Scientists find giant wave rolling through the Perseus galaxy cluster

May 2, 2017 by Francis Reddy
This X-ray image of the hot gas in the Perseus galaxy cluster was made from 16 days of Chandra observations. Researchers then filtered the data in a way that brightened the contrast of edges in order to make subtle details more obvious. An oval highlights the location of an enormous wave found to be rolling through the gas. Credit: NASA's Goddard Space Flight Center/Stephen Walker et al.

Combining data from NASA's Chandra X-ray Observatory with radio observations and computer simulations, an international team of scientists has discovered a vast wave of hot gas in the nearby Perseus galaxy cluster. Spanning some 200,000 light-years, the wave is about twice the size of our own Milky Way galaxy.

The researchers say the wave formed billions of years ago, after a small galaxy cluster grazed Perseus and caused its vast supply of gas to slosh around an enormous volume of space.

"Perseus is one of the most massive nearby clusters and the brightest one in X-rays, so Chandra data provide us with unparalleled detail," said lead scientist Stephen Walker at NASA's Goddard Space Flight Center in Greenbelt, Maryland. "The wave we've identified is associated with the flyby of a smaller cluster, which shows that the merger activity that produced these giant structures is still ongoing."

A paper describing the findings appears in the June 2017 issue of the journal Monthly Notices of the Royal Astronomical Society.

Galaxy clusters are the largest structures bound by gravity in the universe today. Some 11 million light-years across and located about 240 million light-years away, the Perseus galaxy cluster is named for its host constellation. Like all galaxy clusters, most of its observable matter takes the form of a pervasive gas averaging tens of millions of degrees, so hot it only glows in X-rays.

A wave spanning 200,000 light-years is rolling through the Perseus galaxy cluster, according to observations from NASA's Chandra X-ray Observatory coupled with a computer simulation. The simulation shows the gravitational disturbance resulting from the distant flyby of a galaxy cluster about a tenth the mass of the Perseus cluster. The event causes cooler gas at the heart of the Perseus cluster to form a vast expanding spiral, which ultimately forms giant waves lasting hundreds of millions of years at its periphery. Merger events like this are thought to occur as often as every three to four billion years in clusters like Perseus. Credit: NASA's Goddard Space Flight Center

Chandra observations have revealed a variety of structures in this gas, from vast bubbles blown by the supermassive black hole in the cluster's central galaxy, NGC 1275, to an enigmatic concave feature known as the "bay."

The bay's concave shape couldn't have formed through bubbles launched by the black hole. Radio observations using the Karl G. Jansky Very Large Array in central New Mexico show that the bay structure produces no emission, the opposite of what scientists would expect for features associated with black hole activity. In addition, standard models of sloshing gas typically produced structures that arc in the wrong direction.

Walker and his colleagues turned to existing Chandra observations of the Perseus cluster to further investigate the bay. They combined a total of 10.4 days of high-resolution data with 5.8 days of wide-field observations at energies between 700 and 7,000 electron volts. For comparison, visible light has energies between about two and three electron volts. The scientists then filtered the Chandra data to highlight the edges of structures and reveal subtle details.

Next, they compared the edge-enhanced Perseus image to computer simulations of merging galaxy clusters developed by John ZuHone, an astrophysicist at the Harvard-Smithsonian Center for Astrophysics in Cambridge, Massachusetts. The simulations were run on the Pleiades supercomputer operated by the NASA Advanced Supercomputing Division at Ames Research Center in Silicon Valley, California. Although he was not involved in this study, ZuHone collected his simulations into an online catalog to aid astronomers studying galaxy clusters.

"Galaxy cluster mergers represent the latest stage of structure formation in the cosmos," ZuHone said. "Hydrodynamic simulations of merging clusters allow us to produce features in the hot gas and tune physical parameters, such as the magnetic field. Then we can attempt to match the detailed characteristics of the structures we observe in X-rays."

This animation dissolves between two different views of hot gas in the Perseus galaxy cluster. The first is Chandra's best view of hot gas in the central region of the Perseus cluster, where red, green and blue indicate lower-energy to higher-energy X-rays, respectively. The larger image incorporates additional data over a wider field of view. It has been specially processed to enhance the contrast of edges, revealing subtle structures in the gas. The wave is marked by the upward-arcing curve near the bottom, centered at about 7 o'clock. Credit: NASA/CXC/SAO/E.Bulbul, et al. and NASA's Goddard Space Flight Center/Stephen Walker et al.

One simulation seemed to explain the formation of the bay. In it, gas in a large cluster similar to Perseus has settled into two components, a "cold" central region with temperatures around 54 million degrees Fahrenheit (30 million Celsius) and a surrounding zone where the gas is three times hotter. Then a small galaxy cluster containing about a thousand times the mass of the Milky Way skirts the larger cluster, missing its center by around 650,000 light-years.

The flyby creates a gravitational disturbance that churns up the gas like cream stirred into coffee, creating an expanding spiral of cold gas. After about 2.5 billion years, when the gas has risen nearly 500,000 light-years from the center, vast waves form and roll at its periphery for hundreds of millions of years before dissipating.

These waves are giant versions of Kelvin-Helmholtz waves, which show up wherever there's a velocity difference across the interface of two fluids, such as wind blowing over water. They can be found in the ocean, in cloud formations on Earth and other planets, in plasma near Earth, and even on the sun.

"We think the bay feature we see in Perseus is part of a Kelvin-Helmholtz wave, perhaps the largest one yet identified, that formed in much the same way as the simulation shows," Walker said. "We have also identified similar features in two other , Centaurus and Abell 1795."

The researchers also found that the size of the waves corresponds to the strength of the 's magnetic field. If it's too weak, the waves reach much larger sizes than those observed. If too strong, they don't form at all. This study allowed astronomers to probe the average magnetic field throughout the entire volume of these clusters, a measurement that is impossible to make by any other means.

Explore further: The arrhythmic beating of a black hole heart

More information: S. A. Walker et al. Is there a giant Kelvin–Helmholtz instability in the sloshing cold front of the Perseus cluster?, Monthly Notices of the Royal Astronomical Society (2017). DOI: 10.1093/mnras/stx640

Related Stories

The arrhythmic beating of a black hole heart

April 19, 2017

At the center of the Centaurus galaxy cluster, there is a large elliptical galaxy called NGC 4696. Deeper still, there is a supermassive black hole buried within the core of this galaxy.

Colliding galaxy clusters

December 5, 2016

Galaxy clusters contain a few to thousands of galaxies and are the largest bound structures in the universe. Most galaxies are members of a cluster. Our Milky Way, for example, is a member of the "Local Group," a set of about ...

ALMA's ability to see a 'cosmic hole' confirmed

March 17, 2017

Researchers using the Atacama Large Millimeter/submillimeter Array (ALMA) successfully imaged a radio "hole" around a galaxy cluster 4.8 billion light-years away. This is the highest resolution image ever taken of such a ...

Astronomers discover powerful cosmic double whammy

January 5, 2017

Astronomers have discovered a cosmic one-two punch unlike any ever seen before. Two of the most powerful phenomena in the Universe, a supermassive black hole, and the collision of giant galaxy clusters, have combined to create ...

Hubble showcases a remarkable galactic hybrid

March 3, 2017

This NASA/ESA Hubble Space Telescope image showcases the remarkable galaxy UGC 12591. UGC 12591 sits somewhere between a lenticular and a spiral. It lies just under 400 million light-years away from us in the westernmost ...

Image: Hubble showcases a remarkable galactic hybrid

March 6, 2017

This NASA/ESA Hubble Space Telescope image showcases the remarkable galaxy UGC 12591. UGC 12591 sits somewhere between a lenticular and a spiral. It lies just under 400 million light-years away from us in the westernmost ...

Recommended for you

Dawn of a galactic collision

December 14, 2017

A riot of colour and light dances through this peculiarly shaped galaxy, NGC 5256. Its smoke-like plumes are flung out in all directions and the bright core illuminates the chaotic regions of gas and dust swirling through ...

13 comments

Adjust slider to filter visible comments by rank

Display comments: newest first

Jason Chapman
not rated yet May 03, 2017
Isn't this a plot line from Star Trek Generations
Da Schneib
5 / 5 (3) May 03, 2017
Interesting on several levels. Here's one of them: the class of algorithms used to detect this feature are called "edge detection" algorithms, and several different ways have been found to skin this particular cat: https://en.wikipe...etection
antialias_physorg
5 / 5 (7) May 03, 2017
Here's one of them: the class of algorithms used to detect this feature are called

There's a lot more than those. Edge detection is a big thing in medical image analysis. Each year new algorithms are published.
...and I rolled my own (as part of my PhD thesis), to detect edges within the bone (growth plates).

Example of what a growth plate in the knee looks like when it's easily visible (the annotated lighter parts within the bones on the second (small image)):
https://www.quora...ding-age

Example how hard they can be to see (though the algorithm would still able to detect it): First image on the right
http://www.bonesc...html/151

In some cases they even appear to be partially invisible, but with a continuity constraint the algorithm can still find them. (if anyone is interested I can give the paper whichdescribes the alogorithm in more detail)

Captain Stumpy
5 / 5 (2) May 03, 2017
In some cases they even appear to be partially invisible, but with a continuity constraint the algorithm can still find them. (if anyone is interested I can give the paper whichdescribes the alogorithm in more detail)
@AA_P
i'm interested... can ya e-mail the paper ?
thanks!

.

PS - second link is dead (site off-line)
jeremysanders
5 / 5 (5) May 04, 2017
The method and initial image is in our paper: arXiv:1605.02911 (Detecting edges in the X-ray surface brightness of galaxy clusters, Sanders et al 2016). It basically uses a Gaussian gradient magnitude filter to measure the gradient on various different scales. The image is convolved with the gradient of a Gaussian on some scale, both in x and y. The total gradient magnitude is calculated from these. Then we combine together the gradient on different scales.
antialias_physorg
5 / 5 (3) May 04, 2017

PS - second link is dead (site off-line)

Weird. Works for me.

The image is convolved with the gradient of a Gaussian on some scale, both in x and y.

Is there an adjustment for the angle of the major plane of the waves with respect to the projection plane? This may have some minor impact on the gradients in x and y.
Da Schneib
5 / 5 (1) May 04, 2017
Thanks, @jeremy and @antialias. It's very interesting to see how these types of filters are used in practical applications.
antialias_physorg
5 / 5 (2) May 04, 2017

The image is convolved with the gradient of a Gaussian on some scale, both in x and y.

Is there an adjustment for the angle of the major plane of the waves with respect to the projection plane? This may have some minor impact on the gradients in x and y.


Adding some context: In my work I was dealing with medical CT images that had different resolutions in x/y and z direction. Which mean that any gradient approach would be skewed if it weren't adapted to deal with this.
(Luckily the positioning for the structure I was dealing with was well defined in x/y/z due to the way these images are aquired: the patient is lying down with straight legs so the growth plates are, almost, normal to the z axis...so in the end I didn't have to adjust the algorithm)

antialias_physorg
5 / 5 (2) May 04, 2017
It's very interesting to see how these types of filters are used in practical applications

There's really a gazillion of these algorithms. Even though there are a few generic ones - which you will find implemented in photoshop, etc. - in science it's usually better to look at the specific problem and tailor the algorithm to take advantage of any additional information you have about the problem domain.

Even though it upsets the 'artist' in me. I always yearn for this one 'best' algorithm. But I had to learn the hard way that problem specific solutions beat general solutions - always.
antialias_physorg
5 / 5 (2) May 04, 2017
....aaaand just as I was typing it this article popped on the frontpage of phys.org:

https://phys.org/...cut.html
The authors hope that by mathematically proving the futility of universal detection algorithms, they can, according to Larremore "free people up to work on specialist algorithms."


Yep.

jeremysanders
5 / 5 (2) May 04, 2017
antialias_physorg: We have the same spatial resolution in x and y. The technique here is being used to reveal structure - we're not using it very quantitatively. The main problem is the noise in the data, caused by the low number of photon counts.
antialias_physorg
5 / 5 (2) May 04, 2017
The thing I was wondering was: if the real structures are tilted with regards to the projection plane - and assuming you know the tilt - then using a convolution kernel that is not the same size in x and y could enhance results. (We have the same issues when we analyse x-ray images where the structure of interest is acquired at an angle)

For a qualitative picture that's certainly not important. For quantitative results it might be.

Noise is always a doozy. Simulating detector noise helps somewhat, but with a low S/N ratio there's often just not much you can do. Information that's not there is just...not there. I've seen algorithms that are geared to provide a better 'image impression'. But that always comes with a risk of increasing false positives.
Captain Stumpy
5 / 5 (1) May 04, 2017
Weird. Works for me
@AA_P
it's back up and running now - i guess the site was down for maintenance

thanks for the other!

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more

Click here to reset your password.
Sign in to get notified via email when new comments are made.