How Old Is the Mount St. Helens Lava Dome?

by Brian Dunning
March 24, 2009

Today we're going to point our skeptical eye at one of the key players in the debate between geologists and Young Earthers over the age of the Earth. In June of 1992, Dr. Steven Austin took a sample of dacite from the new lava dome inside Mount St. Helens, the volcano in Washington state. The dacite sample was known to have been formed from a 1986 magma flow, and so its actual age was an established fact. Dr. Austin submitted the sample for radiometric dating to an independent laboratory in Cambridge, Massachusetts. The results came back dating the rock to 350,000 years old, with certain compounds within it as old as 2.8 million years. Dr. Austin's conclusion is that radiometric dating is uselessly unreliable. Critics found that Dr. Austin chose a dating technique that is inappropriate for the sample tested, and charged that he deliberately used the wrong experiment in order to promote the idea that science fails to show that the Earth is older than the Bible claims. Yet the experiment remains as one of the cornerstones of the Young Earth movement.

Of most people who have heard of this incident before, that's probably about the total depth of what they've heard. And there's pretty good reason for this: Geology dating is pretty complicated, and if you look at Dr. Austin's paper or at any scholarly criticism of it, your eyes will quickly glaze over from the extraordinary detail and intricacy. So I thought this would be a great place to point Skeptoid's skeptical eye, and see how much of the chaff we can cut through to see what the bare facts of the case really are. Obviously both sides of this debate have agendas to promote, and that means that any summary you're likely to read was probably motivated by one agenda or the other.

Let's begin with a basic understanding of the radiometric dating technique used, K-Ar, or potassium-argon. This dating technique depends on the fact that the radioactive isotope of potassium, ⁴⁰K, naturally decays into other elements, as do all unstable radioactive elements. There are two ways that this happens to ⁴⁰K. About 89 percent of the time, a neutron inside the ⁴⁰K undergoes beta decay, in which the neutron decays into a proton and an electron. This gain of a proton turns the potassium into calcium. But about 11 percent of the time, an extra proton inside the ⁴⁰K captures one of its electrons and merges with it, turning the proton into a neutron and a neutrino, and converting the potassium into argon. In both events, the atomic mass remains unchanged, but the number of protons changes, thus turning the element from one to another. This happens to ⁴⁰K everywhere in the universe that it exists, and at the same rate, which is a half-life of 1.2 billion years. This means that if you have 1000 atoms of ⁴⁰K, 1.2 billion years later you'll have 500, and 1.2 billion years after that you'll have 250. You'll also have 83 argon atoms, and 667 calcium atoms. If I take a sample and measure an argon to potassium ratio of 83:250, I know that this sample is 2.4 billion years old.

However, all of these numbers are probabilities, not absolutes. You need to have a statistically meaningful amount of argon before your result would be considered significant. Below about 10,000 years, potassium-argon results are not significant; there's not yet enough argon created. The 11% of the time that potassium decays into argon and not calcium is also a probability, so this contributes to the result having a known margin of error. In addition, the initial amount of ⁴⁰K that you started with is never measured directly; instead, it is assumed to always be .0117% of the total potassium present, which is the known distribution in nature. This has a standard deviation, so it also contributes to the margin of error. So when my result says the sample was 2.4 billion years old, this is only correct if the sample was at least 10,000 years old to begin with, and it's only correct plus or minus a calculated margin of error, in this example about 600,000 years. The bell curve of probable age starts at about 1.8 billion years, peaks at 2.4 billion, and dips back to the baseline at 3 billion. So whether you call it an exact science or not is a matter of linguistics. Although the exact age can't be known, the probabilities can be exactly calculated.

Since Dr. Austin's sample was known to have solidified in 1986, its argon content was clearly well below the threshhold where an amount of argon sufficiently useful for dating could have been present. And even that threshhold applies to only the most sensitive detection equipment. Potassium-argon dating is done by destructively crushing and heating the sample and spectrally analyzing the resulting gases. The equipment in use at the time at the lab employed by Dr. Austin, Geocron Laboratories, was of a type sensitive enough to only detect higher concentrations of argon gas. Geocron clearly stated that their equipment was only capable of accurate results when the sample contained a concentration of argon high enough to be consistent with 2,000,000 years or older.

And so, by any standard, it was scientifically meaningless for Dr. Austin to apply Geocron's potassium-argon dating to his sample of dacite known to be only six years old. But let's ask the obvious question. If there wasn't yet enough argon in the rock to be detectable, and the equipment that was used was not sensitive enough to detect any argon, how was enough argon found that such old results were returned?

There are two possible reasons that the old dates were returned. The first has to do with the reason Geocron's equipment was considered useful only for high concentrations of argon. There would always be a certain amount of argon inside the mass spectrometer left over from previous experiments. If the sample being tested is old enough to have significant argon, this leftover contamination would be statistically insignificant; so this was OK for Geocron's normal purposes. But for a sample with little or no argon, it would produce a falsely old result. This was undoubtedly a factor in Dr. Austin's results.

The second possibility is that so-called "excess argon" could have become trapped in the Mount St. Helens magma. This is where we find the bulk of the confusing complexity in Austin's paper and in those of his critics. The papers all go into great detail describing the various ways that argon-containing compounds can be incorporated into magma. These include the occlusion of xenoliths and xenocrysts, which are basically contaminants from existing old rocks that get mixed in with the magma; and phenocrysts, which are crystals of all sorts of different minerals that form inside the rock in different ways depending on how quickly the magma cools. 95% of these papers are geological jargon that will make your head spin: Page after page of chemical compositions, mineral breakdowns, charts and graphs, and all sorts of discussion of practically every last molecule found in the Mount St. Helens dacite.

Summarizing both arguments, Dr. Austin claims that xenoliths and xenocrysts were completely removed from the samples before testing, and that the wrong results are due to phenocrysts, which form to varying degrees in all magma, and thus effectively cast doubt on all potassium-argon testing done throughout the world. It's important to note that his arguments are cogent and are based on sound geology, and are often mischaracterized by skeptics. He did not simply use the wrong kind of radiometric dating as an ignorant blunder. He was deliberately trying to illustrate that even a brand-new rock would show an ancient age, even when potassium-argon dating was properly used.

Austin's critics charge that he ignored the probable likelihood that the limitations of Geochron's equipment accounts for the results, just as Geochron warned. They also charge that he likely did not remove all the xenoliths and xenocrysts from his samples. However, neither possibility can be known for sure. Certainly there is no doubt that the test was far outside the useful parameters of potassium-argon dating, but whereas critics say this invalidates the results, Austin concludes that his results certify that the test is universally useless.

If we allow both sides to have their say, and do not bring a bias preconditioning us to accept whatever one side says and to look only for flaws in the other side, a fair conclusion to make is that both sides make valid points. Austin does indeed identify a real potential weakness in potassium-argon dating. However he is wrong that his phenocrysts constitute a fatal flaw in potassium-argon dating previously unknown to geology. In fact, the implications of phenocrysts were already well understood. Yes they are one of the variables, and yes, in some samples they do push the error bars. However, the errors they introduce are in the range of a standard deviation, they are not nearly adequate to explain errors as gross as three or more orders of magnitude, which would be necessary to explain the discrepancy between the measured age of rocks and the Biblical age of the Earth.

Such variables are also a principal reason that geologists never rely on just one dating method, with no checks or balances. That would be pretty reckless. For most rocks, multiple types of radiometric dating are appropriate; and in practice, multiple samples would always be tested, not just one like Austin used. In combination, these tests give a far more complete and accurate picture of a rock's true age than just a single potassium-argon test could. In addition, stratigraphic and paleomagnetic data can often contribute to the picture as well. From many decades of such experience, geologists have excellent data that guides proper usage of each of these tools, and they don't include gross misuse of potassium-argon dating.

What Austin did was to exploit a known caveat in radiometric dating; dramatically illustrate it with a high-profile test using the public's favorite volcano, Mount St. Helens; and sensationalize the results in a paper that introduces nothing new to geologists, but that impresses laypeople with its detailed scientific language. Occasionally scientists do actually make huge discoveries that everyone else in their field had always missed, but such claims are wrong far more often than they're right; and Dr. Austin and his finding that radiometric dating has always been useless is a perfect example.

By Brian Dunning
Follow @BrianDunning