Review of: Short communication: New analytical approach on (U-Th)/He dating of Fe-hydroxide with an example using goethite from the Amerasian Basin, Arctic Ocean

By: Olga Valentinovna Yakubovich et al.

I’m glad to see work on improving (U-Th)/He dating protocol for Fe-hydroxides. This manuscript provides a new sample loading method and the sample preparation technique for Fe-hydroxides (U-Th)/He dating. The sealing samples in quartz ampoules method is useful to solve the loss of U and Th during the sample helium extraction by heating processes. For sample preparation, authors proposed that U may be leached from goethite during the sonication by distilled water and crushing processes. This finding is helpful and may be beneficial for (U-Th)/He dating sample preparation. I think it would fit in this journal. Below I list my comments in more detail and hope it will help the authors improve this paper because I think they provide interesting protocol to contribute (U-Th)/He dating for Fe-hydroxides.

The application of the sealing samples in quartz ampoules

This sample loading method significantly solve the U and Th loss problem, but also bring two disadvantages: one is the relatively high blank of the quartz ampoule, the other is the release of helium gas controlled by helium diffusion through the quartz ampoules. My concern is authors proposed that the characters of helium released from quartz ampoules at different heating stages can reflect the helium retentivity in goethite samples. Although helium easily diffuses through the quartz, the helium release patterns in Fig.5 show all samples (except DG 1015) have similar patterns and release helium at high temperatures (350-1250 °C). This is obviously different with the traditional goethite helium release pattern showing the peak of helium release commonly occur during phase transformation at ca.300 °C. In my opinion, these helium release patterns may be controlled by the helium diffusion parameters of quartz. In this case, I suggest authors may fill helium gas in quartz ampoules and test the helium diffusion characters in quartz. This may help to determine whether the helium release patterns reflect the gas release characters of goethite sample or quartz ampoule, because it is useful to understand helium loss, recrystallisation, and fluid and mineral inclusions in goethite samples. For DG 1015 sample, the helium release pattern is obviously different with others in figure 5, and shows the peak of helium release occur at ~350°C, this is consistent with the goethite commonly release the helium during the phase transformation at ~300°C. In this case, this quartz ampoules may be broken during the heating or improperly sealed quartz ampoule. Finally, I expect authors add all the helium release patterns in figure 5, especially DG969 that show the oldest age in the same goethite clast, then we can identify the reason why some grains yield the dispersed ages.

Goethite U loss during the sonication by distilled water

Investigating goethite U loss during the sonication by distilled water is a very welcomed contribution. Authors design two sonication washing stages during the sample preparation and the results show up to 7.8% U in one vein grain sample can be leached during sonication in the distilled water. As we known, the minerals in this sample are possibly stable in water, because it collected from dredge haul DR7 at 3400 m water depth. In this case, I wonder whether the U is the dissolved ion or occur as suspended colloidal particles in the distilled water after the sonication. In the first case, the U loss may significantly affect the dating results, and this suggests (U-Th)/He isotopic system may be open in water. The second case may not hinder us to obtain the reliable ages. Numerous studies reveal that sonication is commonly useful to isolate smaller size aggregates, but this physical disintegration produce lots of fine particles in water, even after centrifugation and filtration, it also incomplete removal of fine suspended colloidal particles from the supernatant prior to element analysis. The results of leaching experiments in Table 3 show that less than 0.3% U loss occur at first sonication stage, while most of U (up to 7.8%) loss at second stage involved to the crushed and sonicated grain sample. Is it possible that the sonication of the crushed sample increase the suspended colloidal particles in water? In addition, why the U loss of crushed samples with two (first and second) steps is much higher than those with only second step? Authors also find the leachates contain amount of Mn and Fe, why do not list all these results? If possible, the relationship between U-Th and Mn-Fe may provide some information on whether U and Th are from dissolved ions or fine suspended particles.

General comments

80: Table 1 show that minerals in DR7 are composed of goethite, birnessite, and other minerals, this means the sample used to (U-Th)/He dating are not pure goethite, how to evaluate the (U-Th)/He dating results from these mineral mixtures.

90: Figure 2 show multiple generation of Fe-Mn-hydroxides or oxides and small monazite may occur in this sample, these mineral associations could affect the dating results.

115: All the samples used to (U-Th)/He dating were not washed. Why do not carry out some washed samples to test they may yield systematic older ages?

120: How do you seal the quartz ampoule? If use oxyhydrogen flame, is it possible to heat the ampoule and cause the loss of helium?

170, 205-210: The U and Th is not only derived from mineral surface adsorption but also from some suspended particles. Authors may list the concentration of other elements, such as Mn, Fe, Al, P….

216, 218: “Qu ampoule” -quartz ampoule

234-239: How about the genesis of dating sample, are they precipitated from the sea water or hydrothermal fluid?

Fig.5 Why only plot six goethite fragments patterns? How about the patterns of ID 966 and 969?

249-260: Sample DG969 has the lowest U, but sample DG 1031 has the highest U, the two samples yield erroneously old and reproducible ages. In this case, it cannot be simply explained by U-loss of these samples.

This manuscript suggests a new analytical approach to (U-Th)/He dating of goethite samples using the single-aliquot method. Since U can be volatilized during laser heating, the authors propose to encapsulate goethite samples in quartz ampoules which will let helium diffuse through but will retain any volatilized U. The U is then recovered by dissolving the glass ampoule together with its contents. This method is applied to submarine iron oxide nodules.

I appreciate seeing a new approach for dealing with U volatilization in iron oxides. This is a promising technique that could provide a way to deal with U volatilization without having to make changes to the extraction line, which makes the application easier and more accessible.

While the application of (U-Th)/He dating to submarine nodules is exciting, I am wondering whether this was the best choice of sample for an exploratory study into new analytical techniques. In my opinion, samples with more independent age constraints would have been a better test of this method since the efficacy of this method could be verified. However, these samples are sufficiently documented to be a good first demonstration of the technique.

My major analytical concern about this approach is whether the samples can be completely degassed and whether the resulting helium can be completely extracted from the ampoule. Incomplete extraction of helium would lead to an underestimation of the (U-Th)/He age. Have any of the samples been heated to temperatures higher than 1150 °C? Some of the samples seem to still outgas a significant amount of helium in the 1000-1150 °C range (Fig. 5). Even though the amount of helium extracted declines in the 1150 °C step, there could still be helium in more retentive domains. Bimodal release patterns are not unusual and are dependent on the distribution of diffusional domain sizes in the samples.

Samples could be tested for complete extraction by heating previously heated samples to even higher temperatures or melting them with flux to determine whether more helium is released. If the material is homogeneous in age and eU concentration, a separate set of samples could be weighed and completely degassed using flux and/or high temperatures. The ⁴He concentrations obtained from samples heated in the quartz ampoules could be compared against the concentrations from the reference samples to check for complete extraction.

Another issue is the complete recovery of the volatilized U, which could equally affect the calculated ages. I suggest dissolving a number of unheated fragments of each structure and determining U and Th concentrations. The obtained U and Th concentrations of the heated samples can then be compared against these reference values. This would also demonstrate how homogeneous the sample is with regard to parent nuclides.

I think that the efficacy of this method is somewhat overstated. At several points throughout the manuscript, the authors claim that their samples show “no signs of overdispersion”, yet they report an MSWD of 3.4, which is very clearly overdispersed. This should be addressed. Adding additional replicate measurements for the same material and re-evaluating the MSWDs of these larger datasets might increase the significance of the findings.

I think that this manuscript is a good exploration of the quartz ampoule technique for goethite. However, several aspects of the technique, such as complete extraction of helium and full recovery of any volatilized U, need to be demonstrated with a larger number of replicate analyses of samples with more independent age constraints before the method can be claimed to be robust. Such additional analyses could be incorporated into this manuscript to make the conclusions more significant, but if this is deemed to be beyond the scope of this study, I would recommend scaling down the claims as to the robustness and reproducibility of the method.

Overall, I think this manuscript is a good fit for a GChron short communication. This manuscript is a great contribution towards establishing the quartz ampoule degassing method and will be interesting to the (U-Th)/He community at large. However, the issues brought up here and by other reviewers should be addressed before publication. I, therefore, recommend that the manuscript be accepted subject to minor revisions.

Detailed comments:

Line 17: I think “remarkable reproducibility” is overstating the reproducibility of n=4 with MSDW=3.4 and n=2 with MSDW=1.4.

Line 19: Insert “a” between “that” and “significant”.

Line 30: Change “mineral” to “minerals”.

Line 32: What is “sufficient” retentivity? Published studies generally indicate retention in the range of 80-98% for comparable material at Earth-surface conditions. I am wondering whether these samples are comparable to the supergene samples for which diffusion experiments have been performed so far. These samples could benefit from ⁴He/³He diffusion experiments to determine the retention of radiogenic helium.

Lines 34-36: Since these are not exhaustive lists, I suggest adding “e.g.” to these references.

Line 36: Change “successful” to “successfully”.

Line 44: Add citations to support that these methods are “typically” applied.

Lines 74-75: What is the observation that dark-colored goethite has better crystallinity based on? XRD/SEM/both?

Line 108: 5% HNO3 alone might not be enough to stabilize Th in the solution. Typically, a small amount of HF is added to solutions to stabilize Th during the measurement process. Th fractionation can occur in the ICP-MS tubing if it isn’t sufficiently stabilized.

Line 126: Was the sensitivity based solely on comparison with a mineral standard? Was there any internal standardization with an air or other gaseous standard from a tank?

Line 127: What is the repeatability of the 10 measurements of the mineral standard? The manuscript never seems to come back to this, and I don’t see these measurements in the results tables. Please report these results.

Line 131: I’m not sure what “in the camera of the mass spectrometer” refers to. Please clarify.

Line 214, Line 285: The results have MSDWs of 3.4 and 1.4, respectively. These values are both >1, which counts as overdispersed. A value of 1.4 is reasonably close to a univariate normal distribution to claim that the results aren’t overdispersed, but that is for n=2. However, an MSWD value of 3.4 is definitely overdispersed at n=4. (U-Th)/He ages in a variety of minerals are typically overdispersed, so that’s not unusual. It is most likely an effect of zonation and our incomplete understanding of the (U-Th)/He system rather than the measurement process, as long as U volatilization during heating is avoided.

Line 216: The abbreviation “Qu ampoule” isn’t defined. I suggest just using “quartz ampoule” instead. There are several more uses in the same paragraph that should be changed as well.

Line 223: I wouldn’t characterize the O₂ tank as “dangerously explosive”. It is a thick-walled vessel filled significantly below its rated pressure. Since it is filled with pure O₂ and no combustible fuel is present in the tank, it does not present an explosion hazard. Mishandling could lead to damage to the mass spectrometer and extraction line, the risk of which can be mitigated by implementing redundancy and fail-safe measures. Overall, the amounts of oxygen are small, comparable to the natural amount of oxygen present in the air inside a vented extraction line. Pressurized oxygen tanks filled to much higher pressures are common in labs and other settings and can be safely used if handled and stored correctly.

Figure 5: Not all of the performed analyses are shown here. Please add the missing patterns. Additionally, the use of colors to distinguish between sets of samples might present a barrier to accessibility. I suggest using a different symbol for each sample/set of analyses and an accessible color scheme to help visual interpretation. Some of the values given in this figure use dots as decimal symbols while others use commas. This should be consistently applied to the rest of the manuscript as well.

Figure 6: While age-eU plots are a good tool to diagnose radiation damage effects and U loss, the small number of samples and the small range of eU values (2-3 ppm) make the interpretation of this plot difficult. A larger dataset might reveal patterns. Also, use the same symbols and color scheme as suggested for Fig. 5.

Line 259: The presence of zircon and monazite in the sample might affect the helium release pattern. These phases are likely to be more retentive than goethite and would lead to some of the helium being released at higher temperatures. Having the helium release patterns of these samples in Fig. 5 would be helpful. A possible strategy to prevent inference from mineral inclusions would be to pre-screen samples with microCT and only pick inclusion-free grains for analysis.

Line 271: How are you distinguishing between recrystallization of existing material and newly formed material from a second period of goethite crystallization?

General comments

Yakubovich et al. propose an alternative analytical approach for dating goethite by the (U-Th)/He method. To avoid U-loss during heating for He extraction, multiple grains (1-7 mg) are encapsulated in vacuum-sealed quartz ampoules. The quartz ampoules containing the grains are then dissolved in a solution of aqua regia + HF + HClO₄, and U and Th isotopes are analyzed via isotope dilution method.

Two samples of Fe-Mn-crusts collected from the Egiazarov Trough, Amerasian Basin, Artic Ocean, were used to test this alternative methodology. The Fe-Mn-crusts are a mixture of goethite and Mn-oxides and other phases (e.g., lepidocrocite, feroxyhyte, quartz, feldspars, clays, clinochlore, etc.). Goethites from Fe-Mn-crusts are porous and poorly crystallized. Yakubovich et al. propose that U and Th are leached from the samples during ultrasonication and interpret the variable Th/U_leachate and Th/U_grain values as evidence for preferential U remobilization.

It is encouraging to see researchers interested in further refining the (U-Th)/He method applied to the study of supergene phases, as obtaining meaningful ages for these complex materials can be challenging and time consuming. Therefore, Yakubovich et al.’s proposal of an alternative approach to measure the ages of goethites is worth pursuing. However, before the manuscript can be accepted for publication, some important aspects of the work need to be assessed:

• suitability of chosen samples for (U-Th)/He geochronology;
• interpretation of leaching experiments results;
• interpretation of the (U-Th)/He results.

The successful application of a geochronological method depends on sample selection. Since 1999, ~ 2600 goethite (U-Th)/He ages have been produced by four laboratories. The most prolific laboratory, the noble gas facility at Caltech, is responsible for ~2000 of those results. The most important criterion to ensure reliable results is to select pure well crystallized goethite. Powdery goethite and poorly crystalline goethite intergrown with other minerals should be carefully avoided. Early in our studies, to test whether poorly crystalline goethites produce reliable results, we carried out analyses of co-existing crystalline (generally black or brown dense goethite samples) and poorly crystalline (generally yellow, porous and powdery goethite) samples. The combined (U-Th)/He-⁴He/³He experiments show that the latter are unsuitable for geochronology and should be avoided. Yakubovich et al. chose to work with impure, poorly crystalline goethites, the very same type of goethite that should be avoided in (U-Th)/He geochronology. Thus, their results, as interesting as they are, are useful to confirm that impure poorly crystalline goethites are not suitable for (U-Th)/He geochronology. Their findings cannot be generalized to all goethites, particularly to the types of goethites carefully selected for geochronology by most laboratories and analysts.

As demonstrated in this study, impure poorly crystalline goethite suffers from several limitations. One of these limitations is that U and Th may occur both within the goethite crystalline structure but also adsorbed or as a surface precipitate on the high surface area fine grained porous phases (porous goethites, manganese oxides, clays, silica minerals). Most He produced by U and Th adsorbed or precipitated on mineral surfaces will be lost, and the analytical results will provide no information on the sample’s age. In addition, U and Th on those surface sites should be readily desorbed and removed by aqueous solutions, both during their geological history and during sample preparation and cleaning procedures, further challenging the interpretation of results obtained from this type of material. Yakubovich et al.’s results clearly show why these types of samples should be avoided if the major objective is to date a goethite occurrence.

In the interpretation of their (U-Th)/He results, Yakubovich et al. raise the possibility that the Fe-Mn-crusts were formed deep in the Earth’s crust and then acquired their ages after cooling below a certain closure temperature. There is plenty of evidence suggesting the hydrogenetic origin of the Artic Ocean Fe-Mn-crusts (Hein et al., 2017). To avoid misinterpretation of the results, the authors should focus their interpretation of the results in the context of changing environmental conditions at the contact between detrital sediments and sea water. More importantly, the fact the samples are made of a mixture of minerals, not pure goethite, precludes a robust interpretation on the precipitation ages of goethites. Finally, based on our current knowledge about the helium retentivity in powdery, poorly crystalline goethites, the samples analyzed in this work probably lost >70% of their helium, a challenge in interpreting results obtained from poorly crystalline mineral assemblages. These limitations should be incorporated in the discussion.

It would be beneficial if the authors were to compare the results obtained in poorly crystalline and impure samples with an additional set of results obtained from pure crystalline goethite using the exact same analytical approach. This comparative analysis would strengthen the credibility of the approach implemented in this study and provide a more comprehensive understanding of its applicability. Finally, the complex nature of the Fe-Mn-crusts investigated by Yakubovich et al. demands a more detailed characterization of the paragenetic relationships between the different minerals in the samples to determine exactly which phase(s) was(were) dated.

By addressing these points, the authors can enhance the overall quality and impact of their research, making a significant contribution to the field of geochronology.

Specific comments

Lines 37-39: How do Artic Ocean Fe-Mn-crusts (especially DR7-001) differ from Pacific Ocean Fe-Mn-crusts studied by Basu et al. (2006)?

Lines 64-65: What is the percentage of goethite and Mn-oxides in these samples? The indicated values of > 25%, 5-25%, <5% are not sufficient information. The authors should be able to quantify the percentages of the main mineral phases in the samples analyzed by XRD. The reader would benefit from a figure illustrating the X-ray diffraction patterns obtained for each type of material.

Line 115-118: What was the sampling strategy adopted by the researchers? The dark-brown clasts were selected from sample DR7-002 and the yellow-brown vein grains come from sample DR7-001? If so, what is the relationship between the two types of grains analyzed?

Line 116: If (U-Th)/He results were obtained only for sample DR7-001, is this sample also a breccia?

Line 127: Please, tabulate ⁴He amounts measured for calibration standard.

Lines 130-131: ⁴He amounts are retrieved from multiple grains (as shown in Figure 4), not an individual goethite grain.

Lines 182-183: Would it be possible that heating of the grain during sealing of the quartz ampoule caused some U-volatilization as well as He loss?

Lines 195-211: Yakubovich et al. claim that U-loss during sample ultrasonication results in ages that are overly dispersed. In most environments, sorption of U and Th ions/complexes will be less favorable compared to other abundant dissolved metal ions/complexes, suggesting only a minor contribution to the U-Th-He system. U and Th incorporated in the goethite structure during crystal growth would dominate the production of He in a goethite grain. However, the samples used in this study are not pure goethites. Therefore, U and Th can be adsorbed to surfaces exhibiting very distinct characteristics, presenting different adsorption-desorption reaction rates. Additionally, microporosity increases both the surface area of minerals available for sorption reactions and the number of pores to which ions migrate, the sites where they are trapped, and the ease of desorption during ultrasonication. These phenomena may be particularly important in the multi-mineral grains analyzed by Yakubovich et al., but not as significant for well-crystallized, dense goethite masses commonly selected for (U-Th)/He dating. Cleaning mm- to µm-sized goethite grains in an ultrasonic bath at room temperature for 10-15 minutes aims to facilitate the removal of loosely attached fine particles of goethite and other phases (e.g., adsorbed clays or dust) from the surface of the grains to minimize interference on He, U, and Th measurements of the target phase.

Line 214: The calculated MSWDs for the analyses of dark and vein “goethites” show otherwise.

Lines 215-217: Laser heating of encapsulated goethite grain at 950 °C for 6 min ensures complete ⁴He extraction and no U-volatilization from samples suitable for (U-Th)/He geochronology. This is routinely done in laboratories around the world. Now, the quartz ampoule method potentially offers the opportunity to perform helium diffusion studies if it can help to avoid phase breakdown during step-heat experiments.

Lines 238-293: What are the complications in using the quartz ampoule method to derive He diffusion parameters?

Lines 239-240: I do not understand this statement.

No correlation exists between He release pattern and (U-Th)/He ages, reflecting an insignificant thermal loss of ⁴He.

Line 266: Goethite is the most thermodynamically stable iron-oxyhydroxide in the near surface environment. It partially dissolves during interaction with acidic solutions, but its thermodynamic stability does not change. For instance, goethites as old as 70 Ma have been found near the surface of massive sulfide deposits.

Technical corrections

Title: The word goethite in the title is misleading. The samples analyzed by Yakubovich et al. are not pure goethite. In addition, this method can be applied to other phases as well. I suggest changing the title to “A new analytical approach on (U-Th)/He dating by sample encapsulation in quartz ampoules under vacuum: application to Fe-Mn-crusts, Amerasian Basin, Arctic Ocean”.

Figure 1: Enlarge sample photos, indicate on the photo the different types of material (dendrites, botryoidal, glassy, cellular-like, etc.), and show the areas sampled for geochronology and X-ray diffraction analysis.

Figure 2: (B) replace massive with film or cement.

Figure 5: HD⁺vs Temperature and He vs Temperature: What do the patterns of HD⁺ and He release with temperature teach us about sample behavior under vacuum?

Figure 5: These helium release patterns do not say anything about goethite behavior because the grains are not pure goethite.

Line 14: replace Fe-hydroxides with Fe-(oxyhydr)oxides.

Line 19-20: Here are some potentially contributors for overdispersion of (U-Th)/He ages: multiple generations of Fe-(oxyhydr)oxides, He diffusion from poorly crystalline material, alpha recoil, mineral inclusions, and the presence of fine particles attached to the grain surface. The sentence needs to be rewritten.

Line 30-31: Suggestion: “Goethite is one of the most common Fe-(oxy)hydroxide minerals formed during the hydrolyzation of rocks, making it a desirable mineral for dating various surface and subsurface geological processes.”

Lines 34, 36: replace successful with successfully.

Line 45: Shuster and Farley (2005) discuss diffusion experiment results obtained for quartz. Is there an equivalent citation pertinent to goethite?

Line 75: What do replacements mean?

Line 131: camera == chamber?

Line 131: How much HD⁺ was detected during the step-heat experiment? Does the amount of HD⁺ differ between the types of samples?

Lines 215, 218, 220: Qu == quartz?