Decay methods appear to be the chief nail on which geologists hang their hats in proclaiming the ages of the Earth and geological formations which they do.
David Plaisted notes the following:
I found the following statement in an on-line (non creationist) reference, as follows:
“This is possible in potassium-argon (K-Ar) dating, for example, because most minerals do not take argon into their structures initially. In rubidium-strontium dating, micas exclude strontium when they form, but accept much rubidium. In uranium-lead (U-Pb) dating of zircon, the zircon is found to exclude initial lead almost completely.”
[from the Britannica Online, article “Geochronology: The Interpretation and Dating of the Geologic Record.”] So because of this, one can do Rb-Sr dating on micas because they exclude strontium when the micas form. Thus one would know that any strontium that is present had to come from the parent rubidium, so by computing the ratio and knowing the half life, one can compute the age.
In general, when lava cools, various minerals crystallize out at different temperatures, and these minerals preferentially include and exclude various elements in their crystal structures. So one obtains a series of minerals crystallizing out of the lava. Thus the composition of the lava continues to change, and later minerals can form having significantly different compositions than earlier ones. Lava that cools on the surface of the earth is called extrusive. This type of lava cools quickly, leaving little time for crystals to form, and forms basalt. Lava that cools underground cools much more slowly, and can form large crystals. This type of lava typically forms granite or quartz.
A good general introduction to radiometric dating from an evolutionary perspective can be found at http://asa.calvin.edu/ASA/resources/Wiens.html.
I admit this is a very beautiful theory. This would seem to imply that the problem of radiometric dating has been solved, and that there are no anomalies. So if we take a lava flow and date several minerals for which one knows the daughter element is excluded, we should always get the exact same date, and it should agree with the accepted age of the geological period. Is this true? I doubt it very much. If the radiometric dating problem has been solved in this manner, then why do we need isochrons, which are claimed to be more accurate?
The same question could be asked in general of minerals that are thought to yield good dates. Mica is thought to exclude Sr, so it should yield good Rb-Sr dates. But are dates from mica always accepted, and do they always agree with the age of their geologic period? I suspect not.
Indeed, there are a number of conditions on the reliability of radiometric dating. For example, for K-Ar dating, we have the following requirements:
For this system to work as a clock, the following 4 criteria must be fulfilled:
1. The decay constant and the abundance of K40 must be known accurately.
2. There must have been no incorporation of Ar40 into the mineral at the time of crystallization or a leak of Ar40 from the mineral following crystallization.
3. The system must have remained closed for both K40 and Ar40 since the time of crystallization.
4. The relationship between the data obtained and a specific event must be known.
The requirements for radiometric dating are stated in another way, at the web site http://hubcap.clemson.edu/spurgeon/books/apology/Chapter7.html:
“But what about the radiometric dating methods? The earth is supposed to be nearly 5 billion years old, and some of these methods seem to verify ancient dates for many of earth’s igneous rocks. The answer is that these methods, are far from infallible and are based on three arbitrary assumptions (a constant rate of decay, an isolated system in which no parent or daughter element can be added or lost, and a known amount of the daughter element present initially).”
Here are more quotes about radiometric dating from http://www.parentcompany.com/handy_dandy/hder12.htm:
“All of the parent and daughter atoms can move through the rocks. Heating and deformation of rocks can cause these atoms to migrate, and water percolating through the rocks can transport these substances and redeposit them. These processes correspond to changing the setting of the clock hands. Not infrequently such resetting of the radiometric clocks is assumed in order to explain disagreements between different measurements of rock ages. The assumed resettings are referred to as `metamorphic events’ or `second’ or `third events.’ ”
And again,
“It is also possible that exposure to neutrino, neutron, or cosmic radiation could have greatly changed isotopic ratios or the rates at some time in the past.”
It is known that neutrinos interact with atomic nucleii, so a larger density of neutrinos could have sped up radioactive decay and made matter look old in a hurry. Some more quotes from the same source:
a. In the lead-uranium systems both uranium and lead can migrate easily in some rocks, and lead volatilizes and escapes as a vapor at relatively low temperatures. It has been suggested that free neutrons could transform Pb-206 first to Pb-207 and then to Pb-208, thus tending to reset the clocks and throw thorium-lead and uranium-lead clocks completely off, even to the point of wiping out geological time. Furthermore, there is still disagreement of 15 percent between the two preferred values for the U-238 decay constant.
b. In the potassium/argon system argon is a gas which can escape from or migrate through the rocks. Potassium volatilizes easily, is easily leached by water, and can migrate through the rocks under certain conditions. Furthermore, the value of the decay constant is still disputed, although the scientific community seems to be approaching agreement. Historically, the decay constants used for the various radiometric dating systems have been adjusted to obtain agreement between the results obtained. In the potassium/argon system another adjustable “constant” called the branching ratio is also not accurately known and is adjusted to give acceptable results.
Argon-40, the daughter substance, makes up about one percent of the atmosphere, which is therefore a possible source of contamination. This is corrected for by comparing the ratio argon-40/argon-36 in the rock with that in the atmosphere. However, since it is possible for argon-36 to be formed in the rocks by cosmic radiation, the correction may also be in error. Argon from the environment may be trapped in magma by pressure and rapid cooling to give very high erroneous age results. In view of these and other problems it is hardly surprising that the potassium/argon method can yield highly variable results, even among different minerals in the same rock.
c. In the strontium/rubidium system the strontium-87 daughter atoms are very plentiful in the earth’s crust. Rubidium-87 parent atoms can be leached out of the rock by water or volatilized by heat.
All of these special problems as well as others can produce contradictory and erroneous results for the various radiometric dating systems.
And yet as noted many times already we have a convergence upon a value of 4.5 billion years for the age of the solar system from many different "error-prone" methods, many studies, and many source samples. Such is statistically inexplicable in itself if there's nothing really special about 4.5 billion years ago, especially when you consider that no comparable data bias exists for any other answer.
Again, the convergence of many methods, many samples, many studies. Plaisted pretends not to see the issue. He runs down the catalog of mantras. ("This is an issue!" "This is a problem!" "This is frequently unknown!")
It's all a big head-fake, frauds trolling for idiots. Various sources of error have been identified, yes. They can be allowed for. There's more than one test for most purposes. Most of the pitfalls have been identified decades ago. Plaisted is simply pretending that the pitfalls are known only to creationists and that all the published results are spurious.
Again, the real skinny: Radiometric Dating, A Christian Perspective.