How Serious are Errors in Ar40-Ar39 Dates and How Good are Their Monitoring Standards?
Dr. Kevin R. Henke
The following material may be freely copied and distributed as long as the author is properly acknowledged
and the material is not altered, edited or sold.
Most quantitative analytical methods, including any water analyses for organic or metal contaminants (Skoog and West, 1976, chapters 25 and 26), require standards to provide accurate results. With water analyses, a calibration curve is established by analyzing several known standards. The concentrations of the unknowns are then determined by where they plot on the calibration curve. Ar-Ar dating also relies on standards to provide quantitative results. Obviously, good results on unknown samples depend upon having standards with well-defined concentrations or, in the case of radiometric dating, well defined ages.
Over the years, numerous interlaboratory studies have been conducted to test and establish standards or monitors for radiometric dating (as examples, Lanphere and Dalrymple, 1965; Samson and Alexander, 1987; Sudo et al., 1998; Renne et al., 1998; Dalrymple and Lanphere, 1969; Jaeger et al., 1963; Flisch, 1982; Turner et al., 1971; Ingamells and Engels, 1976; Odin et al., 1982). McDougall and Harrison (1999, p. 5) explains why monitors are used with Ar-Ar dating:
'Merrihue (1965) explained that 39Ar generated in a nuclear reactor from 39K in a sample could be measured mass spectrometrically, instead of by a counting technique, after extraction of the argon from the irradiated sample. This 39Ar derived from 39K is designated 39ArK. In addition, other isotopes of argon could be measured in the mass spectrometer, including 40Ar and 36Ar, the latter facilitating correction for nonradiogenic 40Ar present in the gas. From these isotope abundance measurements, the 40Ar*/39ArK ratio could be derived and, as this ratio is proportional to 40Ar*/40K (the 40K/39K being essentially constant in nature), it is also proportional to the K/Ar age. The age was calculated by comparison with the 40Ar*/39ArK ratio found for a standard sample of accurately known K/Ar age, irradiated at the same time as the sample to be dated. Use of a standard sample as a neutron fluence monitor meant that it was unnecessary to know the actual neutron dose received by the samples. It was also unnecessary to measure absolute abundances of either argon or potassium in the sample, the age of which was to be determined.'
The ages of the standards have been independently measured with different radiometric methods, oxygen isotope records (Karner and Renne, 1998, p. 740) and/or astronomical methods (Renne et al., 1998, p. 121-122; Hilgen et al., 1997, p. 2043). In response to this definitive evidence, Woodmorappe (1999, p. 74) unjustly accuses Hilgen et al. (1997) of 'tweaking' the data to make the astronomical and radiometric results agree. The abstract of Hilgen et al. (1997, p. 2043) clear indicates that the discrepancies between the radiometric and astronomical data are trivial and do not support Woodmorappe's irrational crusade:
'Preferred 40Ar/39Ar ages - calculated against TCR [Taylor Creek Rhyolite] sanidine with an age of 27.92 Ma [Ma = million years], intercalibrated to an age of 28.09 +/- 0.10 Ma (1 sigma) for FCT-3 [Fish Canyon Tuff] biotite and an age of 24.99 +/- 0.07 Ma for DRA [Drachenfels, Germany Trachyte] sanidine -are SLIGHTLY but consistently younger than astronomical ages obtained INDEPENDENTLY from the SAME ash beds. The best fit to the astronomical ages is obtained when the age of the TCR sanidine is increased SLIGHTLY to 27.98 +/- 0.19 Ma, the age of the FCT-3 biotite to 28.15 +/- 0.19 Ma and the age of DRA sanidine to 25.05 +/- 0.17 Ma. The ages for the standards arrive SLIGHTLY younger if the 40Ar/39Ar age- of 6.936 +/- 0.006 Ma (1 sigma)- for a single pure sanidine separate of the lower ash - dated astronomically at 6.941 Ma- is considered most reliable.' [my emphasis]
As a typical example, Renne et al. (1998, p. 121-122) testifies to the thorough and independent verification of the age of the sanidine from the Fish Canyon (FC) tuff monitor:
'Sanidine from the Fish Canyon tuff (a large volume ash-flow tuff) was proposed as a 40Ar/39Ar standard by Cebula et al. (1986), who reported its age as 27.79 Ma relative to an age of 518.9 for MMhb-1...[reference omitted]. Subsequent revision of the age of MMhb-1 to 520.4 Ma led to the age of 27.84 Ma for FCs widely used in the literature. Though from an ash-flow tuff, nearly a decade of use (and analysis of more than 3000 individual grains) as a single-crystal standard at the Berkeley Geochronology Center (BGC) has revealed NO evidence of xenocrystic contamination of the sanidine. An age of 28.03 +/- 0.09 Ma has been determined for FCs based on intercalibration with the astronomical time scale...'[reference omitted]. [my emphasis]
McDougall and Harrison (1999, p. 46) further comments on the K-Ar results of the Fish Canyon Tuff:
'Results of K/Ar age measurements are available on phenocrysts of sanidine, biotite, plagioclase and hornblende separated from several samples of the Fish Canyon Tuff. Steven et al. (1967) reported essentially concordant K/Ar ages on all four minerals from a sample collected from the summit of Agua Ramon Mountain, about 7 km north-northeast of South Fork, Colorado. A mean age of 27.9 +/- 0.7 Ma was recalculated from these data...[references omitted] using the currently accepted decay constants. Hurford and Hammerschmidt (1985) published K/Ar ages on the same four minerals from a different sample of Fish Canyon Tuff; the ages are nearly concordant and a mean age of 27.4 +/- 0.4 Ma was given. Lanphere and Baadsgaard (1997) reported K/Ar ages of 27.5 +/- 0.2 Ma on biotite and hornblende from yet another Fish Canyon Tuff. Spell and McDougall (unpublished) measured K/Ar ages on sanidine, biotite and hornblende from two samples of the Fish Canyon Tuff collected at different times from the same locality described by Naeser et al. (1981) adjacent to US Highway 160, 9.0 km south-west of South Fork. These measured ages are virtually concordant and yield a mean of 28.4 +/- 0.2 Ma. This summary of K/Ar results from the Fish Canyon Tuff again serves to highlight the fact that the interlaboratory precision of measurement is not matched by a similar level of accuracy, as the mean age values from the several laboratories differ by AS MUCH AS FOUR PERCENT.' [my emphasis]
Despite interlaboratory analyses involving different radiometric techniques and confirmation with non-radiometric methods, Woodmorappe (1999, p. 73-74, etc.) irrationally refuses to trust the dates of monitoring standards, such as the Fish Canyon minerals. Woodmorappe (1999, p. 73) claims that the choice of monitors (standards) has a significant effect on the Ar-Ar results. Of course, he (1999, p. 73) wants us to believe that depending on what monitors are used, wildly different and inaccurate Ar-Ar results are obtained. Woodmorappe (1999, p. 73) quotes the following section from Renne et al. (1998, p. 118) to support his accusations that the precision of Ar-Ar methods are exaggerated and that these exaggerations create 'serious problems' for the dating method:
'The absolute ages of 40Ar/39Ar standards remains an unresolved issue and as a result, detailed comparison of data from the 40Ar/39Ar system with other geochronometers is not well founded unless appropriate (THOUGH HERETOFORE UNPRACTICED) error propagation is employed.' [Woodmorappe's emphasis]
Woodmorappe (1999, p. 73-74) also claims that the development of error-propagation calculations has exposed 'unrealistic claims' in the precision of Ar-Ar methods. So, how serious are the exaggerations in precision and how helpful are they to the young-Earth creationist cause? In reality, calculations of 'full external errors' on several widely used standards by Renne et al. (1998, p. 117), which includes uncertainties in the decay constants, are approximately +/-1% and less.
Woodmorappe (1999, p. 74) further exaggerates the situation by selectively quoting certain sentences and phrases from Karner and Renne (1998, p. 743):
'Uncertainties in decay constants and the ages of standards have traditionally been ignored in estimating the errors associated with 40Ar/39Ar dates... Systematic errors become particularly important when they are compounded by several standard intercalibrations... It is noteworthy that errors introduced by uncertainties in the values of decay constants and the age of the primary standard magnified by several intercalibration steps are significantly larger than those due to analytical uncertainties alone. The larger magnitudes of these errors...'
Karner and Renne (1998) performed "full external errors" on some Italian volcanic samples. Woodmorappe's quotations are taken from several paragraphs in Karner and Renne (1998, p. 743). Below are the quotations in better context:
'Uncertainties in decay constants and the ages of standards have traditionally been ignored in estimating the errors associated with 40Ar/39Ar dates. EVEN THOUGH THIS PRACTICE CAN BE JUSTIFIED TO SOME EXTENT in comparing 40Ar/39Ar dates based on the same standard, it is appropriate to consider systematic errors when comparing 40Ar/39Ar dates with dates obtained by independent means, as in the present case. [new paragraph] Systematic errors become particularly important when they are compounded by several standard intercalibrations.' [my emphasis]
Notice that the situation is not as serious as Woodmorappe (1999, p. 74) wants us to believe and that Karner and Renne (1998) provide a better solution for defining the errors associated with Ar-Ar monitors. The last part of Woodmorappe's (1999, p. 74) quotation from Karner and Renne (1998, p. 743) is taken from the following paragraph:
'The external errors calculated from equation 4 are included in Table 1. It is noteworthy that errors introduced by uncertainties in the values of decay constants and the age of the primary standard magnified by several intercalibration steps are significantly larger than those due to analytical uncertainties alone. The larger magnitudes of these errors illustrate the need to apply equation 4 routinely, as statistical inference based on comparisons between 40Ar/39Ar dates and those determined by other methods are inaccurate otherwise.'
Notice that Woodmorappe (1999, p. 74) fails to mention that equation 4 in Karner and Renne (1998, p. 743) may be used to quantify the errors introduced by uncertainties in the values of the decay constants and other factors. Nevertheless, how serious are the external errors in Table 1 in Karner and Renne (1998, p. 741)? As a typical example, the 2-sigma analytical precision for sample number R93-15H2 is 251 +/- 8 Ma. If full external errors are calculated for the sample, the error only increases to +/- 11 Ma (about +/- 4% total). In another example, sample R95-16A provided a date of 749 Ma (+/- 14 Ma 2-sigma analytical precision and +/- 24 Ma 2-sigma full external errors). As usual, these errors are too trivial to serve Woodmorappe's agenda!
Rather than properly recognize that full external errors on STANDARDS are usually less than 1% and rarely above 2% (for example, 523.1 +/- 4.6 for standard MMhb-1 in Renne et al., 1998, p. 117), Woodmorappe (1999, p. 74) quotes Obradovich (1993, p. 389) and claims that just switching standards may change the results on samples by 'at least 2%.' Woodmorappe (1999, p. 74) is probably referring to an incorrect date of 91.5 +/- 1.0 Ma from a previous study by Obradovich, which was published in 1988. Obradovich (1993, p. 389) reanalyzed the sample with a new standard and obtained a slightly higher date of 93.3 Ma. Depending on the quality of the samples, and not just the standards that are used, radiometric dates for duplicate samples typically vary by less than 1% to 4% (Renne et al., 1998; Karner and Renne, 1998; Obradovich, 1993). However, how do inconsistencies of 4% or less serve the young-Earth creationist crusade to undermine and destroy the reliability of Ar-Ar dating? How do calculations of full external errors by Renne et al. (1998) and other researchers help young-Earth creationism?
A review of McWilliams (1994), another article listed by Woodmorappe (1999, p. 73), also demonstrates that the errors associated with Ar-Ar dating are typically trivial. Bentonite beds from three localities in North America were sampled. The samples were collected 5 to 80 centimeters above the famous Cretaceous-Tertiary Iridium anomaly layer, so they should be slightly younger than Haitian tektites 64.91 +/- 0.12 Ma (2 sigma) that probably formed from the impact. Not surprisingly, the mean Ar-Ar age of the bentonite beds was a consistent 64.71 +/- 0.09 Ma (2 sigma).
Woodmorappe (1999, p. 74) also cites Kamo et al. (1996, p. 3506) and Vandamme et al. (1991, p. 163) and claims that the use of different standards may result in 'divergent' chronological conclusions. Kamo et al. (1996, p. 3506) is another summary of the dispute between Renne et al. and Dalrymple over the ages of some Siberian igneous rocks that are close in age to the Permian-Triassic boundary (see Kerr, 1995). Woodmorappe (1999, p. 42) misrepresents this dispute when he quotes Kerr (1995, p. 27-28):
'Over time, Dalrymple concludes, some of the argon-40 had leaked out of the trap's rocks, making them look 1 or 2 million years younger than they are. Renne, however, says that he is "very confident about the new data"... they did extensive argon-argon analyses that contradict Dalrymple's conclusions about the alterations of the trap rock. It's not that the trap rocks lost argon, Renne believes; instead, the intrusion carries extra argon-40 picked up before the minerals formed, giving a falsely older age.'
So, how significant is the dispute between Renne et al. and Dalrymple as described by Kerr (1995)? Although Woodmorappe (1999, p. 42) is quick to tell his readers that the discrepancies involve 1-2 million years, which seem large, he does not tell us the ages of the samples. Dalrymple, Renne and their colleagues are attempting to determine if massive 250 million year old volcanic eruptions in Siberia were synchronous with a severe extinction at the Permian-Triassic boundary. In other words, these scientists are arguing over errors of 1-2 million years for events that occurred 250 million years ago. Once more, Woodmorappe (1999, p. 42) is distorting arguments over errors of less than 1% just to make Dalrymple, other geochronologists and radiometric dating results look as bad as possible.
Vandamme et al. (1991, p. 163) discusses some disputes over the use of a muscovite standard by Kaneoka (1980) and a mistake in Kaneoka's error calculations. The problems were resolved in the late 1980s and hardly seem relevant to Woodmorappe's current arguments.
Again, for scientists that want errors well-below +/-1%, the precision and accuracy of Ar-Ar dates for different standards or samples may not always comply with these strict requirements. However, from the perspective of young-Earth creationism, these errors are far too trivial to serve their needs. By hiding the details in his references from his readers and exaggerating the accuracy and precision problems with Ar-Ar monitors, Woodmorappe is only destroying his credibility and discouraging people from adopting his religion.
Cebula, G.T.; M.J. Kunk; H.H. Mehnert; C.W. Naeser; J.D. Obradovich; and J.F. Sutter, 1986, "The Fish Canyon Tuff, a Potential Standard for the 40Ar-39Ar and Fission-track Dating Methods," Terra Cognita, v. 6, p. 139-140.
Dalrymple, G.B. and M.A. Lanphere, 1969, Potassium-Argon Dating, Freeman, San Francisco.
Flisch, M.,1982, "Potassium-Argon Analysis" in Numerical Dating in Stratigraphy, (G.S. Odin, ed.), Wiley and Sons, Chichester, p. 151-158.
Hilgen, F.J.; W. Krijgsman and J.R. Wijbrans, 1997, "Direct Comparison of Astronomical and 40Ar/39Ar Ages of Ash Beds: Potential Implications for the Ages of Mineral Dating Standards," Geophys. Research Lett., v. 24, n. 16, p. 2043-2046.
Hurford, A.J. and K. Hammerschmidt, 1985, "40Ar/39Ar and K/Ar Dating of the Bishop and Fish Canyon Tuffs: Calibration Ages for Fission-track Dating Standards," Chem. Geol. (Isot. Geosci. Section), v. 58, p. 23-32.
Ingamells, C.O. and J.C. Engels, 1976, Preparation, Analysis and Sampling Constants for a Biotite, Nat. Bur. Stand., Spec. Publ., 422, p. 401-419.
Jaeger E.; E. Niggli and H. Baethge, 1963, "Two Standard Minerals, Biotite and Muscovite, for Rb-Sr and K-Ar Age Determinations, sample Bern4B and Bern4M from a gneiss from Brione, Valle Verzasca Schweiz," Min. Petr. Mitt., v. 43, p. 465-470.
Kamo, S.L.; G.K. Czamanske and T.E. Krough, 1966, "A Minimum U-Pb Age for Siberian Flood-basalt Volcanism," Geochim. et Cosmo. Acta, v. 60, n. 18, p. 3505-3511.
Kaneoka, I., 1980, "40Ar-39Ar Dating on Volcanic Rocks of the Deccan Traps, India," Earth Planet. Sci. Lett., v. 46, p. 233-243.
Kerr, R.A., 1995, "A Volcanic Crisis for Ancient Life?", Science, v. 270, Oct. 6, p. 27-28.
Karner, D.B. and P.R. Renne, 1998, "40Ar/39Ar Geochronology of Roman Volcanic Province Tephra in the Tiber River Valley: Age Calibration of Middle Pleistocene Sea-level Changes," GSA Bull., v. 110, n. 6., p. 740-747.
Lanphere, M.A. and H. Baadsgaard, 1997, "The Fish Canyon Tuff: A Standard for Geochronology," EOS Trans. Am. Geophys. Un., v. 78, p. 5326.
Lanphere, M.A. and G.B. Dalrymple, 1965, "An Interlaboratory Standard Muscovite for Argon and Potassium Analyses," J. Geophys. Res., v. 70, p. 3497-3503.
McDougall, I. and T. M. Harrison, 1999, Geochronology and Thermochronology by the 40Ar/39Ar Method, 2nd ed., Oxford University Press, New York.
McWilliams, M., 1994, "Relative Chronology of Events at and near the Cretaceous-Tertiary Boundary," in Abstracts of the Eighth International Conference on Geochronology, Cosmochronology, and Isotope Geology, M.A. Lanphere, G.B. Dalrymple, and B.D. Turin (eds.) U.S. Geologic Survey Circular 1107, p. 214.
Merrihue, C., 1965, "Trace-element Determinations and Potassium-argon Dating by Mass Spectroscopy of Neutron-irradiated Samples," Trans. Am. Geophys. Un., v. 46, p. 125.
Naeser, C.W.; R.A. Zimmermann and G.T. Cebula, 1981, "Fission-track Dating of Apatite and Zircon: An Interlaboratory Comparison," Nucl. Tracks, v. 5, p. 65-72.
Obradovich, J.D., 1993, "A Cretaceous Time Scale," in Evolution of the Western Interior Basin, W.G.E. Caldwell and E.G. Kauffman (eds.), Geological Association of Canada Spec. Pap. 39, p. 379-396.
Odin, G.S. et al., 1982, "Interlaboratory Standards for Dating Purposes" in Numerical Dating in Stratigraphy, (G.S. Odin, ed.), Wiley and Sons, Chichester, p. 123-149.
Renne, P.R.; C.C. Swisher; A.L. Deino; D.B. Karner; T.L. Owens; and D.L. DePaolo, 1998, "Intercalibration of Standards, Absolute Ages and Uncertainties in 40Ar/39Ar Dating," Chemical Geology, v. 145, p. 117-152.
Samson, S.D. and E. C. Alexander Jr., 1987, "Calibration of the Interlaboratory 40Ar-39Ar Dating Standard, Mmhb-1," Chemical Geology (Isotope Geoscience Section), v. 66, p. 27-34.
Skoog, D.A. and D.M. West, 1976, Fundamentals of Analytical Chemistry, 3rd ed., Holt, Rinehart and Winston, New York.
Spell, T.L. and I. McDougall, unpublished data.
Steven, T.A.; H.H. Mehnert and J.D. Obradovich, 1967, "Age of Volcanic Activity in the San Juan Mountains, Colorado," U.S. Geol. Survey Prof. Paper 575-D, D47-D55.
Sudo, M.; K. Uto; K. Anno; O. Ishizuka and S. Uchiumi, 1998, "SORI93 biotite: A New Mineral Standard for K-Ar Dating," Geochemical Journal, v. 32, p. 49-58.
Turner, G.; J. Huneke; F.A. Podosek; and G.J. Wasserburg, 1971, "40Ar-39Ar Ages and Cosmic Ray Exposure Age of Apollo 14 Samples," Earth Planet. Sci. Lett., v. 12, p. 19-35.
Vandamme, D.; V. Courtillot and J. Besse, 1991, "Paleomagnetism and Age Determinations of the Deccan Traps (India): Results of a Nagpur-Bombay Traverse and Review of Earlier Work," Rev. Geophys., v. 29, n. 2, p. 159-190.
Woodmorappe, J., 1999, The Mythology of Modern Dating Methods, Institute for Creation Research, El Cajon, CA.