when velocity is directed away from us). From the peculiar velocity we then get the kinematical redshift component of the large galaxy:

zKlarge = vKlarge / c = -55.8 km/s / (299792.458 km/s) = -0.000186

Cosmological redshift component is (from equation 5 applied to the large galaxy):

1 + zL = (1 + zC) * (1 + zKlarge)

zC = (1 + zL) / (1 + zKlarge) – 1 = (1 + 0.005434) / (1 + -0.000186) – 1 = 0.00562

We could also have derived the cosmological redshift component directly from the redshift independent distance measurement (if we multiply the above calculated zC by c, we get 1685.2 km/s which is quite close to vDIST calculated above). Finally we will also have to calculate the kinematical redshift component of the companion galaxy using the equation 7:

zKsmall = (1 + zS)/(1 + zL) – 1 = (1 + 0.005037)/(1 + 0.005434) – 1 = -0.000395

vKsmall = c * zKsmall = 299792.458 km/s * -0.000395 = -118 km/s

Kinematical redshift component of the companion galaxy turns out to be negative too, but that was expected because the redshift velocity cz was smaller in the companion galaxy suggesting that it has a peculiar velocity towards us. Notice that the calculated value here is almost exactly the value from the measured cz’s = 1510 – 1629 = 119 km/s, so we could have calculated the kinematical redshift component also from measured cz’s directly by:

zKsmall = c * zS – c * zL / c = (zS – zL) * c / c = zS – zL

Now you might notice that this goes exactly against what I said in the beginning, that redshifts are not simply added together. It is true, it goes against what I said, and the reason is this: the equation v = cz is only an approximation that works quite well for velocities far lower than the speed of light. If we would use more precise calculation methods, we would use the relativistic doppler line of sight equation instead of v = cz:

1 + z = sqrt[(1 + v/c)/(1 – v/c)]

Using that equation, the companion redshift component calculation wouldn’t reduce back to zS – zL anymore, and our world makes sense again. 🙂


Davis et al., 2003, AmJPh, 71, 358, “Solutions to the tethered galaxy problem in an expanding universe and the observation of receding blueshifted objects”

1107+036 – QSO near galaxy

Murdoch et al. (1983) discussed 1107+036, a radio source near a galaxy. The radio source is a quasar at redshift z = 0.966. First they discussed the optical identification of the radio source and concluded that the QSO is the most likely identification. Then they discussed the basic properties of the system (the spectrum of the QSO, morphology and the spectrum of the main galaxy). They mentioned that the main galaxy is asymmetric, and seems to be extended towards the QSO. See their figure 6a, the galaxy seems to extend all the way to the position of the QSO.

They included a discussion of this system as a QSO-galaxy association. They noted that there is a low probability for the QSO to be projected so near to the main galaxy. They also said that there is no luminous link between the objects. Their discussion is presented in quite curious form that shows that they have not agreed on the interpretation of the object (one of the authors is Arp). It’s amusing to imagine what kind of discussion actually went on based on this published version.

Figure 1. The objects with measured redshifts near SDSS J111025.10+032138.8. Size of the image is 7 x 7 arcmin. Image is from Digitized Sky Survey (POSS2/UKSTU Blue).


In the DSS image it seems a little as if there would be an optical bridge between the two objects, but the feature cannot be brought up from the noise convincingly enough. Figure 2 shows a brightness and contrast adjusted image that is most favorable to the bridge-hypothesis. However, playing with brightness and contrast doesn’t bring any solid evidence that the apparent bridge would be a real physical connection.

Here is SDSS image of this system. Main galaxy (SDSS J111025.10+032138.8) is at the center, and [HB89] 1107+036 is the nearest bright white-blue object at 3 o’clock from the main galaxy. There doesn’t seem to be a bridge in this image, but adjusting the brightness and contrast brings out a possible connection but it’s hard to say if it’s noise or real feature.

Figure 2. 4 x zoomed in and brightness/contrast adjusted version from figure 1.

Objects and their data

1 SDSS J111025.10+032138.8 Sc, HII 0.030115 16.5 (g) 0
2 [HB89] 1107+036 QSO 0.965617 19.1 (g) 0.315

Object descriptions in NED: object 1, object 2.

Object descriptions in SDSS SkyServer: object 1, object 2.

SDSS image of this system


Murdoch et al., 1983, ApJ, 265, 610, “1107+036 – an unusual QSO-galaxy pair”

0248+430 – Very close quasar-galaxy pair

Pauliny-Toth et al. (1978) identified object 3 in Fig. 1 as a radio source. Kuhr (1977) suggested that it is a quasar, and near a galaxy (objects 1 & 2). Kuhr (1977) also showed an image of the system. Kuhr (1980) measured the redshift of object 3 as z = 1.316 confirming it a quasar.

Junkkarinen (1987) found absorption lines from the spectrum of 0248+430 at main galaxy’s redshift (absorption redshift z = 0.052). This was followed by Sargent & Steidel (1990), who also found absorption lines from the spectrum of object 3. They found some absorption (z = 0.0519) at the redshift of objects 1 and 2. They also found absorption at redshift z = 0.3940. They said:

These images reveal that the interstellar absorption is produced in a part of a complex tidal tail emanating from G0248+4302, which turns out to be a close double galaxy system.

Unfortunately, the electronic version of their paper doesn’t include the pictures they took of the system. They also suggest that there is a faint background galaxy between the quasar and the main galaxy. Womble et al. (1990) also studied absorption lines in the system, and found two absorption systems at main galaxy’s redshift (z = 0.0515 and z = 0.0523), and two at higher redshifts of z = 0.394 and z = 0.451. They also talk about tidal tail and the double nucleus, so in many ways, this is a confirmation of Sargent & Steidel (1990) who also found the 0.394 absorption system. Womble et al. (1990) also note:

We did not find any galaxies at these absorption redshifts; however, we tentatively identify a galaxy 11″ northeast of the QSO at z = 0.240.

The galaxy they talked about is the object 6 in Fig. 1. They also took spectrum of object 7 in Fig. 1 and found out that it is a probable foreground star. Note also that Womble et al. (1990) did not perform a complete search of the field, so it doesn’t necessarily mean anything that they didn’t find any galaxies at absorption redshifts.

Figure 1. The objects with measured redshifts near LEDA 090441. Size of the image is: left panel, 7 x 7 arcmin, right panel, same image but 4 x zoomed in (so roughly 1.75 x 1.5 arcmin). Image is from Digitized Sky Survey (POSS2/UKSTU Red), and it has been adjusted for brightness and contrast to bring out the faint objects in the field.

Borgeest et al. (1991) took new images and spectra of the system. Note how their Fig. 2 seems to show a bridge between the quasar (“Q”) and the galaxy. But let us also remember the tidal tail which is supposedly arising from the merging of the two galaxies, the presence of a tidal tail from such process increases the probability of apparent “bridge” arising to a nearby object. And in this case the “bridge” doesn’t even seem to hit exactly the center of the quasar, which further supports the view that this would be just a background object projected by chance to the direction of the tail. The probability would be roughly 20/360 ~ 0.06, assuming that there’s about 20 degrees out of 360 possible that the tail would point roughly to the quasar. It also should be pointed out that this tail continues far beyond the position of the quasar, see Borgeest et al. fig. 1.

Another interesting thing mentioned by Borgeest et al. was:

Within the positional errors the quasar Q is aligned with A and B.

“A and B” being the merging double nucleus of the main galaxy. Borgeest et al. found absorption line from object 3’s spectrum, the line is at main galaxy’s redshift, supporting previous similar findings. They then analysed the system as gravitational lensing candidate. They first pointed out that this is not exactly a typical system for multiple lensed images because the separation of the quasar and the galaxy is larger than usually in those cases. They discussed the system as macro lensing system (such that the main galaxy as a whole would act as a gravitational lense), but lack of second quasar image makes that scenario improbable.

They also considered micro lensing, in which the stars of the intervening galaxy would cause amplification of the quasar’s brightness when they happen to be between us and the quasar. Difficulty with this scenario is that the amplification events are very brief, so it wouldn’t seem probable that all of our observations would occur at the time of an amplification event. Borgeest et al. argue that with optimistic assumptions it would be possible to detect some lensing variability in this system. They also presented a probability calculation that shows very low probability of the chance projection for this quasar-galaxy association, unless some lensing scenario is at play. But they point out:

However, a strong amplification due to micro-lensing is very unlikely for GC 0248+430, since otherwise the 1950 (POSS) and 1989 magnitudes ought to differ significantly since high amplifications by micro-lensing do not last very long. A large macro-lensing amplification seems also to be very unprobable. From our experience in modelling lens effects we know that highly amplified single images only appear near the centre of the lense.

So, if the lensing is improbable, we are stuck with very low chance projection probability for this system. (But of course, as in all this kind of cases, we need to remember that one cannot do statistics with a single system.)

For other objects in the field, Borgeest et al. noted that object 4 is a F7 star according to their spectrum, and they too think that object 7 is a star. However, they do raise a point about this object; they say that it is possible that it is second gravitational lense image of the quasar (object 3) because of similar color between the two objects, but that there is no signs of emission at correct redshift in object 7’s spectrum for that.

An accompanying paper by same team, Kollatschny et al. (1991), focused on the main galaxy’s double nucleus. They found a redshift for both nuclei (object 1 z = 0.0512 and object 2 z = 0.0507). They also measured galaxy’s rotation curve, which shows disturbances that further advance the merging hypothesis.

Hwang & Chiou (2004) made some observations of the system. Once again we note the bridge-like situation in their fig. 2. Note how the features in the CO-moment map seem to be arising from the main galaxy and focusing to the quasar. Hwang & Chiou found HI absorption at main galaxy’s redshift from the quasar’s HI spectrum. They also pointed out that the absorption derived metal abundances and some other features of the “absorbing cloud” are quite abnormal.

López-Corredoira & Gutiérrez (2006) looked at this system among some other discordant redshift systems, and found that there is another quasar (z = 1.531) further out (object 5 in fig. 1). They said:

The position angle of the first QSO with respect to the major nucleus of the galaxy is -73 [degrees] at a distance of 14.4″ from the major nucleus; and the position angle of the second QSO is -68 [degrees] at a distance of 108″ from the major nucleus. The position angle of the line which joins the two nuclei of the galaxy is -77±18 [degrees] (the error bar is large because both nuclei are very close — 2.7″ — and it was not possible to determine the position of the second centre accurately). Even if we forget the second nucleus and the gas ejection, all in this direction[24], given the low density of expected background quasars, the coincidence of the near-alignment (the difference is 5 degrees) seems unlikely.

Objects and their data

1&2 LEDA 090441 Merger 0.051440 16.11 0
1 AN 0248+43A galaxy 0.051939 16.52 0.075
2 AN 0248+43B galaxy 0.050700 17.36 0.040
3 [HB89] 0248+430 QSO (LPQ) 1.310000 17.45 0.252
4 star F7 probably ~ 0 ~ 0.25
5 quasar quasar 1.531 21.11 (g) 1.8
6 galaxy galaxy 0.240 ~ 0.25
7 star star probably ~ 0 ~ 0.25

NED object list with available redshifts within 10 arcmin.


Borgeest et al., 1991, A&A, 249, 93, “GC0248 + 430 – A possibly micro-lensed quasar behind a tidal arm of a merging galaxy system”

Hwang & Chiou, 2004, ApJ, 600, 52, “A New H I 21 Centimeter Absorber Associated with the H I Deficient Interacting Galaxies G0248+430”

Junkkarinen, 1987, BAAS, 19, 953, “A Search for More QSO Absorption Systems Produced by Intervening Galaxies”

Kollatschny et al., 1991, A&A, 249, 57, “An 0248 + 43 – A cold highly luminous FIR-galaxy with two nonthermal nuclei”

Kuhr, 1977, A&AS, 29, 139, “Optical identification of extragalactic radio sources from the NRAO – Bonn 5GHz survey”

Kuhr, 1980, Ph.D. thesis, Bonn, information about this is given in Sargent & Steidel (1990).

López-Corredoira & Gutiérrez, 2006, AIPC, 822, 75, “Research on candidates for non-cosmological redshifts”

Pauliny-Toth et al., 1978, AJ, 83, 451, “The 5 GHz strong source surveys. IV – Survey of the area between declination 35 and 70 degrees and summary of source counts, spectra and optical identifications”

Sargent & Steidel, 1990, ApJ, 359L, 37, “Absorption lines in the spectrum of Q0248 + 4302 due to a foreground tidal tail”

Womble et al., 1990, AJ, 100, 1785, “CA II and NA I absorption in the QSO S4 0248 + 430 due to an intervening galaxy”

UGC 10806 – a line of galaxies with two different redshifts

Preprint of Melnyk et al. (2009) was just inserted to arXiv. One of the objects in their study is UGC 10806. See their figure 2, which shows UGC 10806 field. Four most obvious galaxies in the pictured field form a line, but the objects in the line have discordant redshifts. Two galaxies are at cz ~ 1000 km/s and two are at cz ~ 7500 km/s. Melnyk et al. (2009) say:

The two faint ”companions” to the west of UGC10806 have radial velocities of +7354 and +7569 km/s; that is, they are objects in the distant background.

The line alignment is somewhat off-center of UGC 10806, but besides UGC 10806, the alignment is very exact.

Melnyk et al. (2009) fig. 2 shows the field of 15′ x 8′. Let us look full 15′ x 15′ field. It is presented here in Fig. 1. There we see the four objects described by Melnyk et al. (2009), the UGC 10806 (object 1) and its three nearby galaxies (objects 2, 3, and 4). However, we note that there are couple of other obvious galaxies in the wider field (for some reason Melnyk et al. have not included them into their sample). Objects 5 and 6 have even higher redshift than objects 3 and 4. Objects 5 and 6 also have similar redshift to each other, so they are probably a physical pair of galaxies. The presence of these two galaxies makes the probability bigger for the apparent line alignment (more objects – bigger probability for any alignments to occur).

Figure 1. The biggest galaxies near UGC 10806. Size of the image is 15 x 15 arcmin. Image is from Digitized Sky Survey (POSS2/UKSTU Blue).

Let us also look at all objects with available redshifts near UGC 10806. They are shown in Fig. 2. This area has been covered by the Kitt Peak Galaxy Redshift Survey, so there’s quite many objects with redshifts. There are no remarkable alignments. Objects 12 and 14 are roughly aligned across the main galaxy. Redshift of object 4 is quite accurately twice the redshift of object 12 (z14/z12 = 2.029).

Objects 11, 13, and 14 have very similar redshifts, their redshift velocity (calculated from relativistic Doppler redshift line of sight equation) is within 300 km/s, so they could be part of a background galaxy group. Object 14 seems to be situated quite far away from other two, though.

Redshifts of objects 9 and 10 are close to each other, and they are also situated near each other. However, their redshift velocity differs by 1250 km/s, which is rather large for a physical galaxy pair, unless they are in a galaxy cluster.

Figure 2. The objects with available redshifts near UGC 10806. Size of the image is 7 x 7 arcmin. Image is from Digitized Sky Survey (POSS2/UKSTU Blue).

Objects and their data

1 UGC 10806 SBdm 0.003109 (932 km/s) 13.73 0
2 LEDA 167108 galaxy 0.003400 (1019 km/s) 16.88 3.126
3 LEDA 167103 galaxy 0.024000 (7195 km/s) 16.27 3.273
4 171812.0+495615 galaxy (7569 km/s) 7.7
5 UGC 10816 Scd 0.033046 (9907 km/s) 15.6 6.396
6 2MASX J17191363+4958366 galaxy 0.033200 (9953 km/s) 15.55 6.608
7 HERC-1:[MKK97] 308890 PofG? 0.003000 (899 km/s) 12.58 0.757
8 HS 1717+4955 PofG? 0.003466 (1039 km/s) 17.7 0.891
9 HERC-1:[MKK97] 308000 galaxy 0.384400 20.05 1.699
10 HERC-1:[MKK97] 307733 galaxy 0.390800 19.65 2.142
11 HERC-1:[MKK97] 307422 galaxy 0.312400 19.46 2.758
12 LEDA 167145 galaxy 0.153900 17.92 2.965
13 HERC-1:[MKK97] 307244 galaxy 0.311200 18.85 3.119
14 HERC-1:[MKK97] 309090 galaxy 0.312200 18.57 3.340

NED nearby objects search for UGC 10806.


Melnyk et al., 2009, arXiv, 0906.1493, “Search For Companions Of Nearby Isolated Galaxies”

SDSSJ092712.65+294344.0 – A high redshift copy of NGC 1275

Komossa et al. (2008) found two sets of emission lines from the spectrum of SDSSJ092712.65+294344.0. They described the object generally:

SDSS J0927+2943 at redshift z = 0.713 is a luminous quasar, observed in the course of the SDSS (Adelman-Mc-Carthy et al. 2007), and was found by us in a systematic search for active galactic nuclei (AGNs) with high [O iii] velocity shifts.

Then they described the spectrum:

Two systems of strong emission lines can be identified in the spectrum, which we refer to as the “red” (r) and “blue” (b) systems. … All lines of the blue system are blueshifted by about 2650 [km/s] relative to the [red system].

(I will use their names (“red system” and “blue system”) for the two emission line systems.)

They also noted:

The difference in velocity of the two sets of emission lines exceeds the peculiar velocities observed in galaxy clusters, and is too large for the two galaxies to form a bound merger. … Consequently, we would then have a very unlikely projection effect, including not just one, but two intrinsically extremely unusual AGNs…

They then suggested a hypothesis that in the past there has been a merger, where central SMBH would have recoiled ejecting the blue system from the core at the velocity of 2650 km/s.

Not a recoil but a massive black hole binary?

Dotti et al. (2008) considered the system as binary black hole system. They first mentioned some shortcomings of the recoil hypothesis of Komossa et al. (2009). Dotti et al. note that in the blue system the redshifts of both narrow- and broadband emission lines are the same, and that it is not possible that emitted gas would emit narrow emission lines. However, they also say that the narrow emission lines in the blue system are not such that are typically found in active galaxy nuclei. They also make some arguments relating to the probabilities in the system if it would be a recoil system. They then describe their hypothesis:

In our model SDSSJ0927 hosts a MBHB surrounded by a circum-binary thin disc feeding the secondary MBH. In the tidal interaction of the binary with the gas, a gap open, i.e. a low density region surrounding the binary.

They explain the blueshifted emission lines (the ones with lower redshift) followingly, first the broadband lines:

…while the b–BELs are produced in the broad line region bound to M2, and can be blue–shifted or red–shifted depending on the orbital phase of the secondary.

“Secondary” here refers to one of the hypothesized black holes. The explanation for the blueshifted narrowband lines:

The b–NELs come from the region of the gap near M2, where forbidden lines can be emitted because of the low density of the gas. SinceM2 and the gas orbiting in the gap at comparable annuli are subject to the same gravitational potential of M1, so that they move with approximately the same Keplerian velocity. Accordingly, the same blue–shift for the b–BELs and b–NELs can be explained if M2 emits an anisotropic ionizing radiation normal to the plane of the discs so that the ionizing photons interact only with the gas in the gap near to M2. In this case, the assumption of an accretion disc around M1 can be relaxed, and no red–shifted NELs are produced even if M1 is able to accrete on a short time–scale all its gas reservoir

But I have to admit that the explanation goes over my head, but I think the basic idea here is that the blueshifts can arise without the black holes having to have huge velocities themselves. They then go on to provide technical details of the system, when considered as a binary black hole system. They conclude:

SDSSJ0927 is clearly exceptional. Either it is the first example of a MBH ejected from its nucleus due to the gravitational recoil, or it harbours one of the smallest separation MBHBs ever detected.

Bogdanovic et al. (2009) suggested the same thing as Dotti et al. in a very similar paper. These two works were probably done in parallel to each other.

Another hypothesis

Heckman et al. (2009) provided some criticism for the binary black hole model:

In this model, the broad lines are made in more or less conventional fashion, even though the second, more massive black hole, is close enough that its gravity must influence the broad line gas—after all, its gravity is strong enough to force an orbital speed for the quasar black hole ≥ 2650 km s−1. The associated narrow lines are attributed to gas flowing between the circumbinary disk and the lower-mass black hole, and therefore traverse a region where, by construction, the characteristic orbital speed is & 2000 km s−1. It is hard to understand how, under these circumstances, the narrow lines could be centered on the velocity of the broad lines and have FWHMs as small as 450 km s−1.

They suggested another possibility for this system:

…that this system is a high-redshift analog to the familiar nearby Seyfert galaxy NGC 1275.

NGC 1275 is similar system that has two emission line systems separated by 3000 km/s, and it is currently thought to be an interacting galaxy pair within a galaxy cluster. The location within a cluster is important, because it makes the suggestion of high peculiar velocities involved more plausible. Heckman et al. (2009) pointed out that the NGC 1275 system and this new SDSSJ0927 system are almost exactly alike. So, in this system we would have a quasar at the redshift of the blue system, and in front of it there would be an infalling galaxy at the redshift of the red system. The higher redshift of the red system would be then caused simply by the velocity of the infall. They went on to make some arguments from the properties of the spectrum of the system, that they are fitting to this view, and they concluded:

Thus, in all respects that we can measure from the existing data, the properties of the redshifted narrow emission-line system are fully compatible with our hypothesis that SDSSJ0927 is a higher redshift version of the NGC 1275 system: a galaxy falling into the deep potential well of a rich cluster of galaxies where it interacts with the host galaxy of an AGN.

But what about the implied galaxy cluster, is it there? They made an initial analysis based on SDSS images of the field, and the photometric redshifts (which are very inaccurate) of the objects there. They found some galaxies that might suggest a presence of a massive cluster, but it is very uncertain. Another outstanding issue is to pinpoint the objects where the two sets of emission lines are produced. They approached this problem by looking at the closeup SDSS images, and found potential candidate for the red system in addition to blue system being the quasar itself. There is an apparent companion galaxy NE from the quasar that might overlap with quasar image. They suggested that further spectroscopy might be able to solve the question.

Shields et al. (2009) also suggested the same thing as Heckman et al. (2009). More from Shields et al. below.

Decarli et al. (2009) performed a photometric study in order to look for the galaxy cluster. They note about the situation:

Such a high velocity difference between the two galaxies is inconsistent with a simple on–going merger, and requires the deep potential well of a rich galaxy cluster.

They then proceeded to do a follow-up study for the Heckman et al. (2009) initial galaxy cluster search. Decarli et al. (2009) took new images of the system and base their study on those. They studied each object within a few square-arcmin field to see which ones are such that they could be a part of z = 0.7 galaxy cluster. They found 11 candidates. They compared the situation by virtually moving couple of known nearby galaxy clusters to the higher redshift (note that there are some corrections involved with this, see their paper for details), and they found that there should be at least 50 candidate objects expected to be detected within their search field in these imaginary cases. So, the simple coonclusion is that there just is not enough candidate objects within this field for massive enough galaxy cluster to be there.

Third emission line system

Shields et al. (2009) obtained a new spectrum of the system, which confirmed the basic properties known so far. In addition to that, they found a new emission line system at intermediate redshift compared to the two previous ones. They note:

The i-system offers an alternative candidate for the host galaxy of the QSO, or it may represent a close companion. In either case, the presence of a galaxy at a velocity close to the broad-line velocity lends credence to the possibility that the r-system may not represent the host galaxy and that the broad line system may be an ordinary QSO. The i-system could also represent gas ejected from the QSO.

They emphasized that finding absorption lines from the quasar host galaxy would be important. They then went on to make similar arguments of the system as Heckman et al. (2009) did, about the possibility of galaxy cluster from SDSS images and photometric redshifts. See also their discussion about other previously suggested explanations for the system (which were discussed above).

UPDATED (September 2, 2009): A new study by Vivek et al.

Vivek et al. (2009) published a new spectra and compared it to a four years earlier taken SDSS spectra. They didn’t find any detectable acceleration between the two spectra, which seems to make the binary black hole hypothesis quite unlikely. They found out that the red system is an extended region of size ~8 kpc, and that would seem to rule out the binary black hole hypothesis. They were also able to determine that the red system is in slightly different position than the blue system, and they concluded that the red system is separated from the quasar by 1 kpc. They couldn’t rule out the recoiling black hole hypothesis or the high redshift analog of NGC 1275 hypothesis. They also confirmed the existence of the third emission line set.

Figure 1. The objects with measured redshifts near SDSSJ092712.65+294344.0. Size of the image is 7 x 7 arcmin. Image is from Digitized Sky Survey (POSS2/UKSTU Red).

Some notes

There’s not much more to say of this system. Nearby objects within the field, shown in Fig. 1, show a line alignment containing objects 1, 4, and 6. Coincidentally, redshift of object 4 is quite accurately twice the redshift of object 6 (z6 / z4 = 1.96), redshift of object 1 is then 2.5 times the redshift of 4. Curious numerical coincidence, which doesn’t mean anything without decent statistical analysis of the situation. Another thing worth mentioning is that the object 7 seems to be very large to be such a high redshift object (see SDSS images), so large that I’m wondering if it’s a mistake in position or somewhere else. [UPDATE (November 23, 2009): My astronomer friend took a peek at the object 7 issue I mentioned and he showed me that the size of the object is not that special, there’s plenty of objects with similar or bigger size at redshift z ~ 0.1 (thank you, John, for the help).]

UPDATED: Here is the link to my website’s entry on NGC 1275

Objects and their data

1 SDSSJ092712.65+294344.0 “blue system” QSO em. line system 0.69798 18.4 (g) 0
2 SDSSJ092712.65+294344.0 “red system” QSO em. line system 0.71279 0
3 SDSSJ092712.65+294344.0 “intermediate system” QSO em. line system 0.7028 0
4 SDSS J092704.52+294401.6 galaxy 0.273648 20.4 (g) 1.788
5 SDSS J092706.01+294503.9 Sc 0.025972 16.5 (g) 1.962
6 SDSS J092656.35+294415.2 galaxy 0.139845 18.5 (g) 3.576
7 SDSS J092728.41+294641.1 Sb 0.103412 17.2 (g) 4.519

NED descriptions for the objects: objects 1-3, object 4, object 5, object 6, object 7.

SDSS object descriptions: objects 1-3, object 4, object 5, object 6, object 7.

SDSS image of the system


Bogdanovic et al., 2009, ApJ, 697, 288, “SDSS J092712.65+294344.0: Recoiling Black Hole or a Subparsec Binary Candidate?”

Decarli et al., 2009, arXiv, 0904.2999, “A photometric study of the field around the candidate recoiling/binary black hole SDSS J092712.65+294344.0”

Dotti et al., 2008, arXiv, 0809.3446, “SDSSJ092712.65+294344.0: a candidate massive black hole binary”

Heckman et al., 2009, ApJ, 695, 363, “SDSSJ092712.65+294344.0: NGC 1275 at z = 0.7?”

Komossa et al., 2008, ApJ, 678, 81, “A Recoiling Supermassive Black Hole in the Quasar SDSS J092712.65+294344.0?”

Shields et al., 2009, ApJ, 696, 1367, “Comment on the Black Hole Recoil Candidate Quasar SDSS J092712.65+294344.0”

Vivek et al., 2009, arXiv, 0909.0018, “SDSS J092712.64+294344.0: recoiling black hole or merging galaxies?”