Last time, I described our work on red spirals.  That’s it for the big results in our paper (*whew*!).

So what’s next?

Well, this was a pilot program, to see what was feasible, what a waste heat search is up against.  There are a few future directions to take:

1. We can try to distinguish natural emission from alien waste heat.  There are many ways to do this.  Two of them, broadly:
   • 

     The Great Galaxy in Andromeda, as seen by WISE. The red, MIR emission is clearly coming from dust in the spiral arms, not the smooth distribution of stars throughout the galaxy.  No obvious K3 here. Image Credit: NASA/JPL-Caltech/WISE Team

     Look at the morphology. Dust follows gas, which is dissipative; it clumps.  Stars tend to be found in a smooth distribution.  Since the waste heat we are looking for should trace the stars, we can use the morphology of the emission to determine its origin.  An elliptical bright in mid-infrared emission everywhere would be very hard to explain naturally.  If stars in a galaxy are well mixed, the most MIR-faint portion of a galaxy is the best measure of its upper limit on alien waste heat.

   • Loot at the whole spectrum. We can use the entire SED of the galaxy to estimate the mid-infrared emission we expect.  A comprehensive but simple model that accounts for star formation history, AGN, current star formation, and dust composition would have a finite number of free parameters that we could fit an SED to.  This is a more quantitative, sophisticated way of looking for MIR-bright galaxies that have no business being MIR-bright, and can be used on galaxies that WISE does not resolve well.
   

   Figure from our second paper. SEDs are for an old elliptical, a typical spiral, and Arp 220, a starburst galaxy. The green and orange curves include the effects of 10% and 35% of the starlight being reprocessed as waste heat. 

   These approaches will let us push our sensitivity to alien waste heat down from 50-85%, where we are now, to something more like 10-30% for spirals, and down to a few percent for dust-free ellipticals. There’s an order of magnitude improvement in sensitivity waiting to be picked up with just a bit more work.

2. We can get more quantitative. We just counted galaxies and expressed things in terms of the fraction of starlight consumed, but we can do better. With resolved objects our limits are really best in terms of surface brightness, and so we should be able to combine the above approaches to put limits of galaxies as a function of their stellar mass, and outputs in physical units (erg/s!) and relate things directly to Kardashev’s and Sagan’s scales.
3. We can look into K2’s—stars in the Milky Way with unnatural MIR excesses.  Right now, this is a difficult exercise because AGB stars, especially carbon stars, and things like obscured AGN show up as MIR-bright point sources, and without a spectrum there is very little to go on to weed them out.  This was the primary difficulty with  Carrigan’s IRAS search, which required painstaking, source-by-source analysis to pick through the AGB stars.But, GAIA is coming, and GAIA will provide parallaxes for anything with an optical counterpart.  So AGB stars and most AGN will have parallax information (that is, small or no measured parallax from GAIA, meaning they’re very distant).  This will allow us to construct a color-magnitude diagram with absolute K magnitude and some color like K-W3.  This is a great way to find debris disks… and that’s about it, because outside of star forming regions nothing else can really mimic a solar-luminosity star with a MIR excess.
4. Of course, we can follow up our weird no-optical-counterpart cluster, our new apparently-starbursts, and our 5-8 UV-faint red spirals.

We’ve got other ideas, too, but these are the obvious next directions for us to go.

It’s going to be fun!

Last time, I discussed the appendix of our paper wherein we examined low- (and no-) surface brightness galaxies, and found that they’re not dim because the stars are all covered in Dyson spheres.

Top: Ellipticals are red, spirals are blue.
Bottom: Except when they’re not. The Galaxy Zoo project identified anomalous galaxies, the blue ellipticals (left) and red spirals (right). Credit: Galaxy Zoo blog.

We also looked at galaxies with anomalous populations of stars. The Galaxy Zoo project identified a new class of galaxy — the “red spiral”.  They have optically red colors, like elliptical galaxies, but clearly spiral morphology.  This is strange: spirals usually have lots of blue stars that dominate their light; why do some lack blue stars?

Blue stars are young stars (they don’t live long), so they trace star formation.  Since red spirals aren’t blue, they are sometimes called “passive” spirals, because they are apparently not “active”ly forming stars.

What does this have to do with SETI?  Well, it has been suggested (and if anyone knows a citation, please send it to me!) that galaxy-spanning civilizations would suppress high-mass star formation, because those stars tend to explode as supernovae, and supernovae are bad for the environment.  One might expect, therefore, that a signature of an arbitrarily advanced civilization would be an unnatural “initial mass function” (basically, the number of stars formed at various masses) that lacked any high mass stars.

So, red spirals seem to fit the bill!  Except that there is another explanation for them:  the color of these galaxies also traces star formation history.  If you produced lots and lots of stars in the past, many of them today will be red giants (like in ellipticals) and if you have enough of those, their light could overwhelm the young blue stars.  So red spirals could just be regular spirals that have an unusually large number of red giants, not an unusually small number of young stars.

And indeed, that’s what Luca Cortese found in his study of the near UV emission of red spirals (which traces star formation). The color formed by comparing the near UV to the r-band emission is consistent with other spirals, even if the galaxies look red in the optical.  Red spirals don’t seem to be deficient in star formation (so they’re not “passive”, they’re just red).

Left: near-UV emission (normalized by r-band stellar emission) on the y-axis (down is more UV emission), mid-infared color on the x-axis (tracing dust and, potentially, alien waste heat; right is more dust/heat). The horizontal line represents a conservative threshold for “passive” galaxies, which are usually ellipticals or very heavily extinguished (by dust) spirals.  Right: same x-axis, but the y-axis traces the W1-W2 color, which should be larger in cases of very warm or very luminous dust/heat. The green dots are the UV-dim galaxies above the line in the left figure.

BUT… not all of them.  We identified eight red spirals from the Galaxy Zoo sample that have very little near UV emission — so they could be red because they lack blue stars.  What’s more, five of these have very high amounts of MIR emission for spiral galaxies, especially in the W2 and W3 bands compared to W1.  Now, some spirals are edge-on, so their UV emission can blocked by dust in the disk, which could explain the low UV emission.  But we checked and all five of these galaxies are at modest inclination, as far as we can tell.

So… maybe aliens?  Here are “red spirals” that seem to lack star formation, and have high, but not off-the-charts, MIR emission.  That’s weird.  In a sample as big as ours (we checked 5,500 red spirals) there are bound to be a few natural outliers, so we didn’t highlight these in our press release, but we do think they’re worth following up, both as SETI targets and as interesting natural objects.

Next time: Next steps!

In the last few posts, I discussed some of the natural sources we found with Ĝ. In this post, I describe some examples of how one could apply the waste-heat approach to “suspicious” galaxies.

In the appendix of our paper, we took a look at two classes of optically anomalous galaxies.  That is, galaxies whose light in visible wavelengths is strange.

The first class are the so-called HI-dark galaxies.  Most galaxies have lots of stars, and so are very bright and obvious.  If you were to enshroud the stars in a galaxy with Dyson spheres, you would lower its surface brightness and it would get darker.  If you completely enshrouded every star in Dyson spheres, it would go totally dark.

In order to create a noticeable decrement in the surface brightness of a galaxy, you would have to cover a lot of the light — around 75% to make it really anomalously faint.  This is the threshold that James Annis used in his trailblazing paper on the topic.  He studied a group of about 100 galaxies all at the same distance, and used the Tully-Fischer relation to determine, from the rotation speeds of the galaxies, how bright their stars should be (their stellar mass content, strictly speaking, then he assumed a mass-to-light ratio).  He then looked for any galaxies that were significantly dimmer than this value.  He found none that would be consistent with 75% coverage by Dyson spheres, similar to our limits from Ĝ.

There is also a whole class of galaxy called a low surface-brightness galaxy (LSB).  These are often very hard to detect because they are so dim.  In principle, these could be examples of the galaxies Annis was looking for.

Of course, if the stars were completely covered with Dyson spheres, the galaxy wouldn’t be dim, it would be black.  But then, you wouldn’t know where to look for them in the first place.  Except, that galaxies don’t just contain stars, they also contain gas, like neutral hydrogen, and such gas can be detected with radio telescopes.

VirgoHI21: a dark galaxy.  “A three-colour (i, r, and B bands) image of VIRGOHI 21 from the Isaac Newton Telescope overlaid with contours of neutral hydrogen density (green) from observations with the Westerbork Synthesis Radio Telescope.” Original image here. Credit Robert Minchin for figure and caption.

Radio surveys have discovered many galaxies by searching for their neutral hydrogen emission, including many in the Virgo galaxy cluster.  But a few of these galaxies are very strange: they don’t seem to have any stars at all!  This is weird, but not impossible: we don’t know why these galaxies don’t seem to have formed any stars with their gas, but there are ideas.  These are called HI “dark” galaxies (HI = “aitch one” = hydrogen in its first ionization state).

But are they Type III Kardashev civilizations?

We went and studied 100,000 galaxies for emission, and if any of the LSB’s or dark galaxies were really dim because alien civilizations were blocking 99% of their starlight, they would have stood out as the very reddest objects in our survey.  Indeed, we didn’t search only known galaxies, but we searched every extended object that was bright in the W3 band, exactly so that we wouldn’t be biased against optically faint galaxies like these.

But just to be sure, in the appendix we specifically targeted a bunch of HI dark galaxies from the literature, and made sure to break out the LSB’s (as identified by SIMBAD) in our search.

We found… nothing among the HI dark galaxies.  No mid-infrared emission.  It turns out we weren’t the first to look, but we were probably the first to look in a SETI context.  So the HI dark galaxies are still strange, but they’re not K3’s.

Figure 11 from our paper. The Low surface brightness galaxies (LSBs) are the blue stars in the upper right panel. They seem to be typical of ordinary galaxies (black points, upper left), and perhaps star-forming galaxies (black points, upper right). Galaxy classifications from SIMBAD.

The LSB’s , it turns out, aren’t especially bright in mid-infrared colors, and they are, indeed, dimmer than typical galaxies, on average. No surprises here: they’re just low on stars.

Next week: Red spirals!

Last time I discussed our discovery of the MIR nebula surrounding 48 Librae.

We also discovered another weird set of point sources.  Our formal search was for extended objects, but these appeared to be a single extended object in our initial cuts because they are so close together.  Also, the blending messed up our initial round of photometry, making them look redder than they really are.

So we really should have just tossed them, but they’re really strange objects for being so high above the Galactic plane (b=+11), and we kind of got caught up in figuring out what they are.

OK, so what am I writing about?  These things:

Mystery sources. Red=W4=22 micron emission, blue=W1+W2, green=W3. Those are some really red objects!

A cluster of 24-micron bright, red point sources.  No big deal, right?  But check out that same patch of sky in the optical:

Mystery sources in optical (B band) light. Nothing! Black sky!

Eric Mamajek. Chalkboard mathy guy FEPS team member, and now apparently a frequent appearer on this blog.

Nothing!  Totally black.  This is the expected signature of a cluster of Dyson spheres (which is not to say that a cluster of Dyson spheres is expected!).  What are these things?  We have a few more clues.  The first is that there is serendipitous MIPS imagery.  The FEPS survey (whose researchers include one of the AstroWright blog’s favorite chalkboard-mathy guys, Eric Mamajek) observed a bunch of young stars with Spitzer.  One of those stars just happened to be at just the right angle when IRAC was looking at it to have the MIPS camera just happen to land on our mystery objects!  Talk about luck.  Anyway, Spitzer/MIPS saw this:

Serendipitous Spitzer MIPS imagery of our mystery cluster.

So there is a lot of structure in there—many sources, including 4 that are labelled, at least three in that bright cluster to the upper right (NE), and maybe many more very faint ones all over the place.

So do we win? Are these Dyson spheres? Well, take a look at that bright NE cluster in the near-infrared (from 2MASS):

2MASS imagery of our mystery cluster. (Note that we’ve zoomed way in here on the bright cluster of sources).

You don’t expect room-temperature Dyson spheres to have emission at 1 micron (J-band) — they would have to be over 1000K for that.  This looks more like extinction than low temperature.

That is, dust blocks the blue light and transmits the MIR very well, with the NIR in between.  That’s what we see: lots from WISE, a bit from 2MASS, nothing from the visible.  If this were thermal emission there would be a much sharper falloff (exponential!) in emission from the WISE to 2MASS bands — 2MASS shouldn’t have seen anything.  Instead, we see a much slower decline towards the blue.

In other words, whatever is blocking the starlight is not totally opaque to NIR emission.  So, not a bunch of solar panels; more likely a lot of dust.  We thought about cosmology, but it’s hard to imagine a scenario that gives us multiple, IR-bright but optically invisible galaxies in a small patch of sky like this.  That means this is probably an embedded cluster of forming stars in the Milky Way.

There is a CO detection in this direction, which supports this conclusion, but the catalog entry is confusing (it looks like it’s been transposed with another entry in Yang et al. (2002)). If the velocity of the CO is correct, then the cluster is at about 850 pc away.  At b=+11, that puts it 170 above the Sun’s position in the Galactic plane, which is pretty high for a molecular cloud.

Anyway, there’s basically nothing in the literature about these things.  It’s not surprising that we would find a previously uncatalogued molecular cloud with embedded YSOs (or something) inside in our search, especially a small one so far away, so that’s probably what it is.

But still, we can’t be sure yet. What we really need is a spectrum.

So… if you’re an IR astronomer reading this, and you have a few spare minutes on a Northern Fall night with a spectrograph, do me a favor and grab me an H-band (or any band!) spectrum of WISE J043329.55+645106.5?  Let me know if they’re some kind of young star.

And if the spectrum looks featureless or otherwise inexplicable…call me right away!

Next time: Back to the SETI targets!

Last time I blogged about our bottom line on Kardashev Type III civilizations.  This time: science of natural sources.

One of the nice things about a search like the one we’re doing is that because we are exploring a new parameter space with a new instrument with orders of magnitude better sensitivity than previous instruments, we’re going to find some strange new things.

First, we found some MIR-bright galaxies that no one had ever noticed before.  They have almost no presence in the literature at all, despite having some of the most extreme MIR colors of any galaxies in the sky.  They are most likely nearby starbursts of various flavors, but until we check more carefully we can’t be sure.

Next, we found a huge nebula around the classical Be shell star 48 Librae.  Now, 48 Librae is very bright, and so is very well known among a certain class of stellar astronomer.  It is a Main Sequence B star (so very hot and massive) that exhibits very low surface gravity features because it is very rapidly rotating — nearly at breakup speeds.  This gives it a highly oblate shape, with the material at the equator nearly flying off.  Indeed, the star is rotating so quickly that it has an excretion disk.  This disk has material that is occasionally seen in absorption as it moves away from the star (the “shells”) and shines brightly in some emission lines (the “e” in “Be”).

48 Librae is known from the IRAS survey to have a midinfrared excess, but that’s not too unusual, since these stars’ excretion disks can have a lot of warm material shining in the MIR.  What is unusual is that the source isn’t the disk, it’s an enormous nebula around it:

The 22 micron Nebula around the Be shell star 48 Librae.

Eric Mamajek. Chalkboard mathy guy and Facebook scientist.

What’s going on?  To find out, I consulted some astronomers.  Mostly by email, but one, in particular, I consulted over Facebook.  Regular readers know that Eric Mamajek helped us figure out the coordinates of the CTIO 1.5m, and taught us astronomers a lot of about the various coordinate systems used to describe the location of things on the Earth.  Eric, true to form, responded to my request to figure out what sort of star 48 Librae is with a lengthy dive into the literature.

So I had to figure out how to cite him, of course.  Fortunately, Facebook has unique URLs for every post, and since my posts are public, that URL serves to point to Eric’s research.  Here’s how the citation looks in the paper:

How to cite a Facebook post

(AAS journals have a “no URLs” rule in the body (and bibliography, I think), but anything goes in the footnotes, so there it goes).

Anyway, we thought at first it must be a reflection nebula.  Lots of B stars, like those in the Pleiades, illuminate the dust around them.  Paul Kalas has a nice paper describing the “Pleiades phenomenon” they discovered around many hot stars in their search for substellar companions with adaptive optics.

In the Pleiades, this dust scatters the blue light of the stars very well, and the hard radiation from these stars both heats and excites the dust, causing it to give of thermal radiation, and the PAHs in the dust fluoresce in the WISE bands (especially W2 and W3).

The problems with this interpretation for 48 Librae are:

• There’s no blue scattered light here.  OK, whatever, maybe you just need more dust to do that?
• There’s no W2 emission, and hardly any W3.  OK, so not much PAH emission, maybe there aren’t many PAHs in the dust?
• The nebula doesn’t look like a randomly illuminated patch of dust. There are arc structures in it.  I fit the arcs by eye to a series of nested ellipses centered on the star, and found that they roughly match up:

I fit the ellipses in by hand using Keynote. They are concentric, centered on the star.

OK, so maybe it’s not illuminated by the star?  The other way to get these nebula is with a bow shock: the star moves faster through the gas than the local sound speed, and shocks it.  The shock heats the gas, which then glows.  The problem here is that 48 Librae isn’t moving very fast. It’s proper motion is quite small, in fact almost entirely due to the Sun’s motion through the Galaxy.  Plus, the nebula doesn’t really look like a bow shock.

So I wondered whether the nebula could be from 48 Librae itself?  The basic physics makes sense:  Be shell stars have excretion disks, and they are known to have winds that carry away material (we see the narrow, blueshifted metal lines in the spectra).  At some distance, this material becomes cool enough to condense into solids (like soot condenses in a flame, I’ve learned from Eric Feigelson). If this happens when the material is still at a sufficiently high density, it will form dust.

This explanation would explain the symmetries of the arcs in the nebula, and provide a mechanism for their emission.  The arcs are centered on the star because they are rings of ejected material, and they might be bright because they have been shocked during collisions either with previously ejected material, or the surrounding ISM.  They dust graints might also might have a size distribution that makes them fluoresce under the light of 48 Librae, but be inefficient at scattering the blue light that makes a reflection nebula.

If this is right, then we would expect the orientation of the ellipses I’ve drawn to correspond to the excretion disk.  I had email consultations with a bunch of Be star experts, including Stan Owocki and Thomas Rivinius, and they pointed me to some of the basics.  Since the disk is seen in absorption, it must be inclined no more than ~25 degrees from edge on, and my ellipses are inclined 20 degrees from edge on.  The actual position angle is known from interferometry to be 50±9 degrees from Štefl et al. 2012, consistent with our rings at 1–2σ, or within the rough precision with which we can define them.  Also, the nebula is not too big to be from the star: at the shell expansion velocity it would only take 30,000 years or so to get that big.

So it all checks out: the arcs are about right to be extensions of the excretion disk.  If this is correct (and I consider it a longshot) then Be stars could be important sites of dust formation, and we should reconsider some of the “Pleiades phenomena” as being something other than a coincidental pairing of stars and dust.

What fun!

Next time: a mysterious cluster of IR sources with no optical counterparts!