Today @ 06:07 PM post located here
TheFiveSolas:

“ wehappyfew:
Did you notice that the "slight decrease in slope" was achieved by throwing out all the data with a steeper slope (steps 1 though 14)? ”

That's not true at all. The measured data itself, as shown in figure 6(a),...

My point exactly... thank you for demonstrating it so clearly, with a graph, even.

Figure 6(a) DOES NOT SHOW ALL THE DATA!
Steps 1 through 14 are missing.
Deleted.
Not there.
Ignored.
Thrown out.

... shows a slight decrease in slope which they are taking into account in their

"least-squares fit of eq. (2) to the New Mexico (Jemez Granodiorite) zircon data "

So, rather than inventing the numbers, as you've asserted they must have done, they were simply taking an average of the data for the range in question (between 440 to 300 celsius, the section of the graph where there is a "slight decrease in slope").

Again, I never asserted that 34.4 or 29.4kcal/mol are "invented numbers". They are perfectly valid curve-fits to PART of the data - steps 15 through 44 for the 34.4kcal/mol value; and all the points in steps 15-44 that are below 440degC for 29.4kcal/mol (ie, steps 17-19,22-26, and 35-43).

Using steps 3 through 8 (also from 300 to 440degC) would give a very different number - much closer to 46kcal/mol.

Fitting curves to part of the data is perfectly reasonable - if there is a compelling reason to. But we can get to that later. All I would like to accomplish at first is to establish to everyone's satisfaction that some of the original data is missing from those calculations.

wehappyfew:
Throwing out data that doesn't fit the model is called data massaging, in my opinion. ”

If that was what they did I would have to agree. However, as I've shown they were providing a least-squares fit for the range on the graph for temperatures between 440 and 300 degrees.

And since the graph in figure 6(a) does NOT include steps 1 through 14, and since steps 1 through 8 are also below 440degC, will you now agree that this represents data massaging?

Just to verify, you can calculate where steps 1 through 4 would be if they were on the graph.

The raw data from Farley's report on page 21:
step . . . . . . . D/a^2 (sec^-1)
1 . . . . . . . . . . 3.78E-11
2 . . . . . . . . . . 2.10E-11
3 . . . . . . . . . . 1.77E-11
4 . . . . . . . . . . 9.34E-11

Now multiply by the effective radius squared ... 30 microns^2... to get the same units as figure 6(a), we now get

step . . . . . . . D (cm^2/sec)
1 . . . . . . . . . . 3.40E-16
2 . . . . . . . . . . 1.89E-16
3 . . . . . . . . . . 1.59E-16
4 . . . . . . . . . . 8.41E-16

As you can plainly see, the left scale of figure 6(a) only goes down to 10^-15, so points 1 through 4 would be off the bottom of the graph!

“ wehappyfew:
The value of 3.74 kcal/mol ... that value is picked from thin air.
Made up...
Fabricated...
Derived from the predictions of the model. They admit it right in the paper on the bottom of page 12.

To repeat, the value of 3.74 kcal/mol was never measured in any physical sample from the zircons in question.
This made-up value is the one used to calculate the "upper limit" of the age of the Earth, and to generate the ridiculous "cryogenic Earth" strawman.

As we will see, the data is not "made-up" or "fabricated." Rather it is extrapolated from the current data!...

{from the article}:
But the slope of the defect line is similar to the slope of points 1, 2, and 3 in both the creation and uniformitarian models of the retention data (Figure 8).

Ah HA!
Now we are getting somewhere. You've dug down into the details and found another trick-with-the-data.

When Humphreys says "the slope of the defect line" he is referring to the Russian zircons - highly radiation damaged - which show a sharp change in activation energy at 390degC. But the Nevada zircons do not have a "sharp knee" even down to 300degC...
"The Nevada and New Mexico data go down to 300ºC (abscissa = 1.745) with no strong knee, implying that the data are on the intrinsic part of the curve."

Nevertheless, Humphreys has fabricated a knee in the data at 197degC, based on "points 1, 2, and 3" in figure 8. It is this knee that requires the made-up value of 3.74 kcal/mol.
Is this knee based on measured diffusion data?

No. There are no measurement of diffusion below 300degC.

Is it based on extrapolating measured diffusion data?

No. Extrapolating any of the three available activation energies - 46, 34.4, or 29.4 kcal/mol would not produce a knee at 197degC, nor an activation energy of 3.74 kcal/mol.

Instead, points 1 through 5 of figure 8 are based on, for the Creation model, ... get this...

...the diffusion values required to reach the the observed levels of helium in 6000 years!

You can read about it on page 9:
"Table 2 lists the resulting values of x, and the values of D necessary to get those values from eq. (14c) using a time of 6000 years..."

The diffusion values for the Creation model are derived from assuming the conclusion, not from any data at all.

In the Uniformitarian model, the diffusion rates are calculated very differently, with a different error used to produce a similar shape. The values are derived by assuming the samples at each depth have been at their present temperature long enough to reach the equilibrium age (called the closure interval by Humphreys). Since the temperatures have been wildly variable in recent geologic history, and are currently rising, this assumption is immediately, and obviously, wrong. Points 1 and 2, which constitute the "knee" in the activation energy, have not had time to even reach equilibrium, not by several hundred billion years or more.

This extrapolation is necessary pending the further testing requested of Farley (see page 6).

I wonder if the test results will ever be published... hmmm....

In addition, it is also pointed out that since the high temperature zircon data agrees with the creation model...

Only if we throw out a third of the data, and fit a curve to a smaller subset of the remaining data - that subset being the most erratic part.

In fact, the diffusion values predicted by the Creation model for points 4 and 5 are higher than some of the measured diffusion values from Farley's experiment. In other words, the Creation model "predicts" that diffusion will INCREASE as temperature goes DOWN. Since this behavior is contrary to all known real-world examples of diffusion, I would contend that the data do not agree at all with the Creation model.

...there is no reason to think the low temperature data won't be as well. With that being said, they are open to the possibility...

Good. I am certainly anxious to see that low temp data. Since the higher temp data fits so poorly to the Creation model, the low temp region should be the final nail in the coffin, since it is extrapolated so precariously.

But such a possibility (that the measured lower temperature diffusion rates would be higher than the extrapolated data) still doesn't seem to help the uniformitarian model since low temperatures conflict with the currently accepted thermal history of the areas in which the zircons and biotite came from.

Not true, you've got it backwards. A high activation energy at low temps means the diffusion rate continues to fall rapidly as temperature decreases. The made-up activation energy of 3.74 kcal/mol means the diffusion rate does not change much at lower temps compared to higher temps. This behavior is possible only with highly radiation damaged crystals, like the Russian zircons. As Farley reports, the level of radiation damage in these crystals is not unusually high, and no such drastic change in activation energy is observed in the actual measurements.

The normal, expected behavior of diffusion in solids is that the rate of diffusion decreases exponentially as temperature decreases. The activation energy of 3.74 kcal/mol represents an almost flat relationship between the diffusion rate and temperature. There are no physical measurements of diffusion data to support such a low activation energy at the temperatures indicated. Without such measurements, the low temp part of the Creation model is wishful thinking at best.

So far, we have:

1. One third of the data is missing in action.

2. The remaining data is selectively culled again, using the noisiest, most erratic part of the data set to "find" a slightly lower activation energy.

3. Activation energy of 3.74 kcal/mol fabricated by assuming the conclusion, and by botching the equilibrium age calculation.

4. Creation model predicts increasing diffusion rate at lower temps of 277 and 239degC, compared to measured diffusion rates at 300degC. This is about as likely as heat energy spontaneously flowing from cold to hot.

5. Amount of helium in crust and atmosphere far lower than predicted by Creation model. Escape into space over billions of years is the only valid explanation.

Wehappy,
It seems that steps 1-14, were a sort of initial start up phase, where the temperature was mononically increasing(unlike the bulk of the experiment). Do you think it is invalid to discount this step.

Secondly,
It seems that when those steps are removed from the "main body of the experiment" where the temp cycles, the numbers for the slope drastically change. This would seem to imply that there is a different phenomenon going on after the initial phase. Based on this, it would seem valid to remove the initial warm up phase, especially if you are not interested in what is happening as temperatures monitonically increase.

CT

There is a problem with the use of the diffusion data. Ken Farley gives his opinion of what the data he acquired most likely mean in appendix C. He doesn't say anything about dividing the pattern into two and fitting a knee, and this procedure is unjustified in the paper. It's hard to avoid the impression that this lower slope is needed to make the data fit the model. As WHF notes, these data do not look especially linear. It would be nice to see a goodness of fit number for both of these fits, just for comparison.

Also, the authors want the data below 450 deg C to lie on a different slope. But they seem to include this data in deriving the 34.4 value. They can't have it both ways - if you exclude this data the slope of the remaining data will increase. What effect would using this slope have on the models?

Perhaps a more subtle point. KF excludes the warm up steps because of his experience and within the paradigm in which he works. Since the report works outside that paradigm, a new justification is needed to discount these points. If the zircon diffusion number approaches the 'more normal' value of around 44 (Appendix C) does not the supposed problem pretty much go away?

The modelling is a very strange way to procede given that deriving thermal histories from helium retention is becoming commonplace. Since you've hired the pre-eminent lab in the field to do your helium measurements, why not ask them to deduce acceptable thermal histories of the zircons and see whether they appear to be outrageous in context?

Yesterday @ 01:47 AM post located here
ChristianTrader:

Wehappy,
It seems that steps 1-14, were a sort of initial start up phase, where the temperature was mononically increasing(unlike the bulk of the experiment). Do you think it is invalid to discount this step.


Secondly,
It seems that when those steps are removed from the "main body of the experiment" where the temp cycles, the numbers for the slope drastically change. This would seem to imply that there is a different phenomenon going on after the initial phase. Based on this, it would seem valid to remove the initial warm up phase, especially if you are not interested in what is happening as temperatures monitonically increase.

CT

Good questions CT,

When Reiners, Farley and other geothermometrists? exclude data, they try to justify it. For example, in Reiners' paper (here), sample 98prgb18 in figure 2 has very erratic diffusion behavior in the initial upsteps. He does not calculate an activation energy, although a line could easily be fit through the mess. They are more interested in finding data that form a straight line on the Arrhenius plot. That means that diffusion is behaving in a predictable fashion. Such data can reasonably be extrapolated beyond the temperature range of the experiment.

But Humphreys excluded the data in steps 1 through 14, even though that is the smoothest, straightest part of the data set. Of what remains, there is another, smaller group that forms a fairly straight line near the end of the run, and a much more erratic group in the middle. It is primarily this middle group that forms the "slight change in slope" with the lower activation energy of 29.4 kcal/mol.

Extrapolating noise is pretty much useless.

If we exclude the noisy parts, on the other hand, and keep the parts that form the straightest lines, the activation energy would still be in the 40-45 kcal/mol range. This value is more reasonably extrapolatable (is that a word?). We have more confidence that the linear shape of the diffusion graph will extend to the temperature range we need to estimate.

There is no "warm-up" phase. It's all good data, but if doesn't make a nice straight line, it's not much use for calculating diffusion coefficients that mean anything. That's why Reniers and Farley routinely use the down steps, because those have most often made the best-fitting lines. The upsteps were noisier. In Humphreys' zircons, the opposite is true.

Monotonically increasing is also not an issue. Most of Reiners' and Farley's published diffsion experiments have monotonically increasing steps, followed by monotonically decreasing steps.

But here's the real clincher that demonstrates how deceitful and deceptive Hupmhreys' "research" is:

Figure 6(a) from the recent article presented at ICC#5 shows, as I've already detailed, steps 15-44 of Farley's data. So Humphreys excluded the data from the initial run up to 500degC, which also happens to be the steepest, straightest data, giving the highest activation energy and the lowest diffusion rates.

The graph also show sample FTC1 from Reiners' paper (linked above). Did Humphreys show all the data from this sample?
No. Once again, he couldn't let us see all the data, he had to pick out the best parts (meaning the data that fits his model).

Did Humphreys use the same criteria when choosing which data points to include in the graph?
Yes and no. He again excluded the steepest, straightest data, keeping the erratic, higher diffusion steps. But this time he kept the initial steps, and threw out the down steps!

In both cases he keeps the data that helps his model, throwing out the data that supports the opposition. Obviously he wasn't trying to find the straightest, smoothest Arrhenius plot, because he excluded the best-fitting data in both sets. He wasn't consistently dropping the initial run-up to 500degC, because he kept it in the FCT sample. A more odious example of scientific hypocricy could not be found, in my opinion.


But WAIT! That's not all. To top it off, he multiplied the Reiners data by 10, moving it up on the graph above the Jemez zircon line.

The total effect of all this data-massaging is to make the measured diffusion data in figure 6(a) and figure 8 appear to line up very neatly with the Creation model derived points 1 through 5 in figure 8. He brags about how nicely they match:

"Figure 8 shows the zircon data from the Jemez Granodiorite, along with the two models. The zircon data are fully consistent with the creation model. These new data are also quite consistent with all published zircon data, as Figure 6(a) shows. As of this writing (February, 2003) we do not have reliable data on the Jemez zircons below 300ºC. But notice that the data have the same slope as the creation model points for samples 3, 4, and 5, and the data nearly touch point 5."

This is all hogwash, as I have shown.

If all the data were plotted on figure 8, especially the smooth straight plots from Reiners' samples, they would line up with the Uniformitarian Model points 1 through 3, and points 4 and 5 would be just below it.

The Creation Model points would be left high and dry, floating above the real diffusion values like a mirage.

WeHappyFew:

Humphreys has published some pre-print information on the internet regarding closure temperature, as well as a phenomenon he calls "re-opening" (scroll down to 10. "Closure Temperature"...):

http://www.answersingenesis.org/docs2002/1030meert.asp

Although I'm admittedly a layperson on this subject, it seems that he not only contradicts himself, but he then uses irrelevant laws to bolster his argument for "re-opening."

From what I understand, Dodson's work (cited by Humphreys) states that as temperture decreases in zircon crystals, the diffusivity D should decrease exponentially. While the crystal is still hot, He will escape as fast as it is produced. However, as the crystal cools and its temperature begins to approach ambient temperature, the loss of He becomes essentially nil, causing the crystal to retain almost all of the He produced through U decay. Measuring the amount of He retained by the crystal will give an apparent age, and the crystal's temperature at that age is the closure temperature.

At this point Humphreys begins to discuss "re-opening." Although I haven't been able to find any information about this process outside of this article, it seems that what he calls "re-opening" is simply the process of He diffusion from a high conentration to a low concentration. However, the process of diffusion requires both a gradient and a diffusivity greater than zero. In this case, he has already cited Dodson's reasearch, which clearly shows that diffusivity is effectively zero after closure temperature, but he appears to then contradict it by claiming that an He gradient alone is sufficient to cause further He loss from the crystals.

Am I incorrect on any of this or missing anything?

Originally posted by Objectivist

From what I understand, Dodson's work (cited by Humphreys) states that as temperture decreases in zircon crystals, the diffusivity D should decrease exponentially. While the crystal is still hot, He will escape as fast as it is produced. However, as the crystal cools and its temperature begins to approach ambient temperature, the loss of He becomes essentially nil, causing the crystal to retain almost all of the He produced through U decay. Measuring the amount of He retained by the crystal will give an apparent age, and the crystal's temperature at that age is the closure temperature.

Hi,
if I am understanding the article correctly, his point is the following:

- "closure temperature" is the temperature at which the helium diffusing out of the zircon is equal to the helium being created by nuclear decay. (Note - *not* the temperature at which no helium escapes)
- once the temperature drops below closure, helium will begin to build up in the crystal. Eventually it will reach ambient temperature and diffusion rate will stabilise. However, the diffusion out of the zircon is *not zero*; helium continues to be lost, but more is being created than lost, therefore the concentration increases.

- However (I think this is the main point), as the helium builds up, the concentration increases; and as the concentration increases, the rate of helium loss also increases (even though diffusion rate remains the same).
- Thus, once the concentration is high enough, the loss rate will rise to once again equal the rate of generation, and thus the system is "open" again; it has stabilised at a certain concentration of helium and it cannot go higher unless the rate of helium generation increases or the rate of diffusion drops.


Does that help?

Originally posted by Tim

Hi,
if I am understanding the article correctly, his point is the following:

- "closure temperature" is the temperature at which the helium diffusing out of the zircon is equal to the helium being created by nuclear decay. (Note - *not* the temperature at whichno helium escapes)
- once the temperature drops below closure, helium will begin to build up in the crystal. Eventually it will reach ambient temperature and diffusion rate will stabilise. However, the diffusion out of the zircon is *not zero*; helium continues to be lost, but more is being created than lost, therefore the concentration increases.

- However (I think this is the main point), as the helium builds up, the concentration increases; and as the concentration increases, the rate of helium loss also increases (even though diffusion rate remains the same).
- Thus, once the concentration is high enough, the loss rate will rise to once again equal the rate of generation, and thus the system is "open" again; it has stabilised at a certain concentration of helium and it cannot go higher unless the rate of helium generation increases or the rate of diffusion drops.


Does that help?

I understand all these points that he makes. However, his definition of closure temperature is incorrect - it is NOT the temperature at which He production rates and diffusion rates reach an equilibrium. I'll cite Dodson, who Humphreys also cites (although he does not cite this exact portion of Dodson's research):

It is assumed that, while the system is near to the temperature of crystallization, the daughter nuclide diffuses out as fast as it is produced by radioactive decay. As the system cools, it enters a transitional temperature within which some of the daughter product accumulates in the mineral and some is lost. Eventually, at temperatures near ambient, the losses are negligible, and the daughter product accumulates without any loss whatsoever...Closure temperature can be given a precise definition namely the temperature of the system at the time given by its apparent age.

This - not He production/diffusion equilibrium - is the accepted definition of closure temperature. Although Humphreys gets this wrong in his paper, this doesn't seem to be the end of his errors. He admits that diffusivity does decrease according to Dodson's equations, and that past closure temperature He levels continue to build within the crystal.

However, without some sort of mechanism to prevent continued buildup of He, Humphrey's entire theory falls on its face. It is at this point that he violates the laws of diffusion. Until this point in his argument, the diffusivity of they zircon is so low that unrestricted He buildup is possible (this is per Dodson's definition of closure temperature - the same work that Humphreys cites). So, Humphreys cites Fick's laws of diffusion as the method through which "re-opening" becomes possible. Fick's laws simply state that, given any concentration gradient, a substance will diffuse from areas of high concentrations to areas of low concentrations. In this sense, Humphreys is right - there is absolutely a diffusion gradient, with the high concentration being inside of the crystal. However, a gradient alone is not sufficient for the He to escape from the crystals - the diffusivity of the crystals must ALSO allow for the He to escape in the first place. Diffusivity is solely a property of the crystal and is NOT a property of the quantity of He within the crystal. As an example: a rubber balloon filled with He has an extremely high diffusion gradient, with pure He inside and air outside. Humphreys would have you believe that this alone is enough for He to escape from the balloon, when in fact the nil diffusivity of the balloon membrane prevents any diffusion from high to low concentrations.

Closure temperature can be given a precise definition namely the temperature of the system at the time given by its apparent age.

OK, but isn't the apparent age, the age at which helium begins to build up in the crystal? If this is the case, then it amounts to the same thing, as the point at which helium begins to build up is also the point at which loss rate drops below generation rate.

However, a gradient alone is not sufficient for the He to escape from the crystals - the diffusivity of the crystals must ALSO allow for the He to escape in the first place. Humphreys would have you believe that this alone is enough for He to escape from the balloon, when in fact the nil diffusivity of the balloon membrane prevents any diffusion from high to low concentrations.

I don't think that's what he means. AFAICT, he was saying that the diffusion is still high enough that helium buildup is not unlimited. Obviously he may be incorrect, I have no idea about that, but I don't think he is claiming that there is helium loss with zero diffusivity.

Originally posted by Tim

OK, but isn't the apparent age, the age at which helium begins to build up in the crystal? If this is the case, then it amounts to the same thing, as the point at which helium begins to build up is also the point at which loss rate drops below generation rate.

You are exactly correct - this is one of the flaws of He dating, that it can yeild a significantly younger age than radiometric dating will.

Originally posted by Tim

I don't think that's what he means. AFAICT, he was saying that the diffusion is still high enough that helium buildup is not unlimited. Obviously he may be incorrect, I have no idea about that, but I don't think he is claiming that there is helium loss with zero diffusivity.

He does claim that there is still helium loss, but he does it in a deceptive way by incorrectly using Fick's laws to justify this supposed "re-opening" when he is clearly aware that diffusion depends on diffusivity and the gradient present. He never clearly states this in his article, but anyone educated on the basic principles of diffusion should be aware of this.

As an interesting aside, on another unrelated message board, one of the members relayed several of my statements to Humphrey about his research, asking for clarification since he himself wasn't familiar enough with the specifics of He diffusion in zircon crystals . In Humphrey's reply, he actually states clearly that diffusion rates depends on both diffusivity and the diffusion gradient, so I know that he is aware of that fact.