# Type Ia Supernova not Standard Candles? I'm confused .

1. Feb 4, 2015

### Garth

Type Ia Supernova are generally thought to be white dwarfs that, either by accretion of mass from a companion (single degenerate), or as a result of a merger with a white dwarf companion (double degenerate), approach and then exceed 99% of the Chandrasekhar limit of about 1.44 M, at which point the star begins a unstable process of carbon fusion that leads to complete deflagration and detonation of the star in a supernova.

Because the detonation mass is more or less fixed the SNe are all about the same luminosity and they can be used as standard candles. They do vary a little and the Phillips relationship has been derived by observation, which relates the peak luminosity to the speed of luminosity evolution after maximum light.

The characteristics of the Type 1a is that its spectrum is rich in carbon and oxygen and deficient in hydrogen. It is for this reason that they are thought to be the detonation of a carbon star, a white dwarf of ~ 1.4 M.

They are very bright and are the principal way of observing the effect of Dark Energy, and one pillar of the consensus standard $\Lambda$CDM model of cosmology.

However an eprint Exploring an Alternative Channel of Evolution Towards SNa Ia Explosion (MNRAS accepted) has recently been published that would suggest they are not all standard candles.

The paper looks at hydrogen and helium contamination of the WD progenitor and finds that a very small amount of hydrogen or a greater amount of helium may cause premature detonation.

From #8 Conclusions:
(Bold mine)

My confusion is that such variation in SNe Ia luminosity should have already been observed in the "standard candle" calibration of these objects.

How would such a variation affect the 'Gold Standard' dataset of distant SNe Ia and the standard $\Lambda$CDM model?

Garth

Last edited: Feb 4, 2015
2. Feb 4, 2015

### marcus

That seems to be a good reason for skepticism concerning the conclusions. But they are cautious, they call this an "exploratory project" and merely say that (because there would be far-reaching implications if confirmed) "the whole subject deserves careful future investigation."

3. Feb 4, 2015

### Garth

Yes, but the likelihood of such a tiny contamination is highly likely IMHO and therefore the variation should be prevalent especially as a New Scientist article Solo supernovae challenge cosmic distance standards quoting one of the authors states:
Susana Deustua, a scientist with the Supernova Cosmology Project comments:
Different beasts with different luminosities?

I have suspicions about the criteria (especially the assumptions underlying them) by which members of the Gold Set are selected....

Garth

Last edited: Feb 4, 2015
4. Feb 4, 2015

### Chalnoth

Not necessarily. The spectra of SN1A's aren't neat and clear. They vary, by a fair amount.

It's been known for some time that some SN1A's are brighter and last longer than others (there is a relatively tight correlation between decay time and brightness), but last time I checked we hadn't yet found any correlation in the spectrum that identified brighter or dimmer supernovae. The search continues, of course, as any additional method of standardizing SN1A's further will reduce the errors on estimates of cosmology that utilize them. It would be very good if this observation held up.

5. Feb 6, 2015

### TEFLing

I understand, that distant SNIa are dimmer and fainter than they ought to be

I.e. we observe distant SNIa and assume that they are standard candles...

We think we know the intrinsic brightness of the SNIa and the red shift of the host galaxy

But we don't receive all of the photons we expect

Some photons don't make the trip... And the father away the host galaxy , the more photons fail to reach earth

Is this correct? I.e. Professor Perlmutter never claimed that his SNIa were too BRIGHT, yes?

6. Feb 6, 2015

### Garth

Hi TEFling,

All photons make the trip apart from those absorbed by dust extinction.

The fact that SNe Ia were fainter than expected (at around z=1) by the 'plain' GR model was evidence that the universe had accelerated under the influence of some 'exotic' Dark Energy/Cosmological Constant.

At even higher redshift they get a little brighter than expected, which gives a precise handle on the model itself - i.e. the $\Lambda$CDM model.

But this 'precise handle' obtained from the 'Gold set' depends on:
1. In the far universe: the Absolute Magnitudes of those very distant SNe Ia being deduced correctly from their apparent magnitudes. This depends on such factors as: the amount of dust extinction, the modelling of their luminosity curves with the delay in detection given their faintness, the correct cosmological geometry, the correct application of the Phillips relationship coupled with cosmological time dilation, any correction because of a secular evolution of metallicity, selection effect (not detecting the fainter members), and probably a few more!
2. In the near universe: the accurate calibration of these supernovae as Standard Candles. Confusion as to different classes of these supernovae and a possible evolution of the ratio of these classes in any set of distant SNe Ia will introduce errors of Absolute Magnitude in the 'Gold set'.
So the point of this thread is to ask the question, given the OP paper, "Is it true that "we know the intrinsic brightness of the SNe Ia" standard candles?"

Garth

Last edited: Feb 6, 2015
7. Feb 6, 2015

### Chalnoth

The really far-away ones are too bright.

But the point of this thread is that there's actually a decent amount of variation in SN1A brightness. If we can produce a physical model for why some supernovae are brighter than others, and can measure the features of that physical model, then we can do a better job of "standarizing" SN1A's to reduce the errors on cosmological parameters.

8. Feb 8, 2015

### TEFLing

I want to clarify, that dust is not the only factor causing extinction... Plain old space plasma would still have some Kramer's opacity, yes? To try to help the focus of the thread, unless someone says otherwise, I will privately assume that a column density of space plasma does have some non zero opacity, as would the same elements when compressed into the interiors of stars

Thanks for the clarifications

9. Feb 8, 2015

### Garth

Hi Chalnoth, thank you.

So are you saying the variation in Absolute Magnitude is already accounted for in the Phillips relationship?

Garth

10. Feb 8, 2015

### Garth

Hi TEFling,

There are two issues as I pointed out in #6, the calibration of the SNa Type Ia Standard Candle in the near universe and the correct application of the cosmological distance modulus (taking into account geometry and expansion rate effects) to the apparent magnitudes of distant SNa Type Ia.

Extinction is only one (but possibly important) correction to be applied to that distance modulus.

My concern in this thread is that if there are several different species of SNe Ia, degenerate, doubly degenerate and now 'single contaminated' SNe Ia, then, (remembering the early error in calibration of the Cepheid Variable standard candle in Hubble's time,) has there been a error of identification of the species in the calibration of SNe Ia?

Secondly, considering the deep time elapse since the detonation of the distant SNe Ia, has there been an evolution in the ratio between these different species?

Specifically as it would take some time for the degenerate, and especially the doubly degenerate, SNe Ia to reach detonation I would expect there to be a greater ratio of the 'single contaminated' types early on in the universe's history.

To see the problem I post the SNe ia Hubble diagram taken from New Hubble Space Telescope Discoveries of Type Ia Supernovae at z ≥ 1: Narrowing Constraints on the Early Behavior of Dark Energy (Reiss et al 2007)

Garth

11. Feb 10, 2015

### TEFLing

For sake of not causing confusion, SNIa at high redshift are not actually too bright, per se. Instead, they are too bright compared to a reference cosmology, namely an empty Milne universe ( rho = 0 ). However, they are still slightly dimmer and fainter than would be expected, for the critically closed cosmology ( rho = 1 ). And, actually, accounting for the attenuation due to electron scattering Kramer's opacity, the high redshift bin at z~1.5 is actually about as luminous as expected.

What seems to be the striking case, is that SNIa at low redshift ( z<1 ) are several tenths of a magnitude too dim, for a critically closed cosmos to accommodate, even with electron scattering included.

For z<1, SNIa appear about 30% too dim, compared to the critically closed case.

12. Feb 10, 2015

### Garth

Actually, the Empty (Milne) model fitted the original Perlmutter data as well as the $\Lambda$CDM model out to z=1. On this 1997/8 plot: Poster displayed at the American Astronomical Society meeting in Washington, D.C., January 9, 1998 (Perlmutter et al., B.A.A.S., v. 29, no. 5, p. 1351, 1997) . Milne model is the (0,0) plot. However going beyond z = 1.5, as in the diagram above, confirms the standard model as the supernova become brighter than that reference Milne model.

However have we got the 'Standard Candle' calibrated correctly?

Garth

Last edited: Feb 10, 2015
13. Feb 11, 2015

### TEFLing

Those data are difficult to reconcile with any critical closed cosmology

Extinction due to optical depth ought to increase with distance, without ever decreasing

But the higher redshift SNIa show expected extinction, whereas nearer ones show MORE extinction ... Like looking through mist and being able to see far but not near

One would have to claim younger SNIa are somehow more obscured by local environments ( since they are closer having less optical depth through intergalactic space )

14. Feb 11, 2015

### Garth

If the change in relative apparent magnitude is due to extinction.....

That of course is one unknown that could affect the result, but as you point out, if the faintness of the z~1 Sne 1a is due to extinction then it seems to work the wrong way round, the more distant SNe should be even fainter than predicted, not brighter. Which is a good reason for finding another factor.

The standard answer is 'Jerk'! That is the transition from a decelerating universe out beyond z = 1.5 or so, to an accelerating universe at around z=1under the influence of $\Lambda$/DE. Whether the universe then stops accelerating in the nearby universe is, I think, open to debate. If the acceleration is due to the Cosmological Constant then, once accelerating, the universe would accelerate for ever as $\Lambda$ continues to dominate. If DE then it all depends on the Equation of State with -1/3 > $\omega$ > -1 to give a non-accelerating recent universe.

The other factors I wonder about:
1. is a secular variation in Absolute Magnitude due to a variation in metallicity,
2. a selection effect in the detection of these distant SNe, for example - your concern #13, perhaps extinction prevents the detection of would-be fainter members of the z=1.65 'bin'.
3. and - following on from the subject of this thread - a secular variation, an evolution, of the ratio of different species of SNe 1a. At z=1.65 in the standard model we are about 4Gyr after BB and the types with rapid pathways to WD detonation ought to be more prevalent than the slower ones as compared to their ratio at about 13Gyr after BB.
Garth

Last edited: Feb 11, 2015
15. Feb 13, 2015

### TEFLing

From the first Friedmann equation, and by normalizing the scale factor and time coordinate by their present values a0,t0...

And by evaluating the first equation at present epoch to notice that the normalized curvature term k ~ ( Omega0 - 1 )...

You derive the relation between

(H0 t0) dt = da / sqrt( OmegaM/a1 + OmegaR/a2 + Omega/\a2 - (Omega0-1) )

Wolfram Alpha example code
integral from 0 to 1 1/sqrt(0.19/a+0.01/a^2+.9*a^2-(1.1-1)) da

For a radiation universe OmegaR=1, the RHS integrates to 1/2...

For a matter universe OmegaM=1, the RHS integrates to 2/3...

And combinations of matter and energy yield values in between...

Thus critically closed matter and energy universes would be young compared to a Hubble time TH=1/H0...

Which would always make the SNIa results hard to understand

Even an empty Milne universe integrates trivially to 1

Only by adding a constant energy density term can you draw out the lifetime age of the cosmos to >1 Hubble time

And the cosmological constant has a very natural interpretation, namely that the fabric of space-time itself has an energy density, as if composed of energy itself, which could be a very unifying concept

The SNIa are calibrated from local SNIa, yes? So to have bright events nearby and far away, with dim events in between, would require three successive populations of SNIa, with the first and third conspiring to have similar luminosities.

Never the less, more precise measurements of these important events would greatly clarify the details of the universe

16. Feb 13, 2015

### TEFLing

Garth makes a persuasive argument for the possibility of two types of SNIa, early and late

And a nearly empty Milne cosmology is consistent with observations out to z~1

What about the angular size effect of GR? Objects at z~1.5 begin to appear bigger and larger on the sky? Could a simple Milne cosmology, with a plausible two types of SNIa, combined with increasing apparent angular size at high redshift, account for and explain observations?

17. Feb 13, 2015

### Chalnoth

As near as I can tell, there has been one example of a SN1A which has fallen outside the expected brightness range. There may well be two types, but one type is extraordinarily rare.