Thursday, August 21, 2008

Why is there nothing instead of something?

Following the laws of gravity, the matter in our universe forms a net of filaments. Voids mostly empty of matter also are a natural outcome of this structure formation. However, it turns out that numerical simulations based on the standard ΛCDM model yield voids that contain more galaxies than we observe. It seems that the voids in our universe are too empty. Though the predictions work well for giant voids, the situation is less clear on smaller scales [1].

[Picture credit: Suvendra Dutta]

The closest void in our galactic neighborhood is called the 'local void'. It is empty except for a lonely dwarf galaxy whose velocity indicates it's trying to move out of the void. That behavior is what one would expect to happen to underdense regions during structure formation. According to estimates following from ΛCDM simulations one would also expect there to be about ten dwarf galaxies in this void [2]. So where are they? Why is the void so empty?

However, what is usually computed in structure formation simulations is not the distribution of visible matter but that of dark matter, and our usual matter follows the dark matter's structures. Thus, what we actually know is that there are too many dark matter dwarf haloes in the simulations as compared to data. It might thus be the problem is not one with the void, but that dark matter haloes just failed to form galaxies. ΛCDM also predicts too many dwarf dark matter haloes as compared with the observed dwarf galaxies [4]. The solution to that puzzle might thus be on the cosmological side - in case there's something about structure formation we haven't yet got quite right - or on the astrophysical side - in case there's something about galaxy formation we haven't appropriately incorporated.

Tinker and Conroy [3] recently extrapolated the halo occupation distribution (the relation between dark matter halos and galaxies) into regimes in which observational data is lacking in order to model the distribution of dwarf galaxies. In doing so, they claim to be able to model the emptiness of voids, which would mean the explanation is on the astrophysical side. Tikhonov and Klypin [1] however point out that to explain the void structures, small haloes with circular velocity Vc > 20 km/s should not host galaxies, which however they do: they include a table with properties of observed isolated dwarf galaxies with circular velocities of about 20 km/s. Tikhonov and Klypin conclude

"We would like to emphasize that the disagreement with the theory is staggering. The observed spectrum of void sizes disagrees at many sigma level from the theoretical void spectrum if haloes with Vc > 20 km/s host galaxies brighter than MB = −12."

There is no bottomline to this post, I'm just trying to summarize some stuff I recently read. I'm still not entirely sure what to make out of the void problem, any comments are welcome.

[1] A. Tikhonov and A. Klypin "The emptiness of voids: yet another over-abundance problem for the LCDM model" arXiv:0807.0924v1 [astro-ph]
[2] P. J. E. Peebles, "Galaxies as a cosmological test"
arXiv:0712.2757v1 [astro-ph]
[3] Jeremy L. Tinker and Charlie Conroy "The Void Phenomenon Explained"
arXiv:0804.2475v2 [astro-ph]
[4] Strigari at al "Redefining the Missing Satellites Problem"
arXiv:0704.1817v2 [astro-ph]


Arun said...


Bee said...


Neil' said...

Hey, I missed the cute Arun/A leading into Bee/B -void series while composing my previous. Maybe you need someone whose name starts with "C" to have a tribute to the "Alpher-Bethe-Gamow" paper!

Arun said...

While thinking about the physics, I would also want to see what assumptions went into the simulation code.

Bee said...

Which one? The LCDM is as far as I know just GR with initial conditions from lin. pertubation theory. The numerical models investigated in [1] and [3] are well documented in the papers.

Neil' said...

BTW I do realize the context of the question is not about the ultimate "why" - but the distribution of matter. I confess I was tricked by the expectation effect into reading it as "Why is there something instead of nothing?" To make amends for pressing the tangential metaphysics, I give some links to alternative attempts to explain attractive behavior and grouping of matter in the universe. Some of these alternative concepts of gravity also claim to explain the Pioneer 10 anomaly. What I don't get is, how can they say that simple mod. of gravity (whether MOND or the newer attempts) is good enough, when I am sure I have seen graphics showing distributions of dark matter as such? It would have been a tip off if the DM densities just symmetrically followed regular matter. AFAIK, DM isn't one-to-one superimposed on matter density, true? (If it was, it would be equivalent to G just being bigger due to lack of "distinction", right?)

In any case, these serious (?) theories show that many investigators feel like looking "outside the box" to explain how our universe turned out.

Anonymous said...

i know some folks at arizona state who are actively involved with the 'where's the dwarf galaxies?' issue - other than that, though, i'm afraid i'm chillin in the dark as well...

Plato said...

For Arun's simulation definition requirement.

Study of structure formation in the Universe is an area of forefront research in astrophysics. The early evolution, when the seed fluctuations are small, can be calculated analytically on a piece of paper without the help of large supercomputers. As the fluctuations grow in their amplitude, the evolution becomes too complex and theorists have to use computers to follow the subsequent evolution.

Now some like Lubos might not like Veneziano's little tidbits of info at Strings 2008, but I would like to compare again his perspective to this looking from "outside the box."

Sean Carroll may not like my comparison to his current assessments of the universe with regard to such topological arrangement. Yet today this maybe relevant from today's perspective while very closely such a string in the formation of the universe less then acceptable can be seen in the WMAP?

But outside the box had to be thought of in relation to what is initiated in the very beginning has some "global perspective" WMAP as the universe unfolds?:)


Andrew Thomas said...

Yeah, New Scientist did an article with a stringy/anthropic theory about the void.

Arun said...

Do 10 blobs of 10^7 solar masses each behave the same (on this scale of resolution) as 10^6 blobs of 10^2 solar masses each? I'm sure someone has done the simulations to check this...

The physics here is of a LCDM with matter consisting purely of black holes. :)

stefan said...

If at some early time, the repulsion now attributed to the "Lambda" was not homogenous/isotropic, because maybe driven by some field that now has diluted out, would that not also have resulted in voids? Are there studies or simulations that try to model this?

Cheers, Stefan

Bee said...

Hi Stefan,

The Lambda term is by definition constant. Even if it wasn't but some more general dark energy component, we know that these contributions have only become important rather recently as observations show. They don't play any role in the early universe. Reason is roughly: the cc term is constant, but the matter term falls like 1/a^3, if they are about the same order of magnitude today, then the dark energy contribution was irrelevant in the early stages. If you want to avoid a similar conclusion for more complicated sorts of dark energy you'll have to fiddle around with the equation of motion quite a lot (and get some sort of tracking solution, in which the densities remain the same order of magnitude). But more generally yes, I would guess that there are some modifications of LCDM that potentially work better with the voids. Somewhat surprisingly though, people don't seem to have examined the theoretical cosmology side on that problem very much. I mean, compared to other deviations from the standard paradigm, like non-gaussianities, scale invariance etc. Best,


Bee said...

Hi Neil,

Thanks for your self-correction. I was very tempted to reply with a comment saying not only didn't you read more than the title, you even misread the title. You first comment was interesting nevertheless ;-)

Thanks also for the links to the articles. I am generally very skeptical about modifications of GR. I too would hope that at the very least one needs only one kind of dark 'stuff', so that attempt seems worthwile. Best,


Bee said...

Hi Andrew,

I'm afraid we're talking about different issue here than was topic of the NewScientist article you mention. I'm not referring to a specific region in the CMB data, I'm referring to the large scale structure of voids. As I mentioned especially the smaller ones seem to be off the predictions. Best,


Bee said...

Hi Anonymous,

That's interesting. Can you give me a name/reference by any chance?


bellamy said...

"Neil' said...
BTW I do realize the context of the question is not about the ultimate "why" - but the distribution of matter."

But I thought it was perhaps implicit there is some sort of relation (or confluence?) least for there to be a successful GUT, mm?

Neil' said...

Bellamy, I was getting at the ultimate question of why anything at all and the problem of existential "favoritism" in one or some "possible worlds" existing (whatever that means, heh) and other PWs not existing etc. I put a big ramble about that earlier which I removed but will resurrect at my own blog soon.

Those kind of totally abstract musings don't answer questions of what happens too well, but the next lower (or higher, dep. on how you count it) level is about things like symmetry and does the universe try to be beautiful etc. Well, the current situation is rather theoretically ugly (no real way to harmonize the current level of DE with other theory, and the wrong prediction of vacuum energy of ten to the dozens power (I forget just what) grams/cm^3 instead of the current low value, etc.)

BTW for anyone who didn't know, the units of Lambda or DE are a/r (acceleration/distance), so it's m/s^2/m which gives MLT of T^-2. That is interesting because we can derive a time from lambda, and so I wonder: what is the ratio of Lambda/Age of universe (the latter ~13.5 x 10^9 years.) and is that "significant" in anthropic terms? It should be interesting, what the universe looks like at t = Lambda^-1/2.

Anonymous said...

I would bet that the problem is related to the finite nature of numerical simulation

Neil' said...

This is a really interesting creative concept about why the "Void" (WITNIOS! - this is the really big Giant Void and not just the spongy voidlets), involving the ever more popular "other universe/s" and referencing hip new observations like "the axis of evil", from our adventurous friends at New Scientist:

Bee said...

Hi Anonymous,

Possibly, I don't know very much about the numerics going into this kind of simulation admittedly. The resolution of the current state of the art (e.g. the Millenium simulation) is however much better than the typical size of the voids. This isn't a matter of resolution then, there had to be something really wrong with the clustering. I find this kind of unlikely given that it works so enormously well for the formation of other structures. Best,


Arun said...


(of ideas about dwarfs).

Sneezy, Sleepy, Dopey, Doc, Happy, Bashful, Grumpy.

Anonymous Snowboarder said...

Hi Bee -

Perhaps dark matter is wrong?

I know you will be happy to hear this Bee - I know of a blog post which states that the LHC and its black holes have made it into a radio ad for a car dealership! They were offering an 'end of the world' discount ...

Arun said...

How about an anthropic explanation?

Plato said...

You might like Stuart Kauffman's "Order for Free?":)

Real Events cannot be ignored in a real universe.

Colors represent the tiny temperature fluctuations of the remnant glow from the infant universe: red regions are warmer and blue are cooler.

Rae Ann said...

I'm sorry that I don't speak the proper scientific language, but I wonder if the discrepancies between the observations and the numerical simulations is something as simple as that we really aren't observing the present conditions of the universe. Maybe if we could actually 'see' the universe as it is 'now' then it would have more galaxies, etc, and less void than it looks like to us. Of course I don't know what processes could transform all the 'dark stuff' we see now into actual matter/galaxies, etc., but I'm just creating my own stupid answer to your question about why there is so much nothing instead of something. Maybe what looks like 'acceleration' now is really just the evidence of whatever processes that are making more matter out of dark, invisible stuff? Oh, nevermind me since I don't know anything anyway. ;-)

Bee said...

Hi RaeAnn,

It is true that the pictures one obtains using simulations (like the one shown in the post, or the Millenium Simulation) do not actually show what we 'see' since it takes a long time for light to travel from e.g. the voids to us. But it isn't hard to find out which stage during the formation of structures we do see, and the problem doesn't only appear 'after' that. You are right though that some things about structure formation in the early universe are puzzles as well, e.g. I just recently came across this article about some weird kind of galaxy in the early universe that seems to have entirely disappeared by now. Best,


Arun said...

Maybe this will be a meaningful comment :)

From here

Recent work on the accurate modeling of the mass function suggests that the mass density of stars and brown dwarfs at the Galactic plane is about 0.05 solar masses per cubic parsec.

Per [1], Tikhnov, Klypin, simulations S1, S2, S3 have particles of 10^6 to 10^8 solar masses in spatial resolutions of around 0.5 to 1.5 kiloparsecs.

My quick estimate says that the simulations thus have minimum densities zero and then 0.035, 0.184, 0.004 solar masses per cubic parsec.

If voids actually have lots of uncondensed matter at less than galactic density, then only simulation S3 can model that; S1 and S2 require the minimal object to have galactic density or higher.

Christine said...

Well, I could be C-void...

a quantum diaries survivor said...

Interesting post and enlightening discussion (apart from the x-void series).

I would like to add that indeed, from the outside (i.e. somebody with little or no background in cosmology, trying to grasp things here and there), these simulations appear like a piece of magic. Beware of simulations tuned to reproduce what is known. The temptation to use them to predict the unknown is huge once one has found the right tweaking of a few handles to make them look in pretty agreement, but they consistently fail outside the range where they've been tuned.

Just my two dimes (sorry, the dollar is so low...)

Bee said...

Hi Tommaso,

There is certainly truth in these two dimes. There is an aspect of this which concerns me occasionally, that is that due to increasing specialization and usually bad documentation at least I find it extremely hard to find out what actually goes into a numerical analysis or simulation. Not specifically for structure formation (I haven't looked into that very closely, at least not yet), but I've come across the same problem in other areas (try to find out data reconstruction at KamLand eg). I find it very unfortunate that it becomes for those on the theoretical side increasingly difficult to figure out what were the assumptions that were made to obtain a specific result. I don't know if that's on purpose or a result of bad publishing habits. Either way, it makes it increasingly difficult to figure out where a problem might originate. Best,


PS: I think you're underestimating the value of the x-void series.

Phil Warnell said...

I didn’t even realize they have expectations to any large degree about these so called holes. All I was aware of is the cosmic background data shows a relatively even distribution at low resolutions with some gaps at higher res. I find it somewhat strange that the concentration of dark matter should form such a focus when it’s not been even remotely assured what it is. I would think it wise to first settle this as to be more familiar with the overall properties before zooming off to speculate what role it plays in large scale structure formation. All this business where you have the most dominant form of matter not being to be able directly observed in part as most of it sits not in our grasp yet being beyond the edge the visible galaxy(s) smacks to much of Ptolemy for me with trying to salvage a bad idea with an even worse solution.

“From man or angel the great Architect
Did wisely to conceal, and not divulge
His secrets to be scanned by them who ought
Rather admire; or if they list to try
Conjecture, he his fabric of the heav'ns
Hath left to their disputes, perhaps to move
His laughter at their quaint opinions wide
Hereafter, when they come to model heav'n
And calculate the stars, how they will wield
The mighty frame, how build, unbuild, contrive
To save appearances, how gird the sphere
With centric and eccentric scribbled o'er,
Cycle and epicycle, orb in orb.”

-From Milton’s “Paradise Lost


Phil P.S. Perhaps three is the charm:-)

Arun said...

Crushed by Tommaso.

That has left a big void in my heart.

Arun said...

I believe the modern title of that work is "Paradise misplaced somewhere in 10^500 universes, but finding it is just around the corner".

Phil Warnell said...

Hi Arun,

“I believe the modern title of that work is "Paradise misplaced somewhere in 10^500 universes, but finding it is just around the corner".”

That could be and if I get the opportunity I will ask Brian Greene who will be the first PI public lecturer this Sept. 25th. I always wanted to ask him if the other of the 10^500 universes are as elegant as our own:-)



a mathematician said...

"Following the laws of gravity, the matter in our universe forms a net of filaments."

It is intuitively clear to me that gravity would cause matter to clump together in some way, but I never would have guessed this filamentary structure.

Why is there this structure, and how generic is this conclusion?

Should I find this surprising?

Arun said...

I'd love to hear Greene's answer, but I expect that it will be that we live in the most elegant of the 10^500 universes. And the anthropic principle is correct because intelligent life will only arise in a universe where the anthropic principle is conceivable and therefore correct.


Neil' said...
This comment has been removed by the author.
Neil' said...

Well, some comments here about multiverses imply it would be on-topic after all for me to explain just why there must be either multiverses and/or "management" to select them, instead of an inexplicably still-popular attitude" "This universe just exists, and that's that." See, WITSION (the usual question, Bee's cute rephrasing being the more direct query here) is not ultimately about applying already-granted laws to their outcomes. It goes way under that, as an ultimate "metaphysical" question about necessary versus contingent existence/being/s, the "principle of sufficient reason" applied to fundamental conditions (why, if there is a universe, is it like this and not some other way to be? etc.), modal realist critiques about whether the very notion of "really existing" as a "material world" is even logically coherent versus all descriptions being equally "real" etc.

Modal realists or nearly, like Max Tegmark have a point: really, it is logically clumsy for a universe like ours to "exist" but not all the others with other properties to exist also. Forget "common sense" or just blowing it off like Sean at CV does - one possible universe existing and not others is just like if the number 23 was reified in brass numerals etc., but not other numbers. Hey, they are all numbers, who or what can "pick out" one or some of them to be "real" instead of the others? Being "real" is too fundamental of an abstract property to logically attach only to universe/s with a restricted set of peculiar principles. (In logic, "peculiar" means properties that aren't logically necessary, like a fine structure constant around 1/137 - and I mean unrestricted "logically necessary" with no cheating based on circular reasoning that recycles what's already like that.)

As far as I'm concerned, either "everything exists" (which causes problems I can explain better later) or "someone/thing" somehow "manages" what exists and what does not. And in that case, even the very huge stringy "landscape" referenced above can't easily explain its "reification" among every describable possible way to be, in an abstractly fundamental sense.

This conference Bee linked to up-blog would be interesting, I wish I could go: Well Bee, maybe you can show them my post and see what they think. (You can tell people that my blog shows up in the top ten Google for "quantum measurement paradox", such as #4 just now FWIW.)

Christine said...

"I never would have guessed this filamentary structure"

It's been some time since I followed the literature, but there are some theories for this, such as the anisotropic merger scenario [e.g. West, M. J., MNRAS, 1994, 268, 79], starting with the formation of dominant galaxies in the center of clusters with elongations that coincide with the filamentary structures in larger scales. I discussed a bit of this question in the case of A3558 cluster of galaxies [see here]. But there must be new discussions on this problem, it's just that I did not follow it for many years now.

Bee said...

Hi Christine,

Of how much relevance is the underlying FRW expansion for these specific structures? I mean, if it was just collapse in a static background, would one get these filaments as well? Best,


Christine said...

Hi Bee,

No doubt that expansion makes a lot of difference in terms of the statistical mechanics of gravitating systems. As Padmanabhan put it -- a "civilising influence" on the collapse. You might want to take a look at his paper Nonlinear Gravitational Clustering: dreams of a paradigm. However, specifically in terms of the filaments, I don't know. Actually, I guess these filaments are somewhat more or less like projections of surfaces of bubble-like structures which envolve the voids. But, as I said, I'm not updated with the literature, as you can see I'm somewhat 10 years late.

Anyway, structure formation is quite puzzling, and as far as I know, one cannot run one simulation alone that agrees with the observed correlation function at all scales. Dark matter bias is also a factor to consider. But if the subject did not advance since the last time I saw it, definitely there are some important ingredients missing in our understanding.


bellamy said...

Bellamy, I was getting at the ultimate question of why anything at all and the problem of existential "favoritism" in one or some "possible worlds" existing (whatever that means, heh) and other PWs not existing etc. I put a big ramble about that earlier which I removed but will resurrect at my own blog soon."

1. I don't ask such a question. It has no meaning, and it'll be interesting if (or perhaps when) physics realises this.

2. anthropism is delusion - or, in my own words, distraxion. (I don't assume free will.)

3. I think my question is more fundamental.

Neil' said...

Bellamy, it may be that the questions (why something instead of nothing etc.) are "meaningless", but just claiming such questions are, doesn't show that they are. Meta-talk claims about what is or isn't meaningful and why are just as metaphysically mushy and debatable as the direct metaphysical questions, that's the irony. Yes I have seen arguments for "falsifiability" etc. but find them lacking in many ways, which I could discuss briefly if you want.

Thomas D said...

... Concerning the question halfway down this thread about, does dark energy make a difference, people at Heidelberg have been looking for some time at the impact of 'Early Dark Energy' which is precisely a tracking behaviour where there is a certain fraction (maybe a few percent) of homogeneous dark energy in the early Universe.

This means that structure has to form earlier, because the dark energy component makes it harder for structure to form but we know how much structure there is just at present. So by looking at high-redshift tracers of dark matter structure - eg Lyman alpha forest - you can potentially measure early dark energy. See authors like Doran, Robbers...

However I don't know that this really has a significant effect on the exact topology of voids or what is inside them. Small scale stuff where the discrepancies are tends to be messy and depend on complications like galaxy formation.

Dr BDO Adams said...

This is big news, most of the papers
I've read on dark energy, usually end with the data being consistent with LCDM, but here we have a simulation, that is not only a little inconsistent with LCDM but very inconsistent. Perphaps we can finally do away with the cosmological constant and find a tangible form or substance for dark energy. I hope there re-do there sim for various models of varying dark energy, so that we can find how much cosmic acceleration has changed over time, d^3 H /dt^3 etc. Can anyone tell me off the top of the heads, whether voids being emptier than LCDM, means dark energy has stronger or weaker, earlier in the history of the universe?

Plato said...

Layman perspective.

There is always the push/pull understanding to the universe. Some do not recognize it's geometrical propensity to what constitutes it's movement "at times" in it's expression(anomalies in the speeding up of the universe?) might have contributed too, from events "within" that very same cosmos?

The map reveals a loose network of dark matter filaments, gradually collapsing under the relentless pull of gravity, and growing clumpier over time. This confirms theories of how structure formed in our evolving universe, which has transitioned from a comparatively smooth distribution of matter at the time of the big bang. The dark matter filaments began to form first and provided an underlying scaffolding for the subsequent construction of stars and galaxies from ordinary matter. Without dark matter, there would have been insufficient mass in the universe for structures to collapse and galaxies to form. See:Three-Dimensional Distribution of Dark Matter in the Universe

Just a thought. At 10 500 such examples are to be contained not as the universe as a whole, but from the perspective that such events are styled "in the universe," contribute to the greater whole?:)


Plato said...

What is your configuration space?

Please forgive the layman perspective here and intrusion. To not think it a sign to attribute a deterioration of this site.

While I am ignorant to a large degree, it does not invalidate, while uneducated according to a criterion, still allows mental faculties to push forward given knowledge. that other's questions and opinion could be of value "by happen stance" too?:)

While not knowing Lee's full intention by previous comments, I hope to see soon, under the heading of "new blog entry" a written work, to combat thoughts about "anonymity" with the understanding, that others in society still remain interested, persevere, where education is concentrated, even with a fake name.


Anonymous said...

looking outside the box may be a good idea. There is some question in mind that our current understanding of gravity relies on empirical observation of it's effect. Though this is very clear I paused to wonder exactly what was meant when someone described the effect of planets on "space / time", what was it a gravity well, not that I like that particular description. I would tend to think of it as a hole (an empty hole)
hmmm brings to mind an interesting speculation should they find the Higgs it will not be where it is meant to be - more rather than less
Trevor - means nothing, 1 month