Excess argon is argon that is incorporated in a rock as it cools or
argon that enters later on, and did not result from the decay of
potassium in the rock. Such argon can artificially increase K-Ar
dates and make them much too old. Since a considerable portion of the
geological time scale is based on K-Ar dating, excess argon, if widely
present, could be making much of the geological time scale excessively
I have been suggesting that excess argon is common, invalidating K-Ar dating on the Phanerozoic strata, where most fossils are found. In his last reply, Dr. Henke states that excess argon is rare, in the following quotations:
Perhaps, Faure (1986) and Dickin (1995) don't mention the method because excess argon is not as much of a problem as Dr. Plaisted and other creationists believe and these 3-D methods rarely need to be used (Alexander, 1983, p 353).Here are some quotations from Woodmorappe (1999, pp. 31-32) about the prevalence of excess argon:
I NEVER said that excess argon never occurs. I said that it was RARE and as I discussed in my last response, McDougall et al. (1969) is an example of excess argon in volcanics. Rapidly quenched submarine pillow basalts may trap argon, at least initially. This is why the centers of the pillows are dated rather than the rims.
NO ONE is saying that excess argon is NEVER a problem. Sometimes, RARELY, it is. That's why K-Ar and 3 -dimensional methods are available. House fires aren't common either, but we have fire extinguishers just in case.
Any component of 40Ar other than 40Ar derived from atmosphere and radiogenic decay, is termed excess argon, and is widely recognized in dated materials ... . Large plutons are particularly prone to contamination by excess 40Ar due to incomplete re-equilibration with atmosphere at the time of crystallization and assimilation of radiogenic 40Ar from the adjacent country rocks during hydro-thermal alteration (Tegner et al 1998 p. 79).Woodmorappe (1999, p. 32) also notes that McDougall and Harrison (1988, p. 11) consider excess argon to be "not uncommon." He mentions that it commonly occurs in igneous rocks over a fairly large geographical area, such as a swarm of dikes in western Arizona (Nakata 1991, p. 25), the Kola Peninsula of Russia's far east (Ivanenko and Karpenko 1988), and rocks in the 1 x 2 quadrangle of west-central Colorado, referred to as follows:
One of the limitations of the K-Ar method for dating minerals is their excess 40Ar, which is responsible for the anomalously old age determined for these minerals (Morozova et al 1997, p. 716).
A significant problem in K/Ar isotopic dating is the siting of "excess" 40Ar acquired by minerals from their environment (that is, 40Ar not produced by radiogenic decay within the mineral) (Cumbest et al 1994, p. 942).
K-Ar dating is based on the decay of potassium to 40Ar. A major constraint is the possibility that a mineral may contain excess radiogenic argon, which results in anomalously high ages (Morozova et al 1996, p. 52).
Potassium-argon ages much greater than are geologically plausible often are observed, and in some cases they are older than the age of the earth (McDougall and Harrison 1988, p. 106).
Indeed the presence of excess argon in amphiboles appears to be much more common than previously suspected (McDougall and Harrison 1988, p. 28).
There have been numerous cases reporting the presence of excess argon in biotite ... . (Hyodo and York 1993).
Where present, this excess Ar yields an anomalously old age and, in certain environments, is a common phenomenon in some minerals, especially biotite (Smith et al 1994b, p. 808).
Nevertheless, many Proterozoic and Tertiary rocks have produced Paleozoic and Mesozoic dates. For the Tertiary rocks, all of which are plutonic, the abnormally old K-Ar dates probably can be attributed to excess argon (Wallace 1995, p. 6).Woodmorappe (1999, p. 31) also mentions the title of a GSA abstract of Poths et al (1993): Ubiquitous Excess Argon in Very Young Basalts.
Dr. Henke notes that geologists can infer some things about the history of a rock from their effects. Heating, weathering, and chemical changes in a rock can often be detected. However, inherited argon can be present in a rock without any evidence of weathering, metamorphism, or chemical change, and all the diagnostic tests for it have limitations. Therefore there is no way that we can be sure that the argon in a rock is not excess argon.
The existence of excess argon is usually not inferred from isochrons or the spectrum test or any similar test, but generally simply because the ages are older than expected (Woodmorappe, 1999, p. 32):
Also, in most cases the explanation for the excess argon is given without any real proof, while the identification is usually based either on comparing the results with the precisely established geological age or by comparison with other isotopic methods (Rublev 1985, p. 73).Another possible evidence that argon in minerals is excess argon is simply the fact that minerals with less potassium tend to give higher K-Ar ages, suggesting that the argon in the rock did not result from decay of potassium in the rock at all (Woodmorappe, 1999, p. 94):
That is, minerals with low potassium contents, and inheriting variable amounts of argon, would tend to give higher ages (than high-K minerals such as biotite) by virtue of this fact, and this would be especially true of rocks having very low potassium contents (the ultra-mafic rocks -- which are indeed particularly prone to give absurdly high K-Ar dates: Phillips et al 1998).Perhaps at one time it was believed that excess argon was rare, but this no longer appears to be true.
Where did all of this excess argon come from? It could have come from the same place as the argon in rocks that have significant argon but no potassium, or the argon in many rocks acknowledged by geologists to have excess argon, whatever that source is.
The prevalence of excess argon on the Phanerozoic and the centrality of K-Ar dating for the geological time scale together imply that the entire Phanerozoic time scale may be in jeapordy. It wouldn't surprise me if geologists decide that excess argon is invalidating the whole Phanerozoic time scale, and start over from scratch. It is true that other dating methods than K-Ar are used for this time scale, but a re-evaluation of K-Ar dates might cause these others to be reinterpreted as well.
Now, Dr. Henke argued that lava cracks as it cools, so that argon from inside the lava can escape through the cracks, implying that excess argon should be rare. He used the analogy of a balloon, from which air escapes rapidly through a small opening. However, the surface of a balloon is flexible so that escaping air can enlarge the hole. Cooled lava is rigid, so escaping gas can only enlarge the cracks by blowing away pieces of rock, which rarely happens except in volcanic craters during eruptions.
Dr. Henke also argued that sediment is permeable to argon, so that argon should be able to escape from intrusive magma (that is, buried in sediment). However, the flow of a gas is proportional to the gradient of partial pressure and to the permeability of the medium, I believe. Assuming the permeability of sediment is roughly constant, for the sake of simplicity, and that most of the argon 40 is diffusing up from the mantle, this implies that at equilibrium the upward flow of argon is about the same everywhere. This means that the gradient of partial pressure of argon is about the same everywhere, which implies that the partial pressure of argon is highest in the mantle, has some value at the base of the sediment, and then decreases linearly as one travels upward through the sediment, finally reaching the atmospheric partial pressure at the top. The higher partial pressure of argon at greater depths would tend to imply that magma cooling at greater depths would inherit more argon, and thus look older based on K-Ar dating. This argon traveling up from the mantle might also enter rocks (such as glauconies) even after they cooled. Of course, the true situation is much more complex, but this shows that the permeability of sediment is not inconsistent with high partial pressures of argon there.
Isochrons can be used to test the amount of daughter product present
initially, and thus can in principle correct for excess argon and can
also be applied to other dating methods. Unfortunately, isochrons can
also be caused by mixings. There is a mixing test to detect this; if
a certain correlation is present, the isochron may be caused by a
mixing. However, even if the correlation is present, it does not mean
the isochron is caused by a mixing, and even if the correlation is
absent, the isochron could still be caused by a more complex mixing
(Woodmorappe, 1999, pp. 69-71). Therefore this test is of
I proposed a mixing that could lead to Rb-Sr isochrons of positive but not negative slope. Even though Dr. Henke did not seem to understand it, it is correct. The example can be generalized to permit an appreciable amount of Sr87 in the second source, as long as the proportion of Sr87 is less than in the first source. This will cause the isochron line to pass above the origin. The Sr87 in the first source could have originated from radioactive decay of Rb87 or fractionation.
I also proposed a mixing scenario that could explain old K-Ar isochrons in lava. This was challenged by Dr. Henke. My reasoning was as follows: Lava is often thoroughly mixed, which implies that the concentrations of potassium and argon should be relatively uniform. If the lava does not cool very slowly, then there will not be time for various substances to form crystals, so the concentrations of potassium and argon will remain relatively uniform even in the cooled lava, and there can be considerable excess argon, leading to an age that is much too old. In order to get an isochron, it is necessary to have a mixture of argon 40 and argon 36 diffuse into the mixture, more in some places than others. This can happen, for example, if there is a pocket of gas containing argon in the magma chamber, or the magma encounters some such pocket of gas on the way up during the eruption. This can also happen because of atmospheric argon diffusing in; even though more argon will diffuse out than in, individual gas molecules move in both directions, so some argon will diffuse into the lava as it cools.
There is also an interesting change of my text in Dr. Henke's reply. My original article said the following:
The assumptions of this process also seem natural in lava which does not cool slowly, as one would expect the concentration of potassium to be relatively uniform.Dr. Henke quoted me thus:
DAP: The assumptions of this process also seem natural in lava, which does not cool slowly, as one would expect the concentration of potassium to be relatively uniform.Dr. Henke added a comma, which significantly changes the meaning of the quote.
I earlier cited a study from Woodmorappe (1999, p. 70) showing that of 69 isochrons examined from the literature, only 9 had good statistical and geological quality. The criteria for good quality did not even include the lack of a mixing correlation. Of course, this lack does not preclude mixings, but it helps. It would be interesting to see what kind of dates these 9 isochrons give, and what kind of a Phanerozoic time scale one obtains if all dates are restricted to come from isochrons with good statistical quality and without mixing correlations.
I had stated that the geological time scale is based on fewer than 800
dates, as given in Harland et al (1990). Dr. Henke disputed this,
citing other dates that have been measured, and other time scales that
have been constructed. Dr. Henke states,
Harland et al. (1990, p. 79) state that EACH one of the 700 or so dates in their table are special in that they are well "bracketed" stratigraphically, which allows independent confirmation of the dates with fossil, paleomagnetic or other data.I realize this, but it does not change my point. These other dates do not influence the geological time scale, even though they may (or may not) be grounds for believing in the validity of radiometric dating. As for the previous geological time scales, they undoubtedly used fewer dates than Harland et al, since Harland et al attempted to include all the dates that they could from earlier studies.
There appears to be a trend for K-Ar dates to be older as one goes
deeper in the geological column, and K-Ar dates from the same layer
may often be roughly the same everywhere. If real, this trend
suggests that the K-Ar dates are measuring true ages, but the trend
might also be explained by some other mechanism, as I have indicated
before. A more convincing justification of the geological time scale
on the Phanerozoic would be a correlation between K-Ar dates and
non-K-Ar dates there.
It is therefore important to know whether there is a correlation between K-Ar dates and non-K-Ar dates on the Phanerozoic, since such a correlation would at least suggest that something real is being measured and that not all dates are due to excess daughter product. Such a correlation is not conclusive, however, because it could result from common chemical or physical behavior among several isotopic systems. In my last response I was trying to analyze the number of concordances among different dating methods to see if it gives evidence of such a correlation.
I gave a hypothetical statistical distribution of dates to show that if the non-K-Ar dates were random, then it would not be surprising to obtain the observed 30 concordances among the over 700 dates in Harland et al (1990). Thus the number of concordances is not impressive, and not evidence of a correlation between K-Ar and non-K-Ar dates. However, even non-K-Ar dates on formations with no measured K-Ar dates can be seen as concordances, if they agree with the expected ages of their formation obtained stratigraphically. Therefore all non-K-Ar dates can be seen as concordances, and need explanation. So the analysis of concordances should be done in the following way: How many non-K-Ar dates would have to be measured to obtain the observed number agreeing with the expected age of their geological layer to within 10 percent? The same analysis yields the result that for every such non-K-Ar date, 19 others would have to be measured and discarded, on the average. Even this seems reasonable, because for example glaucony is considered unreliable for K-Ar dating, and yet many K-Ar glaucony dates are in the table. Thus there must be many K-Ar glaucony dates that were excluded. Since glaucony can also be dated by Rb-Sr dating, there might be even more Rb-Sr dates that were excluded. This is not even counting dates obtainable by other methods on other minerals. Also, since only about 30 locations have dates by more than one method in the table in Harland et al, the great majority of locations have only dates by one method in the table. It is only reasonable to assume that geologists tried to date many of these formations by several methods, but most of these dates did not make it into the table in Harland et al.
However, the number required might be less than 19, since there are typically correlations among the dates obtained by a given method at a given location or in a given geological layer. Whatever processes are determining these dates, the same processes are likely to work in more than one sample at a given location, and in more than one location in a given geological layer. Once a date is obtained that seems reasonable, geologists would naturally attempt to measure more dates using the same method in the same geological layer and also in the same location, in order to get a more accurate estimate for the geological time scale. This would increase the chances of finding more dates that agree with expectations in a smaller number of attempts, even assuming the non-K-Ar dates are uncorrelated with the K-Ar dates.
The dates that were measured but excluded from the table could either have been published and considered anomalous and excluded from the table, published and explained as inherited or the result of a later heating event, or not even published. An example of the latter possibility is given by the following quotation from Woodmorappe (1999, p. 39):
... there was a wide range of glauconite dates from the same bed, and Odin and Curry (1985) arbitrarily picked certain ones and ignored others that tended to give older ages. In other cases, they selectively ignored dates that did not agree with their preconceptions, even though these came from areas with well-established stratigraphy (Prothero and Schwab 1996, p. 455).On another matter, I had stated that if 100 dates had been measured at a given location, there would be 5000 possible pairs of dates that could be concordant at a variety of ages. Dr. Henke says, "Obvious geologists don't have the money to obtain 5,000 pairs of dates for 10 samples, so they can pick the ones that they like." Of course, they do not have to measure all 5,000 pairs separately, but only 100 dates.
A number of quotations from Woodmorappe (1999, pp. 29-30) and elsewhere show that different methods are typically discordant. This also casts doubt on whether the geological time scale is real, rather than a pattern imposed on the data by geologists. Phrases are used such as "rarely do all the calculated ages agree," "Natural zircon typically displays an inconsistency (discordance) of age values obtained," "The isotopic systematics of zircon populations from most SCT granitoids can be described with a single word: discordant," "When determined by several methods ..., radiometric ages for coexisting minerals in metamorphic or igneous rock generally differ because of different closure temperatures," and "Rb-Sr whole rock dates tend to be younger than U-Pb dates from the same rock." Many similar quotations could be given, showing that dates often do not agree with the expected value for their geological layer. There are also a number of quotes in the geological literature about anomalous dates being discarded and not published (one of quite a few referenced in Woodmorappe (1999) is given above).
Zircons are often used for U-Pb and related dating methods.
Woodmorappe (1999, pp. 82-83) gives a number of quotes showing that
inherited (xenocrystic) zircons cannot always be optically
distinguished from others. He suggests that ion-microprobe analysis
may do a better job, but expresses some reservations about it.
Magnetic separation may be used to select undamaged zircon crystals, to increase concordance. I don't have any objection to this, if it is used consistently and if all published dates are corrected in the same way.
As I noted before, concordia-discordia methods on zircons lead to few data points of significance on the Phanerozoic. It would be interesting to know which of the zircon dates in the table in Harland et al used magnetic separation or concordia-discordia methods, and for which of them ion-microprobe analysis was used to diagnose inherited zircons.
Xenocrysts are older rock that is carried along in lava or magma
without melting. This can give erroneous isotopic dates. The
question arises whether xenocrysts can be identified by criteria other
than a date that is too old, and whether such a date is arbitrarily
blamed on a xenocryst without other evidence. We give one among quite
a few quotes from Woodmorappe (1999, p. 48) showing that
identification of xenocrysts can be very difficult:
In volcanically active regions where tens of [or?] hundreds of eruptions can occur over geologically short time intervals, entrainment of older volcanic material in younger flows can be problematic. Identification of those contaminant (xenocrystic) populations can be extremely difficult (Karner and Renne 1998, p. 740).Another evidence of this difficulty is a controversy about the central Brooks range in Alaska, cited in Woodmorappe (1999, p. 39). One group of geologists argued for a Devonian age and another argued for a Cretaceous age. The first group said that the Cretaceous results were "rejuvenated" while the second group said that the Devonian dates were from zircon xenocrysts.
Slusher (1981) asserted that a sliding branching ratio was used to
make K-Ar dates agree with others, based on a statement from Cook
(1966). Based on Harland et al (1990), the correct value for the
branching ratio has been known and used since the mid-fifties.
Dr. Henke has accused me of not correcting my articles on this point.
My Radiometric Dating Game article has several links to an article
containing a correction on this point, and the version at the True
Origins site has a correction as well. It is also interesting that
Cook repeated this information in 1966, when it should have been
outdated already by about ten years. Slusher's first edition was in
1973, not long after Cook's work, and Slusher may not have made many
changes to the second edition.
I am also not certain when the (nearly) correct value for the branching ratio began to be used universally. Just because the correct value was agreed upon in the mid-fifties does not mean that all geologists immediately began using it. Some dates using the wrong value may still be cited from the literature of that time without correction. The whole incident also has historical value.
Dr. Henke questions what is the probability of an intense beam of
radiation from a nearby supernova hitting the solar system and causing
an increase in decay rates. It may be zero, for all I know; we need
to learn a lot about supernovae and other astronomical objects before
deciding. But if there is a nearby pulsar with the proper
orientation, the probability could be much higher. The sun could also
have unleashed a storm of gamma rays for some unknown reason.
Dr. Henke questions why this radiation would not have sterilized the earth. Of course, it could have caused a massive extinction. Organisms in the sea would have been protected from most of the radiation. It would require more study to find out how intense a beam of gamma radiation (or other radiation) is needed to influence decay rates, and what its biological effects would be.
The Oklo natural reactor might be evidence that the rate of fission of uranium has not changed, as Dr. Henke suggests, since fission is what leads to uranium chain reactions. However, this would not necessarily imply that the rate of alpha and beta decay, on which most dating methods are based, has not changed. Furthermore, radiation that influenced decay rates may not have penetrated deeply enough to influence the Oklo reactor.
Cook, M. A., 1966, "Prehistory and Earth Models," Max Parrish,
Cumbest, R.J., E. L. Johnson, and T.C. Onstott. 1994. Argon composition of metamorphic fluids: Implications for 40Ar/39Ar geochronology. Geological Society of America Bulletin 106:942-951.
Dickin, Radiogenic Isotope Geology, 1995.
Faure, Principles of Isotope Geology, 1977.
Harland, W.B.; R.L. Armstrong; A.V. Cox; L.E. Craig; A.G. Smith; and D.G. Smith, 1990, "A Geologic Time Scale 1989," Cambridge University Press, Cambridge.
Hyodo, H. and D. York. 1993. The discovery and significance of a fossilized radiogenic argon wave (argonami) in the earth's crust. Geophysical Research Letters 20(1):61-64.
Ivanenko, V. V., and M. I. Karpenko. 1988. 40Ar-39Ar data on excess argon-40 in nepheline from the Kovdor Massif, Kola Peninsula. Geochemistry International 25(1):77-82.
Karner, E. B., and P. R. Renne. 1998. 40Ar/39Ar geochronology of Roman volcanic province tephra in the Tiber River valley. Geological Society of America Bulletin 110(6):740-747.
McDougall, I. and T. M. Harrison. 1988. Geochronology and Thermochronology by the 40Ar-39Ar Method. New York: Oxford University Press, 212 pp.
Morozova, I. M., et al. 1996. Inheritance of radiogenic argon by newly formed minerals during glauconite transformation. Transactions (Doklady) of the Russian Academy of Sciences: Earth Science Sections 344(7):52-57.
Morozova, I. M., et al. 1997. Radiogenic argon as an indicator of the inheritance of material during glauconite hydrothermal transformations. Geochemistry International 35(8):716-723.
Nakata, J. K. 1991. K-Ar and fission-track ages (dates) of volcanic, intrusive, altered, and metamorphic rocks in the Mohave Mountains area, west-central Arizona. Isochron/West 57:21-27.
Phillips, D. et al. 1998. A petrographic and 40Ar/39Ar geochronological study of the Voorspoed kimberlite, South Africa. South African Journal of Geology 101(4):299-306.
Poths, J., H. Healey, and A. W. Laughlin. 1993. Ubiquitous excess argon in very young basalts. Geological Society of America Abstracts with Programs 25(6):462.
Prothero, D. R. and F. Schwab. 1996. Sedimentary Geology. New York: W. H. Freeman and Co., 575 pp.
Rublev, A. G. 1985. The possibility of correcting for excess argon in K-Ar dating. Geochemistry International 22(4):73-79.
Slusher, H.S., 1981. Critique of Radiometric Dating, Institute for Creation Research, Technical monograph 2 (2nd ed.), 46 pp.
Smith, P. E., et al. 1994b. A laser 40Ar-39Ar study of minerals across the Grenville Front: investigation of reproducibe excess Ar patterns. Canadian Journal of Earth Sciences 31:808-817.
Tegner, C., et al 1998. 40Ar-39Ar geochronology of Tertiary mafic intrusions along the East Greenland rifted margin. Earth and Planetary Science Letters 156:75-88.
Wallace, A. R. 1995. Isotopic geochronology of the Leadville 1 x 2 quadrangle, west-central Colorado -- Summary and Discussion. U. S. Geological Survey Bulletin 2104, 51 pp.
Woodmorappe, J., The Mythology of Modern Dating Methods, Institute for Creation Research, 1999.
Back to home page.