Dear RC1 and Editor

Please find below our responses to RC1 comments and the new appendix  (TL and OSL data) as attached file

The manuscript by Tissoux and co-authors deals with the relationship between impurities and the OSL sensitivity of quartz, which is an emerging method in sediment provenance analysis and sediment tracing. Previous studies demonstrated that the OSL sensitivity of quartz crystals in parent igneous and metamorphic rocks is relatively low compared to the OSL sensitivity of quartz from mature quartz-rich Quaternary sediments. However, the variability of quartz OSL sensitivity is in crystals from parent igneous and metamorphic rocks is significant and this has been attributed to crystallization conditions such as fluid composition and temperature which determines the amount and type of defects influencing charge trapping and recombination leading to OSL. Previous studies show ambiguous results regarding the relation between impurities and quartz OSL. Finding a clear relationship between impurities type and quartz OSL sensitivity opens the possibility of using OSL to track sediment source lithologies. In this context, the study by Tissoux presents relevant information to improve applications of quartz OSL (and TL) sensitivity to sediment provenance analysis. Specifically, the manuscript shows ICP-MS, ESR, OSL and TL data from quartz sediment grains from a watershed with well-characterized geological sources (parent lithologies). Besides the interest of specialists in luminescence methodos, I believe that the study is very attractive for researchers interested in the development and applications of sediment provenance methods. Two major points that I recommend to be addressed for publication are:

 I. Clarify that the quartz OSL and TL sensitivity appraised in the manuscript is in the low variation range compared to the range observed in sediments and; 

All values were already indicated in section 5.1 but we understood that it wasn’t clear enough for the reader.

For this reason, we therefore propose adding a table with sensitivities for OSL and TL measurements in the main text in section 5.1 and report OSL and TL curves in an appendix. We provide the values we found in the different types of bedrock and also now providing OSL and TL110 curves for each bedrock family. These OSL values are low (10 to  250 cts/Gy/mg) but in the range of previous published values for other bedrock OSL sensitivities (e.g. Constantin et al., 2025 in addition to other references we already cited) ranging  between ~10 to 500 counts/Gy or slightly higher values for siliclastic rocks. This is now clarified in the revised text 

II. Discuss if the observed low OSL and TL sensitivity patterns related with quartz impurities are preserved or influencing the high OSL and/or TL sensitivity of quartz in mature sediments. Is it possible to use OSL or TL sensitivity as proxy for source lithologies even if surface processes increase the sensitivity in several orders of magnitude? 

We see the point raised by the reviewer, however the main of our study is to investigate the bedrock OSL/TL sensitivity and not its evolution during transport driven by surface processes. We acknowledge that some of our sentences were misleading as explicitly dealing with sediments, so we rephrased them and only mention this point in our discussion. One important point is that our catchment is small, so the sensitivity changes by surface processes should be different than for other published studies dealing with large river catchments.

Specific comments

Abstract

Line 22-23: Clarify “sediment dynamics”.

The sentence is changed into “understanding sediment provenance and surface processes processes”.

1 Introduction

Line 46: Specify “quartz OSL sensitivity”. Revise entries throughout the manuscript.

Done in the revised manuscript.

Line 50: Add reference for this information.

Liu et al. (2022) correlated irradiation sensitivity of quartz Al and Ti centers and baking temperature by volcanic lava flow, the reference has been added at the end of our sentence as follow : “Similarly, ESR sensitivity (i.e. amount of signal growth per unit dose,) can be used to understand the quartz grain history (Liu et al 2022)”.

Line 69-71: This statement about sediment dating is not matching very well with the previous sentences. Same comment for lines 74-75. This information creates the expectation that OSL sensitivity on sediment dating will be discussed in the manuscript.

We understand that the sentences may have caused confusion, as the topic of our paper is not related to sediment dating or sediment tracing. We have therefore rephrased lines 65–71 and 74–75 to better emphasize the core focus of our study: the sensitivity of bedrock OSL/ESR signals (we removed mentions to Quaternary dating from our sentences).

See our revised manuscript. 

Lines 96-98: Clarify how the mentioned results are related to sediment dating. Or exclude information on sediment dating.

We agree this was confusing and have excluded the sediment dating information. See our revised manuscript.

2 Overview of OSL and ESR methods as sediment/bedrock tracers

Lines 101-104: The statement “This large variability was linked to magma-crystallisation processes and/or pressure-temperature conditions during metamorphism” has no support in the cited reference because the quartz from bedrocks have low OSL sensitivity (one to two orders of magnitude).

The text has been rewritten to distinguish between the variability of sensitivity within magmato-metamorphic rocks and the difference in sensitivity between these rocks and sedimentary rocks and present-day sediments from the literature.

Lines 113-114: Most papers dealing specifically with this subject did not observe quartz OSL sensitization due to sediment transport. See Sawakuchi et al. (2018), Capaldi et al. (2022), Magyar et al. (2024) and Parida et al. (2025). The downstream increase in quartz OSL sensitivity presented in Pietsch et al. (2008) can also be attributed to changes in source lithologies  (volcanics to sandstones).

We thank the reviewer for this suggestion. We revised our manuscript and included the proposed references. 

This is our goal here to present the different hypothesis for variation in quartz OSL sensitivity. We reported here the conclusions of the different authors, including Pietsch et al 2008, without preconception.  Moreover, our study will not deal with quartz sensitization in sediment. 

4 Methods and protocol

I understood that luminescence, ESR and LA-ICPMS measurements were performed in different aliquots. If this is the case, how does the variability among aliquots affect the comparisons or how homogeneous are the aliquots regarding the measured signals?

LA-ICPMS analysis was carried out on total rock thick section: On each section, the quartz grains present were located and targeted beforehand by microscopy. All quartz grain types of sufficient size and comparable to those measured by ESR and OSL were analysed. ESR and OSL measurements were carried out on different aliquots, but all aliquots were taken from the same prepared sample. Variability between aliquots was tested in OSL and ESR prior to measurement and found to be satisfactory.

4.2 OSL measurements and sensitivity analysis

Revise titles and text in order to inform that both OSL and TL (luminescence) analyses were performed. Titles and sentences are informing only OSL.

We changed titles for section 4.2 to “OSL and TL measurements for sensitivity analysis” and for section 5.1 to “OSL, TL and ESR measurements” and for sections 6.1  and 6.2.1.

For the requested information below, all data are now provided in Table 2, in the main text or in an appendix.

Line 246: Inform wavelength used for bleaching.

For bleaching we used the LED at 458nm, 90mW/cm² for 100s, and detection at 380 nm.

Line 248: Inform heating rate used for TL.

with 5°C/s to 125°C, information added in the manuscript on line 247.

Line 250: Inform peak temperature and/or show example of TL glow curve to make sure that the 75-125oC interval is encompassing the “110 TL” peak.

We are now providing graphs with TL and OSL measurements in an appendix.

Line 256: Inform the integration interval for the late background. 

90-100 sec , information added in Table 2.

5 Results

I strongly recommend showing examples of OSL decay and TL glow curves for the studied samples. They are very informative for specialists in the luminescence dating fields, who I believe will be interested in this study.

The TL curves and OSL decay are provided in an appendix.

Figure 4 shows significant IRSL signals from the studied samples, suggesting the presence of feldspar. How pure are the quartz concentrates? Please, clarify if the OSL (1s) and TL (75-125oC) signals could be influenced by feldspar luminescence signals.

Despite long H2SiF6 baths, it was not possible to eliminate the IRSL signal in the quartz for those samples. This is probably due to inclusions within the grain. The samples were crushed, so the grain size fractions were artificially created, and the remaining inclusions could not be reached by chemical etching.

Figure 4G presents the IRSL signals that we measured on step 4 in table 2. However, we see no  influence of the IRSL signal in samples #2, #9B and #33, as OSL signal used for sensitivity was done after an IR stimulation. So, the sensitivity provided in the paper is poorly influenced by the IRSL signal. It shows that IRSL treatment prior to OSL measurement eliminates  all feldspar-related signal, making it an important step for our OSL sensitivity.

If samples #2, 33, and 9B were significantly influenced by the IRSL component, the BOSL signal should exhibit much higher integrated values and a noticeably slower OSL decay — neither of which is clear in our data  (see appendix). Only for #9B me may suspect on the OSL decay curve some feldspar contamination but  its decay rate is similar to the two other samples from the same family (ie. #9A and 10). The TL-110  curves show no shift.

Furthermore, when looking at figures 4B and 4G , the samples with the highest OSL values are showing the lowest IRSL signals (#17, 20, 29,44,45), showing that the two signals are not correlated.

Figure 4: Consider the linear correlation (r2) with caution because of clustering of data points (low homoscedasticity). 

On Fig 4  A and D we  consider that the points are relatively well distributed along the linear relationship. In those cases, we are rather confident with the presence of a relation between the observed parameters and the r2 should be rather reliable. F4 B in particular may indeed be problematic because most of the points are clustered towards the lower values apart from a few points. These few points located away from the main cluster carry more weight in the calculation of the linear regression and the r2. Even though the r2 value is the best of all our figures it is not the most reliable.

We need to tone down our comment on the good correlations in paragraph 5.3. especially for Fig4 B, which is the least reliable (with a short explanation of why this is not reliable). We will indicate within the text that that our dataset only suggests these relationships, and that supplementing it with additional analyses of other rocks from other locations will enable us to confirm or refute these initial observations. 

Discussion

Line 389: Clarify “composite” metamorphic bedrocks.

The word composite has been removed, as unclear and unnecessary.

The relationship between impurities concentration and OSL/TL sensitivity seems complex and beyond a direct and deterministic connection among impurities, electron/hole traps and luminescent charge recombination. Other studies, for example, did not find correlation between impurities and luminescence sensitivity. A discussion on the potential changes in traps and/or radiation and non-radiative centers would be interesting in the discussion section. The luminescence model and sensitization processes described by Bailey (2001. Towards a general kinetic model for optically and thermally stimulated luminescence of quartz. Rad. Meas.) is a starting point for this discussion.

We agree with the reviewer’s comment about this point. However we don’t think we have enough experimental data to open and feed this discussion. 

We thank the reviewer for this missing reference, which is now included in our revised manuscript by adding the following sentence on line 431: The relationship between luminescence models and sensitization processes remains poorly understood, due to the complex interplay between impurities, electron/hole traps, and luminescent charge recombination (Bailey, 2001) .

Lines 473-474: Differentiation of the effects due to crystallization temperature and heating temperature after crystallization is needed. Changes in luminescence sensitivity due to heating by laboratory experiments correspond to effects by heating after crystallization.

We agree with the reviewer. For granites, we indicate the crystallization temperature, whereas for metamorphic rocks, it's the metamorphic temperature (post-crystallization) that is indicated. the formation temperatures of the initial rocks are not known, but quartz are recrystallized by metamorphism and initial temperatures are not necessary. We have precised this information within the text and we have highlighted that laboratory experiments do align with metamorphic processes. A still open question is to what extent metamorphism can reduce or change the ESR/OSL sensitivity in quartz since metamorphism may not completely led to re-crystallization.

Lines 481-482: The statement disagrees with many previous studies showing the sensitization of the OSL fast component and TL 110oC peak due to heating. The cited references are referring to crystallization temperature.

 I think that there is a confusion here between crystallization temperature (metamorphism) and heating (annealing) after crystal formation. 

To avoid any confusion, we deleted this sentence. 

Line 485-486: Quartz from paragneiss is recrystallized by metamorphism. Thus, it is not expected to keep the previous sedimentary characteristics. 

We do not completely agree, although paragneiss will have newly crystallized grains and thus low ESR/OSL sensitivity, some grains may be inherited from the parent material (sediments). This is an open question and would depends on the metamorphism degree and conditions (pressure/temperature), as well as on the initial sediment chemical composition. We thus prefer to leave this sentence.

Lines 509-515: See recent literature about sensitization due to sediment transport. Recommended papers are Sawakuchi et al. (2018, EPSL: doi.org/10.1016/j.epsl.2018.04.006), Capaldi et al. (2022, Quat. Geoch.: doi.org/10.1016/j.quageo.2022.101317), Magyar et al. (2024, Quat. Geoch.: doi.org/10.1016/j.quageo.2024.101629), Moyaed et al. (2024, Rad. Meas.: doi.org/10.1111/arcm.12826) and Parida et al. (2025, EPSL: doi.org/10.1016/j.epsl.2025.119267).

We have included a discussion note and have cited the proposed references.

Line 513: Double-check “remediation”. 

Thanks for the correction, this has been changed in the revised text.

Conclusions

Line 540: This conclusion requires the assumption that quartz from metamorphic rocks will not be sensitized by surface processes. This seems beyond the results of the study and seems not reliable because settings hosting quartz with very high quartz have metamorphic rocks as major parent rocks such as in Australia and Brazil.

This conclusion referred to dating difficulties that we have encountered elsewhere (other studies) on sediments taken from the foothills of highly metamorphic zones, where the sediments are not very mature (few bleaching-irradiation cycles due to proximity to sources). We have rephrased the text.

Line 545: Regarding the role of hydrogen content and water, see Sharma (2017, Proc. Ind. Nat. Sci. Acad.: doi.org/10.16943/ptinsa/2017/49024) and Stalder et al. (2025, Sed. Geol.: doi.org/10.1016/j.sedgeo.2025.106853).

Thank you for the references, those are included in the conclusions: “Additionally, the significance of hydrogen content (see Sharma and Stalder et al., 2025) undetectable by ICP and currently being determined by other means, warrants further investigation.”

Dear RC#2 and editor-in-chief, please find our answers to your comments.

General comment:

Present work tries to address an important issue related to variation in quartz OSL and ESR sensitivities by comparing their respective intensities with each other and estimated trace elemental concentration in the samples. In this work author restrict the study to source rocks in order to see the variabilities in intensity at source of sediments. Although, the work tries to address important question, but I am disappointed that no major conclusion is deduced. Majority of conclusion are in line with existing ones. Besides this, I find there are some serious scientific and measurement flaws which need to be addressed. I am giving majority of major comments below. In addition to this, minor and major comments are also given as comment in the attached pdf file. 

The originality of our study lies in combining for the first time different trapped-charge methods together with elemental analysis on quartz derived from various lithologies. It also aligns with recently-published works and our new datasets allow to further discuss or confirm/infirm proposed mechanisms to explain the diversity of ESR/OSL-TL signals related to rock type/quartz geochemistry. Previous works have indeed proposed similar mechanisms but only considering one individual trapped-charge method (ESR or OSL-TL), one specific lithology, or on very different rocks not from one catchment. The fact that our results mostly align with previous works is also an important outcome, reinforcing the potential of trapped-charge method to be used for source-to-sink studies.  We have an extended discussion and propose further investigation/perspectives to be tested in regard to our results. We have carefully checked our data and measurements (graphs have been updated when necessary), in order to be confident in the discussed measurements and correlations between ESR/OSL signals and quartz elemental concentrations. We provide below specific responses to the reviewer comments.

Comments

Page 8, Line 245-255: Why authors measured sample weight after measurements? any specific reason. Generally, silicon oil tends to evaporate during heating which can reduce the weight, thus mass normalization will be improper. In addition, it’s also shown that quartz crystal contain molecular water and it evaporates during heat treatment. This will also result in improper weight estimation.

Regarding the impact of the silicone mass or water, there are several points to consider:

–    The silicone is homogeneous, light (density = 0.5) and we applied it with a small pad, so there was very little added to the disc. It is thus a minor contribution to the measured weight compared to sample mass.

–    Measuring our discs before analysis could have biased our estimates: If the silicon really had an influence on the total weight, we would have overestimated the total quartz weight (during analysis quartz grains can be lost). By measuring after analysis and heating, we are closer to the real sample weight that was analyzed.

–    Concerning the role of water, we heated at “low temperature” ie. 190°C and we don't think that molecular water could have evaporated as it could have done at temperatures above 260°C.

“The aliquots were first bleached for 300 s at 90 mW/cm² and 125°C before any given dose.”

Does this step introduces any sensitivity variation? Have you checked?

We have not verified this on our samples, but according to Wintle & Murray (1999), who also bleach at 125°C, there is no sensitization at this temperature, even after several hours. Bleaching at 125°C is also a classic step in other published works looking at quartz sensitivity.

“a given dose of 50 Gy was given to all aliquots to allow,”

It will be better to refer it as a test dose or regeneration dose whichever is appropriate than to say it a given dose. It creates confusion. 

Yes, we agree that the term might be confusing. We used “regeneration dose” instead of “given dose” in the revised manuscript. 

“is investigated to evaluate its natural sensitivity and possible correlation with the OSL sensitivity”

Authors mentioned above that they have already bleached this peak and performed a natural measurement, then how are you referring it as natural sensitivity?

Again, our phrasing is confusing and we have changed “natural sensitivity” by “regenerated sensitivity” 

Some works (Chauhan et. al. 2015) have suggested that the 110 C TL signal and OSL signal are correlated only up to certain small dose, not for larger dose. 50Gy is quite large dose. Did authors check the correlation explicitly? 

We choose a moderate regenerated dose (i.e. 50 Gy) because the brightness of quartz was very low and difficult to measure above the background at low dose (i.e 5 Gy). We haven’t checked the influence of the regeneration dose on the OSL and TL signal, depending of the dose. The goal of our study was to show a potential correlation between sensitivity and bedrock type. However, Chauhan et al., 2019 (Fig. 1) showed a linear correlation up to high regenerated dose (i.e 95 Gy), and our chosen dose lies in the mid-range of the tested doses for this study (cf. g figure response#1 in attached  file)

 

“255 second similar irradiation dose but measured at room temperature (BOSLf25°C), which”

How does it help? 

This experiment was previously proposed in earlier publications. We replicated it to investigate potential variations and assess the correlation between the OSL signals at 125°C and 25°C. Our results reveal a notable linear correlation between OSL at 125 °C and 25°C, comparable to the previously observed relationship between OSL at 125 °C and TL at 110°C. While this type of correlation has already been reported for sediments—and only occasionally for bedrock samples (e.g., Sawakuchi et al., 2020)—our new dataset for bedrock may provide valuable insights and be of particular interest to the OSL research community. Furthermore, investigating the BOSL signal at “room” temperature allows to think about field investigation of quartz OSL with the Portable Luminescence reader (analogue measurement conditions) so this is insightful.

does it not get effected by retrapping by 110°C peak?

Yes, quartz OSL measurements at room temperature might be affected by contribution of unstable OSL components. However, the correlation observed with the BOSL125°C confirms that OSL sensitivity at low temperature (as low as 20-25°C)  could be used for quartz fingerprinting, a promising protocol to further investigate.

Page 9, Line 262-277:

“was exposed to light (SOL2 solar simulator - HONLE) for 2000 hours“

How does it compares with bleaching using blue led for 300s in previous experiments?

The aim of the experiments was not to compare bleaching mechanism of ESR and OSL but to reset the different paramagnetic centers both in ESR and OSL, which may not be the same. For each method, the commonly accepted protocol is applied, which allows maximum bleaching for each method independently (i.e. Toyoda et al. 2000 for ESR; Sawakuchi 2000 for OSL).

“We reasonably assume that both Al and Ti centres have reached saturation before sample processing and measurements because source bedrocks are older than 250 Ma.”

what is typically saturation dose of Al/Ti centres and what is the dose rate of source rocks?

According to single exponential dose response curves, the saturation dose of the Al centre is reported in literature from 3000 Gy to 6000 Gy, and that of the Ti centre was between 2700 and 4700 Gy for quartz (Rink et al., 2007; Duval, 2012). Fitted with a single saturating plus linear function (Duval 2012) the saturation dose of the Al centre range between 1900 and 6500 Gy).

Annual doses were calculated from laboratory gamma-ray spectrometry measurements on rocks sampled in the Strenghbach basin (Ortec HPGe spectrometer, MNHN,France ). Average annual doses on granites are of the order of 4,000 µGy/year (4Gy/1,000 years). For an age of 250Ma (see text), the dose received is estimated at around 1,000,000 Gy. For gneiss, annual doses are a about 3000-3500 µGy/year, and doses received are therefore higher than 1,000,000 Gy. For sandstones, the calculated annual dose is around 1000µGy/year. Considering the Triassic age of the deposits, the total dose received is higher than of 300,000Gy.

Page 9, Line 282-299:

“ To prevent quartz from breaking under the laser beam, analyses were carried out on thick sections made from whole rock samples from bedrocks. A total of fifteen samples were analysed for trace elements in quartz (including Al, Li, and Ti) with between 28 and 60 spots on the various quartz grains present on the thick section.”

Normally for ESR and OSL analysis, quartz is separated from the rock or sediment by rigorous chemical and physical processes. In this case whole rock is being used, so how reliable this data set will be for comparison with ESR or OSL data?

How accurate would these be for comparing against OSL/ESR data?

How the spot positions were chosen?

The comparison between quartz grains used for ESR/OSL and for LA-ICP-MS is indeed a very important question. The ideal would have of course been to analyze the same set of grains in OSL/ESR and LA-ICP-MS. As the preparation for the quartz separation is already taking a long time and there is an additional preparation time to be able to analyze loose grain by laser ablation (grains need to be mounted in resin so that they cannot move during the ablation), we decided to start by analyzing the quartz directly on whole rock “thin” section made thicker than usual because of quartz properties under the laser beam. We did that so that we could start LA-ICP-MS analyses quickly and compare them with ESR and OSL results as soon as possible, but making LA-ICP-MS analyses on the purified quartz grains has also been planned from the start. 

We took a particular care into the choosing of the samples used to make thin sections. They have been chosen to be as representative as possible of the rest of the rock sample that was used to prepare quartz OSL and ESR aliquots. On each thin section, the quartz grains present were located and targeted beforehand by microscopy. The location of the analyzed spot was chosen to provide a sampling as exhaustive as possible of the quartz within each thin section and trying as much as possible to be representative of quartz diversity (when there were several types of quartz). To be comparable with ESR and OSL analyses, quartz grains of size larger or comparable to those in the aliquots were favored. Grain borders were also avoided, partly to avoid chemical signature mixed with neighboring phases other then quartz, but also to avoid analyzing parts of the quartz grains that may have been removed in the OSL and ESR aliquots by the chemical treatment of the grains. Major mineral inclusions in the quartz grains were avoided but not micro-inclusions as they would have not been removed by the quartz extraction process.

Even though we took as much care as possible into the choice of the samples and spot location, we cannot completely exclude the possibility of an operator bias, of a sampling bias considering the size of the total analyzed volume in comparison with the total quartz volume used for ESR and OSL analyses, and of a preparation bias due to the preparation protocol that has not been undergone by the quartz analyzed in LA-ICP-MS. However, meanwhile the writing and submission of the present work, we could perform the LA-ICP-MS analyses on the quartz grains that were prepared for OSL and ESR. We want to keep these new results for another publication but one main observation when comparing these new results with analyses from the whole rock thin sections is that there is no major difference between both results except for one sample (gneiss 02) which Al content is much more homogeneous (towards lower values) in the new data (cf. g figure response#2 in attached  file). This is because the contribution of a certain quartz type in the rock has been overestimated in the whole rock thin section analyses.

 

“ The median concentration for each considered element was calculated for each sample to allow comparisons with ESR and OSL data”

It would be much better to use quartz grains in pellets form or use other method for elemental analysis. I am not sure, how reliably we should compare whole rock data with separated quartz grains data.

We used whole rock samples for the LA-ICP-MS analyses but these are very different from whole rock analyses. Using the laser beam we could focus only on quartz grains present within the whole rock samples. We'll take a closer look at this measurement in the methodology section and rewrite the paragraph to make it clearer for the reader.

The representativeness of our results is indeed a very good question. As explained in the response of the previous comment, we did our best to keep our LA-ICP-MS sampling as representative and unbiased as possible. We also tried to have reasonable statistics for each sample so that the median value would be a reasonable representation of the quartz sample concentrations. Despite all that, a bulk ICP-MS analysis of each quartz aliquot would have probably provided a more reliable representation of the average concentrations. However, by using laser-ablation, we want to explore the chemistry variability of quartz within each sample and compare it to the ESR and OSL response. This aspect is not particularly investigated in the present study but will be explored in more details in a future manuscript.

Page 10, Line 304-315:

“On the one hand, low sensitivities ……………. 100 and 250 cts/Gy/mg. The”

It will be better to give typical sensitivities in table format and shine down curves comparison. How the shape/decay constant of the two types of quartz varied. It can be a useful information. 

We added a table (Table 3) in the main text presenting the average BOSL125°C and TL110°C sensitivities. The curves for TL and OSL measurements are also now presented in an appendix. 

“TL110°C peak sensitivities present…………. higher values than other bedrock samples.”

Are there any peak shift and peak width changes observed for the different sensitivity quartz?

See our appendix for sample TL110°C curves - We do not see a peak shift for sample #29 but indeed a different peak shape with a small tail between 110 and 130°C, not observed in samples #30 and 44. However, the amplitude of the main peak is at the same position (85-90°C) but is much higher (x4 to 5) for sample #29.

“sensitivity values related to BOSL25°C measurements follow the same pattern with two groups of low and high BOSL25°C intensities”

I really don't understand the purpose of BOSL 25, why is it required? How it can help in understanding the posed objective?

See response to the reviewer for comment L255 above.

“Comparison between BOSL25°C and BOSL125°C ………..plot along this linear fitting trend (Fig.2A)”

Although positive correlation is expected, but the nature of regression may not be defined. Why authors say it’s a linear correlation? The R2 is low.  

We do not understand this comment as the reported  R2 is 0.89, which is far from a low correlation. Also the nature of regression is linear and was reported in Fig 2A (see the equation in Fig 2A). However, for a better vizualisation of the data, the graph is shown in a semi-log scale.

Page 10, Line 325-335:

“measurements: low Ti-Li and………remaining granites and sandstones.”

Is there a reason why such a specific trend is observed?

This is an observation from our analysis, discussions and hypothesis are presenter latter in the text.

“Ti-H. Third, intensity …………. granites/sandstones group”

Why is there no correlation between these quantities as they are expected to be correlated to luminescence properties of crystal?

Does that mean that there is no point of using correlation between BOSL and Ti-Li together? 

Several earlier works have suggested the correlation of Al centre with OSL of quartz, but present work suggests opposite. Which one to believe?

We agree with reviewer, for instance McKeever et al considered that the [H3O4]0 and [AlO4]0 recombination centers were responsible for the 380 and 470 nm emissions, respectively; Martini et al present evidence linking the 380 nm band to [AlO4]0 centers.  According to Poolton et al 2000, “the most obvious EPR defects to compare with the OSL data are the [TiO4/Li+]0 and [TiO4/H+]0 donors, since we have shown that these bleach at the same rate as the AlO4]0 recombination centers, and are thus possibly directly involved in the luminescence process” . 

In our manuscript, we present measurement results obtained on quartz extracted from source rocks in our study area, in which direct ESR intensities and  OSL sensitivities do not correlate directly which does not rule out a more complex link between ESR and OSL centers. These surprising results are discussed in our discussion section but the exact reason for these results need to be further explored.  

Page 11,12, Line 345-355:

Why there is no trend of different centers? As there is expected correlation of these centers with OSL, there should be a trend. Is it because the measurements were done on thick slice of samples rather than on the quartz grains?

We do not understand this comment. This paragraph focuses on trace element content, and at this stage we are not yet comparing ppm contents with ESR intensities or OSL sensitivities (see section 5.3). In this sentence we only note that element concentrations highlight the same groups of rocks.

Page 12, Line 363-371 section 5.3:

“First, comparison between Al contents and BOSLf125°C sensitivities show a good correlation (Fig. 4B, R2=0.719).”

Is it not the self inconsistent statements. In previous sections authors show that BOSL can be linked to rock types but Al and other centres can are not representing any trend with rock type, but here again authors are claiming that they are having good correlation. please check for consistency in the text.

We think that the reader might misunderstood our statements, where we sometimes refer to Al content (ppm) and sometimes to Al-center-ESR-intensities . We have checked in our manuscript, Al concentrations is named “Al content” with values in ppm, whereas Al-center-ESR-intensities are named “Al intensities” with values in a.u.

Here (Page 12, Line 363-371 section 5.3) Al content (ppm) is regarded, not the Al-center-ESR-intensities. As mentioned in the text, we don’t observe any correlation between Al-contents and Al-center ESR intensities (Fig4C) nor correlation between BOSL and Al-Center-ESR-Intensities (Fig2G), but interestingly there is  a good correlation between Al contents (ppm) and BOSLf125°C sensitivities (Fig4B).

section 5.3

In general, the correlation drawn in this section appear uneven. As discussed above, the centres intensities are not showing correlation with BOSL or systematic trend. Sandstone has high OSL but low Al content although others seem to have a good correlation. 

We do not agree, as in Fig. 4B sandstone samples (29, 30, 44) are showing both high Al contents and high BOSL 125C sensitivities.

No linear relation between Al and ESR Al is observed. Is it not because the measurement is incorrect. LA-ICPMS is measuring whole rock, which implies aluminum in the matrix can also be estimated and obviously it will be significantly different from Al content of the quartz grain. Possibly same is the reason for other uneven behavior. I encourage authors to do measurements on the separated quartz grains rather than bulk matrix to obtain consistent results. 

LA-ICPMS analysis was carried out on total rock thick section: On each section, the quartz grains present were located and targeted beforehand by microscopy. All quartz grain types of sufficient size and comparable to those analyzed by ESR and OSL were analyzed. No matrix or other mineral were targeted by the laser.

A test analysis of LA-ICPMS was conducted in a second time on quartz grain extracted for ESR /OSL (plots). Analysis gave similar median concentrations (see our detailed response and graph to the previous reviewer’s comment).

Page 13, Line 385-396:

“and (ii) the received natural dose is high enough to be considered identical for all samples.”

How can you considered natural dose same for all samples? Even if it is old, the variable doserates and the variable ambient temperature conditions will result in equilibrium doses to be different. This is wrong statement

We partly agree with the reviewer, as sensitivity for ESR is ΔE/ΔGy and is not correlated with saturation intensities. The text has been modified and  reference to “ESR sensitivity”  replaced within the text by “ESR intensity”.

However, we disagree with the comment that ambient temperature conditions differ between samples. Those have all been collected in a catchment at “surface” with low exhumation rates (<0.1 km/Ma) since the Cenozoic. Furthermore, dose rates are relatively similar between samples (see our detailed response to previous comment).

“(ii) two main groups can be obviously distinguished based on the analysis of different measured signals.”

Will it be true for present samples or for any other sample?

This statement is valid for our dataset, this is now specified in the main text.

“but no correlation was observed between ESR-Al and BOSLf125°C sensitivities”

This is clearly inconsistent statement w.r.t. previous statements.

See our response to previous comment and confusion between Al concentration and ESR-Al intensity. We mention in lines 334 and 394 that our observations on ESR-Al intensities and BOSLf125°C sensitivities are not correlated in our samples.

Page 13, Line 404-407:

“positions. First, we report a significant linear correlation (R2=0.891) between quartz’s Al and Li contents so that two groups of quartz-bearing bedrocks can be distinguished in the Strengbach catchment (Fig. 4A).”

Isn't this true for other silicate minerals also, why specifically quartz here?

The reviewer might be true, but in this study, only quartz grains were analysed by LA-ICPMS in order to better understand quartz behaviour regarding ESR and OSL. 

Page 14, Line 410-417:

“A good correlation is observed between Al contents and BOSLf125°C (Fig. 4B R2=0.719), as already reported in Alonso et al. (1983) and Preusser et al. (2009), whereas it is much weaker between Al contents and TL110°C (Fig. 4F R2=0.137) signals.”

This is inconsistent, OSL and TL are correlated, Al and OSL are correlated but TL and Al are not. why is it so?

We have checked and re-graphed all plots shown and discussed in Figure 2 and 4.

“Li+Na+K). On the other hand, …..quartz lattice itself.”

How much was feldspar contamination in separated quartz grains?

Despite long H2SiF6 baths, it was not possible to eliminate the IRSL signal in the quartz for those samples. This is probably due to inclusions within the grain. The samples were crushed, so the grain size fractions were artificially created, and the remaining inclusions could not be reached by chemical etching.

Figure 4G presents the IRSL signals that we measured on step 4 in table 2. However we see no  influence of the IRSL signal in samples #2, #9B and #33, as OSL signal used for sensitivity was done after an IR stimulation. So the sensitivity provided in the paper is poorly influenced by the IRSL signal. It shows that IRSL treatment prior to OSL measurement eliminates  all feldspar-related signal, making it an important step for our OSL sensitivity.

If samples #2, 33, and 9B were significantly influenced by the IRSL component, the BOSL signal should exhibit much higher integrated values and a noticeably slower OSL decay — neither of which is clearly evident in our data  (see appendix). Only for #9B me may suspect on the OSL decay curve some feldspar contamination but  its decay rate is similar to the two other samples from the same family (ie. #9A and 10). The TL-110  curves show no shift.

Furthermore, when looking at figures 4B and 4G , the samples with the highest OSL values are showing the lowest IRSL signals (#17, 20, 29,44,45), showing that the two signals are not correlated.

Page 14, Line 423-430:

“assumed to be caused by the fact that not all substitutional Al atoms are present in the form of paramagnetic centres.”

How the two groups are giving two different hypothesis? In present case I strongly suspect that it could be due to measurement on rock slice instead of quartz grains. Is there any way to prove or differentiate between substitutional Al showing non paramagnetic behaviour?

In present case the dose given is sufficiently high as authors claim.

As only the quartz grains were measured by LA-ICPMS, all the aluminum concentration does indeed come from quartz. The lack of correlation between aluminum concentration and ESR intensity of the Al center, on the other hand, indicates that not all the aluminum present in quartz is paramagnetic. Gotze (2004) explains that saturation of the ESR-aluminum centers in quartz was not reached in his study. In our case, we consider that saturation has been reached, so either our hypothesis is wrong, or other phenomena need to be taken into account. This needs to be further investigated.

“The increasing amount of Ti content may cause a general increase of the Ti-Li ESR signal, and an increase of substitution of Si by Ti.”

How this is happening? This is hypothetical. Why should this happen?

This is an empirical hypothesis “The increasing amount of Ti content may cause a general increase of the Ti-Li ESR signal, i.e. an increase of substitution of Si by Ti although the mechanisms involved remain unexplained". Literature revue indicates an increase of Ti substitution with temperature, not with Ti concentration. 

Page 15, Line 455-459:

why only quartz vein results are explained, why not other samples?

This paragraph is focused on the vein quartz sample, to discuss our own results with literature observations on the same material. 

.Page 15, Line 460-483:

“Importantly, not only are these observations somehow inconsistent with each other (e.g. contrasting behaviour, different threshold temperatures…), but they also only partly fit our data. On”

why you want to fit inconsistent observations with your data?

This is correct, and was not our aim.  The sentence has been modified as follow : “Importantly, these observations are somehow inconsistent with each other (e.g. contrasting behaviour, different threshold temperatures…). In our study, the formation temperature… “.

“If one would have expected the highest ……..gneissic samples (Fig. 2E). On”

This indicates that something other than temperature is playing a role. Constant high temperature helps the ions to defuse uniformly in the crystal. But cooling rate decides the trapping of defects in crystal. A faster quenching enable trapping of impurities, while a slow cooling lead to more pure crystal structure. Thus, it will be good if analogy can be extended for explanation.

Thank you this remark; we added a sentence in the revised text : “This indicates that  other external parameters than temperature could play a role, for instance cooling rate or pressure (see sections below).”

“Also, low OSL sensitivities are not only observed for the gneisses but also in granitic bedrock which undergone deformation (Bilstein Fault zone for samples #9A, 9B, 10). This”

This is contrary to laboratory observations. It is often observed that annealing up to 500 C in laboratory result in several order change in OSL/TL intensity.

The changes in luminescence sensitivity due to heating by laboratory experiments  correspond to effects by heating after crystallization of the quartz, in our case, the heating (and pressure) undergone by the sample lead to re-crystallization of quartz grains. See also our response to reviewer#1. 

 Page 15, Line 495:

“ This suggests that pressure can be one of the prevailing factors driving changes in OSL/ESR sensitivities”

what will be the mechanism for this? 

This is still unknown, see for instance Odlum et al 2022, Hu et al 2024, papers already cited in our manuscript.

Page 15, Line 515:

“the two other sandstone samples (#30 and 44) present similar OSL sensitivities than those of the granites.”

What is distance of transport from source for these sediments?

The source of these sandstones is unknown but the shapes of the grains and sorting of grains sizes indicate a long transport distance, that could have sensitized quartz grain to an OSL point of view. By the way, sample #29 present the highest OSL sensitivity (Fig 2A). Since we don’t have quantitative estimates for transport distance of the sandstones, we prefer not to discuss this point. 

Other PDF comments: 

Line 33. " (ii) enhanced OSL sensitivity in mature and recycled sediments underscores the impact of sedimentary transport and reworking."

This is not clear.

 This sentence has been reworked as unclear.: “enhanced OSL sensitivity in mature and recycled sediments (sandstones) highlight the importance of sedimentary transport and reworking on OSL/ESR signals”

Line 44. " time (Tsukamoto et al., 2011;...".

include Goswami et al. 2024

Ok, thanks for the reference, added to the list. 

Line 54. " (e.g. Sawakuchi et al...".

include sharma et al. 2017

Added to the references, we thank the reviewer for this suggestion: Sharma, S. K. et al  (2017). Understanding the reasons of variations in  luminescence sensitivity of natural Quartz using spectroscopic and  chemical studies. Proc Indian Natn Sci Acad 83 No. 3 September 2017 pp. 645-653

Line 246. "a given dose".

test dose 

“Given dose” as been modified for “regeneration dose” as proposed by reviewer #1.

  Line 250. " The TL peak".

TL peak sensitivity.

 Changed as suggested.

Line 251. " After IR stimulation at 60°C".

Is the effect of IR stimulation on BOSL estimated? can't it change the sensitivity?

IR stimulation have been done at low temperature (60°C), minimizing changes in sensitivity. This is specified in Table 2.     

Line 302. " ~10 to ~250 cts/Gy/mg."

how much is the background?

It depends on the samples. Generally the background is between 10 to 50% of the first second of the OSL signal. The OSL sensitivity has been corrected from the background signal (integrated between 90 and 100 sec - see table 2).

Line 309. " Second, Third, TL110°".

If you want to make these as bullet points then properly make these bullets or numbered.

 We have rewritten this section to classify our results.

 

Line 315. " A good linear correlation is also observed between TL110°C and BOSL125°C (R2= 0.647, R2 = 0.804 without #32 outlier; Fig.2B).".

comments similar to figure 2A,

it is difficult to see correlation via, semi-log plot. only thing one can say is that both the quantities will have positive correlation. Correlation coefficient is also not quite great, which represents linear correlation.

The representation of the data on a regular scale do not show the difference between bedrocks with very low vales as most of them are clustered on very low values. The R2 is significant for a linear correlation.

 Line 348. " Al contents do not discriminate between bedrock types, but median values seem slightly increasing from the quartz vein, gneisses and altered and deformed granites to those of granites."

the reason for this should lie in formation mechanism of different rock types as well as way the measurements are conducted.

See our previous response regarding measurements. For formation mechanism this is latter discussed in the discussion.

 

Line 352. " granites (except #17) and sandstones."

Why it should be high in sandstone?

This is our observations from LA-ICPMS measurements, we have no expectation.

Line 354. " scattered. The ratio between the median Al and Li contents shows a strong positive correlation (R2=0.891; Fig.4A), highlighting again two main groups of bedrocks as previously reported with the trapped-charge methods (ESR/OSL)(Figs.2A, 2D and 2G)."

what could be reason for existence in correlation for these two but not for others?

A distinction must be made here between the correlation coefficients and the plotted values shown in figures 2 and 4. It is the plotted values that almost systematically allow us to distinguish the two rock families. The sentence was rewritten to avoid this confusion of meaning.

 

Line 366. " sandstones, except for sample #17 showing rather low Al content."

Sandstone had high BOSL, but low Al content, does that mean correlation that is being drawn is not correct.

There is a misunderstanding, sample 17 is a granite so the “low Al content” is not referring to sandstone.

 

Line 367. " - no linear correlation is observed between Al contents and ESR Al intensities (Fig.4C)."

How is this possible?

We do not know and this is our observation. One assumption is that not all the Al in the sample is paramagnetic

Line 379. " 0.804 without #32 outlier,".

Why 32 was outlier?

We consider sample 32 as it cannot be attributed to one specific bedrock group (it depends on the measured signal/element).  Although the correlations can differ for our presented graphs in Figure 4, the two groups of bedrocks are still distinguishable, see for Figure 2 and 4.  Sample 32 has characteristics that sometimes place it in the granite domain, sometimes in the gneiss domain. The latter was sampled in granites intrusive into gneisses, right at the boundary between granite and gneiss. It has acquired characteristics specific to both formations, as all the quartz grains were not recrystallized. 

Line 380. " observation widely acknowledged in the literature".

For low doses its true, but for high doses, such relation need to be checked for samples being used.

See our response to previous comment.

     

Line 398. " 6.2 Relation sensitivities"

 The section title has been rephrased.

 

Line 461. " of Si- and O-vacancies".

what is more important higher temperature or rate of cooling?

We do not know, this is an interesting comment but an open question and we have no data/estimate to discuss the influence of cooling rate. 

 

Line 474. "If one would have expected the highest increase in Ti-Li and Ti-H ESR sensitivities for these metamorphic rocks, we observe the contrary as granitic bedrocks (i.e. samples#15,17,20,45; the “undeformed” granites) have higher ESR sensitivities than the gneissic samples (Fig. 2E)."

This indicates that something other than temperature is playing a role. Constant high temperature, helps the ions to defuse uniformly in the crystal. But cooling rate decides the trapping of defects in crystal. A faster quenching enable trapping of impurities, while a slow cooling lead to more pure crystal structure. Thus, it will be good if analogy can be extended for explanation.

We agree, this is why we also suggest that pressure can play a role (see section 6.3.2). 

 

Line 482. " suggesting that heating do not favour OSL sensitisation".

See previous comment

See our response to previous comment.

   

Line 532. "affect ESR Ti-Li and Ti-H centre intensities as well as BOSLf125°C signals".

How geological processes effecting the defect centres?

Those processes that we suggest are  linked to temperature/pressure changes (sections 6.3). 

Line 534. "especially, the pressure could be one of the prevailing factors driving changes in OSL/ESR sensitivities."

mechanism not clear.

 See our answer or previous comment.

 

Line 546. "warrants further investigation."

most of the conclusion are general ones, derived based on previous studies, but no novel finding in the present study.

See our detailed answer to main comment.