RSS Feed
iTunes Link

Episode 166 - Stellar Evolution, Age of the Universe, and Young-Earth Creationism

Download the Episode

Recap: The half-truth and misconception addressed in this episode is whether the age of a star with error bars that overlap with the age of the universe means all of science is wrong and the universe was created 6000 years ago.

Additional Materials:

Episode Summary

Claim: I tend to not like to lead with such an obvious bias in my intro, but the topic for this episode really can be summarized by that run-on sentence. For a bit more detail before we really get into it, recently, Answers in Genesis, one of the three major young-Earth creationism institutions involved in confronting nearly every aspect of modern science, recently published a blog post written by Danny Faulkner, one of their primary writers in astronomy. In the post, Dr. Faulkner points to a journal publication about the age of a star which has a calculated age that is older than the calculated age of the universe. His conclusion is that modern science is so model-dependent that it's pretty much useless and nothing can be trusted. Ergo, young-Earth creationism. The format for this episode is going to be very much like some of the original episodes, discussing the appropriate background material first, then getting into the subject matter, and reassessing the claim.

Background

So first off, the paper in question. Howard Bond, a scientist working at the Space Telescope Science Institute, published a paper in 2013 along with four co-authors. They published the paper in The Astrophysical Journal Letters, one of the primary astrophysics journals, so this checks the appropriate argument from authority boxes - or at least sanity check boxes - that the authors are mainstream academics and they published in a mainstream peer-reviewed journal. That's not to say that every academic has correct ideas nor is it to say that anything that makes it through peer-review is accurate, but at least it isn't something self-published on something like the Above Top Secret forums.

The paper discusses the star HD 140283. That's the 140,283rd star in Henry Draper's star catalog from many œons ago but is still used today. I'll just call it "the star" at this point. The star is very old. They say that it's very old for a few reasons.

Reason 1: The star is very metal-poor. To an astronomer, the universe is composed of hydrogen, helium, and everything else is a "metal." I know that's going to break most people, but just go with it. So yes, that means oxygen is a metal so far as an astronomer is concerned. Anyway, based on models of how the universe formed, it was almost all hydrogen, some helium, and then a very very little bit of extremely light elements like lithium and beryllium. That means that only those elements would have been around for the first stars, meaning that the first stars would be very metal-poor.

We can measure the ratio of metals to hydrogen in stars by looking at their light, spreading it out, and seeing the effects of different elements absorbing different wavelengths. If there's a lot of a certain element, then the wavelengths of light that that element absorbs will be almost entirely missing from the star's light, but if there's just a little, then we'll see most of the light at those wavelengths from the star.

So, these astronomers did that, or used other work by other astronomers who did that for this star. In this case, they used published data by other astronomers for this star. In fact, this star was one of the first to have been shown back in 1951 to have a metallicity much lower than the sun's.

Anyway, to circle back, the star is metal-poor meaning that it formed when there were very little metals in the universe, soon after formation.

Reason 2: It's in a region of the Hertzsprung-Russell Diagram (H-R Diagram) where the brightness of stars is extremely well correlated with the age of the star. I don't think I've talked about the H-R Diagram before except maybe in episode 124 on the astronomical distance ladder. So, what is this key piece of stellar astrophysics?

The H-R Diagram was developed at the same time, independently by - oddly enough, a scientist with the last name Hertzsprung and another with the last name Russell. They graphed a star's color on the x-axis and a star's brightness on the y-axis. When doing that, they found that almost all stars fell along roughly a diagonal line from the upper-left to the lower-right, meaning that most really bright stars were also very blue, and most really dim stars were also very red. There are various branches and clusters of other stars on the H-R Diagram, but this has very much turned into one of the key tools in the study of stars.

The lifecycles of stars also progress along the H-R Diagram in various ways, and many models of stellar evolution will include an output H-R Diagram showing how the star moves through it. Now, almost of course, there are almost always exceptions to the rule, but this star in particular falls in a very well understood region of the H-R diagram where it's true brightness is very diagnostic of its age.

Reason 3: So, the third reason that they think this is a very old star is that they figured out how far away it is. That may seem like a non sequitur, or perhaps even trivial, but it's not. The astronomers used observations from the Hubble Space Telescope to track the star's motion across our vantage point as Earth moves in its orbit around the sun. They were able to see it moving slightly, and when they did the analysis, they were able to get a result that is 5 times more precise than the previous measurement by a European mission from about two decades ago.

The only way to know a star's true brightness is to know how far away it is. I discussed this idea at length in Episode 124, so just in brief here, the idea is that I can shine a light towards you and it may look bright, but absent any other information, you have no way of knowing if the light looks bright because I'm very close to you and it's faint, or if it looks bright because I'm really far away from you and it's extremely bright. If I tell you how far away it is, now you can calculate how bright it is at its source, or at a standard distance. In astronomy, the standard distance is 10 parsecs, which is related to Earth's distance from the sun, which is again talked about at length in episode 124, so for the sake of this episode, 10 parsecs is about 32.6 light-years.

By determining the distance a factor of five times more precisely than has been done before, they can determine its brightness much more precisely. By determining its brightness much more precisely, they can determine its position on the H-R Diagram much more precisely. By determining its position on the H-R Diagram much more precisely, they can determine its age much more precisely.

The Result

The result from all of this is they get an age of 14.46±0.8 Gyr. Pretty old.

Age of the Universe

The currently accepted value for the age of the universe is 13.77±0.06 Gyr. Pretty old. But younger than the star if you take the numbers at face-value and don't look at them as a scientist would. More on that in a bit.

First, how do we know how old the universe is? These days, the number is based on several completely different lines of evidence. These include measuring the expansion rate of the universe and then literally just running that expansion rate backwards, and also measurements of the Cosmic Microwave Background Radiation.

I couldn't quickly find an episode where I talked about the CMB, so really quickly, the CMB is effectively the light from the universe as soon as it turned transparent. What that means is that the universe was so hot early on that the entire universe was basically like a star. Meaning that light could not travel freely within it, the light kept bouncing around and getting absorbed and emitted and it was a real mess. But as the universe expanded, it cooled. That's basic physics where if you increase the volume while holding the amount of material constant, temperature must decrease. And it cooled enough and got low enough density that light finally could stream freely about.

That pattern of light in the sky is known as the Cosmic Microwave Background Radiation, and it gives us our first picture of the early universe. Based on it, we can figure out various things, including how old it is.

It's possible I'll do a whole show on this, but in the meantime, suffice to say, the age of the universe now is very well constrained, and that's reflected in the small error bar.

Revising an Old Argument without Consideration for Error Bars

Which gets us back to the age of this star. What Dr. Faulkner is effectively doing is revising an argument that real scientists were having just 20 years ago, before we had good measurements of the CMB and before we knew a lot of the nuclear and stellar physics we know now.

Just 20 years ago, there was a big problem in astronomy where the models for the ages of the oldest stars put them at up to 20 billion years old, but models for the age of the universe put it at as young as 10 billion years. Clearly, something was amiss.

And we figured it out! As I said, we have learned a lot in the last two decades. The Hubble Space Telescope 20 years ago was really just a few years into its mission and we were still using it to pin down the astronomical distance ladder which was revising the ages of stars DOWN. We were learning a lot more about atomic and nuclear physics because computers were getting better and laboratory equipment was getting better and we were able to revise the ages of stars down, again, based on incorporating that updated physics. Things like how quickly stars can fuse certain elements together at a certain temperature and pressure, or even what that temperature and pressure would be in different parts of the star -- all of those factors affect the models of stellar evolution and the theoretical H-R Diagrams that come out of these models.

Similarly, the Hubble Space Telescope with helping us get a handle on that distance ladder and therefore brightness of objects and distances to objects and therefore the expansion rates of the universe was helping us to revise the age of the universe UP. The COBE satellite and then WMAP satellite and then Plank satellites all got us progressively better and better maps of the CMB which also helped to considerably better understand the age of the universe, revising it UP.

As I said, now-a-days, the age of the universe is pretty well known. Stellar evolution models are also fairly good and unlikely to be radically changed moving forward. But, there are still uncertainties, and that's what error bars are form and those need to be recognized when looking at this star, with a model age of 0.69±0.8 billion years older than the age of the universe.

Dr. Faulkner pretty much admits that in his article, stating, "Most astronomers have ignored this difficulty, assuming that the errors in the computed ages of both HD 140283 and the universe could be interpreted in a way that solves the problem." And that's true. But it's also misleading when he wrote earlier in the article, "It would require both sets of errors to conspire in a very convenient way." This is simply not how you're supposed to work with error bars, it's not what they mean.

Where Dr. Faulkner also shows the creationism side, and ignores how science results are to be read, states in the very next sentence, "However, that may be unlikely, and it hardly solves the need for the existence of even older stars to explain the metal content of HD 140283."

That introduces a new problem that I'll get to momentarily. But clearly, he's trying to get you to look at the numbers without the attached error bars.

In science, error bars are key. I have reviewed several papers, grants, and other things where if I see graphs or numbers without error bars, it's something I always try to call out for fixing. I even do this in science fairs where, to me, it can make the difference between getting first place and not.

The reason error bars are so important is that they make the difference between a significant result and a non-significant result. Let's do a pretty simple "for instance." Let's say that someone did an experiment and discovered that half of all people are psychic. "Wow," you might think. But, the error bars on that 50% number of people are ±50%. Meaning that anywhere from 0% to 100% of people are psychic. Suddenly, it doesn't seem nearly as significant as it did before.

For a lot more on error bars, see episode 82 in how to design a hyperdimensional physics experiment.

From that contrived example, we come to this situation. The error bars that are quoted are one standard-deviation or "sigma," meaning that there is a 68% chance that if the experiment were repeated, the number would be within that range. If you assume the error bars are Gaussian, meaning they follow the classic bell-shaped curve, then if you double the error bars they encompass 95% of the repeats. In physics, we pretty much always quote 1 standard-deviation, in particle physics we use 5, but in statistics, they often use 95%. In my own sub-field of crater studies, several of us are trying to get people to move from 1 standard deviation to using 2. But that's a separate story.

Anyway, the 1-sigma error bars overlap: 14.46±0.8 versus 13.77±0.06. That means that there is no conflict with this star being older than the universe, so far as pretty much any scientist would interpret those numbers. The only people who would interpret this as a problem for basic physics is those who have an agenda or those who don't know basic statistics. The authors themselves state, "There is a remarkable accordance (within their respective uncertainties) between the age of the Universe inferred from the CMB, the age of the chemical elements, and the ages of the oldest stars."

That also doesn't take into account the fact that the authors of the paper even state that the ±0.8 on their age for the star is a minimum, that there are other sources of uncertainty which are not included. Meaning that there's even less of a non-issue.

The authors also state that an implication of their work is to use the star in the distance ladder, but if they do, it changes the distance to another object which makes their distance to that object smaller than almost all other distance measurements. A potential fix is to make their star a bit redder, that there may be some dust between us and it which they didn't account for. If they do that, it makes the age of the star even younger by about 0.65 Gyr, meaning that it's in even better agreement with the age of the universe.

The First Stars

Before I end, I said there was one other thing to address here, and that is the first generation of stars. Dr. Faulkner stated that we need older stars to explain the metal content of this one.

In fact, the authors even state this, in their Implications they wrote: "It is not quite a primordial star, given its low but non-zero metallicity."

The issue here is that you need that first generation of stars to make heavier elements, like gold or lead or nickel or iron or uranium or even oxygen.

But, that's also not a problem here. The first stars likely started to form when the universe was only about 200 million years old, and these first stars were about 20 times the mass of the sun up to several hundred times the mass of the sun. The larger they are, the faster they go through their material, and despite having so much more of it, they can die out and go supernova in under a million years. That million years is in the noise for this number, so we can subtract about 0.2 from the age of the universe for the time of the first stars and we STILL get comparison of the ages that are within just slightly more than 1 standard deviation of each other, meaning that they are still practically identical so far as a scientist is concerned. And if we add that age difference of -0.65 Gyr from the extra reddening, they are very easily compatible with each other.

6000 Years?

The whole point of the Answers in Genesis article, if you read it from start to finish, is to strongly cast doubt on the science and to promote a "GodDidIt" answer. It's filled with insinuations of scientists not really answering the questions posed and even trying to skew numbers to match a desired outcome, or at least that's how I interpret his article.

The end is also a dead giveaway as to its true intent: "If one changes the model, the problem may not even exist. Certainly, within the creation model, this isn't a problem."

Which is true: If you assume that you have something that can basically erase all the confusion, all doubt, and anything that doesn't make sense to you, then there are no problems at all.

But that's not how science works, and this is why I and many others argue that young-Earth creationism is a "science stopper" -- its basic premise, its raison d'être, is to promote a supernatural answer to everything and thus any science finding that gets away from a divine whim goes against it.

Provide Your Comments: