the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 18 Mar 2025
Interlaboratory reproducibility of ID-TIMS U–Pb geochronology evaluated with a pre-spiked natural zircon solution
Abstract. The highest precision and accuracy in U–Pb geochronology is achieved using isotope dilution thermal ionisation mass spectrometry (ID-TIMS), a technique which owes its reliability to precise Pb and U isotope ratio analysis, a largely unified framework of lab protocols, and common isotopic tracers with accurately determined compositions. However, while hardware and protocol developments have steadily improved the analytical precision, the level to which ID-TIMS U–Pb dates from different laboratories agree remains largely unquantified. To better assess both internal repeatability and interlaboratory reproducibility of this method, we have conducted an experiment in which a large batch of natural zircon was dissolved, mixed with a newly prepared 205Pb–233U–235U tracer, and distributed as solution to participating laboratories. Thus prepared, pre-spiked, homogeneous PLES535 solution underwent the full sample preparation and analysis process separately in each lab, allowing a maximally unbiased comparison of the entire analytical procedure on a sample of unknown age. The results from 14 instruments at 11 institutions demonstrate internal repeatability of individual labs at 5 to 10 U–Pb analyses, with MSWD values generally indicative of single age populations. Lab weighted-mean 206Pb/238U and 207Pb/235U ages for the 337 Ma zircon solution agree within 0.05 % and 0.09 % (two standard deviations), respectively. This underscores the reliability of the participating laboratories for precise and accurate zircon U–Pb analyses, while highlighting the need for continued exchange on lab protocols and method improvement. We identify likely reasons for the remaining interlaboratory bias and discuss ways forward toward the goal of 0.01 % reproducibility.
Competing interests: Ryan Ickert is a member of the editorial board of Geochronology.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this preprint. The responsibility to include appropriate place names lies with the authors.- Preprint
(1230 KB) - Metadata XML
-
Supplement
(1847 KB) - BibTeX
- EndNote
Status: open (until 18 May 2025)
-
RC1: 'Comment on egusphere-2025-1001', Anonymous Referee #1, 06 Apr 2025
reply
This manuscript presents a well-documented test of precision and reproducibility of ID-TIMS U-Pb dates using pre-spiked solution of natural zircon, undertaken by the EARTHTIME community. The paper is well written, and in principle can be published in the present form. Still, it can be improved by greater attention of details in presenting analytical data and interpretations.
First of all, I would like to express my opinion about the design of this study. In my opinion, some design decisions made at the start of the study were not optimal, and significantly reduce the value and utility of this test.
- First and foremost, the choice to use a 205Pb-233U-235U spike without 202Pb. It is mentioned in the text several times that this decision was driven by low availability of 202Pb and ET2535 mixed spike. However, without the numbers showing the size of the remaining stock of ET2535, and the amount of spike used in this test, it is impossible to see whether this decision was justified. The data presented in this manuscript clearly show that instrumental isotope fractionation of Pb causes the greatest component in the age uncertainty (and this is emphasized in the text many times), but without Pb double spike, it cannot be properly quantified and accounted for. In my opinion, the authors must present complete quantitative information that lead them to the decision to use single Pb spike. If there was any chance to use ET2535 without critically depleting the remaining stock, I would consider the choice of spike made in this study an unforgivable mistake. Using the spike containing 202Pb and 205Pb would provide a lot of additional valuable information, without losing any currently available information, because the Pb isotope analyses with 202Pb-205Pb can be reduced with both internal and external fractionation correction.
- The decision to distribute only the pre-spiked solution. In my opinion, a better way would be to split the solution into two portions, pre-spike and equilibrate one of them, and then provide each participating lab with two aliquots: spiked and unspiked. The participating labs should have been allowed to use the spike of their choice for analysis of solution provided without spiking. This approach would have at least two advantages compared to the one used in this study. First, it would allow to determine the magnitude of errors related to sample-spike homogenisation (and to spike calibration, in the cases where spikes other than ET525 and ET2535 are used). Second, it would allow participation of the labs that are involved in the measurements of natural 238U/235U, and hence avoid handling enriched 235U. The only downside is the increasing the number of analyses, but additional 10 (or even 20) analyses is a fairly modest burden for a lab that specialises in U-Pb dating and performs hundreds of analyses (in some cases many hundreds) each year.
- The third is the decision not to accompany analyses of zircon solution with analyses of synthetic age solutions, e.g. ET100 or ET500. I consider this a missed opportunity to check whether any systematic differences between the labs vary in the same way for two or more age solutions.
- The fourth problem is insufficient supporting technical information from the labs that does not contain many potentially important pieces: useful ion yields for Pb and U analyses (should be included in the supplementary excel table), gain and baseline history, any determinations of cup efficiency performed on the same instrument, type of the ion counting multiplier (including manufacturer and the model), details of the deadtime calibration and linearity assessment, any analyses of the interference patterns (if performed), any in-run correction of oxygen isotope fractionation in UO2+ analyses. If the same instrument is used for high-precision (ppm level) isotope analyses of other elements, these data could also be useful. With these data, the evaluation of uncertainties related to baseline, gain, linearity, cup efficiencies and the like would be much better constrained, and not as speculative as in the current version. Fortunately, it is not too late to request these data from participating labs, and include them in the paper.
More specific questions are marked in the attached annotated manuscript. They are the inherent part of the review, and I encourage the authors to consider them as seriously as the text above.
-
RC2: 'Comment on egusphere-2025-1001', Brian Jicha, 07 Apr 2025
reply
Szymanowski et al present the results of an interlaboratory experiment on a pre-spiked zircon solution to evaluate the thermal ionization mass spectrometry for U-Pb analyses and the inherent corrections associated with them. The manuscript is impressive, clear, and well-written. All steps of the process, including preparation of the solution, are clearly explained. The manuscript can be published after some minor modification. The novelty of the community experiment is a little oversold as they have eliminated many variables from the typical U-Pb process. It is often very smart to eliminate variables in an analytical protocol to understand the limitations of some of the steps in the process, however, the manuscript needs to do a better job of discussing the limitations of this experiment. For example, on line 87 and again on line 90 and later in the discussion, the text states that they have chosen to avoid local tracer addition and avoid sample-spike equilibration. Why? Are there recent data available to suggest that this step is an issue? If so, please cite it or explain further why they have opted to eliminate this fundamental step from the experiment. Moreover, the reasons for the differences between the labs could be better conveyed or displayed so that people can see what analytical protocols produced the best results. Right now, the reader can only deduce what instrument/collector configuration produced the best results. The manuscript could serve as a very useful guideline for instrument operators if more of the instrument parameters were given.
On line 163 it says that the dead time calibration and baseline settings have been left to the decision of the participants. These parameters for each lab/instrument should be added to Table 1 so that other users of these instruments can see how they are being utilized. Thus, Table 1 should probably then be subdivided into two tables or a Table 1a and 1b with one showing additional instrument parameters (dead time, baseline settings, acquisition times, # of cycles, # and frequency of standards (ET 2535 or SRM 981) used for Pb mass fractionation correction, etc) and another table with blank information.
Line 260: The text states that by using a pre-spiked solution, it offers the opportunity to exclude geological bias. While this statement is true, the limitations of using a pre-spiked solution are not stated and should be. This would be another place to add a brief discussion of the limitations of the experiment.
As noted by the authors, in Figure 3, there are large and systematic differences in raw ratio precision among labs. They suggest that there are systematic differences in terms of Pb ionization efficiency or acquisition time. To further illustrate this point, a plot of raw ratio precision vs acquisition time would be beneficial within Figure 3. This plot could potentially address if some of the differences are a function of acquisition time vs. ionization efficiency. The same thing can be done for Figure 2. The text states that the large differences in the precision of 208Pb determinations reflect reduced 208Pb counting times preferred by some labs. Again, the counting times of each lab could be shown with an additional panel in Figure 2 (instrument vs counting time OR 208/205 precision vs counting time) or as a column in Table 1, preferably both. The “Next Steps” #4 says that common protocols for ion detector performance are needed. If all these parameters are given for each lab in this manuscript, it will become clearer to the community what protocols may be needed.
Lastly, I will offer a comment. The group has chosen a 337 Ma zircon solution. This age represents a sweet spot for U-Pb geochronology as there is adequate parent and daughter available for analysis. I would encourage any future intra- or inter-laboratory experiments to be conducted on zircons that are closer to the extremes of Earth’s history to fully assess some of the limiting factors on laboratory protocols. For example, instead of a 337 Ma zircon solution, the group should consider ~3 Ma zircon or ~3000 Ma zircon solutions.
This is a solid manuscript. Nice job to all involved.
Brian Jicha
University of Wisconsin-Madison
Citation: https://doi.org/10.5194/egusphere-2025-1001-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 296 | 45 | 7 | 348 | 12 | 3 | 6 |
- HTML: 296
- PDF: 45
- XML: 7
- Total: 348
- Supplement: 12
- BibTeX: 3
- EndNote: 6
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1