Expert Voices

'One of the greatest damn mysteries of physics': We studied distant suns in the most precise astronomical test of electromagnetism yet

The sun. (Image credit: NASA)

This article was originally published at The Conversation. The publication contributed the article to Space.com's Expert Voices: Op-Ed & Insights.

Michael Murphy, Professor of Astrophysics, Swinburne University of Technology

There's an awkward, irksome problem with our understanding of nature's laws which physicists have been trying to explain for decades. It's about electromagnetism, the law of how atoms and light interact, which explains everything from why you don't fall through the floor to why the sky is blue.

Our theory of electromagnetism is arguably the best physical theory humans have ever made — but it has no answer for why electromagnetism is as strong as it is. Only experiments can tell you electromagnetism's strength, which is measured by a number called α (aka alpha, or the fine-structure constant).

The American physicist Richard Feynman, who helped come up with the theory, called this "one of the greatest damn mysteries of physics" and urged physicists to "put this number up on their wall and worry about it."

Related: How many stars are in the universe?

In research just published in Science, we decided to test whether α is the same in different places within our galaxy by studying stars that are almost identical twins of our sun. If α is different in different places, it might help us find the ultimate theory, not just of electromagnetism, but of all nature's laws together — the "theory of everything."

We want to break our favorite theory

Physicists really want one thing: a situation where our current understanding of physics breaks down. New physics. A signal that cannot be explained by current theories. A sign-post for the theory of everything.

To find it, they might wait deep underground in a gold mine for particles of dark matter to collide with a special crystal. Or they might carefully tend the world's best atomic clocks for years to see if they tell slightly different time. Or smash protons together at (nearly) the speed of light in the 17-mile (27 kilometers) ring of the Large Hadron Collider.

The trouble is, it's hard to know where to look. Our current theories can't guide us.

Of course, we look in laboratories on Earth, where it's easiest to search thoroughly and most precisely. But that's a bit like the drunk only searching for his lost keys under a lamp-post when, actually, he might have lost them on the other side of the road, somewhere in a dark corner.

The sun's rainbow: sunlight is here spread into separate rows, each covering just a small range of colors, to reveal the many dark absorption lines from atoms in the sun's atmosphere. (Image credit: N.A. Sharp / KPNO / NOIRLab / NSO / NSF / AURA, CC BY)

Stars are terrible, but sometimes terribly similar

We decided to look beyond Earth, beyond our solar system, to see if stars which are nearly identical twins of our sun produce the same rainbow of colors. Atoms in the atmospheres of stars absorb some of the light struggling outwards from the nuclear furnaces in their cores.

Only certain colors are absorbed, leaving dark lines in the rainbow. Those absorbed colors are determined by α — so measuring the dark lines very carefully also lets us measure α.

A view of the sun captured by the Daniel K. Inouye Solar Telescope in Hawaii. (Image credit: NSO / AURA / NSF, CC BY)

The problem is, the atmospheres of stars are moving — boiling, spinning, looping, burping — and this shifts the lines. The shifts spoil any comparison with the same lines in laboratories on Earth, and hence any chance of measuring α. Stars, it seems, are terrible places to test electromagnetism.

But we wondered: if you find stars that are very similar — twins of each other — maybe their dark, absorbed colors are similar as well. So instead of comparing stars to laboratories on Earth, we compared twins of our sun to each other.

A new test with solar twins

Our team of student, postdoctoral and senior researchers, at Swinburne University of Technology and the University of New South Wales, measured the spacing between pairs of absorption lines in our sun and 16 "solar twins" — stars almost indistinguishable from our sun.

The rainbows from these stars were observed on the 3.6-metre European Southern Observatory (ESO) telescope in Chile. While not the largest telescope in the world, the light it collects is fed into probably the best-controlled, best-understood spectrograph: HARPS. This separates the light into its colors, revealing the detailed pattern of dark lines.

HARPS spends much of its time observing sun-like stars to search for planets. Handily, this provided a treasure trove of exactly the data we needed.

The ESO 3.6-metre telescope in Chile spends much of its time observing sun-like stars to search for planets using its extremely precise spectrograph, HARPS. (Image credit: Iztok Bončina / ESO, CC BY)

From these exquisite spectra, we have shown that α was the same in the 17 solar twins to an astonishing precision: just 50 parts per billion. That's like comparing your height to the circumference of Earth. It's the most precise astronomical test of α ever performed.

Unfortunately, our new measurements didn't break our favorite theory. But the stars we've studied are all relatively nearby, only up to 160 light-years away.

What's next?

We've recently identified new solar twins much further away, about half way to the center of our Milky Way galaxy.

In this region, there should be a much higher concentration of dark matter — an elusive substance astronomers believe lurks throughout the galaxy and beyond. Like α, we know precious little about dark matter, and some theoretical physicists suggest the inner parts of our galaxy might be just the dark corner we should search for connections between these two "damn mysteries of physics."

If we can observe these much more distant suns with the largest optical telescopes, maybe we'll find the keys to the universe.

This article is republished from The Conversation under a Creative Commons license. Read the original article.

Follow all of the Expert Voices issues and debates — and become part of the discussion — on Facebook and Twitter. The views expressed are those of the author and do not necessarily reflect the views of the publisher.

Join our Space Forums to keep talking space on the latest missions, night sky and more! And if you have a news tip, correction or comment, let us know at: community@space.com.

Michael Murphy
Professor of Astrophysics, Swinburne University of Technology

I obtained my PhD in astrophysics from the University of New South Wales in 2003. I was a research fellow at the Institute of Astronomy, University of Cambridge, from 2003 to 2007. In 2007 I joined Swinburne University of Technology where I am currently an Australian Research Council Future Fellow and Deputy Director of Swinburne’s Centre for Astrophysics and Supercomputing.

  • rod
    "From these exquisite spectra, we have shown that α was the same in the 17 solar twins to an astonishing precision: just 50 parts per billion. That's like comparing your height to the circumference of Earth. It's the most precise astronomical test of α ever performed. Unfortunately, our new measurements didn't break our favorite theory. But the stars we've studied are all relatively nearby, only up to 160 light-years away."

    My observation. Various reports through the years show efforts to test alpha, the fine-structure constant and find some variation perhaps :) Here is an example from 2004:

    New Quasar Studies Keep Fundamental Physical Constant Constant, https://ui.adsabs.harvard.edu/abs/2004eso..pres....5./abstract, March 2004. "...PR Photo 07/04: Relative Changes with Redshift of the Fine Structure Constant (VLT/UVES) A fine constant To explain the Universe and to represent it mathematically, scientists rely on so-called fundamental constants or fixed numbers. The fundamental laws of physics, as we presently understand them, depend on about 25 such constants. Well-known examples are the gravitational constant, which defines the strength of the force acting between two bodies, such as the Earth and the Moon, and the speed of light. One of these constants is the so-called "fine structure constant", alpha = 1/137.03599958, a combination of electrical charge of the electron, the Planck constant and the speed of light. The fine structure constant describes how electromagnetic forces hold atoms together and the way light interacts with atoms. But are these fundamental physical constants really constant? Are those numbers always the same, everywhere in the Universe and at all times? This is not as naive a question as it may seem. Contemporary theories of fundamental interactions, such as the Grand Unification Theory or super-string theories that treat gravity and quantum mechanics in a consistent way, not only predict a dependence of fundamental physical constants with energy - particle physics experiments have shown the fine structure constant to grow to a value of about 1/128 at high collision energies - but allow for their cosmological time and space variations. A time dependence of the fundamental constants could also easily arise if, besides the three space dimensions, there exist more hidden dimensions..."

    There is always the issue of fine-tuning of the universe and the origin of the universe we see today in astronomy.

    How do we know the fundamental constants are constant? We don't., https://forums.space.com/threads/how-do-we-know-the-fundamental-constants-are-constant-we-dont.59359/
    Reply
  • KirkSutton
    The question of the universe's fine-tuning and its creation are perennial problems in astronomy.
    Reply
  • Helio
    I'm curious which stars were found to be solar twins. At one time, 18 Sco was the best solar twin, but another one was found to be better.

    Observing these stars will demonstrate the Sun's non-yellow (or orange as in this article) color. But, at times, they can also appear with a tint of yellow or gold. These color variations are strong indicators of high atmospheric particle counts. I suspect this may corelate to seeing conditions, which is helpful to know before serious observing.

    Is this likley?
    Reply
  • KirkSutton
    These stars can be used to demonstrate the Sun's non-yellow (or orange, as it is described in this article) tint. But occasionally they might also have a golden or yellow hue to them. Strong signs of high atmospheric particle counts are these colour variations. Before engaging in any serious watching, it would be nice to know if this had anything to do with the seeing conditions.
    Reply
  • Helio
    KirkSutton said:
    These stars can be used to demonstrate the Sun's non-yellow (or orange, as it is described in this article) tint. But occasionally they might also have a golden or yellow hue to them. Strong signs of high atmospheric particle counts are these colour variations. Before engaging in any serious watching, it would be nice to know if this had anything to do with the seeing conditions.
    I agree. 😏
    Reply
  • COLGeek
    Helio said:
    I agree. 😏
    I would think so. ;)
    Reply
  • Dragrath
    Helio said:
    I'm curious which stars were found to be solar twins. At one time, 18 Sco was the best solar twin, but another one was found to be better.

    Observing these stars will demonstrate the Sun's non-yellow (or orange as in this article) color. But, at times, they can also appear with a tint of yellow or gold. These color variations are strong indicators of high atmospheric particle counts. I suspect this may corelate to seeing conditions, which is helpful to know before serious observing.

    Is this likley?
    Its tricky to define the Sun's solar relatives after all as far as we can tell relatively massive "lower mass stars, i.e. F and G type main sequence stars, like our Sun only form in dense star formation clusters the type where very massive stars form something we have evidence to support in terms of the presence of fossilized chemical bonds from short lived radioisotopes from the Early solar system with different chemistry to their daughter products (basically elements in chemical bonds that the current element wouldn't form but which a radioisotope that decays into that element does suggesting that when the molecules in "primordial" asteroid, meteor and or cometary derived material formed the short lived radioisotope was present).

    Adding to that work which has combined GAIA astrometry with spectroscopic data has shown that the Sun formed near the tail end of the Milky Way's most intense starburst event linked to the disk plunging collision of the Sagittarius Dwarf Spheroidal galaxy. If the GAIA sample of stars is broadly representative of the galaxy at large that suggests that well over 50 percent of the stars that the Milky Way galaxy has ever formed formed during that billion year starburst episode in this case differentiating the Sun from other stars by age is likely a daunting task. Its even quite possible that the Sun may have formed in a now tidally disrupted super star cluster like those we observe in the LMC during its own starburst episode.

    Given that the Sun likely has tens of thousands of siblings at least even though the vast majority of the remaining stars will be less massive K and M type main sequence stars which will naturally fail to meet the solar twin criteria there will still be a lot of G type stars that meet that criteria and because the chaotic inhomogeneous conditions in starbursts even stars from the Sun's birth cluster may vary considerably in chemical composition.

    Then of course you get to the issue of stellar migration since we have some evidence to suggest the Sun likely formed closer to the Galactic center than it is found now which isn't surprising given that stars are locked within an angular momentum constrained random walk of sorts which over time via stellar encounters and interactions with passing density waves tidal perturbations etc. which all can cause shifts in a star's orbit around the Milky Way. This makes it difficult to trace back the Sun yet alone any of its siblings to their place of origin as there have been nearly 5 billion years of jumbled mixing with the general disk population. Its not impossible per say especially if you have a big enough data set of say every star's spectrum in the Milky Way but it is tricky

    Hence its a far more complex question than people tend to think. On one hand though if the Sun has millions of siblings finding stellar twins might be surprisingly easy its just verifying them would be hard.
    Reply
  • Helio
    Dragrath said:
    Its tricky to define the Sun's solar relatives after all as far as we can tell relatively massive "lower mass stars, i.e. F and G type main sequence stars, like our Sun only form in dense star formation clusters the type where very massive stars form something we have evidence to support in terms of the presence of fossilized chemical bonds from short lived radioisotopes from the Early solar system with different chemistry to their daughter products (basically elements in chemical bonds that the current element wouldn't form but which a radioisotope that decays into that element does suggesting that when the molecules in "primordial" asteroid, meteor and or cometary derived material formed the short lived radioisotope was present).
    Yes, but yet there are "solar twins" that are surprisingly close in all features, including composition. This would allow a close spectral comparison.

    Roughly 4% or 5% of the stars in the galaxy are G-type. If only 4% or 5% of those are G2, then there may be as many as 200M G2 stars out there. Maybe 1/3 of those can be observed, leaving maybe 50 million or so. Deduct from those the time frame, as you mentioned.

    So, some may have very similar composition. Those that are similar (not very similar) are called solar analogs. The very similar are the "twins".

    Given that the Sun likely has tens of thousands of siblings at least even though the vast majority of the remaining stars will be less massive K and M type main sequence stars which will naturally fail to meet the solar twin criteria there will still be a lot of G type stars that meet that criteria and because the chaotic inhomogeneous conditions in starbursts even stars from the Sun's birth cluster may vary considerably in chemical composition.
    There was a report I read a while back, but I don't think I saved it, that explained that the no. of siblings might be between 1000 and 3000. The upper constraint was due to the claim that the outer planets would have encountered more orbital instability due to perturbations from the close neighbors.

    The IMF (Initial Mass Function) seems to show about 60 stars out of 3000 that would have masses between 0.9 and 1.1. So, perhaps there are a few solar twins from our own original GMC.

    Then of course you get to the issue of stellar migration since we have some evidence to suggest the Sun likely formed closer to the Galactic center than it is found now which isn't surprising given that stars are locked within an angular momentum constrained random walk of sorts which over time via stellar encounters and interactions with passing density waves tidal perturbations etc. which all can cause shifts in a star's orbit around the Milky Way. This makes it difficult to trace back the Sun yet alone any of its siblings to their place of origin as there have been nearly 5 billion years of jumbled mixing with the general disk population.
    All good points. I do recall U of Texas, however, finding one sibling, though I don't recall its type.

    I must admit I don't really understand the crux of their argument. There are over 25,000 lines in the solar spectra. How many must they match with another star? Can't they pick the ones that match the most evident atoms, say a certain iron ion, to get their verification? If the constant they wish to confirm is not constant, wouldn't it be seen in those dozen lines or so for one atom?
    Reply
  • Dragrath
    Helio said:
    Yes, but yet there are "solar twins" that are surprisingly close in all features, including composition. This would allow a close spectral comparison.

    Roughly 4% or 5% of the stars in the galaxy are G-type. If only 4% or 5% of those are G2, then there may be as many as 200M G2 stars out there. Maybe 1/3 of those can be observed, leaving maybe 50 million or so. Deduct from those the time frame, as you mentioned.

    So, some may have very similar composition. Those that are similar (not very similar) are called solar analogs. The very similar are the "twins".

    There was a report I read a while back, but I don't think I saved it, that explained that the no. of siblings might be between 1000 and 3000. The upper constraint was due to the claim that the outer planets would have encountered more orbital instability due to perturbations from the close neighbors.

    The IMF (Initial Mass Function) seems to show about 60 stars out of 3000 that would have masses between 0.9 and 1.1. So, perhaps there are a few solar twins from our own original GMC.

    All good points. I do recall U of Texas, however, finding one sibling, though I don't recall its type.

    I must admit I don't really understand the crux of their argument. There are over 25,000 lines in the solar spectra. How many must they match with another star? Can't they pick the ones that match the most evident atoms, say a certain iron ion, to get their verification? If the constant they wish to confirm is not constant, wouldn't it be seen in those dozen lines or so for one atom?
    Not necessarily after all the evidence we have that alpha isn't constant always comes from particle accelerators which seems to suggest the strength of alpha is energy dependent at least at the high energies of quark gluon plasma, if there is an energy dependence than selection of stars based on similar spectra could be biased since those stars would also presumably be of similar temperatures of which both spectral line expression and black body radiation curves are dependent. I.e. elemental spectral lines depend not only on the abundance of an element but the amount of that element getting energetically excided enough to emit or absorb light. Thus its possible that you might be implicitly biasing results.

    And yeah I think there have been a few such candidates found over the years or at least reported candidates not sure if any have ever been veritably confirmed(if we are even capable of such a confirmation).
    Reply
  • Helio
    Dragrath said:
    Not necessarily after all the evidence we have that alpha isn't constant always comes from particle accelerators which seems to suggest the strength of alpha is energy dependent at least at the high energies of quark gluon plasma, if there is an energy dependence than selection of stars based on similar spectra could be biased since those stars would also presumably be of similar temperatures of which both spectral line expression and black body radiation curves are dependent. I.e. elemental spectral lines depend not only on the abundance of an element but the amount of that element getting energetically excided enough to emit or absorb light. Thus its possible that you might be implicitly biasing results.
    Yes, temperature is very important, and maybe I’m being too pedantic for a general article, but wouldn’t a solar analog work for their purposes? I assume the absorption lines are not dependent on the other lines of other elements. But perhaps density and other parameters will vary with composition.
    Reply