Could this be mostly common Pb (207/206 = 0.917) and varying U concentrations? Did you also measure Pb 204?

Whilst the 204 mass range was measured during laser ablation analysis, the majority of the 204 signal comes from the mercury (204Hg) present in the carrier gas, rather than the 204Pb in the calcite. Consequently, once the 204 signal has been corrected for the 204Hg background contribution, the remaining 204(Pb) counts are right at the lower limit of detection, and are frequently negative. This is somewhat exacerbated by the fact that the calcite I'm dating in this sample is geologically very young (> 0.5 Ma but < 3.0 Ma old), with low U (1-2 ppm).

From my experience, the common Pb intercept for carbonates tends to be 207Pb/206Pb = 0.83 +/- 0.1, however, this sample appears to show a rather higher common Pb ratio, as you pointed out. Also, in samples showing only common Pb, i tend to see very low 238U/206Pb ratios (e.g. < 100), which given that I'm working Quaternary age material isn't that radiogenic. What's strange in this example, is that we have varying U/Pb ratios, but relatively invariant common Pb.

U concentrations in this sample typically varied by 0.5-1.0 ppm across growth layers, as well as vertically throughout the stalagmite.

If all analytical artifacts (downhole fractionation, Hg correction...) can be ruled out, your data would mean that the sample is very young ("zero age"), contains only common Pb and variable U/Pb ratios. You might want to try 230Th-234U-238U U-series method to date this speleothem.

probably this sample is so young, thet do not contain measurable radiogenic lead, even at high 238U/206Pb up to 1100-1200. The variations in isotopic compositiom of common lead for speleotherms can be due to its variations in waters and aerosols from different locations 207Pb/206 from 0.806 to 0.906 (R.Millot, e.a., Chem Jeol 203 (2004) pp 75-90).

"Radiogenic Pb-loss"

From my experience in zirconology, this usually could be due to inheritance or radiogenic Pb-loss or combination of both. Low levels of radiogenic Pb in young materials also make them susceptible to over-correction and large error propagation (Large uncertainty in 207Pb/235U due to low amount of radiogenic 207Pb).

I rather propose radiogenic Pb-loss for your discordant ages with a trajectory to 0 Ma (see the attached figure). You should also note that the 207Pb and 206Pb are chemically identical and loss in exactly same proportion. In other words, 207Pb will not lose in preference to 206Pb or vice versa. The line (trend in your figure) is therefore straight because the loss of 207Pb is always directly proportional to the loss of 206Pb.This is why you have variable 238U/206Pb ratios with invariant 207Pb/206Pb ratios. U gain also affects the sample in the same manner as Pb loss that complicates the task of interpretation. But U loss in less common than Pb loss because Pb will occupy a site damaged by the alpha decay, making its diffusion out easier.

The Tera-Wasserburg concordia cannot show data with a large range of U-Pb ages, and impossible to show a zero-age point. You might want to try Wetherill Concordia as well to see if you can draw discordia line/s and find intercepts (you need to filter your data a bit) or change your method as Leonid proposed.

You can also use other trace elements + REEs to figure out more on the geochemical history of your speleothem calcite. Look for more papers on Pb-loss issue in your research field.

Good luck!

Hi there,

I think you are fighting a losing battle here. The 7/6 ratio is clearly modern common Pb. Even with your max age 3Ma predicted you have a U-Pb ratio in the 10^-4s. Which means that for your 1ppm U you are gonna want a pretty amazing mass spec with your laser, and I doubt you are ever going to see through that common Pb anyway.

If you can do some sort of pre filtering/imaging of your samples to find the highest U and lowest Pb in the sample you might get somewhere, but judging by the amount of analyses there you probably hit just about everything you could. I would suggest directly contacting a good speleothem group such as UniMelb and read some of their literature e.g Woodhead 2006 (you probably have already) and see how they pick through these nightmares!

Thanks very much for all your answers and contributions!

As most of you have suggested, this sample appears to be very young (geologically speaking), and so has had insufficient time for a large quantities of radiogenic Pb ingrowth.

In response to Leoonid Neymark's recommendation for U-Th dating, this particular stalagmite along with all other samples from the same region were first dated by U-Th methods, and were shown to be in secular equilibrium (>500 ka), and therefore potential candidates of U-Pb dating. Samples were also screened by digital autoradiography in order to identify the regions with the highest activity and therefore U concentrations. Care was take to avoid the basal/sediment contact zone, as activity was often highest here, and was more likely to be as the result of detrital 232Th activity rather than 238U activity. I also used the detrital 232Th concentration as an estimate of the 204Pb/238U ratio using the Pb/Th of 1.6 and 204Pb abundance of 1.4%, suggested by Woodhead et al. 2006. However, in more recent papers Woodhead et al have acknowledge that this method is not all that reliable for estimating 204Pb/238U, but since it required only a simple calculation it seemed prudent to do so.

Richard Taylor - thank you for your suggestion to contact Jon Woodhead at UniMelb. I am actually already working with him and the rest of the Melbourne group as part of my PhD research. The 3.1 Ma sample that I mentioned was was given to Matt Horstwood (NERC Isotope Geosciences Laboratory) and myself for U-Pb LA MC-ICP-MS dating at NIGL, and has formed part of a inter-lab/methodology comparison.

Ultimately, it would seem that the inability to data this particular sample stems from the fact that the sample is too young for insufficient radiogenic Pb ingrowth. However, this problem can be somewhat countered by samples with higher U concentrations, as I have had success with dating other speleothems by U-Pb LA MC-ICP-MS that are 250 ka (up to 25 ppm 238U), 350 ka (up to 80 ppm 238U), 1400 ka (up to 8 ppm 238U) and 3100 ka (up to 2 ppm 238U).

Rasoul Bezenjani - Thank you for your suggestion about radiogenic Pb loss. Presumably, radiogeneic Pb that would be lost from the calcite matrix in preference to initial/common Pb? As the common Pb would be locked into the calcite matrix, whilst the alpha recoil of daughter products would lead to radiogenic Pb often occupying the proe spaces between calcite grains, and therefore more susceptible to leaching? Rn loss can also be problem leading to radiogenic Pb below the U-Th chronometer, correct? But to my understanding, this is only a problem for highly porous calcite, and because this material is very dense, any radiogenic Pb loss, if indeed it has occurred here is unlikely to result from this processes. Correct?

Hi Chris,

When you see fixed 207Pb/206Pb associated with variable 238U/206Pb in zircon, it is a diagnosis of recent Pb loss (or U addition which is quite uncommon) as Rasoul Bezenjani pointed out. I am not familiar with U-Pb in speleothems but your 238U/206Pb ratios go really high and I wonder if it is not because U and Pb have been fractionated (during leaching?). This would be equivalent to recent Pb-loss.

So, in your samples you can either have a prevailing common Pb signature (not enough radiogenic Pb) or a preferential loss of the radiogenic Pb. But if you had preferentially lost radiogenic Pb, I guess the high 238U/206Pb ratios should be associated with lower 207Pb/206Pb ratios indicative of the age of your sample … unless Pb concentration is too low to be accurately determined. I guess a starting point would be to identify if the increase in 238U/206Pb ratios is due to an increase in U concentrations or a decrease in Pb concentrations.

A last thought: would it be possible that your sample got contaminated somewhere between sampling and analysis?

Dear Christopher,

Generally analytical uncertainty and initial Pb couldn’t make such a horizontal trend in your sample. We should consider that any disturbance in U-Pb system due to open system behavior would be entirely random (not systematic). This means either Pb (or intermediate daughter) loss or U gain would be required to create a trend like you have. In the case where Pb-loss (or U gain) occurs, the data give an age that is younger than the real age. Increasing in 238U/206Pb ratio is highly likely due to 206Pb-loss because you have a range of 238U/206Pb and even two analyses close to the Concordia! So there might have been enough radiogenic Pb. You might want to filter much of your discordant data (i.e. >10% or >20%). It could give you a better illustration.

I am not sure that calcite is a very dense material. Maybe I am biased by zircon! But we have Pb-loss in zircon as well. You should look for possible evidence of alteration, recrystallization, re-dissolution, aragonite inheritance etc. How about thick sections petrography?

Regarding the Rn loss you are right; if occurs it results in a young discordant age. Rn is the only gas in the decay chain and this is why you are concern about it. But I think due to its low mobility and more importantly its short half-life it has a negligible effect on your data. If the Rn loss was the case, you should have had your 207Pb/206Pb ratios more scattered vertically. If occurs, Rn gas diffusion makes your data randomly scattered (Your data are linear and nearly horizontal). 

I have a few questions for you regarding the digital autoradiography. 

How reliable is this method in order to choose the right spots to aid your laser beam?  Is it a qualitative or quantitative method that you have used? And are you able to discriminate between different types of radioelement? Otherwise, the hot radiogenic zones in your radiographic images might show the most porous areas as well as high U rich zones. You could get your 3100 ka age with 238U up to 2 ppm. Then the low amount of U in not the real problem here, so why not just choose the pure crystalline (inclusion free) prisms and zap them with your laser.

Dear Taylor,

You stated that "The 7/6 ratio is clearly modern common Pb" How could it be true in this sample with such a wide variation in 238U/206Pb?

I think no one is fighting a battle here. At least I am not interested! It's all about science and sharing our understandings (no matter if we make mistakes). We all know that we have different background and want to know more from others.

In each of your locations, i would look " Upwind " to find the source of

the material that is contaminating the speleothems.  Since Speleothems

take about 5,000 years to grow each inch of layers, it is important to obtain

cores ( I assume you are coring vertically down the axis of the stalagmites ),

and it is important to use cooling water that is not contaminated while coring.

Once you have the core back at the lab, cut it down the axis using distilled

water as a cooling agent to obtain two uncontaminated faces.

Note that while sampling from the cut face, each year will contain a winter, and

a summer layer about  one ten thousandth of one inch thick each.

I would also think you would need to get back in time past when

man started messing with nuclear weapons, and coal fired power plants,

and possibly sample caves on the western edges of continents as the wind blows

around the earth from west to east.   The individual layers in the 

stalagmites record what was in the air, and got rained out on the 

limestone, and the elements in the water, and in the calcium carbonate

gets combined together into the stalagmite.

I had advocated coring cave formations back in 1968, but the US passed

the cave protection Act, and squashed that idea in the USA.  It was not until

recently that I realized that coring verticaly is important to isolate the

minerals in individual layers that were at the top of the stalagmite in each

year that they formed.  This is important because it samples 

the minerals carried in the water thru the rock above for a single

year, or half year, and not a random mixture of data from several

years along the outer surface of the stalagmite.

Who knows, you might become the world leader in the sampling and

dating of the Earth using vertically cored stalagmite chronology.

Tree rings cover thousands of years, but cave formating cover millions

of years to tens of millions of years.

What you would be looking for is signature events like volcanic eruptions,

or meteor impact events literally recorded in stone with specific

numbers of yearly layers between specific signature events.

These might be region wide, or they might be world wide events.

Remember to plug the core hole so the formation can continue growing.

Use a non shrink grout, and a small plug of  formation, no oils, or it

will not continue growing.

One of the things you need to consider is the age of the limestone that 

the cave formation formed in.  In geology there were six geologic time

periods when reefs were froming because the sea level of the planet

was fairly stable, but still becoming gradually deeper. Reefs grow in periods

when they can grow upward and forward, and they die when the oceans 

suddenly drop several hundred feet in elevation ( oceans accessing a 

dry basin, and fill it up rapidly ).

When you are sampling you speleothem, perhaps you should also 

core the walls of the limestone to obtain some background information

about what was being incorporated into the limestone at the geologic

time period of reef formation.

Your cave formation will be a blend of older materials from the reef,

and the younger materials carried in the rainwater that is dissolving the 

reef.   

There may also be problem with how caves form.   The caves will tend to

be much older than the formations.  they can be associated with the 

stagnant water containing H2S which forms H2SO4 which creates

Gypsum which is removed from the cave as sea levels drop.  At a much 

later date Carbonic Acid, and Humic Acid dissolve the limestone,

and then redeposit the limestone in cave fromations inside the cave passages

which are subject to evaporation and run-off as air passes over them.

If the formation is very deep inside the cave, then only barametric 

pressure changes affect evaporation, but if the formation is 

close to an entrance, and there is a lot of air exchange, the formation

is becoming covered with dust  carried on the wind.  A moist surface is a dust

magnet.

Since the Borneo Caves tend to be some of the largest caves in the

world, and have multiple entrances, I suspect you have a natural

cave fromation contamination problem carried on the wind.

Pick small caves where the formations are very far from the entrance,

and there is not a lot of air exchange way back in the cave.

A good location would be formations near cave lakes that have false

floors growing on the surface of the lake.  This implies stable 

inflow - outflow conditions controlled by evaporation rate equal to

drop by drop inflow rates, stable over hundreds of thousands of years,

so no air flow contamination of the nearby stalagmites.

Dear Rasoul,

I am not sure what you mean by your comment about science and understanding. You seem to think i was being negative. I was simply pointing out the scientific problems which this dataset is addressing.

The "battle" i am referring to is that of common Pb in U-Pb dating. As you have said your experience is with zircon, this means you are spoiled in that with zircon you are almost always winning this battle, as it essentially contains zero initial Pb. Therefore even with discordant zircon you are likely to have meaningful trends on the concordia diagrams.

The plot Chris has shown has several distinct features that catch ones attention. Firstly the 207/206 ratio is consistent but geologically meaningless (unless it is a solar nebula speleothem!), giving a modern day common Pb ratio. Secondly the trend is very long extending down to very young 238/206 ages (~6Ma), but obviously not approaching concordia due to the consistent 7/6. Finally the error ellipses look like they are (barring a few balloons) fairly consistent, somewhere in the 6-8% region across the whole range (Chris might be able to expand on this).

What these features tell me is that their is a fairly consistent composition being measured (modern common Pb). Their is obviously some variation being produced by increased U content, but that every analysis is so swamped by the common Pb that the dataset is unable to produce a trend down towards concordia - hence "losing the battle".

Dear Christopher,

I am slightly puzzled by the fact that this particular stalagmite was shown to be in secular equilibrium (>500 ka) by U-230Th method. A simple calculation shows that if you start accumulating radiogenic lead in a speleothem initially having common Pb (206/204~18.5, 207/204~15.6, and 208_204~38.5) and 238U/206Pb=50, then in 3 Ma in a closed system its 207Pb/206Pb will be =0.68, which is a bit lower than your lowest measured 207Pb/206Pb value. Because 238U/206Pb value decreases with time, measured values of ~1000 in your sample may indicate that initial 238U/206Pb was even higher. For the initial 238U/206Pb=1000 207Pb/206Pb will be =0.15 in 3 Ma, which you do not see. Therefore, there are at least 2 natural reasons to explain your data:

(1) Your speleothem is younger than ~ 50 ka

(2) You deal with open U-Pb system. The first scenario would require COMPLETE loss of radiogenic Pb, which is highly improbable. The second scenario would require very recent considerable U gain, which would result in young 230Th/U ages.

Therefore, my suggestion would be (1) to double check U/Th dating results and confirm that the same materials were analyzed for U-Pb and (2) carefully check the accuracy of your U/Pb ration LAICPMS measurements during the analytical sessions when these data were collected.

I hope this would help in resolving your puzzle.

Best,

Leonid