Radiometric dating is not a simple topic. Chances are, you learned a simplified version of its techniques at one point—if you remember your chemistry teacher discussing isotopes, half-lives, and hourglasses, well, that was it—but have since tossed the lesson into a box labeled “High School Amnesia” in some dark corner of your brain. If you’re reading this now, however, you might be curious to reopen that box in an effort to follow my argument as I answer the title of this post (or, if nothing else, to avoid admitting that chemistry was “not really your thing”). But whatever your passion for decaying metals and your level of chemical comprehension, I want to share my confidence that you can follow along just fine. Anyone can learn technical jargon (queue Wikipedia page for Potassium Argon Dating); reading this post only requires a knack for scientific reasoning.
Before I begin, there is one set of terms you should be able to distinguish: radiometric dating is a method of estimating the age of geological events using radioactive isotopes in minerals; radioactive dating is a romantic relationship between colleagues at the nuclear power plant and has very little to do with geology. Confusion of these terms is a common sign of unfamiliarity with the topic, so now that you are better prepared, let’s continue!
Introduction to the controversy
Over the years, Answers in Genesis has committed to undermining the credibility of radiometric dating techniques. The underlying motivation is obvious: all dating techniques consistently yield age estimates that are far greater than 6–10,000 years, conflicting with their purported age of the Earth. Despite technological improvements to various methods, which resulted in “adjusted” age estimates and the discovery of some invalid assumptions, scientists have long been confident in ascribing a 4.5-billion-year age to our planet. Believing the Earth is young in the 20th century requires casting serious doubt on scientists, whose age estimate must be wrong by a factor of more than 400,000.
About ten years ago, the Radioisotopes and the Age of The Earth (RATE) team was formed to combat the traditional interpretations of geochronologists (people who develop/practice radiometric dating methods), and reopen a case that seemed already to be closed. To their credit, they have remained prolific and optimistic about their findings, despite widespread ridicule from academia and the general public. Unfortunately, I feel their optimism is unwarranted and believe the sheer length of their individual publications serves to hide the invalidity of their interpretations from faithful readers, who depend on them for answers.
So where to start? I imagine that I will return to this topic some time in the future, as AiG has published a number of articles and books that discuss radiometric dating methods (and the Potassium-Argon method in particular). For now, I wanted to consider an older article, only one page long, entitled “How do you date a New Zealand volcano?” Although the article was not published by any member of the RATE team, it provides a simple example of AiG’s critical approach:
- Remind readers that several assumptions are inherent to radiometric dating methods.
- Provide a case-in-point where at least one of those assumptions was falsified.
- Extrapolate the proven uncertainty to the rest of geochronology without qualification.
- (optional) Advise readers that anyone defending radiometric dating methods is trying to undermine God’s clear teaching of a young Earth and, consequently, the gospel itself.
Regardless of the article’s technical level, AiG and RATE team authors will make all 3 or 4 points in their publications, so I decided this jargon-free, ‘short enough to read on your smoke break’ commentary about New Zealand volcanoes was a great place to start. And in case you don’t actually have smoke breaks, then keep in mind that simply reading about natural wonders that smoke is a healthy substitute.
Overview of the Potassium-Argon (K-Ar) radiometric dating method
Many minerals, such Feldspar and Mica, contain significant quantities of potassium (K). Less than 1% of this potassium occurs as 40K, which is the radioactive isotope. For any given element, an isotope refers to the forms with different numbers of neutrons, while the number in front of the element (in this case, 40) refers to the total atomic mass of that particular isotope. Since neutrons have no charge, they don’t affect the chemical behavior of an element (besides its mass). Therefore, any mineral that contains potassium (K) will contain a mixture of all its isotopes (39, 40, and 41).
When 40K decays radioactively, it produces both 40Ar (argon) and 40Ca (calcium), with a half-life of 1.25 billion years. For example, imagine you crsytallize a rock with 1 gram of 40K (and no 40Ar). If you came back after 1.25 billion years, and assuming nobody has heated the rock or altered it chemically, you would find 1/2 grams of 40K and 1/2 grams mixture of 40Ar/40Ca. After another 1.25 billion years, you should find 1/4 grams of 40K and 3/4 grams of 40Ar/40Ca. Therefore, one can use the measured ratio of potassium to argon in a given mineral to infer the time at which the mineral crystallized and began to accumulate argon. Typically, one assumes that no argon (or a negligible amount thereof) was initially present, because argon is a noble gas and can easily diffuse out of minerals that are still hot. If any excess argon becomes trapped in the mineral during crystallization, the mineral will appear older than it actually is; if any argon is lost after the mineral crystallizes, the mineral will appear younger. As you can imagine, the chaos of Earth systems can produce both scenarios, so geochronologists have developed techniques to verify (test) each assumption.
Before moving on, I must clarify one thing. Most people (geologists included) think of radiometric dating methods as a means to assign absolute ages to rocks/minerals. This definition can be misleading, however, without some qualification. First, all radiometric dating methods are scientific models used to estimate the age of geologic events, provided a number of physical assumptions regarding the rock’s history are met. In this sense, it is much like estimating the origin of a cannonball in flight, using a set of physical observations and the laws of gravity. Secondly, radiometric dating methods (K-Ar in particular) do not estimate the age of a rock, but the time at which a mineral in that rock was last near a given temperature (called the closing temperature). For any given rock, each kind of mineral will yield a different age, depending on how quickly the rock cooled. If the rock was reheated at any point, the method no longer provides a straightforward interpretation of the cooling age (hence all dates are termed “apparent ages”). While such geological complexities pose additional challenges to geochronologists, even “bad” dates can be very useful. My hope is to convince you this is the case through the following example, and that AiG had prematurely discredited the K-Ar dating method.
How do you date a New Zealand volcano?
Robert Doolan of Answers in Genesis concluded that the K-Ar method is not a valid option in response to the above question. His argument goes like this:
- We can use the distribution of vegetation, tree rings, carbon-dating from wood samples buried in ash, and even historical reports to date a number of recent volcanoes in New Zealand.
- Therefore, even by “evolutionist” standards, we know from multiple lines of evidence that the volcanic eruptions occured between 50,000 and 300 years ago (i.e. they are recent in either paradigm).
- However, radiometric dates, using the K-Ar (potassium-argon) method, yield ages of 145,000 to 465,000 years for the youngest volcano!
- Since we know these are false ages due to excess argon being trapped in the cooling lava flows, we should not trust the K-Ar method to date volcanic rocks.
At first glance, Mr. Doolan provides a convincing case against the credibility of K-Ar dating. I remember reading similar reports years ago, which cited numerous cases of historical lava flows (historical meaning recorded by humans) that were dated radiometrically to be a few hundred thousand to millions of years old. I was originally quite persuaded by the discrepancy and concluded that radiometric dating methods were fundamentally flawed: “If radiometric dating methods are so wrong when the age is known, how can we trust them when the age is unknown?” Now, my concern is that to the non-scientist (or even to the experienced scientist that doesn’t regularly work with geochronology), this reasoning may seem plausible and end the debate without warrant. But if the failure of the K-Ar dating method is so obvious, why do scientists still spend so much money on running samples? Is there a grand conspiracy to hide the flaws, which are so simple to point out?
As you might expect, it is never that simple. So here is my analysis:
1) The author is either ignorant of his source or is being intentionally deceptive
Mr. Doolan first explains that the largest volcano is the youngest. This is true, but he does so in a way that would make you think scientists either doubted that young age, or figured they could use K-Ar dating to come up with a “final answer.” (If you did not get this impression from the first paragraph, then my point here is invalid, but I’ll continue nonetheless) First, Mr. Doolan says: “In the late 1960s, scientists from the Australian National University in Canberra dated numerous volcanoes in Auckland using the potassium-argon method…Results seemed to show that Rangitoto was not a few hundred years old as it appeared to be.” Then he notes that “In every case the potassium-argon dates were clearly wrong to a huge extent,” so “If the real dates were not fairly well established by other means, who could have proved that the potassium-argon dates were so wrong?” Unfortunately, it appears that he expected his readers to assume that since he quoted a reputable journal source, he must have done his homework. Anyone with access to a university library system can check whether this is true, but most AiG readers (and I don’t blame them) wouldn’t care to take the time to find a 40-year old journal article. Since I always have a search engine open for journal articles, I was able to find it rather quickly. One only need to read the abstract to get my point here, where McDougall writes:
“Because of the good age control, this area was chosen for a detailed study to test whether the K-Ar dating method could be used for dating such young basaltic rocks.”
In other words, the original study was not by any means carried out to determine the age of volcanoes, but rather to test to the model assumption that no argon would be present in newly crystallized rocks. As you can see, they found this assumption to be false—all samples contained between 1–5 x 10-13 moles/g of argon. So did the original authors warn others to reject the K-Ar method and that it could never give us insight to the true age of rocks? Of course not.
But why not? The first reason is that the amount of initial Argon, while detectable, is very very small. I realize that to most people an error of 400,000 years seems substantial, but if the volcano you are dating is actually 500 million years old then it makes no difference that the “clock” started at 400,000. The second reason is more profound, but before I elucidate, I would conclude that Mr. Doolan has clearly misused the source he cited: McDougall et al. (1969) did not attempt to date the volcanoes by the K-Ar method but rather used well-dated volcanoes to improve the use of the K-Ar method on volcanic rocks.
2) Despite an attempt to discredit the K-Ar method, the author cited a source that actually proved the effectiveness and reliability of the K-Ar method, even when model assumptions are invalid
Now the more important question becomes “Why is there argon trapped in cooling lava and why doesn’t that invalidate the model?” The first answer is simple: while argon can easily diffuse from minerals when temperatures are high, the partial pressure of argon in the atmosphere often causes trace amounts to remain trapped in the crystal structure (or in fluid inclusions, which are pieces of melt that get trapped inside of crystals—think of an air bubble being trapped in a piece of ice). This process is obviously more important in rocks that solidify at the surface (i.e. lava flows), but may be true for minerals that crystallized deep in the Earth as well. In order to get to the surface, magma must make its way through thick portions of country rock, which is much older and may contain a significant amount of radiogenic argon. While the magma rises to the surface, it heats up the surrounding rock and argon is transferred to the melt. This argon is potentially captured in the crystal structure of forming minerals, which would give it an anomalously old date.
Geologists have known about this problem since the K-Ar method was first put forth, and McDougall et al. (1969) is a prime example. Therein, the authors explain contamination processes and consider the isochron method to solve the problem (I won’t describe that process in detail here; their results are given on p. 1507 (Figures 5-8) in the article for those interested). The isochrons plot well, and form statistically significant lines (i.e. they are internally consistent). Each isochron yields “apparent ages” from ~70,000 to 534,000 years (anomalously old). While they disagree on the “age”, all isochron lines point consistently to a high initial 40/36 argon ratio. In other words, there was more initial argon than might be expected from a purely atmospheric source.
McDougall et al. (1969) continued to investigate the source of excess radiogenic argon by analyzing the mineralogy of the samples. Scattered fluid inclusions in olivine/pyroxene could be found throughout the sample. Xenolithic quartz was also found, indicating that the magma had incorporated foreign material during its assent to the surface. While 87/86Sr ratios and whole rock chemistry indicate that argon enrichment from the surrounding sedimentary rocks was very small, it would have been sufficient to provide the excess radiogenic argon that resulted in dates that were obviously too high. Thus McDougall et al. (1969) concluded that one may not rely solely on the K-Ar method when dealing with “young” volcanic basalts, even when the isochrons are internally consistent. This method should always be combined with thorough petrographic analyses to constrain the degree of contamination from the atmosphere and wall rock, as well as other radiometric dating methods like radiocarbon and uranium-series disequilibrium.
While the resultant isochron ages were obviously “false”, they did produce internally consistent isochron ages. This means that if one came back in 100 million years and dated the same volcanic rock, it would produce an age that is 70-535,000 years too old, or 100.07-100.535 million years, as opposed to 100 million years. In other words, this study confirms the use of K-Ar isochrons in older volcanic rocks (since the true age would be within the uncertainty range of your model age). To summarize thus far:
- The purpose of McDougall et al.’s study was to test the reliability of the K-Ar method in basaltic rocks that are known to be very young. The authors found the method to be unreliable in isolation, but readily explained the discordant ages using thorough petrographic and geochemical analyses.
- While it may sound comforting to AiG readers that the ages appeared much older than they actually were (half a million years vs. 300 years), it should not. As documented in the McDougall paper, the excess radiogenic argon had to come partly from sedimentary rocks surrounding the magma chamber. This means radiogenic argon had been accumulating in those rocks for hundreds of thousands, if not millions of years before it was incorporated into the erupting basalts. Even if all K-Ar ages are invalid, one must still deal with the physical reason they are invalid.
- Lastly, the article cited was published in 1969. Radiometric dating techniques were rudimentary, in that they required large samples and a steady hand. McDougall et al. predicted that fluid inclusions in olivine/pyroxene could have been largely responsible for the excess argon, but had no way to test this, since fluid inclusions were far too small to be analyzed individually. So wouldn’t it be nice if someone used more recent technology to test such predictions? That brings me to my next point.
3) AiG needs to update their article database
Cassata et al. (2008) published a paper in Earth and Planetary Science Letters in 2008, entitled “Laschamp and Mono Lake geomagnetic excursions recorded in New Zealand”. It is fairly concise, but in depth. The most important statement relative to our discussion is this:
Most experiments yielded concordant or weakly discordant age spectra that revealed little evidence of excess or inherited argon, indicating sample preparation procedures sufficiently removed potential sources of extraneous argon. To confirm this, experiments were conducted on lava that erupted during historical time at Rangitoto, which had previously resulted in discordant K–Ar ages ranging from 146 ± 12 ka to 465 ± 11 ka (McDougall et al., 1969). The purified groundmass from this lava resulted in an age spectrum in which all but the highest temperature steps yielded a zero age owing to negligible quantities of radiogenic argon. Similarly, the highest temperature steps for all samples occasionally yielded apparent ages significantly older than the plateau age and distinct from the isochron array, most likely a result of excess argon in melt inclusions within fragments of incompletely removed olivine and pyroxene (McDougall et al., 1969; Esser et al., 1997). These steps were imprecise and were not consistently reproduced in successive experiments on additional subsamples, indicative of incomplete removal of phenocrysts and hence a small degree of sample heterogeneity. (p. 82, emphasis mine)
In short, Cassata et al. (2008) used the same method to analyze rocks from the same volcano, but with newer technology. They determined that according to the K-Ar method, the age of the volcano was indistinguishable from zero (i.e. consistent with the historical record), except when the minerals were analyzed at the highest temperature. Release of excess argon at high temperatures suggests the presence of contaminants within the mineral, and explains the anomalously old ages obtained by earlier studies.
Fortunately, we are now able to obtain elemental and isotopic ratios from much smaller samples with far better precision than 40 years ago. Contamination can often be accounted for, along with the loss of argon. Now, instead of dating a whole sample of rock/mineral, one can obtain a number of “apparent ages” from various locations within a single crystal, quantifying the diffusive loss of argon in the sample and accounting for microvariations in potassium content from zoning. Such technology has been used in recent years to confirm early suspicions of geologists that tried to interpret discordant dates.
In this particular article, I think it is clear that Mr. Doolan has employed argumentation that is misleading, if not deceptive (though I will grant the benefit of the doubt, and assume he is not intending to deceive anyone). Granted, I picked an article that was rather short and old, so I will aim to verify whether newer publications by the RATE team have added anything qualitatively to the debate. On the other hand, while I disagree with the scientific conclusions of Mr. Doolan and others at AiG, I am not hesitant to point out where they have raised valid points (in this case, for example, the presence of excess radiogenic argon is a valid problem when using the K-Ar dating method, and warrants objection to its accuracy in some studies). Overall, I only wish to encourage consistency in scientific arguments, especially when those arguments are aimed at more pertinent matters than historical geology.