Quick radiometric dating question- misused techniques

I caused the jump. I know that Coyote can give precise details on C14 dating but not K-Ar dating. There is a strong parallelism between them as far as the wording of reporting. We'll have to wait until someone can give other method based reports.

This is all interesting, and the creationists missing out a greater-than sign is pretty funny. But I wonder if someone could answer my question:

quote:

---

is it simply a matter of observing that you have received the minimum date for your chosen method and therefore you need to try a method more suitable for a younger sample?

---

Getting into a creationist's brain again (scary place), I'd be thinking (knowing next to nothing about real science) that a scientist dating a sample of unknown provenance (which is uncommon, I realise) and using K-Ar would "go with" the date they were given (presumably 100,000 years). Though if the result came back saying <100,000 years, that's clear enough -- is that what happens?Or, thinking of what Austin did, would you simply get wildly inaccurate results from this method because there's too much argon in young rock? In which case, if you got a sample from a lava field and didn't know it was very young (again, unlikely I realise), how would you know that a figure of millions of years was actually wrong? Is there a way of first being able to measure excess argon?Edited by LindaLou, : No reason given.

You'd have to see the actual lab report to see what was claimed in this case. Was the measured amount of argon consistent with measurement background? If so, the result should have said "less than" a background date. If not, the result should have indicated a date range. The way to measure excess argon is by the 39Ar-40Ar method. (Also called the Ar-Ar method.)

That's the one pirates use, right?*ducks and runs*Edited by Perdition, : No reason given.

Only for a vast set of old rocks

FYI, here's a simplified description of how the 39Ar-40Ar method works:Basic K-Ar dating relies on the decay of 40K to 40Ar in formerly molten rock. Assuming that all of the original argon diffused out while the rock was molten, the present amount of argon can be measured (by heating the sample and allowing the Ar to diffuse out) and the age since the rock was molten can be calculated.If the rock cooled too fast or cooled under pressure (e.g. underwater lava flows), argon can be trapped in the rock. It is also possible for argon to diffuse into the rock over time. In either case, K-Ar dating will give an incorrect result due to the excess argon.But there is a physical difference in the location of this excess argon and the argon resulting from radioactive decay. As the mineral grains crystalize, any initial argon will be forced out of the grain bulk and to the grain boundaries. Likewise, diffusion rates are much higher along grain boundaries than through the grain bulk, so any later diffusion will primarily add argon at grain boundaries. In contrast, radioactive decay will generate argon in the grain bulk.So we need some way to distinguish where the argon resides (grain boundaries or bulk), and this is what the 39Ar-40Ar method provides.The 39Ar-40Ar method first neutron-activates 39K to 39Ar in a nuclear reactor. This creates 39Ar where the 39K was, i.e. in the mineral grain bulk, not on the grain boundaries. Now the sample is heated in a mass spectrometer and both the 39Ar and 40Ar are measured. The initial 40Ar/39Ar ratio consists of argon diffusing from grain boundaries so does not reflect the true age. As time progresses, argon will start to diffuse from the bulk and the ratio will stabilize at a value reflecting the true age.The parameters that are needed for the age calculation are:
1) a good characterization of the neutron dose in the reactor (often done with concurrent calibration samples of known age)
2) knowledge of the original or current 39K/40K ratio in the rock

In the particular cases of Austin and his Mt. St. Helens dacite, and Snelling and his lava from Mt. Ngauruhoe, they were absolutely sure of one thing; the samples they chose were a mixture of old and recent material, and therefore the K-Ar method would return a date that is older than the recent material is, and most likely would return a date much older than the minimum that the K-Ar method could measure. See Young-Earth Creationist 'Dating' of a Mt. St. Helens Dacite: The Failure of Austin and Swenson to Recognize Obviously Ancient Minerals. And, from Snelling himself in ANDESITE FLOWS AT MT NGAURUHOE, NEW ZEALAND, AND THE IMPLICATIONS FOR POTASSIUM-ARGON "DATING":

quote:

---

A second representative set (50-100 g from each sample) was sent progressively to Geochron Laboratories in Cambridge (Boston), Massachusetts, for whole-rock potassium-argon (K-Ar) dating - first a split from one sample from each flow, then a split from the second sample from each flow after the first set of results was received, and finally, the split from the third sample from the June 30, 1954 flow. ...

​

Because the sample pieces were submitted as whole rocks, the K-Ar laboratory undertook the crushing and pulverising preparatory work. ...

​

Steiner [90] stressed that xenoliths are a common constituent of the 1954 Ngauruhoe lava, but also noted that Battey [7] reported the 1949 Ngauruhoe lava was rich in xenoliths. All samples in this study contained xenoliths, including those from the 1975 avalanche material. ...

---

{Emphasis addded}Whole-rock = crush the entire thing and test the result without separating anything.Xenolith (literally "foreign rock") = a piece of rock that is older than the lava flow that contains it. They may be easy or difficult to separate, but in this case they weren't separated.Smoking gun. Well, if you have an agenda like Snelling or Austin does, you pick a method that can be fooled, even though it's been almost completely supplanted by Ar-Ar that's much more difficult to fool, and even though other methods are significantly more widely used. Then you pick samples that are guaranteed to fool the method.If you are really interested in finding the age of a sample with no prior information, you do it with several methods and you do them on individual minerals separated from the rocks. Then if the methods agree you almost certainly have a good date.In the real world, as you know, you essentially always do have prior information.The K-Ar method (and almost all radiometric methods) boils down to measuring a quantity of stuff. (The best methods actually measure the ratio between the quantities of two different stuffs, and then use other information about the quantity of one to calculate the quantity of the other.) When a rock is old, there's a lot of stuff to measure. When the rock is young, there's not much stuff to measure. A K-Ar lab will start by cleaning as much previous sample material and argon from the equipment as is feasible. (If you tell them you think your sample is relatively young, they will take extra care in the cleaning process. And charge more.) Then they will measure your sample and the readouts will tell them there's Y amount of argon.Every few runs they will run the equipment with nothing in the sample holder. The readouts will say there's a non-zero amount of argon, call it X, which they can calculate as being equivalent to an age of Z years.If the amount they measure in your sample, X, is pretty much equal to Y, then they will report the age as Z years or less. If X is greater than Y they will report an age older than Z with appropriate error bars. If X is less than Y they will try to figure out whats going on.

Expanding a little on Dr. Bertsche's explanation ...I think of the Ar-Ar method as two methods. One, irradiating the sample, allows measuring the ratio of ³⁹Ar to ⁴⁰Ar instead of measuring the ratio of ⁴⁰Ar to ⁴⁰K. Measuring the ratio of two isotopes of the same element is easier and more accurate than measuring the ratio of two different elements. You could just vaporize the sample and calculate an age; but the other method, step-heating, allows compensating for various interfering phenomena as Dr. Bertsche explained.It's common to present the results of Ar-Ar analysis as step-heating diagrams. The horizontal axis is the heating step, from 0 to 100%, and the vertical axis is the age measured at that step. From Radiogenic Isotope Geology, the on-line version of a standard reference work (unfortunately the equations are unreadable), here's an great step-heating diagram for two Texas tektites in which there's no interfering phenomena and K-Ar would give the same result as Ar-Ar: If there's excess argon, you get more argon out early on, initally showing an erroneously old age but flatttening out into a "plateau" after the excess argon is exhausted (the sstep-heating diagram is on the bottom): If the material has been heated, leading to a loss of argon from places where it's loosely bound, you get an erroneously low age initially: if the sample has been heated a lot, you may not get a plateau, but you can tgry to get an estimate by modeling the argon loss:

kbertsche and Jon, I think you've answered all of my questions and then some. I've learned a lot more about science from debating creationists than I did in school. Thanks very much

This is what I was dreading: a creationist who has been delving into scientific papers on radiometric decay rates, who doesn't understand them himself but can cherry-pick quotes from them that in his opinion suggest that scientists are unable to accurately determine the half-lives of radioactive isotopes. I would be able to say things about the correlation of dates using different methods, and that it looks to me like scientists are simply using more accurate techniques to refine the decay rates, but I'd like to be able to directly address some of the claims. Could anyone here who understands this stuff help me out?The first of the two papers is this PDF:
Strontium-92 half-life measurement through accurate -ray spectrometryHere's what the creationist quoted:

quote:

---

... For these reasons, the half-life of 92Sr has been measured to solve a recently observed inconsistency with the quoted value in the nuclear data libraries: T1/2 = 2.71 0.01 h. In this work, a new value is proposed: T1/2 = 2.594 0.005 h. A better accuracy is achieved compared to previous evaluations. It also shows a good agreement with the most recent studies: T1/2 = 2.627 0.009 h.

---

quote:

---

A weighted average from the three values was determined
for each fuel pin. Nevertheless, as the discrepancies were
sometimes beyond their reported uncertainties, the associated
uncertainty is calculated from the standard deviation between
the three values, with a multiplication factor tp() = 1.32 [15]
that arose from the low degree of freedom of the sample.

---

quote:

---

...For these reasons, the half-life of 92Sr has been measured to solve a recently observed inconsistency with the quoted value in the nuclear data libraries: T1/2 = 2.71 0.01 h. In this work, a new value is proposed: T1/2 = 2.594 0.005 h. A better accuracy is achieved compared to previous evaluations. It also shows a good agreement with the most recent studies: T1/2 = 2.627 0.009 h.

---

He thinks this shows "unacceptable uncertainties." I would tell him myself that the recent work has simply refined the figures but that's all I would be able to do. I really hate it when they cite stuff like this when it's clear they haven't a clue what it means themselves (and I'm struggling too!).The second citation is simply an abstract:
Half-life evaluations for 3H, 90Sr, and 90YAnd naturally these quotes have been cherry-picked:

quote:

---

A recent paper has reviewed methods for the evaluation of discrepant sets of data and demonstrated the results of applying these methods to the published half-life data of 90Sr and 137Cs [MacMahon, T.D., Pearce, A., Harris, P., 2004. Convergence of techniques for the evaluation of discrepant data. Appl. Radiat. Isot. 60, 275-281]. The half-life data for 3H has been subject to a comprehensive review and critical evaluation by Lucas and Unterweger [2000. Comprehensive review and critical evaluation of the half-life of tritium. J. Res. Natl. Inst. Stand. Technol. 105, 541-549].

---

quote:

---

...resulting in a recommended half-life of 4497(4) days. MacMahon et al. [Convergence of techniques for the evaluation of discrepant data. Appl. Radiat. Isot. 60, 275-281] highlighted problems in the evaluation of the discrepant half-life data of 90Sr, in particular the worrying upward trend in the data, where the weighted mean of all the measurements increases, on average, by 35 days each time a new measurement result is added.

---

This is CTD, who migrated from this forum to the other one where I'm talking, which is run by creationists and where he gets to be a moderator, as opposed to, say, talking to real scientists here and being shown what ignorant nonsense he's spouting. He seems to see "inconsistencies" and "discrepancies" and my hunch is that they are small but I don't understand this data well enough to say how insignificant they are.added in edit
The author of the second paper, McMahon, is citing a previous paper of his own when mentioning "discrepant half-life data." The abstract is here: Convergence of techniques for the evaluation of discrepant dataIt looks to me like this is saying that when more data is obtained, the more obvious the "true" value becomes on a graph:

quote:

---

The various evaluation techniques should converge towards the ”true’ value as the number of data in a data set increases, and the ”robustness’ of each evaluation technique can then be tested by the rate at which that technique converges.

---

I think I could do an OK job of convincing CTD that he should leave things alone that he doesn't understand but it would be nice to try to address some of the above more specifically.Edited by LindaLou, : No reason given.Edited by LindaLou, : No reason given.Edited by LindaLou, : No reason given.Edited by LindaLou, : No reason given.

That first study on strontium 92 has a spread of 5% between values, and with a half-life of under three hours, that isotope is 1) rare and 2) REALLY HOT to handle. Errors in decay rate measurement for things like carbon-14 will be smaller.And 5% error in half-lives fails to be much comfort for a YEC, anyway. They need five million percent errors.Edited by Coragyps, : fumblefingers

I just noticed you're debating with CTD. He was here for about a year up until about a year and a half ago, and I've also encountered him over at Evolution Fairy Tale. He has a bit of a frustrating style in that he keeps throwing new data and arguments at you that seem to pop into his head and that often don't relate to the topic, and he'll keep throwing general ad hominem against evolutionists into the mix. The discussion can begin to feel a bit fragmented.But he's sincere and really believes what he's saying, and I've found he will concede on points that have been successfully demonstrated and that don't directly compromise his position. Be patient, stick to your points, make your arguments as clear as you can, realize that you're going to be learning a lot more during this process than he is, and that science is a lot more messy and uncertain than we would wish, and don't hold out any false hopes that he's going to be persuaded.--Percy

Well, you might ask him how old the Earth and life are if we are wrong about all decay rates decay rates by a factor of 100 …Decay rates are known to within a few percent or better. One of the several reasons that U-Pb concordia-discordia dating is so often used is that the decay rte of uranium is known to much better than a percent. Bombs and reactors tend to attract lots of research money.The classic study on uranium is Jaffey A. H., Flynn K. F., Glendenin L. E., Bentley W. C., and Essling A. M. (1971). Precision measurement of half-lives and specific activities of ²³⁵U and ²³⁸U. Phys. Rev. C4, 1889—1906. Alas, this is not available online, but from Begemann, F., Ludwig, K. R., Lugmair, G. W., Min, K., Nyquist, L. E., Patchett, P. J., Renne, P. R. Shih, C.- Y., Villa, I. M. and Walker, R. J. (2001). Call for an improved set of decay constants for geochronological use. Geochim. Cosmochim. Acta 65, 111--21:

quote:

---

The decay of 238U and 235U to 206Pb and 207Pb, respectively, forms the basis for one of the oldest methods of geochronology. While the earliest studies focused on uraninite (an uncommon mineral in igneous rocks), there has been intensive and continuous effort over the past three decades in U-Pb dating of more-commonly occurring trace minerals. Zircon in particular has been the focus of thousands of geochronological studies, because of its ubiquity in felsic igneous rocks and its extreme resistance to isotopic resetting.

​

No decay constant of any radionuclide used for geochronology has been (or, arguably, can be) more-precisely measured than those of 238U and 235Ua consequence of the mode of decay (alpha), favorably short half-lives, and the availability of large quantities of isotopically pure parent nuclides. The most recent measurements by Jaffey et al. (1971) (Figs. 1, 2) quote precisions (recalculated to 95%-confidence limits) of 0.11% for 238U and 0.14% for 235U, with the somewhat cryptic statement that systematic errors, if present, will no more than double the quoted errors.

---

There's been no studies since then that upset these findings.The bottom line is that, while we are improving our knowledge of decay rates all the time and some of our values may be off by as much as a few percent, there's no way that we are far enough off to make a YEC scenario conceivable, by thousands of orders of magnitude.And, indeed, the consilience of dates obtained with multiple methods using different isotopes with different decay modes is powerful evidence that our dates are correct to about the specified precision.Some useful links on consilience are Radiometeric Dating Does Work! (note that Table 2 includes the Hell Creek formation, from which the dinosaur bones with "blood cells" and "soft tissue" were extracted), Consistent Radiometric dates, Are Radioactive Dating Methods Consistent With Each Other? (one of my favorites), Are Radioactive Dates Consistent With The Deeper-Is-Older Rule?, Radiometric Dating, Radiometric Ages of Some Early Archean and Related Rocks of the North Atlantic Craton, and Radiometric Ages of Some Mare Basalts Dated by Two or More Methods.

You're right about CTD, Percy. I've seen some of your posts at the Evolution Fairytale forum when I've been browsing and admire your restraint (though it's doubtlessly necessary on a forum like that which is run and moderated by creationists). It's the continual ad hom that can be a wind-up. I've been debating with him for about 2 years now I reckon, and yes, there never will be any convincing him. Thanks for the advice, it's good.

Thanks JonF and Coragyps, your info here has been very useful to me in constructing a reply. Without a grasp of some of the info you've given here, it's hard to know where address an attack against radiometric dating and decay rates. I am of course not surprised to hear that the decay rates of isotopes used to date rocks are known within a small margin of error. One of CTD's favourite arguments (which he hasn't tried on me yet) is that we're "guessing" because we can't be around for billions of years to make sure that, e.g., uranium continues along its predicted decay curve to lead and doesn't . . . I don't know, jump sideways and dance at the 2.36 million year mark or something. I think a good response to this would be to show that we can see isotopes decaying in ancient starlight -- I'm assuming we can, though I haven't researched this yet.