General comments:

In trapped charge dating, U, Th, K contents are routinely measured to calculate the environmental dose rates. In the manuscript of Kolb et al. (2021), the authors have systematically studied the performance of the µDose-system in determining U, Th, K contents by a combined alpha and beta counting technique. The accuracy and reproducibility of the µdose-system is tested by repeated measurements on IAEA U, Th, K standards and two loess standards. The minimum measurement time needed for reliable results with the µdose-system is obtained by comparing the results from different measurement durations. Comprehensive inter-laboratory comparisons have been carried out by the µDose-system and other commonly used methods (TSAC, ICP-OES, gamma spectrometry) on ~ 50 natural sediment samples from various environments. Finally a conclusion is drawn that the µDose-system, as a fast and cost-effective new method, is capable of measuring radionuclide contents with good accuracy and precision.

This manuscript talks about the new technique in determining the environmental dose rates for trapped charge dating, which fits the scope of Geochronology journal. The data produced in this manuscript is profound and the conclusion is solid. The good performance of the µDose-system in determining the environmental dose rates would be of interest to researchers expertized in luminescence dating and electron spin resonance dating. It would also give guidelines for the researchers who are currently using or going to use the µDose-systems in the luminescence or ESR dating labs. Therefore, I recommend this manuscript to be published in the Geochronology journal after minor revision. 

Specific comments:

• Because of the algorithm used in uDose-system, the U and Th contents are negatively correlated. We can see from Figures 3, 4, 7, 8 that, when the Th content or activity is higher (assume it is overestimated), the corresponded U is lower (underestimated), and vice versa. I think it might be more helpful to add the bulk U+Th activity as another parameter for comparison. Even though the individual activities of Th and U are deviated from the expected values, as long as the bulk U+Th activity is close to the expected value, it might still be treated as a successful measurement regarding the calculation of environmental dose rate. The conversion factor from activity to beta dose rate is higher for U and lower for Th, while the conversion factor from activity to gamma dose rate is higher for Th and lower for U. They compensate each other, and the total dose rate does not vary much with the exact Th/U ratio (e.g. section 4.3.1 in Aitken 1985; Li and Tso, 1995). For example, in thick source alpha counting (TSAC), sometimes we can simply assume that the sample has equal activities of U and Th series. So, I think the bulk U+Th activity would be another evidence for the reliability of µDose-system to accurately determine the environmental dose rate.
• Maybe, it would be even more straightforward to calculate the final environmental dose rates for comparison. Dose rates can be calculated according to the true settings of individual samples (grain size, mineral, water contents, etc). Alternatively, dose rates of all samples can be simply calculated based on etched quartz with a fixed diameter (e.g. 150-200 µm) by U, Th, K measured from µDose and other methods (TSAC, ICP-OES, gamma spectrometry), and just assume constant cosmic ray and water content. Comparison of dose rates can directly give the readers an impression about the performance of µDose in determining the environmental dose rate.

Aitken, M.J., 1985. Thermoluminescence dating. London. Academic Press.

Li, S.H., Tso, W.M., 1995. Systematic error from Th/U ratio in luminescence and ESR dating. Nuclear Science and Techniques 6, 113–116.

• To study the impact of measurement duration on the results, the authors have made repeated measurements with different durations. As the data can be stored during the measurements, I have a concern: are the short-measurements separate measurements or are the short measurements the former parts of long-measurements? I guess the former strategy is more reasonable, otherwise there would be correlation between the short and long-measurements.

• When comparing the results of the 47 natural sediment samples, the discrepancy of several samples is attributed to the disequilibrium in U and Th decay chains. For example, fluvial flood plain sediments may have strongly alternating ground water levels which can increase or/and decrease specific radioactive daughter nuclides in the U and Th decay chains (line 483). In the beginning, I thought you meant the radon-loss induced disequilibrium, then I got confused. Because for the TSAC and gamma spectrometry measurements, the samples have been stored for 4 weeks before measurements and for µDose-system the samples have not been stored before measurements. If the discrepancy is caused by radon, the problem would exist for all samples. Now, I guess you meant the disequilibrium caused by long-lifetime daughter nuclides, right? Could you give examples of the daughter nuclides that might be influenced by the ground water level change, and if possible, list a reference? That may help the readers to better understand what you mean.

• In the sample preparation step (line 195), the samples were pulverized in a ball mill (29.5 Hz for 45 minutes) and then dry sieved restricting the grain size diameter to < 63 µm. Do you assume that grain size < 63 µm would be fine enough for alpha counting? And usually how much sample would be left coarser than 63 um after being pulverized in a ball mill for 45 minutes? I am a little worried that this sieving step may cause fractionation of the sample component. For example, if the quartz is more difficult to grind than feldspar, the left residue of > 63 um will contain more quartz and the fine powder will have higher K (as well as Th, U) contents. Or maybe, the left residue of > 63 um contains more heavy minerals which have high Th, U contents (e.g. zircon), and the fine powder will have lower U, Th contents. Would it be better that we extend the grinding time and avoid the sieving process?

Technical corrections:

Line66: ‘disk’ is used here while ‘disc’ is used in line 197. Please make it consistent.

Line79: ‘These pairs are the result of…’ change ‘result’ to ‘results’.

Line174: ‘16.5 ± 1.5 mg/kg for K’, change to 1.65 ± 0.15.

Line 197: ‘sample carrier’, could you please indicate in Fig. 1a which one is sample carrier? And also give the name for that metal base.

Line 385: ‘of only view hours’, change ‘view’ to ‘a few’.

Line 443: I think it is not necessary to use a separate paragraph here.

Table 4: Maybe it is better to also convert the activities of Murray et al. (2018) into the concentrations. That would give the readers a direct comparison of the results measured by different methods or labs.

Review of gchron-2021-20 The mDose-system: determination of environmental dose rates by combined alpha and beta counting – performance tests and practical experiences by Kolb et al.

In this study Kolb and colleagues present the results of performance tests and experiments undertaken on the mDose system (e.g. Tudyka et al., 2018) using IAEA calibration standards, two loess samples used by the luminescence communities as standards (Nussy and Vokegem), and a range of natural samples taken from the environment.  The authors present results of experiments which seek to test the efficacy of the mDose system in measuring IAEA reference materials of known radionuclide composition, and of loess standards which have had radionuclide compositions determined at a number of different laboratories. Two further experiments are then run to i) better understand appropriate measurement times on the mDose and ii) comparing mDose measurements with results from environmental samples measured previously at different labs with different approaches.

The mDose system has only recently been developed, and offers a new means of radionuclide determination from emission counting which can be used for trapped charge dose rate calculation purposes. Whilst thorough testing has been undertaken by the manufacturers, this study offers a broader examination of how well mDose measurements compare with previously made measurements, and investigates appropriate measurement times. The manuscript is reasonably written, structured, and executed and has the potential to make positive contributions to the growing body of literature which uses the mDose system for experimentation, but also the limited literature where cross-comparison of dose rate determinations are made.

It is my view that this manuscript could be suitable for publication in Geochronology. Before I can recommend it fully, I have some comments for the authors to consider, as a well as a number minor corrections, which I hope will help the authors provide better clarity in the manuscript. I would like to urge the authors to review carefully the paragraph at line 524 about µDose measurements potentially calculating reliable dose rates for samples showing disequilibria. I detail my reservations below, but feel strongly that this paragraph should be heavily amended or better still, deleted. You can’t calculate reliable dose rates for trapped charge dating purposes where disequilibria is present without extensive measurement of the decay series (e.g. with HRGS) and with numerical modelling (even then, you can only hypothesise about disequilibria history).

Comments/Queries

Paragraph starting at L40: It’s worth distinguishing here between emission counting and geochemical techniques, which may be relevant for considering data produced as part of the inter-laboratory comparison.

L88: You mention secular equilibrium here, can you clarify whether the mDose can be used to identify disequilibrium? I guess not because the two U series are only sampled once, and your two Th pairs are fairly close together in the decay series.

L104: Can you clarify please how repeated use of the standards will increase the accuracy of the calibration? To my mind, the accuracy (e.g. how well known the true value is) cannot be changed by repeated measurements, but precision perhaps can be monitored.

L105: Can you also clarify what you mean about the repeated measurements of the calibration? You mention later on in the manuscript that the calibration is repeated approximately every 6-8 months – is this what you mean here, or are you suggesting that the standard should be run after a fixed number of samples?

Sections 3.2 and 3.3: In these sections when reporting the various determined values, there is a switch between mg.kg-1 and Bq.kg-1. These come from the original publications, but it doesn’t make it easy for the reader to navigate. It would be very helpful to convert one unit to the other, and this could be included table 4. Also, you present a number of different measurement datasets, but don’t provide a rationale for why you chose the dataset that you did – is it simply a case of going for the first published values, or is there a different rationale?

Table 4: I think it would be really helpful here to provide a dose rate for these samples. I can see from the concentrations of U/Th/K that dose rates will be high, but off the top of my head, I don’t know how high – this will provide context for later measurement time experiments.

Table 5: I suggest strongly to move this table to the appendix/SI. It’s two pages of data which isn’t necessary for the paper and interrupts the flow. For completedness  and good reporting practice, can the sampling location data for Heidelberg and Gliwice be included?

L195: Was the sieving of the samples necessary after 45 minutes of milling? If you’re filtering out sand size particles, it would be preferable to extend the milling time rather than sieve away these resilient mineral grains because you may be introducing bias.  

L231: do you mean for samples with average dose rates?

L248: I advise to change ‘acceptable’ to ‘desirable’ in this line – longer measurement times cannot be used as an excuse for not measuring a sample.

Section 4.3.3: this section is long and is passive in terms of not offering any value to the paper. Tabulating it would offer an easier means of digestion for the reader (e.g. columns for homogenisation techniques and determination of the radionuclides and/or dose rates). Alternatively, move this section to the appendix/SI.

L321/325/334/351 (and potentially elsewhere): please avoid the use of excellent when making reference to accuracy/precision/results/reproducibility – it’s subjective, descriptive, and meaningless. E.g. what is excellent accuracy? You’re better off letting the data speak for itself.

Figure 2: can you offer some further explanation for the data in figure 2 please? When you say repeated measurement, do you mean literally repeat measurements on the same sub-sample, or do you mean that you re-sample the standard and re-measure? Clearly the K is reproducible, but there’s more variability in the Th and the U. Is this due to heterogeneity in the sample if you’re resampling each time, or is this a reflection of your measurement uncertainty? Is there a reason that the Th shows more variability than the other two radionuclides?  

Section starting at L335: I agree with your conclusion at the end of page 17 (L351/2) that the IAEA measurements suggest the µDose is reproducible, and don’t think that at first glance there is a problem with µDose (L346). However, there is a deviation of your results for Nussy and Vokegem loess standards and the published values you’ve chosen for comparison, and this should be further considered. Did you resample and remeasure your standards to assess intra-sample variability? This should be the first step. Then you can consider whether the discrepancies are due to ‘methodological problems’ (L349) associated with other techniques, and why these discrepancies might exist. It’s not enough to state that ‘more than 10%  are neither unusual for dosimetry measurements …’ (L350), especially without any references (see works by Hossain et al., 2002, De Corte et al., 2007, Williams et al., 2010). Once you’ve considered the sources of these discrepancies, placing the deviations in the context of calculated dose rates would be a positive way of reassuring the reader that you’re talking about minor absolute variations in U/Th, and that these don’t have a significant impact on overall environmental dose rate, and therefore, age calculation.

Section 5.2: This is a really interesting section, and likely very useful for µDose users. As a general comment, I’m a bit lost understanding how the number of counts relates to i) time and ii) dose rate. Of course, these are sample specific, but for example knowing the dose rate of the loess standards and how long it took to reach 1k, 2k, 3k counts (for example) would be very helpful to the reader.

L379: You start to explain the experiment here, but I don’t quite understand how these measurement duration experiments were undertaken. Were you making only one long measurement (up to 5/7k counts) and integrating counts for the shorter count times, or were you making numerous measurements (e.g. 0-250 cts, then 0-500, 0-1000 etc)? The former is preferable in my view.

Paragraph starting L477: can you say more please about suspected disequilbirum issues, drawing from the literature about how disequilibrium might manifest in fluvial sediments. You have pretty high Th values for these samples. Olley et al.’s 1996 study finds that U238 disequilibria is more common for their modern fluvial samples than U235 or Th. For all the samples you mention, Th is higher from the µDose system than Giessen TSAC/ICPOES – why would this be? Also, can you comment on the data for sample Gi455, where the Giessen and your data diverges for all three radionuclides. K won’t suffer from equilibria issues, yet your measurements is 4 times greater than the Giessen one. Is this an experimental issue? Why do you think there’s such good agreement for this sample when you look at the comparison for HRGS in Figure 8b? Please offer a fuller explanation at L506 when you return to this issue

Paragraph starting at 525: I suggest most strongly that you delete this paragraph, it’s not correct.  Dose rates calculated in the context of disequilbria are not accurate for trapped charge dating, no matter how precisely you can determine radionuclide concentrations, because ionisation is assumed to be constant through time. You haven’t discussed how COL-UGW1-4 samples have been identified as being in disequilibrium or discussed the extent to which this results in excess/loss across the U/Th chains and how this might impact final dose rate calc.

L533: “At the moment, we cannot decide whether our preliminary results are only an odd anomaly or an indicator for the µDose-system’s capability to produce reliable dose rate estimates even for samples suffering from radioactive disequilibria”. µDose is a measurement tool for determining radioactivity. As far as I’m aware, it cannot be used to identify disequilibria in the U/Th decay series, and even if it were, there’s no way of knowing of reconstructing the radioactivity history of a sample throughout it’s burial history.

L534: “In order to give a final answer, further detailed and systematic investigations are required, including the question whether the magnitude of radioactive disequilibria is a decisive factor for the µDose-system’s capability to determine correct values”.  I am entirely convinced that µDose can accurately and precisely determine U/Th/K and therefore infinite matrix dose rates. However, these infinite matrix dose rates rely on two fundamental assumptions i) of equilibrium and ii) of an infinite matrix.

Table10 and the paragraph starting at L537: I personally don’t think this table or section is relevant. Why wouldn’t µDose be able to handle samples from a broad range of depositional settings? The problem only comes when the two dose rate assumptions mentioned above cannot be applied.

Minor corrections/typos

L23: hyphenation not required, instead “heating events, and for ESR dating, the precipitation of minerals”

L30: replace “minerals are not stimulated any more” with “minerals are no longer stimulated”

L30: replace “are still” with “remain”

L31: remove comma midway through the sentence

L32: insert “the” in front of palaeodose

L44: replace “as well as” with “additionally”

L51: remove “here”

L58: I think you mean detection rather than determination

L71: Rewrite sentence to “..an Analogue to Digital Converter (ADC) samples and transforms the …”

L74: Replace “allowing to discriminate” with “allowing discrimination”

L75: Rewrite sentence to: “..pulses, as well as the elimination of background pulses caused by interfering variables”

L103: Replace “reveal rather” with “have”

L204: Replace “might have a tampering effect on” with “may impact”

L365: Correct ‘acitivity’ typo

Table 7: is “1.04 ± 0.003” a typo for the Nussy loess K% on Ahnert value?

L368/9: replace ‘rather fast’ with ‘relatively rapid’

L369/70: amend sentence to “without the need for storage for specific periods of time”

L371: do you mean precision instead of quality here?

L385: replace “view” with “a few”

L557/558: should be one paragraph

L565: these pieces of software should be referenced