Andrew A. Snelling started his “‘Excess Argon’: The ‘Achilles’ Heel’ of Potassium-Argon and Argon-Argon ‘Dating’ of Volcanic Rocks” with the following:
For more than three decades potassium-argon (K-Ar) and argon-argon (Ar-Ar) dating of rocks has been crucial in underpinning the billions of years for Earth history claimed by evolutionists. Critical to these dating methods is the assumption that there was no radiogenic argon (40Ar*) in the rocks (e.g., basalt) when they formed, which is usually stated as self-evident. Dalrymple argues strongly:
The K-Ar method is the only decay scheme that can be used with little or no concern for the initial presence of the daughter isotope. This is because 40Ar is an inert gas that does not combine chemically with any other element and so escapes easily from rocks when they are heated. Thus, while a rock is molten, the 40Ar formed by the decay of 40K escapes from the liquid.1
However, this dogmatic statement is inconsistent with even Dalrymple’s own work 25 years earlier on 26 historic, subaerial lava flows, 20% of which he found had non-zero concentrations of 40Ar* (or excess argon) in violation of this key assumption of the K-Ar dating method.2
The historically dated flows and their “ages” were:
[Ma is a million years]
Hualalai basalt, Hawaii (AD 1800-1801) 1.6±0.16 Ma; 1.41±0.08 Ma
Mt. Etna basalt, Sicily (122 BC) 0.25±0.08 Ma
Mt. Etna basalt, Sicily (AD 1972) 0.35±0.14 Ma
Mt. Lassen plagioclase, California (AD 1915) 0.11±0.03 Ma
Sunset Crater basalt, Arizona (AD 1064-1065) 0.27±0.09 Ma; 0.25±0.15 Ma
Far from being rare, there are numerous reported examples of excess 40Ar* in recent or young volcanic rocks producing excessively old K-Ar “ages”:3
[Snelling gives a list of some.]
…
1 G.B. Dalrymple, The Age of the Earth (1991, Stanford, CA, Stanford University Press), p. 91.
2 G.B. Dalrymple, “40Ar/36Ar Analyses of Historic Lava Flows,” Earth and Planetary Science Letters, 6 (1969): pp. 47-55.
I take severe exception of Snelling’s characterization of Dalrymple’s writings. In first work which Snelling quoted, Dalrymple did briefly mention, in the same paragraph quoted, the problem of excess argon. Far more importantly lets look at that 1969 paper of 26 historic lava flows that Snelling quoted. Dalrymple concluded in it:
With the exception of the Hualalai flow, the amounts of excess 40Ar and 36Ar found in the flows with anomalous 40Ar/36Ar ratios were too small to cause serious errors in potassium-argon dating of rocks a few million years or older. However, these anomalous 40Ar/36Ar ratios could be a problem in dating very young rocks. If the present data are representative, argon of slightly anomalous composition can be expected in approximately one out of three volcanic rocks.
Thus while Snelling implied that Dalrymple found severe problems with K-Ar dating when the truth is quite the opposite. Dalrymple found that they are reliable. Two-thirds of the time there is no excess argon at all. And in 25 times out of 26 tests there is no excess argon or there is so little excess argon that it will make only a tiny error, if any, in the final date for rocks millions of years old. Thus Dalrymple’s data is not consistent with a young Earth whatsoever. Indeed, if Dalrymple’s data is representative, 3 times out of 26 the K-Ar method will give a too young date (though by only an extremely trivial amount for a rock that is really millions of years old). The one case that would have produced a significant error, the Hualalai flow in Hawaii, was expected (see the previous essay). Even that significant error is only 1.19 million years (and not the 1.60 million years that Snelling claimed). If the identical rock had been formed 50 million years ago, the K-Ar would give a “false” age of a little over 51 million years. Thus this data is strongly supportive of mainstream geology. Here is a list of the 8 (out of 26) cases, which had excess argon, scanned from the original paper:
And remember that the 18 (out of 26) lava flows not listed above had no excess argon and thus dated correctly to zero million years.
Thus if the young-earth creationists, like Snelling, are trying to use the problem of excess argon to show that radiometric dating is wrong, they have failed miserably. The need to show that K-Ar consistently, not sometimes, gives bad dates. They need to show that K-Ar’ bad dates make the rocks look too old and not sometimes too young. And they need to show that the quantity of the errors is often several thousand times greater than the errors shown here. And then they would have to show the same for all the other many other forms of dating.
Let’s go on to another false claim by Snelling. In the second sentence he claims that both K-Ar and Ar-Ar methods assume there is no 40Ar* in the rock when it formed. That is not true. The original K-Ar method does make that assumption and the majority of the time it is a reasonable assumption. However the Ar-Ar method, while it uses the same radioactive decay scheme as K-Ar, does not make that assumption. The Ar-Ar makes no assumptions about the argon content of the rock when it formed. Furthermore, unlike K-Ar, Ar-Ar does not assume the rock is a closed unit. Ar-Ar can sometimes date rocks that have lost argon over time. (Though sometimes the Ar-Ar procedure will just indicate that the system has been disturbed and will not give a clear date.) For a brief explanation on how this works see this and for a more advanced treatment see that.
Now let’s look at a real life example. “40Ar/39Ar Dating into the Historical Realm: Calibration Against Pliny the Younger” was written by P. R. Renne et. al. and published in Science 277: 1279-1280 (1997). They tested Ar-Ar dating by checking it against the 79 A.D. eruption of Vesuvius that is famous for destroying Pompeii. Now they note that “Analysis of single crystals, for example by laser fusion, can obviate xenocrystic contamination, but single crystals are seldom large enough to yield measurable quantities of 40Ar* through radiogenic ingrowth in the Holocene [i.e. last 12,000 years].” Would Ar-Ar work such recent material?
It did. They got an age of 1925±94 years. The true age was 1918 years when the test was done. The test was off only 7 years and was correct within the margin of error. Seven years is only 0.36% of the true age. The reported margin of error for the measurement was a bit under five percent of the true age.
The material dated in this experiment did have excess argon. If the creationists had sent a sample to a K-Ar dating lab like they did for Mount St. Helens, they would have gotten a date that is too old. This is one of the reasons that Ar-Ar test is now preferred over K-Ar in spite of K-Ar’s lower cost.
The scientists concluded:
Thus despite the presence of excess 40Ar, a sample less than 2000 years old can be dated with better than 5% precision, validating 40Ar/39Ar dating as a reliable geochronometer into the late Holocene. These results also demonstrate that excess 40Ar can be identified in volcanic sanidine, and while perhaps negligible in pre-Holocene rocks, it has important consequences for sample at the limit of the method’s applicability. Further improvement in precision of 40Ar/39Ar analysis of historically dated samples may lead to welcome refinements in the ages of neutron fluence monitors, currently a limitation on the accuracy of the 40Ar/39Ar method. Our results also substantiate validity of the 40Ar/39Ar method in establishing the eruptive histories of populated active volcanic regions, where such information is vital to volcanic hazard assessment.
In short, the method works.
So why is Snelling telling us about problems in the K-Ar method and not tell us that Ar-Ar largely fixes them. And why does he tell us in several articles that both the K-Ar and Ar-Ar methods assume the rocks had no initial argon when only the K-Ar method makes that assumption?
Two months after Snelling published his article quoted above he continued his nonsense in “Potassium-Argon and Argon-Argon Dating of Crustal Rocks and the Problem of Excess Argon.” That article tells us:
An Ar-Ar “dating” study of high-grade metamorphic rocks in the Broken Hill region (New South Wales) found widely distributed excess 40Ar*.13 Plagioclase and hornblende were most affected, step heating Ar-Ar “age” spectra yielding results up to 9.588Ga. Such unacceptable “ages” were produced by excess 40Ar* release, usually at 350-650°C and/or 930-1380°C, suggesting excess 40Ar* is held in sites within respective mineral lattices with different heating requirements for its release. Thus at crustal temperatures, which are less than 930°C, some excess 40Ar* will always be retained in those trapping sites in minerals where it is “held” more tightly. A viable interpretation of these Broken Hill data was only produced because assumptions were made about the age of the rocks and of a presumed subsequent heating event (based on Pb-Pb and Rb-Sr dating), when it is conjectured that accumulated 40Ar* was released from minerals causing a significant regional Ar partial pressure of ~3 x 10-4atm.
…
13. T.M. Harrison and I. McDougall, “Excess 40Ar in Metamorphic Rocks from Broken Hill, New South Wales: Implications for 40Ar/39Ar Age Spectra and the Thermal History of the Region,” Earth and Planetary Science Letters, 55 (1981): pp.123-149.
This paragraph is pure misrepresentation. People who don’t understand how Ar-Ar works will be impressed that the age spectra had an 9.588 Ga date. (Ga is a billion years). This is far older than the Earth. So has Snelling disproven Ar-Ar merely by telling us this? Consider the just mentioned dating of Vesuvius had apparent ages up to 0.5212 Ma, which is almost 272 times the true date of the eruption as recorded by Pliny the Younger. And yet the test gave the correct result.
Let’s examine the paper they cited and see what really happened. Previous researchers had dated when the Broken Hill rocks underwent metamorphism. One researcher used the rubidium-strontium (Rb-Sr) method to get a result of 1660±10 Ma. Another researcher used lead-lead (Pb-Pb) whole rock isochron method to get 1660±16 Ma. Please note that Rb-Sr and Pb-Pb are completely independent dating techniques and they were done by different researchers. If these techniques did not work, the odds that they would give the same results by chance is very small.
Now let us look at one of the many Ar-Ar tests that this paper reported. Sample 79-171 had excess argon and apparent ages up to 3.987 Ga (3987 Ma). Lets look at the results in graphical form:
 |
 |
I will provide an over-simplified version of what these two figures mean. Ar-Ar dating is a variant of K-Ar dating. The sample is put into a nuclear reactor that changes some of the 39K into 39Ar. The 39Ar, which serves as a measure for potassium, and 40Ar can be measured in a single experiment that uses far less material than regular K-Ar methods. Now lets pretend that the rock started out with some excess argon at time zero. When the rock is found and measured in the laboratory millions of years latter it will have the initial argon plus the 40Ar generated by the radioactive decay of the potassium. But we have the problem of how we can tell the difference between the initial argon and the argon generated by the decay of potassium. The solution comes from the fact that for every part of the rock the amount of argon generated radioactive decay will always be the identical multiple to the amount of potassium. That is not an ad hoc assumption but a simple observed fact of how radioactivity works. Any initial argon or argon gained or lost through time will not act in that way. There is no reason for them to do so except through a rather unlikely coincidence. Thus any excess argon, or non-radioactive argon loss or gain, will be revealed by checking multiple parts of the rock. One method is to use a laser and get dates for a bunch of tiny spots. The method in the paper in question is called an age spectrum. The sample is heated incrementally. This liberates its argon in stages because some argon is more tightly bound than other argon. If there was no excess argon and the rock was closed (did not gain or loose argon to the environment) the apparent age of each heating stage will be the same. If these assumptions are not true then the age spectrum will be more complicated and the geologist can often determine from the spectrum what happened. Thus the Ar-Ar dating can be used with rocks with complex histories that the K-Ar method would be inappropriate method.
Often the age spectrum from an Ar-Ar procedure will show a sequence of stages that the apparent date was the same. The age spectrum of 79-171 shows such a pattern that is called an age plateau. When a distinct plateau is formed that plateau usually indicates the age the researcher wants. This can be further tested by means of what is called the isochron plot that is shown above to the right of the age spectrum. For the Ar-Ar method, the isochron plot plots two argon isotopic ratios (39Ar/36Ar and 40Ar/36Ar). For information of isochron methods in general click here. 36Ar is not involved in the radioactive decay. By dividing both the parent (measured via 39Ar) and the daughter (40Ar) by 36Ar, it can be shown mathematically that age is determined by the slope of the resulting line and the 40Ar/36Ar at time zero is determined by the y-intercept (not assumed). If the data does not form a line, the assumptions have been violated, and the test will not generate an age. The odds that a line will form a line if the assumptions of the method are not true is very small. Thus this isochron, like other kinds isochrons, is self-checking. Sample 79-171 had the six stages that formed its plateau plotted in an isochron plot. The six points fall on almost perfect line. The isochron plot of these six stages is very strong evidence that they have not been affected by excess argon. The slope gives an age of 1573±5 Ma. This age is one we can have confidence in.
Now remember that the Rb-Sr and Pb-Pb dates were both 1680 Ma. The Ar-Ar date is 1573 Ma. Now if these dates were all supposed to reflect the same event then there would be a significant error. (Though the young-earthers would not have anything crow over a 6% difference dates since they need discrepancies of thousands of percents.) However these particular dates do reflect the dates of different events. The Rb-Sr and Pb-Pb dates are when the rocks were experienced metamorphism. These dating systems became closed at about 800°C. However sample 79-171 was done in a mineral called hornblende that does not trap argon until it cools to around 500°C. The Ar-Ar age is for when the complex, which the sample came from, cooled down to that temperature. That means the temperature cooled at about 3 degrees per million years. (The mass that was cooling was huge.) By using different minerals and dating systems that close at different temperature, the researchers were able to deduce a bit of the geologic history of the rocks. (Joe Meert provides a great example of this in his “Radiometric Dating, Paleosols and the Geologic Column: Three strikes against Young Earth Creationism” and “Consistent Radiometric dates.”) A sample of muscovite, which has a closing temperature of about 350°C, gave an Ar-Ar date of 1507±7 Ma. Notice the pattern as cooling proceeds, the date becomes younger as should be expected. Two Rb-Sr dates for this same system gave 1490±20 Ma and 1495±20 Ma. These three measurements agree quite well. The mineral biotite was used to demonstrate a heating event around 520 Ma and fission track dating showed that it had cooled to about 100°C around 280 Ma.
And so on and so forth. Using these and other indicators the researchers were convincingly able reconstruct some geological history. The results are self-consistent. And yet Snelling wants us to believe that the researchers are just making up assumptions to explain away the data. That is simply not the case. This is simple a case of the application of techniques that work.
Also see:
“Rate:
More Faulty Creation Science from the Insitutute for
Creation Research” - Bad radiometric dating done
by young-earth creationists