blue Review to “ Technical Note : Uncovering the influence of methodological variations on the extractability of iron bound organic carbon

Fisher et al. investigated how modifications of the frequently used citrate-dithionite buffer (CDB) extraction for iron-bound organic carbon influence the respective results. The CDB method is widely applied in soil and marine sciences to extract iron (Fe) and co-precipitated or adsorbed organic matter. Despite its common application, the method has some drawbacks that are, according to the authors, often neglected or at least not properly discussed. In this sense, this study reminds me of the recent publication by Hepburn et al. in Chem. Geol.: “The use of operationally-defined sequential Fe extraction methods for mineralogical applications: A cautionary tale from Mössbauer spectroscopy” and the study by Oonk et al. (2017, Chem. Geol.): “Fraction-specific controls on the trace element distribution in iron formations: Implications for trace metal stable isotope proxies“. As in these previous publications, Fisher et al. try to tackle the problem that wet chemical extractions lead to operationally defined fractions that are not entirely specific to distinct minerals. The authors set up experiments where they varied the strength of the chemical extract as well as the composition of the sample that‘s to be leached. They also tested whether a longer duration of the CDB treatment leads to higher Fe and Fe-OC yields. Studies like this are urgently needed to achieve comparability of datasets even though they unfortunately never result in a crystal clear recipe that is to be preferred for all kind of samples. However, this article will make researchers more aware of the shortcomings of the CDB method so that they are put into a better position to judge in which way they should apply it and discuss their data. The manuscript is for the most part well written and easy to understand. The figures and tables are adequate and the discussion is supported by the presented data.

As mentioned above, the usual CDB extraction includes crystalline phases like goethite and hematite that might not be so relevant for OC. I am missing a statement concerning how the (maybe in your case unintended?) leaching of more crystalline phases potentially skews the FeR : OC relationship. The typical amount of highly reactive Fe (amorphous phases) in shelf sediments is, I would say, less than 1 wt% (so by far lower than what you were testing for). So I would therefore be a bit hesitant to transfer the results of your experimental data to real marine sediments and it's good that you included tests with Antarctic sediments in this study.
Yes, CBD can leach goethite and hematite in addition to, e.g., ferrihydrite, and we refer to this in the now modified section in the introduction by including values for hematite leaching from the Thompson et al., 2019 study (Line 52). Hematite can play a role in OC binding, so in the same section we refer to the study by Adhikari and Yang on OC release from hematite-humic acid complexes (Line 54). The reviewer is correct that leaching of Fe phases not associated with OC can skew the OC-FeR relationship and this is shown in the Barber et al., (2017) study. We have not previously discussed this in this manuscript as we only include ferrihydrite in our experimental system (in part to get around problems such as these), so leaching of other phases is not directly relevant for this study.
Lines 324-328: "We postulate that freeze drying-induced aggregation of sediment particles could result in reduced Fe extractability compared to non-dried samples since grain size is a known key factor in limiting determination of bioavailable Fe (Raiswell et al., 1994)." I'd actually argue the other way around. I am wondering about the potential differences between grinded and non-grinded natural samples. You typically freeze-dry samples to be able to grind them and make them more homogenous. I would assume that the freeze-drying itself might result in a transfer of Fe from a more reactive into a less reactive pool. But at the same time I would guess you reduce effects of grain size differences or clogging/shielding of grains (coatings) by grinding the samples. Ok. I see that you mention this in the following sentence. (Add a space before "The influence. . .") As grinding is what's typically done, I'm not convinced that the aggregation plays the dominant role. I'd rather think that the amorphous Fe compounds aren't stable during the processing (freeze drying). Would be worth checking whether there is a transformation of ferrihydrite during and after drying...
We have previously shown that the complexes we produce are definitely still 2-line ferrihydrite by XRD of the freeze-dried precipitate in Fisher et al., (2020) so transformation was ruled out as a reason for the difference between Fe recovery from dried and non-dried precipitates. We provide a copy of this XRD in this review file (response to Peter Kraal) for reference. We have added a sentence to confirm transformation of the Fe phase did not occur during drying.
Lines 311-313: "An alternate hypothesis, that mineralogical transformation of ferrihdyrite during freeze-drying may lead to reduced Fe recovery, was ruled out by X-ray diffraction (XRD) characterisation of a representative freeze-dried sample, which confirmed the identity to still be 2-line ferrihydrite (Fisher et al., 2020)." We acknowledge later in the manuscript that grinding is typically performed but to an undefined grain size ("finely ground" or similar), which can introduce error.
Lines 308-311 " McKeague and Day (1966) similarly report that finer grinding of sediment resulted in an increased extraction of Fe. These findings indicate that particle size is a critical parameter in determining the amount of Fe extracted, however, current methods do not define the particle size of "finely ground" sediments." Line 335-336: "The alternate tested method of using wet samples has largely been avoided, with only a few studies (e.g. van Bodegom et al., 2003;Chen et al., 2020) reporting the use of a wet slurry sample in soils and none for sediments." Suggestion: The alternative method of using wet samples has largely been avoided, with only a few exceptions in soil studies (e.g. van Bodegom et al., 2003;Chen et al., 2020). Your statement is not true when you don't limit your view to the Fe-OC extraction by dithionite but also consider the many studies focussing on the Fe or P. The Poulton and Canfield (2005) method that includes a similar dithionite step is often applied to wet sediments. Check papers by Natascha Riedinger, Laura Wehrmann and Katja Laufer (2019, Reactivity of Iron Minerals in the Seabed Toward Microbial Reduction -A Comparison of Different Extraction Techniques). The same is true for Fe-P extractions with CDB (Kraal, e.g. 2017 GCA paper). One reason for people sticking to the freezedrying and grinding is that with lots of samples, that's the only option. I'm thinking of IODP material (usually pretty hard mud rock) or black shales.
We thank Dr Henkel for highlighting these papers, we agree it would be suitable to include these as examples of wet sediment treatments. However, not all the suggested papers actually use wet sediment (e.g., the Kraal 2017 GCA paper explicitly uses freeze-dried "ground sediment of 50 and 100 mg for sequential extraction of Fe"). While Laufer et al., (2019), Wehrmann et al., (2014) and Riedinger et al., (2017) do use wet sediment, these studies all freeze sediment at -20°C which we later go on to discuss may introduce freezethaw aggregation.
We have now cited these studies alongside our use of Arctic freeze-thawed samples.
Lines 316-320: "Wet thawed samples have been used more widely in the sequential extraction of Fe (e.g. Wehrmann et al., 2014;Riedinger et al., 2017;Laufer et al., 2020) and the Arctic marine sediment sample used in our analysis was similarly thawed following freezing on collection. We find that our thawed sample shows no difference in its recovery for Fe compared to the dried variant of this sample." Lines 370-373: "As we observed incomplete Fe extraction (Fig. 1) for all our samples, a range of CBD extraction times were trialled to understand whether increasing the length of a reaction would increase Fe liberated, as seen for other chemical Fe extractions; oxalate, for example, is known to continue to extract Fe beyond a standard 1 hour treatment (McKeague and Day, 1966)." Okay, but it does not make too much sense to compare the CDB method to the oxalate method, because the oxalate extraction works differently. The extraction is actually catalyzed by dissolved Fe2+. So the longer the extraction continues, the more Fe2+ is in solution and the stronger gets the extraction (well described in Oonk et al., 2017, Chem Geol. and references therein).
We have removed the reference to oxalate.
Line 415-417: "We suggest that if future studies were to increase Na dithionite addition in the CBD method this should be followed by a similar increase in trisodium citrate to ensure the entire reduced Fe pool is complexed, preventing precipitation of Fe before quantification." I ran some tests with citrate myself with Fe contents that are comparable to natural occurrences and found that it's usually not limiting. The citrate concentration can in fact be reduced compared to original protocols (I tested the Poulton and Canfield method) as long as you work under strictly anoxic conditions (Henkel et al. 2016, Chem. Geol.). Might be a good alternative.
We have added this as an alternative earlier in the discussion where we first mention increasing citrate. Line 277 now includes: "However, Henkel et al. (2016) found that a reduced concentration of citrate is sufficient to fully complex the reduced Fe pool when the extraction is performed under anoxic conditions, which may remove the need to further increase the addition of citrate as an organic reagent." More specific technical remarks (sentence structure, typos etc.): Line 16: I suggest to use "synthesized sample" instead of "sediment".
Changed to synthetic sample Line 36: Delete "important" before "for water retention" as it is an unnecessary repetition.

Comma removed
Line 118-119: "To achieve this, we mixed the precipitate with a marine sediment 'carrier' material as described by Fisher et al. (2020), using the same original carrier sample and similarly treated to liberate OC and inorganic carbon." Weird sentence structure. In order to make it easier for the reader I suggest to include one or two sentences to what the carrier material is. I guess the original FeR contents are known? It's fine to refer to the previous publication, but the reader shouldn't be "forced" to look it up.
Replaced with Line 128 onwards: "To spike marine sediment with the experimentally produced OC-FeR coprecipitates, different amounts of OC-FeR were added to aliquots of a marine sediment sample from the Barents Sea (sediment core depth 33.5 cm;station B6, E40;cruise JR16006;see Hopkins (2017)). This sediment was freeze-dried, ashed (650 °C, 12 hrs) to remove OC, and fumigated with HCl vapour to remove inorganic carbon. The resulting material was predominantly siliciclastic in nature with a Fe content of 16.33 mg/g." Line 126: Replace "A" by "The".

Corrected
Lines 164-166: "Initial concentrations of Fe in synthetic samples were obtained by digesting ∼2 mg of dried sample in 1 mL 12N HCl at room temperature followed by a 10-fold dilution with 1% HCl solution. Further dilutions were made as necessary, dependent on Fe content, using MilliQ water to produce a subsample within the detectable window (1-10 ppm Fe)." I'm not quite sure about the fraction that is intended to be dissolved here. Bulk Fe? I guess it's okay in case that the synthetic sample does not contain Fe-bearing silicates. As mentioned above it would be good to add what was used as "carrier" for the Fe oxide-OC spikes.
Details on spike content has been added with the previous comment. The digest here is to determine the initial Fe content (i.e., the ferrihydrite we add), 12N HCl is sufficient to dissolve the spiked ferrihydrite and any reactive Fe in the synthetic sample (i.e. anything that can be later extracted by CBD).

Corrected
Section 2.6 about ICP-OES analyses: Generally (for future), I would recommend using an internal standard for correction of different ionic strengths.
Atomic absorption spectroscopy (AAS) was used for the analysis of Fe rather than ICP-OES in this study.
Line 171-173: "Extraction of Fe was calculated by subtracting the amount of Fe lost in the control experiment from Fe lost following extraction, then subtracting this from the initial Fe of each sample." The formulation of this sentence seems more complicated than necessary and I don't fully get it. How about: "The recoveries of the extractions were determined as extracted Fe compared to the initial Fe content." (I suppose you made sure that the carrier sediment that you spiked did not contain any Fe?) Line 179 Changed to:" The recovery of Fe following sample extractions was calculated by subtracting the control corrected loss of Fe from the initial Fe content of the sample." There was a small amount of Fe in the carrier sediment, as now detailed in the spiking section of the methods. However, this only makes a minor contribution to the overall Fe pool once spiked and sequential extraction of multiple cores from the same region show that the iron mineralogy at this depth is dominated by stable crystalline phases, which would not be particularly susceptible to a short neutral pH CBD extraction. We have no indication that the baseline Fe content affected any of the results, even if this Fe content was to have an affect it would be the same effect on all aliquots and iterations involving the sediment carrier.
Line 188: ". . .requires a 0.25 g addition relative to 0.25 g of dried sediment sample" Recommend to use "per" instead of "relative to".
Corrected : "All samples show incomplete reduction of Fe regardless of Na dithionite addition, with those samples containing the least Fe proving extractable for the greatest proportion of Fe." Unnecessarily complicated formulation. How about: "highest recovery of Fe in samples with low OC-Fe contents".

Changed
Line 199-201: "From this, we can deduce the maximal %Fe in sediment extractable by 0.25 g Na dithionite lies between a 20 and 30% OC-FeR mix, equivalent to 7-10 wt% Fe content in the sediment." I would slightly reformulate the last part of the sentence to not imply that this is total Fe you're talking about. And wouldn't it make sense (for practical reasons) to translate your "20 and 30% OC-FeR mix" into an absolute amount of Fe (e.g. in mmol or mg) that can be liberated? Line 208 Added: "Therefore, assuming a 0.25g sample size, the absolute amount of Fe extracted would be between 17.5 and 25 mg. " Also changed wt% Fe to wt% FeR to clarify this is reactive not total Fe.
Lines 203-205: This is about LECO data, right? I wonder whether you could avoid confusion by just calling it "extracted OC" or OCFeR. Calling this fraction OC-FeR is a bit confusing as I would intuitively translate it as "OC-bound reactive Fe". But you mean "reactive Fe-bound OC". As for the Fe I assume that your carrier did not contain any further OC? This is about measuring OC, the LECO data, yes. OC-FeR is the fraction of organic carbon bound to reactive iron, not reactive iron bound to OC so the reviewer translation is correct. And yes, the carrier is OC free, this has been clarified in response to a previous comment in the methods.
Line 255-256: "For the four synthetic samples we subjected to dithionite reduction, these differed in composition (7-24 wt% Fe, 20-50% initial OC-FeR content)." Didn't you also have a batch with 100% OC-FeR??? (See Table 2.) The greater than 50% OC-FeR values are only used for the wet vs dry comparison as that is a more mechanistic relationship. In this section (4.1), we are discussing the concentration of Na dithionite used to extract OC-FeR from samples so the concentrations are closer to what may be found in a natural sediment.
Lines 256-257: "The concentration of Fe in these samples results in an effective dithionite to (wt) Fe reduction reaction ratio of 1:0.07-0.24." Replace "concentration of Fe" by "Fe contents", "results" by "resulted" and replace "effective dithionite to (wt) Fe reduction reaction ratio" by "effective dithionite to Fe mass ratio".

Changed
Line 261-262: "This has the potential to drive wt% Fe higher in small samples of sediment such as those treated by the method (0.25 g)." Recommend to replace "treated by the method (0.25 g)" by "typically used for the CDB extraction".

Changed
Lines 266-269: "Maximal extraction here is defined as the point from which further addition of Na dithionite does not increase the extraction of Fe beyond the amount of Fe extracted under the previous dithionite addition mass ± error. For example, the 20% OC-FeR sample subject to 0.25 g dithionite is removable for 88.79% ± 3.55 of FeTotal while 0.375 g addition extracts 90.94% ± 3.64; . . ." What is meant by "is removable for"??? Unnecessarily complicated formulation. Use "yields" or "liberates". Delete "beyond the amount of Fe extracted under the previous dithionite addition mass ± error" and add a "further" before "increase". How this is meant is getting clear through your example.

Corrected
Line 276-278: "This finding demonstrates that the OC-FeR composition would not be correctly determined following the method of Lalonde et al. (2012) for these OC-FeR rich sediments, and the overall extent of OC-FeR in the marine sediment pool would be underestimated. You can delete the "pool". It kind of implies that you're talking of a specific fraction of the marine sediment, but here you mean the sediment itself (bulk). Why would you limit this to marine sediments? Couldn't you say this is a general outcome of your study no matter which sediment (fluvial or marine or soil) is used? (Now, again, it would be nice to know the composition of your carrier material.) I would write "amount" instead of "extent".
Changed from marine sediment to sample and extent to amount.
Line 290-291: "If the increased strength dithionite treatment increases dissolved Fe beyond the complexing capacity of citrate, then excess Fe likely precipitates out of solution before measurement." This can be avoided when performing the extraction under anoxic conditions (e.g. Henkel et al. 2016).
See previous comment, Line 277 Added: "However, Henkel et al. (2016) found that a reduced concentration of citrate is sufficient to fully complex the reduced Fe pool when the extraction is performed under anoxic conditions, which may remove the need to further increase the addition of citrate as an organic reagent." Line 293-295: "Measurement of OC-FeR extracted for the concentration of Na dithionite at which maximum Fe is extracted showed incomplete OC-FeR loss (Fig. 1)." I would replace "loss" by "liberation" or "recovery".
Changed to recovery Line 294-295: "The similarity of OC-FeR and raw Fe extraction values indicates that OC and Fe are reductively released from the sediment in comparable proportions, as is expected due to the low molar OC:FeR ratio of the coprecipitate (∼0.7:1) ." What is meant by "raw" values? Raw data is typically used in another sense.

Deleted raw
Lines 297-298: ". . .could benefit from using increased strength Na dithionite compared to the 0.1 M treatment currently used." Or shorter: "compared to the conventional 0.1 M treatment".

Changed
Line 346-347: However, the use of wet sediments is likely to be inappropriate for some analyses or sample sites. Yes! You should add one or two sentences to that. I believe it's for most cases not as if people using these methods are not aware of its shortcomings.
Line 337 Added: "While application of the CBD method to slurried samples could increase the extracted proportion of Fe associated with OC, such an approach may not always be practical; either due to practical considerations, such as the difficulty in transporting heavy wet sediments, or when there is a need to preserve the sediment profile, for example, protecting anoxic sediments from oxic biological transformations." Line 354: You can delete the "method" after "storage". "Any storage" is enough.

Deleted
Lines 373-375: "Additionally, as previously mentioned, some iterations of the CBD method have been repeated multiple times in succession to extract the full FeCBD pool, but it is unclear whether time or reagent concentration limit full extraction of this pool on the first treatment." By you or others? It's not getting clear here. "Iteration repeated multiple times in succession..." Here you say the same thing twice (or actually three times).

Line 352
Changed: "In some applications of the CBD method, the extraction stage is repeated multiple times for the same sample in order to fully extract FeCBD (e.g. Aguilera and Jackson, 1953;Mehra and Jackson, 1958), but it is still very difficult to attribute full CBD extraction to this multiple extraction protocol because the parameter that prevents full extraction of CBD extractable Fe in the first place is unknown." Line 379: ". . . concluding that an increase in chemical exposure time has no difference on Fe extractability." Replace "has no difference on" by "has not enhanced" or "has no effect on".

Changed
Line 381: "We would perhaps not expect any benefit from increasing the length of CBD treatment as dithionite, . . ." You don't seem to be very convinced by your data. Replace "would perhaps not" by "do not".

Changed
Line 382-383: ". . . with a rapid second order rate constant (K2) of 3.0 (g-molecule/L)-1 min-1 at 79.4 •C, . . ." The unit is written in an unnecessarily complicated way. I guess it should be L/(mol*min)? Please check! This was originally taken from the, rather old, citation so can be updated. Gram molecules (g-molecule) are equivalent to moles so this has been simplified to 3.0 mol L -1 min -1 .
Lines 421-422: "Freeze drying induced aggregation appears to reduce Fe liberation in synthetic coprecipitates that were freeze dried relative to slurried, however, we were unable to replicate this increased extraction for natural samples." Suggestion: Freeze-drying induced aggregation appears to reduce Fe liberation from synthetic coprecipitates. However, we were unable to confirm this reduced Fe extraction for a set of natural samples.

Changed
Lines 422-424: "While we speculate this may be due to the use of freeze thawed samples, which can introduce aggregation in itself, it is hard to see a practical implementation of this adjustment for marine sediments due to the difficulty in transport of pristine samples." Replace "which" by "where freeze thawing", otherwise your reference isn't fully correct. (You'd refer to the samples and not the process of thawing.) And I believe you can delete the "in" before "itself".

Changed
Line 427: Period missing after "extraction". Changed Figure 1: I have difficulties understanding your black and blue symbols. Shouldn't the percentage of extracted Fe be equivalent to the extracted OC-Fe? Or is the data behind the blue symbols the LECO-data? Do you really need the separate axis with the different scale??? It's just (at first glance) confusing that e.g. the blue diamond is so much further up the fitted curve. And the offset in "dithionite added" between blue and black symbols (equivalent to maybe 0.1 g) is odd, too.
The percentage of Fe would only be equivalent to OC-Fe if there was a 1:1 OC/Fe ratio. The blue symbols are the LECO data. Blue symbols were originally offset on dithionite concentration to make the symbols easier to distinguish but these can be put back to maintain accuracy. They do need separate axes as they are different measurements (one Fe, one C) regardless of the different scales. (Note these are red in the revised figure) Figure 2: You don't need 3 different patterns if you distinguish between the different OC contents by different colors (gray scales). So reduce the complexity of this graph by just using 3 colors for the three differing OC batches and filled vs. hatched bars for dry and wet. I would also (for clarity) change the figure a bit so that it doesn't appear as if the OC-Fe to total sediment ratio was 5% for the lowermost wet batch 3 COOH mix and close to 30% for the lowermost dry batch 1 COOH mix. You know what I mean? Those extractions all belong to the 20% test, right? Figure  Colours have been changed. The y axis represents the original mix of the sample (precipitate to sediment), they are not all the 20% extraction but the 20,40,60,80 and 100% extractions. So the y axis is what was added and the x is what was extracted, added groupings to the y to make it clearer that this is categorical and not continuous data.  Table 1: Use format "left-aligned" in the first column. Table 2: I find the expression "%OC-Fe:sediment" a bit confusing. I guess you mean % of OC-Fe coprecipitate to total sample". It's inconsistent because when you write "Sediment (mg)" you mean the carrier only. Response to review by Tom Jilbert.

Changed
Original comments in black, responses are in blue.
Fisher et al. present results of experiments into the extractability of OC-FeR (that is to say, sedimentary organic carbon bound to reactive Fe) during treatment with citrate bicarbonatedithionite solution. CBD extractions are a commonly applied method, either stand-alone or as part of sequential extractions, for investigating elements associated with reducible phases in marine sediments. As the authors state, there is much heterogeneity in the details of applied CBD extraction protocols, even within the narrower context of studies into OC-FeR. This has led to difficulty in comparing results and the possibility that the currently used protocols may be sub-optimal for their stated goals. Therefore there is clearly a need for studies like this one, to eventually improve/harmonize the approaches used in the community.
Overall I had the feeling that the study delivers some interesting results although some of the interpretations are left only lightly justified. This leads to the idea that a more developed set of experiments could have yielded a more useful step forward. For example, the conclusion that complexation by citrate may be limiting the recovery of Fe in the highest-OC-FeR experiment deserves to be tested through a concentration series similar to the dithionite series the authors report. This is especially the case considering the comments of the first reviewer (Henkel) questioning whether citrate limitation is a feasible explanation for the observations in Fig. 1. I do not demand that the authors produce such additional data before publication but it is clear that their conclusion would be more robust if it was available, and therefore the overall impact of the study would be greater. A similar criticism could be leveled at the interpretations of the experiment comparing freeze-dried and wet samples, although I would say this is a complex topic that warrants a separate study.
We thank Dr Jilbert for their review and for seeing the value in our study. In the first review of this manuscript, Dr Henkel did raise the issue of whether citrate limitation could explain the reduced loss of Fe at high wt% Fe contents and in turn directed us to Henkel et al., (2016) which suggested the use of anoxic extractions instead. We happily included this point in our revisions to the manuscript; however, both anoxic extraction and increased citrate address the same issue of Fe precipitation in different ways so there appears to be some consensus on the underlying issue here. It is difficult to think of any other reason why Fe losses would be minimised, as there is nowhere else for the Fe to go besides staying in solution or precipitating.
We did not conduct experiments with increased citrate as the method we test is used for OC extraction and we were therefore very cautious about adding C containing compounds to the reaction. Indeed, we did not have a solution to this problem until Patzner et al., (2020) very recently incorporated a correction for the background DOC levels in the reaction.
We agree there is more work to be done on freeze-dried vs wet samples. Particularly around the effects of freeze-thawing, we unfortunately were unable to conduct these types of analyses as the sediment samples we had access to had all been previously frozen (as is common). However, our study provides a new perspective on this issue; for example, we are not aware of any studies which have quantified the effect of variable water mass on introducing error in determining the dry weight equivalent of sediments, or any studies which have aimed to compare the same synthetic samples under different preparation methods.
We have modified our interpretation of the freeze-drying experiment on the natural samples following the comments raised by Peter Kraal.
Another important point is that it seems that some of the content here may be an overflow from the authors' recent Chem. Geol. paper (cited Fisher et al. 2020), which is not a criticism as such but in some cases I had the feeling the reader is being referred there to explain what is going on in the experiments presented here, which should be avoided as this ms. must also be a stand-alone study, even if it is a Technical Note. I am specifically referring to the interpretation of the results of the experiment in which the degree of carboxylation of OM in synthetic OC-FeR is varied (Fig. 2). The discussion of the mechanisms here (Paragraph from Line 358) is too thin and the reader cannot understand why the degree of carboxylation makes such a difference to Fe extractability without accessing the other paper.
We have expanded our manuscript to provide an explanation for the carboxyl relationship identified in the previous study. We also expanded the methods section in response to a similar comment by a previous reviewer.
Lines 327-333: "Even though such samples consisting exclusively of ferrihydrite-associated OC are unlikely to occur in nature, on the whole our results confirm previously identified relationships between Fe extraction and carboxyl content, where Fe extraction using the CDB method is more efficient for OC-FeR that contains more carboxyl rich OC (Fisher et al., 2020). This is attributed to the greater amorphicity of ferrihydrite coprecipitated with carboxyl rich OC, i.e. the resultant mineral phase is less crystalline than ferrihydrite coprecipitated with less carboxyl rich OC, and is therefore easier to reductively dissolve" Unfortunately I read Susann's review only after making my own comments on the original ms., then later noticed that she has done a very comprehensive job and found several of the same issues that I wished to highlight. It is good to see that the authors have responded thoroughly to Susann's comments and this will undoubtedly improve the next version. Therefore my list of additional comments is comparatively short.
We are grateful for this response to the changes we have already made to the manuscript and agree it has been significantly improved by the comments we have received.
General (in addition to the above; all Line numbers refer to the original ms): -The Introduction can be better worded and arranged: First I suggest to move the para. starting Line 59 to directly above the short para. starting Line 81. This way you first describe the problems with the existing methodologies, then set out how you intend to solve them.
Next, check a few key sentences: e.g. Line 26 "Understanding in which environments organic carbon (OC) persists": please clarify that you are referring to preservation of OC in sediments; Line 54 "fully reduce all solid reactive Fe phases and associated carbon"... I could not find this phrase in Lalonde et al. 2012, although it is presented here as a quotation. Please check.
We have moved this paragraph to the suggested location. The quote from line 26 is no longer part of the manuscript but the focus of the start of the introduction has been changed to be more specific to marine sediments.
The quote from Lalonde et al.
(2012) appears on the first page of the PDF of that manuscript. On checking this, the original paper doesn't include the word "fully", so this has been removed from our quote but this does not change the meaning of the quote as the original article uses "all".

Original (Lalonde): (The CBD method) "dissolves from the sediment matrix all solid reactive iron phases and the organic carbon associated with these phases (OC-Fe)"
Our amended Line 61: These findings contrast the previous understanding of the CBD method performed in an experimental context which states that this extraction will reduce "all solid reactive iron phases and the organic carbon associated with these phases" (Lalonde et al., 2012) -Throughout: the terminology in this field is easily misunderstood. E.g. the first reviewer thought for the whole time that % OC-FeR refers to % of total sediment, when in fact it refers to % of total OC. I also had major difficulties to get this upon first reading. So I suggest to clarify terminology early on, and modify figures and captions to make this easier to follow. E.g. I note that Barber et al. 2017 use more descriptive terminology in tables and figures e.g. "OC bound to Fe (% of total OC)" in their Table 1 and "Fraction of total sediment Fe" in their Fig. 4. Also check that CDB/CBD is used consistently. Both current appear.
We agree that the terminology can be misunderstood in these type of studies. A key difference between our study and that of Barber et al., (2017) is in the total OC content of samples. In our study, the natural sediment matrix has been made OC free, and we have added this aspect to the methods to emphasise how this sediment was prepared in response to the first review. The only OC source in our spiked sediments comes from the coprecipitation of organic acids with ferrihydrite (OC-FeR), while natural sediments contain various forms of unbound OC. Any addition of OC-FeR would represent 100% of the total sample OC pool regardless of how much was added, so this is not a useful metric. When we refer to a sample containing e.g. 20% OC-FeR that is not to say 20% of the OC present is Fe bound, but that 20 wt% of that sediment is the OC-FeR coprecipitate complex. Following this review and the previous, we have decided to refer to quantities in wt% as much as possible.
We have modified the definition given to OC-FeR on line 39 to "OC bound to reactive iron".
Additionally we have removed the use of %OC-FeR following the previous review which asked us to consider the use of content vs concentration. Therefore, the amount of OC-FeR added is now referred to as the OC-FeR content in terms of wt%. This has also been changed in the text, e.g. from "20% OC-FeR " to "the sample containing 20 wt% OC-FeR".
Where reference is made to the total OC pool (when referencing other studies) this is explicitly stated; e.g. line 248 has been changed to "OC-FeR has been observed at contents exceeding 40% (of total OC) in terrestrial environments".
In table 2, which describes the content of spiked samples, we have added "wt%" to describe the OC-Fe content of the sample. Table 2 also includes the actual composition of the sediment in mg to make it as clear as possible.
The caption to figure 3 has been updated to be more descriptive so they can be understood without reference to table 2. The CBD/CDB interchangeability has been picked up on and corrected.
-Description of Fig. 1 results. The phrase "maximal extraction" is used repeatedly when describing the results, but it is only explained in the Discussion (Line 266). The best place for this description is actually Methods, because you can already state how you intend to use the data to estimate this value. That will make reading the paper a whole lot easier overall.
The definition for maximal extraction has bene moved to the end of section 2.6 in the methods. We have rephrased line 251 to retain the example of maximal extraction and ensure this is not lost.
"Maximal extraction is defined as the point at which further additions of dithionite do not increase Fe liberation. Here, the sample containing 20 wt% OC-FeR liberates 88.79% ± 3.55 of FeTotal with a 0.25g dithionite addition, and increasing this to 0.375 g of dithionite only marginally increases Fe liberation to 90.94% ± 3.64; as these values are within error of each other, we conclude that maximal extraction is achieved with the lower 0.25g dithionite content." Specific -Line 67: maybe qualify with 'partial hydrolysis' or similar. For significant digestion of OC from sediments, either very low pH (and use of specific oxidizing acids) or very high pH are required.
Added "partial" to line 67 and subsequent occurrences.
- -Line 244-246: Does this mean that the natural sediment samples in these experiments were freeze-thawed before the experiments? If so it will be important to state this in Section 2.2.
This refers to the sediment samples which we performed the extractions on rather than the carrier sediment that was spiked with the coprecipitate. Section 2.5 is the corresponding method section for this and includes the line "half of which was freeze thawed and half was freeze dried"  (Poulton and Canfield, 2006)

or in Fe-Mn nodules (Hein et al., 1997)"
Line 293-294: This is a confusing opening sentence to the paragraph. Rephrase to make more concise.
Line 258 Rephrased to: "At increased OC-FeR contents, 0.25 g of Na dithionite seems to be no longer sufficient to achieve maximal extraction" Line 304-305: It is not clear to me how an increase of DOC (citrate) during the extraction would impact on the quantification of OC after the extraction., if this is done on the solid phase. Can the sample not simply be rinsed before the drying and analysis?
A triplicate rinse stage is conducted as citrate is known to be retained in the solid matrix, however, in the Lalonde et al. (2012) study this was not a significant problem (<0.08% of dry weight was citrate). This problem is heightened in synthetic studies such as ours, as the OC/Fe ratio is much lower than for natural samples. Therefore, the ferrihydrite surface has much more space available for sorption of organics such as citrate. Further, citrate has 3 carboxyl groups and we previously showed that 3 COOH containing organics were very difficult to remove from ferrihydrite with the CBD method in Fisher et al., (2020). This would suggest that should citrate become bound, it is unlikely to be removed by the rinse.
To clarify this point we have added a sentence to the section highlighting this is of most concern for synthetic samples. "This is particularly acute for the reduction of synthetically precipitated samples where low OC/Fe ratios leave more of the mineral surface available for sorption compared to Fe phases associated with natural organic matter." Line 326-328: Clauses of the sentence are not well constructed Restructured into two sentences.
Response to review by Anonymous Reviewer 3.
Original comments in black, response in blue.
The manuscript is an extended footnote to an earlier paper Fisher et al. paper published this year on the effect of organic acids on reactivity and solubility of iron oxides. The manuscript addressed three aspects of a modified citrate-dithionite-bicarbonate (CDB) extraction method used for determining Fe content associated with various iron oxide mineral phases. Variables tested were the dithionite concentration, freeze-drying versus wet extraction, and time of extraction.
We thank the reviewer for their review of our manuscript. On the point that they feel this is an "extended footnote", it may be useful for us to clarify the differences between our previous study and this new submission for any readers unfamiliar with the earlier work. The reviewer is correct to say that it follows an earlier paper we published and which we refer to in this technical note; however, the only similarity is in methodology. In our previous study we established a system by which we could create synthetic OC-Fe compounds with known OC and Fe concentrations, allowing us to determine the true recovery of Fe phases by chemical extractions. We used this system in the previous study to probe chemical interactions related to carboxyl content of organic matter. Here, we utilise this method to investigate the efficiency of the extraction methods. This manuscript contains entirely new data from a series of 4 experiments (time, dithionite concentration, sediment preparation methods for synthetic samples, and extraction of Arctic sediments). We believe this is presents original data and is sufficiently different from our previous work to warrant publication, and we felt the technical note format was most suited to the type of study we conducted.
Starting with the title, the goal of the study and the actual study is mismatched. The paper is about extraction of iron oxides, and the entire discussion revolves around the efficiency of the dithionite method toward the extraction iron oxides at pH, not the organic carbon that is extracted.
While we appreciate that the carrier phase of organic carbon (i.e., ferrihydrite) is dissolved in our experiments, the method we are testing was designed to release organic carbon bound to this carrier phase. We argue that the current title and study goal reflect this fact. In other words, we are not trying to quantitatively dissolve iron oxides, but we are testing possible iterations of a published method designed to quantitatively liberate iron-oxide associated organic carbon.
My reading of the Lalonde et al. 2012 article is that they were employing a more gentle (i.e. circumneutral pH) treatment in order to not overestimate the loss of organic carbon due to hydrolysis. That is perfectly reasonable, as they did not want to overestimate OC losses from the main pool due to hydrolysis. Their goal was not to accurately quantify the Fe content, but to dissolve most of the iron oxide fraction and thereby release iron oxide bound organic matter. Here, the authors imply that this approach is not quantitative. It is not clear at whom or at which samples this study is aimed.
We agree with the reviewer here, the method is aimed at extracting (as quantitatively as possible) the Fe-associated OC pool, not Fe itself -which also addresses the issue raised in the reviewer's previous comment. However, for a system where OC is quantified via the reductive release of Fe, all associated Fe must be reduced to fully liberate the OC pool. We show both in this study (Fig.1) and in Fisher et al. (2020) that full reductive release of OC is not achieved under our experiment conditions. The aim of our Technical Note is to provide colleagues who conduct this type of extraction with an awareness of possible factors (if any) that may affect the efficiency of their extractions. Additionally, given the wide variation in applied methodologies we wanted to understand whether this was necessarily a problem for reproducibility, i.e. how robust the method is to certain (but admittedly not all the possible) variations in the protocol.
The authors only considered ferrihydrite. So is this applicable only for modern sediments? What about sediments or rocks containing greater concentrations of goethite or hematite?
The benefit of the method we deploy to investigate OC-Fe interactions is that the geochemical system is simplified as far as possible, as such we only consider ferrihydrite here. We realise that this approach also brings limitations, i.e. it does not consider goethite, hematite, magnetite etc. However, research progresses incrementally, and our study does not claim to give all the answers to all the questions in the OC-Fe realm. But regarding the issue of more crystalline iron oxide phases, we show that the tested method cannot fully extract OC and Fe in our ferrihydrite based system, so more crystalline phases are almost certainly going to be even more resistant to the chemical treatment. Indeed, Adhikari and Yang (2015) have already conducted these type of experiments in a hematitie based system (coprecipitated with humic acids) and show incomplete (<50% of Fe liberated) reduction of similar synthetic compounds. As the reviewer suggests, the experiments we conduct are therefore most applicable to modern sediments where ferrihydrite concentrations are high (e.g., near hydrothermal vents, in acid mine drainage deposits), however, this is also where current research suggests the largest fraction of the OC-Fe pool to reside. Recently it was also shown that goethite and hematite likely become less important in terms of OC association due to the desorption of OM during phase transformation (Jelavić, Mitchell and Sand (2020) Geochem. Persp. Let).
The authors claim that no study has thus far has performed a determination of the reductive capacity of the dithionite method (Lines 74-75). The authors, however, also do not clearly provide the criteria for "reductive capacity". It is only implied in their approach of using varying "weight percentages of OC-Fe" that simulates a titration of sorts. And as described below, there are methodological problems with this approach. It is not clear that this a substantial step beyond the Fisher et al., 2020 paper.
In the Fisher et al., (2020) paper a dithionite addition of 0.25g was used throughout, following Lalonde et al., (2012), so no investigation into the reductive capacity was made or can be determined from that study. The purpose here was to take account of the fact that many different iterations of this method have used different masses of dithionite to extract the OC-Fe pool, and a compilation of these is shown in Table 1. None of these studies, to the best of our knowledge, have quantified how much Fe can be extracted based on the varying amount of dithionite they use, but simply recommend a mass of sample which can obviously vary widely in Fe content, especially given the range of samples this method is applied to. We accept the reviewer's comment that we have not provided a definition for reductive capacity and have amended this in the abstract.
"We provide an assessment of the reducing capacity of Na dithionite in the CBD method (the amount of Fe reduced by a fixed amount of dithionite)" Line 16 The authors pose the question of whether there is a one-size-fits-all solution (line 95), or should the extraction be adjusted to fit the set of samples and exact research question. But they do not really answer this question. For instance, the effects of freeze drying on wet chemical extractions of sediments as extensively discussed in Section 4.2, has long been known (e.g., Rapin et al., 1986, ES&T;and more recently Raiswell et al., 2004, Chem Geol.). This discussion is superfluous. The authors point out the problems faced by every sediment biogeochemist, but offer no new insights of their own, or at least none that have not been already considered by other studies. They propose no solutions to any of these aspects, except to say that methods employed should be rigorously documented. As the authors point out, analyzing freshly collected wet sediments is not practical for most studies. One has to ask if the efforts to improve the dithionite method are even worth the effort, if freeze-drying is out of the question. The topic of study is an important one in sediment biogeochemistry, especially in how do we deal with examining organic carbon concentrations and speciation in complex matrices. But does this paper bring about a consensus on how to proceed? Unfortunately, I have to answer, no it does not, outside of stating that when using wet chemical sediment extraction methods, that geochemists should carefully consider the type of sediment being analyzed, the amounts of reactants in the methods, sample storage and the exact question being investigated.
The reviewer raises an important point here about the overall value of the study, and we are glad that in principle they agree that research in this area is important. We accept that some of the implications made in the introduction about our aim and the overall conclusions do not match up, particularly about whether the method should be improved/replaced by a new method. We wish to be clear that this is not a 'modified protocol' study; our aim here is not to replace that set out by Lalonde et al., 2012.
This has now been clarified by a restructure of the introduction with a new clear focus Line 62-65 "Given the incomplete extraction of OC-FeR in synthetic samples and the apparent sensitivity of the method to changes in OC composition and Fe phase, we investigated whether differences in method parameters can alter the amount of OC-FeR extracted from a given matrix, in order to better understand the robustness of the method and to determine how methodological variation in the CBD extraction can affect the comparability of OC-FeR results". We have also modified the discussion to be less "superfluous" by concentrating around the newly outlined focus. We have added clarity in the introduction that this is not designed to be an all-encompassing study, and yes other studies have in isolation considered some of (not all) these factors for their influence in various extraction protocols but rarely do studies such as ours pull these variable parameters together or make it the focus of a discussion. We hope in creating this synthesis this results in a valuable technical paper for colleagues utilising the CBD method, increasing the awareness of pitfalls in reproducibility, particularly given that OC-Fe studies continue to evolve their methodologies.
The new scope of this study is formed on the basis that ,both historically and continually, a vast range of modifications have been made to the CBD method for OC-Fe and these are largely uncalibrated and difficult to compare to one another. Within this scope, rather than suggesting a replacement method, the discussion is now framed around understanding whether the variability in these pre-existing methods is a barrier to comparability and reproducibility between data sets. In other words, does it matter if different extraction times/ dithionite strengths/ sample preparation methods are used? We fully appreciate that we do not present all possible variations of all parameters, but clearly a few of the most critical ones.
We believe by narrowing the focus of the study we can retain its technical importance but avoid the apparent disappointment arising from comments regarding lack of completeness from reviewers who have, fairly, expected that this study is designed to replace or modify an existing method.
In addition, as pointed out in review 2 by Tom Jilbert, some of the topics we pick up on could be papers in their own right, e.g. freeze drying vs wet sediments for OC extraction, and we hope that this can lay the groundwork for future studies.

Methodology
The high iron oxide contents used in these experiments are problematic. First of all, it is not entirely clear what exactly is being measured (see comments on the term "OCFe"). Let's assume that it is %Fe.
Correct, %Fe is measured. Other reviewers raised a similar point and in response, we have added to section 2.6 of the methods to clarify that Fe was measured and how this is calculated.
Line 179-181 "The recovery of Fe following the extractions was calculated by subtracting the control-corrected loss of Fe from the initial sample Fe content. Maximal extraction of Fe is defined as the point from which further addition of Na dithionite does not further increase the extraction of Fe. " For example, the 20% OC-Fe sample contains 0.2x 0.25 g artificial sediment, or 0.05 g Fe.
In the 15 mL of reaction solution, this gives a. 0.06 mole/L Fe. The dithionite solution of 0.25 g Na2S2O4 in 15 mL yields 0.093moles S2O42-anion per L. Assuming that upon dispropotionation of the dithionite inwater yields two reducing equivalents, which is probably overestimated due to sidereactions with oxygen and other S decomposition products, we would have <0.18 mol/Ldithionite reducing capacity. Dithionite is barely in excess of the reactive iron fraction,which is a poor starting point for a quantitative extraction. It certainly becomes worseor untenable at 50% or 75% or 100% OC-Fe contents. If the OC-Fe weight% refers to FeOOH, things improve. But only by a third. It is not surprising that the method failsmto reduce these high Fe oxide containing slurries. These high Fe concentrations are actually not realistic (see also my comment below). As the authors point out in Line 259, wt% Fe contents in most sediments are usually less than 10%.
These details can be found in Table 2, where the terminology used has been simplified in response to previous reviewer comments -we apologise for the confusion that has led the reviewer to conduct these calculations however, they are incorrect. The 20% OC-Fe sample contains 0.2 g of artificial sediment and 0.05 g of the OC-FeR coprecipitate. That is to say 0.05 g of the resultant complex from OC coprecipitation with ferrihydrite, not 0.05 g of pure Fe. We give the Fe value for this sediment sample as ~7 wt%, equivalent to 0.0175 g in a total mass of 0.25 g, which is less than a third of the amount assumed/calculated by the reviewer and very much in line with common Fe contents in natural sediments. Hopefully this removes some of the concern the reviewer had regarding samples at the 20% OC-FeR content.
The reviewer is correct to say that the Fe contents become very high in the 50%+ OC-Fe samples (roughly equivalent to 17wt%+ of Fe), and it is for this reason we did not use these high Fe samples in the dithionite concentration experiments or time experiments. The 50-100% OC-FeR samples are only used in the experiment to determine whether sample preparation methods affect Fe extraction from these samples. These were suitable samples to use for this experiment as this is an intracomparison of dried vs wet samples; the actual Fe extraction values are not of critical importance, but mainly the different between dried vs wet sample aliquots. Further, by performing an experiment on samples containing 100% OC-FeR (i.e. only the coprecipitate) this serves as a control to ensure that interactions between the coprecipitate and the added sediment are not responsible for the differences observed here.
We are grateful that the reviewer has acknowledged that we explicitly give a comparison of the true Fe content of sediments in the paper. In response to a comment from Dr Jilbert we have further reinforced this point by removing a reference to Fe clustering increasing sample Fe content. We have made an effort to be as transparent as possible in which samples relate to environmental content; however, the purpose, and benefit, of experimental studies such as this is to be able to create artificial conditions way beyond what is found in natural samples, providing a better understanding of fundamental chemical mechanisms.
Lastly, we would like to highlight that there are numerous instances where Fe contents of a sediment exceed the value for "normal" marine sediments, we have added detail of this in section 4.1 line 246-249 "While reactive Fe content in bulk natural sediments is usually below 7 wt% Fe (e.g. Canfield, 1989;Raiswell and Canfield, 1998), spatial and temporal variation in Fe fluxes to the seafloor can result in Fe rich sediments e.g. near hydrothermal vents (Poulton and Canfield, 2006) or in Fe-Mn nodules (Hein et al., 1997)." The high iron concentrations used in these experiments exacerbate another problem with the experimental set-up. As far as I can tell, the samples were not shaken. I assume that the precipitates sank to the bottom of the reaction vessel (details on the reaction vial type and geometry are missing). Over time the reaction rates will become diffusion limited without shaking. This also renders the results of the time-course experiment somewhat difficult to interpret.
We apologise for this omission. The samples were in fact vortexed following the addition of dithionite (and likewise for the controls), and this important detail has been added to the methods. Shaking throughout the extraction for this extraction protocol is impossible because the samples are incubated in a heated water bath. However, for the time experiments these were manually shaken every 15 minutes, as each sample was removed. We thank the reviewer for noticing this omission in the methods and have corrected accordingly. Reaction vial type was a centrifuge tube (see line 127).
There appear to be no replicates for each dithionite addition. This makes interpretation of the results, especially in Figure 2, difficult. This is discussed in line 153-157. Essentially, the synthetic OC-FeR complexes are limited by production yields with a 5 L reaction volume producing only 5 g of precipitate. This was already scaled up from a usual 2 L reaction volume, but it is not logistically possible to conduct larger scale reactions, and recover the products, without industrial scale laboratory facilities. Therefore, careful choices were made over how this finite amount of sample should be used.
Firstly, behind every data point there are two samples given that a control extraction is conducted alongside each reduction. Secondly, as stated in the text "Repeats of samples across the content gradient are in lieu of direct replicates for each unique sample condition." Performing direct replicates from multiple precipitations is problematic as the actual amount of ferrihdyrite precipitated, and the amount of OC coprecipitated, varies between precipitations, i.e. you get a slightly different product each time you precipitate. So while the trends within each series of dilutions from the same precipitate are the same, averaging multiple different samples to represent one data point is likely to introduce an artificial error given we are not experimenting with absolute values. In the context of Figure 2, to maximise the data set, "replicates" occur at different dilutions of OC-FeR (y axis) as opposed to individual point replicates. In this figure we see the same trend (= Fe recovery in wet extractions exceeds that of dry extraction) across all samples. In addition, the reviewer can be further assured by the inclusion of three different organic acid based precipitates, and the trend between wet and dry samples persists across three independent precipitations. Finally, while the experiment in Figure 2 is not designed to investigate the role of different organic acids (due to inclusion of high Fe contents as noted by the reviewer), there is strong agreement in the overall trends between the results shown here and those in our previous study (which does investigate the effect of organic acids). This provides a good check and balance that the system here is performing as expected based on earlier experiments. We give an expanded explanation of the comparison between the two studies in an addition to section 4.2 following a previous reviewer's comment.

Style and Readability
The manuscript would be better served by a radical reduction in length. This is a technical note describing three relatively short comparison experiments that are an extension of the Fisher et al. 2020 paper. For instance, the first two sentences of the manuscript (lines 25 to 27) are obvious to readers of Biogeosciences. There are details (Lines 166-174) about diluting samples for AAS analysis that do not need to be repeated in such detail. The reader assumes that the authors have a basic understanding concerning the basics of the instrumental analysis. Section 2.7 appears superfluous because there is no where in the Results where organic C is discussed.
Following our response to the reviewer's previous comment, where we suggested to refocus the scope of the manuscript, this will certainly shorten the introduction and discussion. However, both the other reviewers and this reviewer (see next comment) have asked for expansion in places to remove reliance on the Chem Geol paper which we agree is important and have been happy to fulfil (see response to Reviewer 1 and 2). In addition, we think that it is the purpose of a technical note in particular to report all technical aspects of the analysis in detail, to leave no doubt about the procedures.
The reviewer is mistaken to say section 2.7 is unnecessary, as organic C measurements are included in Fig.1 and directly discussed in the results section 3.2. It is possible that the reviewer did not notice this was a commentary on the OC results as it is referred to as OC-FeR, (organic carbon bound to reactive iron). In response to reviewer 1 we have added a more explicit definition of this term at the start of the manuscript which we hope will remove any misunderstandings regarding terminology.
On the other hand, the any clear description of the carrier material was lacking, and I had to read the Fisher et al. 2020 Chem Geology article to understand how this key component had been treated.
This point was similarly made by reviewer 1 and 2, we have expanded the methods section relating to treatment of the carrier material to fully describe this.
I am confused by the use of the term "OC-FeR". What exactly is this? Organic C associated with iron oxides, as per Lalonde et al., 2012? Or is this Fe that is somehow made unreactive by Organic Carbon? Or is this simply the total iron oxide content? Or perhaps, the reactive iron content, whatever that may be? Are they referring to %dry weight Fe? Or are they referring to %dry weight FeOOH, or perhaps Fe2O3?,or perhaps %weight of whatever happens to precipitate including the organic fraction added?
Yes, this links to the point made above, and we apologise for the confusing terminology. We have added a definition "OC bound to reactive iron (OC-FeR)" in the introduction following comments made by previous reviewers. We have also made a substantial effort to reduce terminology where possible, including a much wider use of wt% for dry masses and removal of unnecessary formulae such as Fe2O3.

Further Comments
Line 183 The clause in the first line of the Results has no meaning. The manuscript is plagued by ill-defined discussion of reactivity. There are sentences such as "associated OC has a large influence on Fe reactivity." Towards what?
References to reactivity in this manuscript refer to the reactivity of Fe when reduced by dithionite, and this has been clarified for all in text mentions of reactivity. We have removed the first part of line 193-5 to make this sentence more concise.  1994 (Chem Geol). The authors must be referring to the extraction of organic matter.
Correct, this was a reference to OM extraction. This no longer stands part of the manuscript given we shifted the focus from attempting to improve the method.
Line 46: This is not surprising as hematite has been shown to be only partially dissolved by CDB method (see Kostka and Luther, GCA, 1994).
The method used in the study cited by the reviewer is a pH 4.8 dithionite extraction, these extractions are fundamentally different to the neutral pH OC-Fe extraction we conduct. In the Lalonde et al. (2012) iteration of the dithionite extraction (which we are working from), the neutral approach is said to reduce "all solid reactive iron phases and the organic carbon associated with these phases" (line 61). Therefore, while we agree with the reviewer that the outcome is not necessarily surprising, it is notable that other studies have come to other conclusions when applying this specific extraction to hematite samples.
Line 48 What do the authors mean be Fe reactivity here? Is this the goal of the study? Or the extractability of organic compounds.
See response to the first comment replied to in the "Further comments" section.

Removed developed
Line 120 This is not a concentration gradient. First of all, the authors are referring to contents, not concentrations (there seems to be confusion about the terms concentration and content throughout this manuscript). Secondly, a gradient implies a change in concentration over some property (e.g. depth, distances, density, etc..) We have checked and modified our use of content vs concentration throughout the manuscript in response to a similar point by reviewer 1. Gradient has been substituted for series.
Line 122 Confusing. Was the carrier material freeze-dried before or after mixing (or not all)?
Before mixing, clarified in the expanded methods section on carrier treatment. Line 128 "This sediment was freeze-dried then ashed (650 °C, 12 hrs)

to remove OC, and fumigated with HCl vapour to remove inorganic carbon."
Line 140 This is repetition of the Lines 80 and following.
Lines 80 onwards discuss how other studies have varied methodological parameters; Lines 140 onwards discuss the specific changes we make in our method, Line 170 Samples that were highly concentrated were diluted only 10 times while the more dilute samples were diluted 20 times?
To clarify, the samples diluted 10 times (or not at all) were the supernatants from the seawater rinses which follow collection of the extraction supernatant. The highly concentrated samples from the extraction were strongly diluted as appropriate to fit the AAS calibration window. The seawater rinses have very low concentrations of Fe and hence do not need much (if any) dilution. We have clarified this in text.
Line 172-3 "Dilutions of initial samples, in addition to the extraction supernatants, were conducted using MilliQ water to produce a subsample within the 1-10 ppm Fe range." Line 180 It's not clear that the authors differentiate here between a standard that is used for calibration and a secondary standard used as control.
For this purpose, they are the same thing, i.e., there is no control of accuracy with an international reference material. The machine is calibrated by sampling different masses of the manufacturer's standard. Then, during the sample run, this standard is resampled to check for drift. This is an important check in our method due to the use of decarbonated samples where drift can be a sign that the filter requires changing due to chloride accumulation. For analysing Fe concentrations by AAS at such high levels in experimental samples, the use of international reference materials is not common practise.
Line 190: This does not show the reductive capacity of the dithionite. If, for instance, dithionite is in excess, then 100% Fe extraction cannot show the reductive capacity of the dithionite.
The reviewer is correct here, but since we never reach 100% Fe extraction, this is not an issue in this application.
Line 224 Freeze-thawed samples? This experiment is not mentioned in the methods. Furthermore, this sentence (which is also discussion/interpretation) does not make sense.
What "previous one". The sentence refers to Figure 1. There are no freezethaw or freeze-dry samples in Figure 1.
Please see Line 167-9 of the methods "Arctic Ocean seafloor sediment was collected (…), of which half was freeze thawed and half was freeze dried." The sentence refers to the experiment between slurry and freeze-dried samples (shown in Fig. 2). We have added a clarification.
Line 226 "This experiment only compared freeze-dried vs. freeze-thawed (not slurry) samples" Figure 1: What do the fits represent and how is the fitting done? It looks to me like if you added more dithionite, eventually the %Fe recovery would start to decrease at some point.
Also, the blue symbols representing maximal Fe extraction do not match the corresponding curves for the black symbols.
The trend shown is a second order polynomial nonlinear fit. We address the issue of decreased Fe reduction as being a feature of Fe precipitation, an area of the manuscript we have expanded on in response to the review by Dr Henkel. The blue symbols represent maximal OC extraction rather than Fe, the OC/Fe ratio of <1 explains why they do not line up. A revised version of Figure 1 can be found in the response to Dr Henkel. Figure 2 is difficult to interpret. Firstly, the dependent variable is plotted on the x-axis, which is confusing for the bar chart depiction. Secondly, outside of the observation that freezedried sediments tend exhibit lower extractability than the fresh samples at high Fe contents, it is difficult to ascertain any kind of trend. Given the lack of replicates for each sample, and the large degree of variability in extractabilities, I find it difficult to be able to say anything concrete about these results.
We refer the reviewer to the improved Figure 2 created following constructive input from Dr Henkel which can be found in our reply. The issue of replicates has been replied to in detail earlier in this response.

Removed
Line 261. This is a red herring type of argument. One of the reasons that sediments are dried or freeze-dried and ground, is to avoid the problem that very small sample sizes and heterogeneity incur in solid phase analytical chemistry, when comparing average samples within a study. If one is interested in very small scale Fe-C heterogeneity, then a wet chemical extraction method is not the right approach.
This section was removed in an earlier edit in response to comments from Dr Jilbert.
Substituted with "reinforced" Line 284: Just state that the reagents were no longer in excess (see comments above).
Unclear what this comment refers to.
Line 294: I don't believe that the authors mean to say that organic carbon is reduced and released into the solution phase. Interestingly, sulfite incorporation into carbonyl groups may promote organic carbon solubility.
Replaced "released from the sediment" with "released from the coprecipitate complex".
Line 303: If increasing reagent additions make more problems, then what is the point?
We have addressed this point in detail thanks to a suggestion from Dr Henkel about the potential of incorporating anoxic methods as an alternative to deal with Fe precipitation.
Line 399: There is no "standard" method against which to calibrate This line no longer stands part of the revised manuscript following earlier edits.

Response to review by Peter Kraal (Reviewer 4)
Original comments in black, responses in blue.
With interest I have read this manuscript, in which the authors explore the impact of adaptations to an established chemical method (reductive dissolution of Fe(III) by dithionite) to extract OC associated with Fe oxide minerals. This topic is of interest, because the impact of Fe-OC interactions have a bearing on the environmental fate (and possible global budgets) of both Fe and C. However, the manuscript seems to imply that no adaptation is necessary for the majority of marine sediment samples, and the full analytical impact of some suggested changes was in fact not explored. As such, in my honest opinion, I do not really see the added value of this "technical note". I do not think that the rather loose suggestion, that increasing dithionite concentration during OCFe extraction might be useful with the caveat that it may have negative consequences that were not investigated, is particularly useful to the geochemical community. Furthermore, I think that the authors overlook some key points that can be taken from the data and do not properly consider the relationships between the findings for the poorly ordered synthetic Fe(III) precipitates (and their properties) and natural samples.
We thank Dr Kraal for their review. In response to the overall value of the study, reviewer 3 raised a similar point which we responded to in detail in that review. Essentially, we have reframed the introduction and discussion around addressing the issue of comparability between extractions conducted by differing methods which we felt provided a better focus. We also think that within this context the synthesis we provide is more useful to colleagues utilising this methodology. The citrate issue is somewhat of a red herring and we detail why any potential implications of increasing citrate are likely to be unproblematic in response to Tom Jilbert, we also include an anoxic alternative to citrates increases thanks to suggestions by Susann Henkel. We are grateful for the comments around how we can better discuss the data we present; this was particularly useful for the discussion on freeze drying of the Arctic sediments which we have adapted substantially.
Below are detailed comments. Note: I prepared this review and afterwards read the excellent, extensive reviews by Susann Henkel and Tom Jilbert. I apologize for a limited degree of overlap.
Key points General language: unnecessarily verbose and at times rather vague, essentials are buried in winding sentences from which the reader has to deduce the actual information. For a technical note, the experimental section is poor. There seem to be some errors in the use of % and wt.% which are a bit confusing, please carefully check units for Fe and OC concentrations and extraction efficiencies.
This has been corrected following previous reviews; the use of "content" and "concentration" has been modified, and wt% has been used much more widely to avoid confusion.
Regarding the choice of organic compounds, I understand why the selection was made to have compounds with different amounts of carboxylic functional groups. It would be good if the authors could also explain the choice for these compounds in general, from a point of view of representing natural organic material in marine sediments. As mentioned by Tom Jilbert, the discussion on the impact of type of organic compound (L358 ff) is weak and too dependent on other study by Fisher et al.
The compounds weren't chosen to represent marine OM, but to be the simplest types of compounds capable of bonding with reactive iron. We have expanded on the selection of these compounds in the methods (line 120) through explicit mention of the bonding mechanism.
"These acids differ in their carboxyl group content (pentanoic-1 COOH, hexanedioic-2 COOH, 1,2,4-Butanetricarboxylic-3 COOH), a factor thought to influence their binding to FeR via bonding between carboxyl groups and mineral hydroxyls (Karlsson andPersson, 2010, 2012;Mikutta, 2011)." Additionally, as Tom Jilbert also commented on this matter, we added the extra section (below) to expand on the role of OC compounds here and remove the reliance on the previous publication.
Line 327-333 "Even though such samples consisting exclusively of ferrihydrite-associated OC are unlikely to occur in nature, on the whole our results confirm previously identified relationships between Fe extraction and carboxyl content, where Fe extraction using the CDB method is more efficient for OC-FeR that contains more carboxyl rich OC (Fisher et al., 2020). This is attributed to the greater amorphicity of ferrihydrite coprecipitated with carboxyl rich OC, i.e. the resultant mineral phase is less crystalline than ferrihydrite coprecipitated with less carboxyl rich OC, and is therefore easier to reductively dissolve." Coprecipitation is known to affect the structure of Fe(III) precipitates, was any mineralogical characterization of the Fe(III) precipitate performed? The impact of coprecipitation will perhaps be minor for 2-line ferrihydrite, i.e. a poorly ordered ordered Fe(III) precipitate would likely become a bit more disordered. However, there are that such minor changes in structure can result in large changes in reactivity (and thus solubility). And the very high OC:Fe ratio might in fact result in a Fe(III)/OC coprecipitated that is not 2L ferrihydrite but some organic-rich, amorphous hydrous ferric oxide. There is no mention or discussion on the likely characteristics of the Fe(III) precipitates and their relation to natural counterparts anywhere in the paper. Also, I wonder why only 2-line ferrihydrite was used? This poorly ordered Fe(III) precipitate often transforms very rapidly into more crystalline Fe(III) precipitates such as lepidocrocite or goethite (for which limited solubility in CBD would be much more relevant?). Overall, it would be good if the authors could spend some words on their exclusive choice for 2-line ferrihydrite (or the Fe(III) precipitate that would actually form under the experimental conditions): is it representative for the Fe-OC pool in soils and/or sediments?
Mineralogical characterisation was performed on these same samples in the previous study; we include the XRD analysis shown in sup fig. 1 of that paper below. The reviewer is correct that the mineral becomes more disordered and this is now mentioned in text (see response to previous comment). This XRD analysis shows that the resultant sample remains as ferrihydrite but the peaks soften with the decreasing crystallinity of the structure. Following the review by Susann Henkel, where the suggestion that the ferrihydrite may be transformed during freeze-drying was raised, we added a reference to this XRD data in text.
Line 310 "An alternate hypothesis, that mineralogical transformation of ferrihdyrite during freeze-drying may lead to reduced Fe recovery, was ruled out by X-ray diffraction (XRD) characterisation of a representative freeze-dried sample, which confirmed the identity to still be 2-line ferrihydrite (Fisher et al., 2020)." We have added a section to expand on the use of 2-line over 6-line ferrihydrite. For the production of ferrihydrite we worked from the method of Schwertmann and Cornell (2000) Iron Oxides in the Laboratory. They describe the difference in production of 2-line vs 6-line ferrihydrite as "Rapid, forced hydrolysis of Fe III salt solutions under very acidic conditions at elevated temperatures (e.g. 80 "C) for a short period of time leads to 6-line ferrihydrite whereas rapid hydrolysis at RT and close to neutral pH produces 2-line ferrihydrite (Chukhrov et al., 1973;Schwertmann and Fischer, 1973)". Additionally, 2-line ferrihydrite acts as a precursor for hematite and goethite, mirroring mineralogical transformations in sediment.
Regarding the choice of ferrihydrite, we know that in most natural environments the association of organic matter with Fe occurs during Fe 2+ oxidation (e.g. Sodano et al., 2017, Riedel et al., 2013, and often (i.e., in many soils and most marine sediments) under low temperature, near netural pH conditions. These conditions are therefore much closer to the synthesis of 2-line than 6-line ferrihydrite, hence the choice. Additionally, 2-line ferrihydrite is used exclusively for OM coprecipitation as the high temperature, low pH conditions needed to precipitate 6-line ferrihydrite would make sorption of OM impossible due to hydrolysis. It would, without doubt, be interesting to observe what happens to the ferrihydrite-OC coprecipitate during further aging/crystallisation, but that is beyond the scope of this Technical Note.
In section 2.1 of the Methods, we added: "To produce synthetic OC-FeR samples 2-line ferrihydrite was chosen to represent FeR as it is readily precipitated in low temperature, oxic, circumneutral pH conditions and is therefore ubiquitous in soils and sediments (Zhang et al., 2018). In addition, it has an established ability to be experimentally coprecipitated with organic matter (e.g. Eusterhues et al., 2008;Eusterhues et al., 2011;Eusterhues et al., 2014)." Would have been interesting to see OC extraction efficiency for all treatments with the Fe-OC/sediment mixtures, not sure whether results from the variable dithionite experiment justifies assuming 1:1 relationship between Fe and associated OC extractability across the experiments conducted in this study. In fact, it would have also been very useful to have OC data for the dithionite/FeOC concentration range in the first experiment shown in Fig. 1: currently, we are not really given much to go by to understand how OC extractability varies as function of extraction conditions: one number for OC and then the assumption that Fe extractability is a perfect proxy for Fe-associated OC extractability for synthetic precipitates and natural samples alike. I find this a bit meagre (particularly for a technical note).
Unfortunately, it was not possible to produce OC data for all the points. The take away from Fig.1 is that even when Fe extraction is stretched to its maximum (by alteration of dithionite content), OC recovery is still incomplete. Therefore, at any point below this dithionite concentration, OC is certainly also incompletely extracted. While we do not state that Fe is a perfect proxy for OC loss, a strong similarity between the two values is to be expected given the 0.7-1:1 molar ratio between OC and Fe.
The impact of freeze-drying (section 4.2) is not represented properly. It shows a strong negative effect on the extractability of synthetic fresh Fe(III) precipitate, which is to be expected and reported regularly for poorly ordered Fe(III) precipitates (e.g. Kraal et al., Chemosphere, 2019). But is showed to have no effect on Fe extractability from the Arctic sample! However, in the discussion this important observations is ignored and there is a winding, unfocused and partly incorrect (see also Susann Henkel's review) discussion on the practices and challenges of freeze-drying.
We added to this section following Susann Henkel's review to include the suggestion of anoxic conditions. The inclusion of the Arctic sample is not meant for direct comparison with the artificial precipitate, but shows the effects of different dithionite additions on a natural sample representing what is often available to these types of studies. . It is also worth noting that our synthetic system is not meant as a direct analogy for real marine sediments, but plenty of studies perform mineral synthesis and utilise these samples for extractions, so even in synthetic samples this is an important difference. We did expand on the practicalities of marine sediment treatment in new sections added to 4.2 e.g.
Line 316-9 "Wet thawed samples have been used more widely in the sequential extraction of Fe (e.g. Wehrmann et al., 2014;Riedinger et al., 2017;Laufer et al., 2020) and the Arctic marine sediment sample used in our analysis was similarly thawed following freezing on collection. We find that our thawed sample shows no difference in its recovery for Fe compared to the dried variant of this sample." Line 337-340 "While application of the CBD method to slurried samples could increase the extracted proportion of Fe associated with OC, such an approach may not always be practical; either due to practical considerations, such as the difficulty in transporting heavy wet sediments, or when there is a need to preserve the sediment profile, for example, protecting anoxic sediments from oxic biological transformations." I have some reservations about the discussion on the impact of dithionite concentration on extraction efficiency in relation to increasing concentrations of Fe and OC. Firstly (L273-281), the authors present data from sediment/synthetic Fe-OC mixtures with Fe and OC concentrations that are incredibly high and as far as I know definitely not, as the authors claim, common (20-30 wt% OC and > 10 wt% Fe are not representative of normal marine sediments, shallow or deep).
We do not claim that we are trying to replicate normal marine sediment conditions here as this is a mechanistic study, see response to reviewer 3 for our discussion on high Fe contents.
Secondly, the authors focus a lot of attention on the relationship between dithionite concentration and OC content (L272-281), while I would expect that it is the ratio between dithionite and Fe (which is discussed later in the discussion, L282-291). I would argue that the bulk OC content of a sediment has relatively little to do with extractability of OC-FeHR, and I think this discussion on the role of dithionite concentration should better reflect the processes by which OC-FeHR is liberated. It actually seems like the "standard" (Lalonde) procedure works well for most sediments.
We agree that the ratio between dithionite and Fe is the most important fact; we do not discuss bulk sedimentary Fe since our samples contain only OC-FeR. The Lalonde method does likely work well for most sediments so we removed the direct criticism here, instead saying (Line 378-380) "While a 0.1 M Na dithionite concentration appears to be sufficient for most marine sediments with average reactive Fe contents, data produced with lower dithionite to solid ratios should not be compared to those extracted by a greater concentration of dithionite." However, many alternate methods have differing dithionite contents, and we show dithionite content can be an important factor Fe extraction.
The only adaptation really put forward in the current technical note is increasing the dithionite concentration, with two huge caveats: it seems only necessary for extremely Ferich samples (an observation not properly represented, as mentioned above) and the authors suggest that jacking up the dithionite concentration may have negative consequences on the performance of the extraction method, without actually exploring those potential negative impacts. I then wonder what to take away from this technical note?
As expanded on in the response to prior comments, the discussion around the increase of dithionite has been expanded to include a non-damaging anoxic method. The overall take away from this paper should be that methods which have already used differing dithionite conditions in the past are unlikely to be comparable in their results. That increasing extraction time, as done by recent methods, is largely pointless and that the important differences noted in drying of sediment is potentially important and requires further consideration. We hope this will become more apparent by the reframing of the discussion.

L29. CO2
Changed L87. HCl Changed L100-115. Would be nice to report the Fe/OC ratios during coprecipitation and the rationale behind choosing the concentrations. Also, how long was the precipitation allowed to occur?
Added: "The resultant slurry was rinsed 5 times in 5 L of DI water over 4 days" Added (Line 123-126): "The mass of organic acids used was determined through batch coprecipitations with varying organic contents, and the masses used here represent the saturation point, where a greater addition of organic molecules did not result in an increased OC association with ferrihydrite." Added (Line 122-123) "The coprecipitations produced three OC-FeR samples with an increasing number of carboxyl groups, resulting in increasing molar C/Fe ratios of 0.04 (pentanoic), 0.25 (hexanedioic) and 0.70 (1,2,4-Butanetricarboxylic)."

L111-112. a factor
Changed L113. binding association? Seems repetitive. It's also not type of binding, as all organic compounds have carboxyl groups. So. . . do you mean denticity, i.e. number of groups with which ligand binds to atom?
Changed to "strength of OC-FeR association".
L112-115. Please rephrase this, it is unclear (what is "weak" binding in this context?) and the link to slurry and dry sample is not explained.
Removed from the revised manuscript.
L117. "to explore whether mechanistic trends persisted", please explain what this means here. In other words, use less fancy words and provide more concrete information about what you want to test by varying Fe-OC content Removed from the revised manuscript L206. At which dithionite concentration?
Added: "at the maximum dithionite content each sample was subjected to." L208. "varying C content" is highly confusing here, you mean three types of OC, right?
Changed to "varying by organic acid" L207-212. I am confused again. From the methods (section 2.1), I gather that the FeOC coprecipitates used in the dithionite concentration experiment reported in Figure 1 contained hexanedioic acid (' 2 COOH') and was performed with freeze-dried FeOC. In Figure 1, it shows a decrease in % Fe (and presumably % OC) extracted with increasing Fe-OC wt% from 90% to 40% between 20wt% to 50 wt% Fe-OC. But in Fig. 2, Fe extraction % decreases from 50% to 30% for 20 wt% to 40 wt% Fe-OC for freeze dried Fe-OC with 2 COOH as C compound. Am I misunderstanding something, or was there a big difference in Fe extraction % for the two experiments under supposedly similar conditions?
This is broadly correct, they are similar precipitates in that they were made in the same way, however, they were the product of separate precipitations (as one precipitation does not give a large enough yield). This results in variable Fe and C contents in the product of each precipitate, and the variation becomes stark when percentages are implicated (particularly at lower contents such as 20 wt%) due to the small numbers produced. So a relatively small absolute change can lead to a large % change. This illustrates why independent repeats would not be useful, information should be taken from the trends within each figure rather than a cross comparison between two different products. As noted, the relative trend (decrease in Fe recovery with increase in OC-FeR content) is the same, even if the scale differs.
As an aside, it would be very useful if the authors could mention at which dithionite concentration the second experiment was conducted (maybe it is tucked away somewhere in the Materials and Methods, please repeat in this section).
Yes, everything outside of the varying dithionite concentrations experiment was conducted according to the Lalonde method so the dithionite content returns to 0.25 g. Added "Recovery of Fe following extraction with 0.25 g of Na dithionite" to methods section 3.2 L215. With 2 out of 5 treatments not showing this trend, I would consider removing "typically".
Removed in earlier review.
L219-220. What is implied here? The sediment contains 20 wt% Fe-OC, that is not a trace amount where analytical limitations would interfere with trends (it is mentioned earlier that this is equivalent to 7 wt% easily reducible Fe). Please explain how the comparatively low (emphasis on comparatively, as the whole range is strongly biased towards really high Fe-OC concentrations) Fe content could obscure a trend.
It is correct to say these are not trace amounts, the issue with low concentrations in in the quantification of Fe. At 20 wt% OC-Fe, the Fe supernatant requires a 400x dilution to measure Fe in the 1-10 ppm range. Therefore, any small difference in the absolute amount of Fe measured would be multiplied by a large factor which results in the bias shown. This serves as an additional explanation for the difference in % terms referred to in the earlier comment.
"Extraction of Fe from freeze-dried samples was…" L225-227. No doubt this will be dealt with in the discussion (having read the discussion, I now know it is not addressed), but there is a large difference of the impact of freezedrying on freshly precipitated poorly ordered Fe (for which it is established that freezedrying decreases reactivity, likely by aggregation) and natural sediment from a depth of 22-23 cm in the sediment in which you will not find any labile, freshly formed Fe unless the sediment forms fresh minerals as an artifact of sample treatment (or am I wrong? There is in fact no information on the chemistry of the studied sediment at all). In this sense, there seems to be a mismatch between the synthetic sample and the natural test material.
We have now published this data (Faust et al., 2021). A truncated sequential extraction was performed on this core up to 20 cm, at which depth we can confirm there is essentially no ferrihydrite. We have changed the focus of this section to reflect the difference in mineralogy.
e.g. Line 344-346 "this finding was not replicated for naturally aged Arctic Ocean samples, where FeR content was dominated by more stable phases such as hematite and goethite, indicating that older natural sediments are less influenced by freeze-drying compared to freshly precipitated ferrihydrite." Line 319-321"Natural aging processes within the sediment could explain this lack of freezedrying effect in older sediments due to both the physical effects of aging on Fe minerals, and the transformation of poorly crystalline reactive ferrihydrite phases to more stable phases at 22cm depth (Faust et al., 2021)." We appreciate the overall point about the difference in Fe content of deeper sediment samples compared to surface sediments and have adjusted wording to remove any idea that the sediment sample is comparable to the synthetic minerals. . So there is the issue of OC hydrolysis at low pH. But to quantify OC in sediments, they are commonly decalcified to remove inorganic C, using dilute HCl. What do the authors think about this? Does low pH have to be avoided at all costs during CBD extraction, to then submit the sediment to low pH during decalcification?
This is a good point that we agree with, for that reason in this paper we chose to decalcify the samples by HCl fumigation as oppose to HCl rinses (as done in other studies). Since nothing is discarded from the sample when it is treated by fumigation then we reduce any possibility of OC loss.
L239. You targeted physical aspects? Such as. . . the concentration of dithionite??? This is chemical, surely.
Changed L244-246. As mentioned before, the authors should probably also consider the large difference between the impact of freeze-drying on freshly coprecipitated Fe/OC and the impact of freeze-drying on relatively old and stable sediment. I do not think these are comparable.
See response to the final comment on the same point.
L258. Is the unit wt%? The context implies that this is the recovery efficiency (% of added Fe that is subsequently extracted), not the Fe content in the sample. (Should be checked in other instances as well, for instance y-axis of Fig. 1) Checked and changed, Y axis of Fig. 1 doesn't include wt%.
Changed to liberates L271-281. Because Fe and OC co-vary in the treatments, it is hard to judge what factor determines decreased extractability of OC-FeHR: is it the increase in OC or the increase in Fe (up to 24 wt% Fe!)? Merely looking at the trend in OC-Fe extractability in Fig. 1 does not answer that question. I would expect OC-FeHR to be liberated by dithionite because the Fe is reductively dissolved, and so the efficiency of OC liberation scales with efficiency of Fe reduction and not necessarily OC content. This would in fact mean that the bulk OC content of the sample is irrelevant, and the focus is on the wrong parameters in this section. The text in L293-306 supports this, the authors need to rethink their focus and wording here to better capture the chemical processes that occur when treating a sample with dithionite to reductively dissolve FeOx and associated compounds/elements. Again, regarding the units: the concentrations should probably be wt.% rather than % for the reported OC-Fe contents?
We are unclear exactly what this comment is getting at, we suspect this is an issue around phrasing which we hope to have resolved following earlier comments. We fully agree that Fe extraction is the critical parameter here, however in our experiments Fe is intimately linked to OC following coprecipitation. The bulk OC of the samples is the same as the amount of OC bound to FeR, which the comment agrees "scales with efficiency of Fe reduction".
wt% has been now used throughout to refer to OC-FeR contents.
Also, the authors mention that "many samples exist in the 20-30 [wt.]% range". And "the average value for marine sediment OC-FeHR composition is greater than 20 [wt.]%". These numbers and statements surprise me. Studying coastal and deep-sea sediments myself, I usually find TOC concentrations of 1-10 wt.% in a range of marine environments. The authors seem to claim that extremely organic-rich sediments are in fact very common (even though the phrasing "in many samples" is very different from "in many marine systems". . .), and then provide just one example, one value, from the Equatorial Pacific. I would like to have this point discussed a bit more: what kind of depositional environments host these very OC-rich samples, are we talking about modern or ancient (black shales and such, where the issue again appears that testing with labile OC-rich HFO makes little sense)?
The samples we are referring to here as having over 20% OC-FeR contents mean that 20% of the TOC pool of a sediment is bound to reactive iron, not that 20 wt% of the sediment is OC-FeR.
L293-306. So, now the authors are saying that the existing method works fine for most samples, that increasing the concentration of CBD could work for some (extreme) samples but may also have negative effects that are not explored. . . So what are we to do with this technical note?
This narrative can be shifted by changing the focus of the paper slightly. We can instead say that the wide range of dithionite concentrations used in previous methods (Table 1) can have an effect on Fe extractability, and therefore studies with different dithionite masses can produce non-reproducible results, rather than focusing on achieving maximal extraction. The point about citrate as a negative was addressed in response to Susann Henkel and Tom Jilbert, essentially we now include the prospect of anoxic conditions for high Fe samples (which Henkel et al., (2016) shows to work). The issue with citrate contamination is also of minor importance for natural samples, it is only synthetic ferrihydrite which has a much greater free surface (less OM coverage) where citrate contamination can become a big problem. Lalonde et al., (2012) showed that citrate contamination for natural sediments is at most 0.08% of dry sediment weight so increasing it is unlikely to significantly change this. We have toned down the caution given to citrate in the manuscript and included the anoxic alternative.
L307-367. This is an excessively lengthy paragraph on freeze-drying. It is actually strongly detached from the findings and just meanders along various aspects of freezedrying. The main issue is that the authors do not correctly represent their own results: a decrease in Fe extractability was found in the synthetic samples, but not in the Arctic samples! As I mentioned before, it is to be expected that freeze-drying a much stronger effect on a fresh, poorly ordered Fe(III) precipitate than on a rather old sediment sample. In fact, there are findings that show that freeze-thaw cycles can increase extractability of elements (the authors also touch upon this, and Susann Henkel also hints at some inconsistencies in this section). Note: this is a change in extractability, not content; the text became rather confusing when the authors started to speak about freeze-drying as a treatment that can increase the contents of for instance OC and metals. . . (L351-353). Overall, this whole section fails to address the key point, i.e. the discrepancy between the results for the synthetic and environmental samples (or, in broader terms, the difference between artefacts in fresh and old Fe minerals, whether they are synthetic or natural), and instead presents long and rather unfocused and at times confusing literature review on freeze-drying.
Despite my sharp tone (work in progress), I trust this review is fair and constructive.
We have rewritten this part of the manuscript to remove the extended discussion on freezedrying. We state that the dominant Fe phases here are crystalline phases unlikely to be extracted by CBD treatment under any circumstance. We also consider the fact that in deeper sediments such as ours the freeze drying effect we observe for freshly precipitated ferrihydrite has occurred through natural aging of the sediment (due to transformation and aggregation with diagenesis). This allows us to conclude that extraction of surface sediments with high ferrihydrite contents is likely to be improved through use of non-freeze dried sediments while for those deeper in the sediment profile these effects are less important.
Line 344 "However, this finding was not replicated for naturally aged Arctic Ocean samples, where FeR content was dominated by more stable phases such as hematite and goethite, indicating that older natural sediments are less influenced by freeze-drying compared to freshly precipitated ferrihydrite." We have removed the section relating to changes in absolute C/metal contents. We hope the changes we have made in this section satisfy the request for a clearer discussion on the differences between synthetic/environmental samples.