Potential of common extraction procedures to evaluate organic carbon association with different species of metal(loid)s in soils

Amenkhienan, Bright E.; Dijkstra, Feike A.; Warren, Charles; Singh, Balwant

doi:10.5194/egusphere-2026-2843

Preprints

https://doi.org/10.5194/egusphere-2026-2843

Preprints

25 Aug 2026

| 25 Aug 2026

Potential of common extraction procedures to evaluate organic carbon association with different species of metal(loid)s in soils

Bright E. Amenkhienan, Feike A. Dijkstra, Charles Warren, and Balwant Singh

Abstract. A strong association of organic carbon (OC) with metal oxides, especially iron (Fe) and aluminium (Al) oxides has been reported in several studies. However, the role of various species of Fe, Al, manganese (Mn) and silicon (Si) in the stabilisation of OC in Fe-rich and phosphorus (P) deficient soils, common in tropical and sub-tropical regions, remains poorly understood. To address this, we collected topsoil (0–20 cm) and subsoil (20–40 cm) samples from 37 sites across agricultural regions of New South Wales, Australia. Soil samples were subjected to three separate chemical extractions, i.e., sodium pyrophosphate (PP) to extract organo-metal complexes, ammonium oxalate (OX) to extract poorly crystalline and short-range order (SRO) minerals and dithionite citrate bicarbonate (DCB) to extract total metal oxides, to determine OC associated with each of the extractions. Soil organic carbon (SOC) was extracted in the sequence: C_PP>C_DCB>C_OX,with a mean of 62±0.16 %, 41±0.19 % and 28±0.21 %, respectively, of the total C extracted from soils. The extraction sequence for Fe was: Fe_DCB>Fe_OX>Fe_PP, with a mean of 49±0.18 %, 9±0.05 % and 3±0.03 %, respectively, of the total Fe extracted from soils. Aluminium was extracted in the same sequence as Fe, with a mean value of 4.0±0.02 % for Al_DCB, 3.9±0.02 % for Al_OX and 2.0±0.01 % for Al_PP of the total Al. Manganese extraction occurred in the sequence - Mn_OX>Mn_DCB>Mn_PP, with an average of 78±1.58 %, 65±1.10 % and 43±1.23 %, respectively, of the total Mn extracted. All extractants dissolved <1 % of the total Si. All extractable forms of Fe and Al showed significant positive correlations with extractable C, which suggested their potential role in the preservation of SOC. The large fraction of OC extracted by PP suggested that organic-Fe/Al complexes constituted a large fraction of the total SOC in the studied soils. The significant amount of OC associated with OX extraction also indicated that a substantial portion of the SOC was associated with SRO and poorly crystalline Fe oxides, while a relatively small proportion of the total OC was extracted by DCB suggested a limited role of crystalline Fe/Al oxides in the OC stabilisation.

Received: 18 May 2026 – Discussion started: 25 Aug 2026

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.

Download & links

Preprint (PDF, 1571 KB)

Supplement (832 KB)

Download & links

Bright E. Amenkhienan, Feike A. Dijkstra, Charles Warren, and Balwant Singh

Status: final response (author comments only)

RC1:
'Comment on egusphere-2026-2843', Anonymous Referee #1, 30 Sep 2026
The manuscript presents results on the extraction of metals, Si, and C from (sub) tropical soils with frequently used methods. I agree that the (sub) tropical soils under study have not intensively investigated yet so the authors aimed to fill a gap in knowledge. However, there are some major issues that must be clarified prior to potential publication.

My main concerns relate to definitions, lab procedures and data interpretation.
Several times, the authors mentioned “Fe/Al oxides” without defining this phrase. Do they mean the sum of individual Fe and Al oxides? Or do they mean mixed oxides, which contain both Fe and Al so that the “pure” Fe and Al oxides, respectively, are not included? Clarification is necessary, as it determines the interpretation of extraction data.
Several times (see below), the authors claimed that Al oxides are dissolved (extracted by) DCB. This is wrong and shown by many studies/reviews, which the authors even cited. On the other hand, sometimes the authors interpreted DCB extraction of Al oxides correctly. Again, clarification and uniform correct interpretation is necessary.
Several times, the authors mentioned SRO and poorly crystalline minerals. Does it mean that they distinguished between SRO and poorly crystalline? If so, which minerals are SRO, but not poorly crystalline and vice versa?
The authors referred to Wagai et al. (2013), who suggested filtration of PP extracts with a 0.025 µm filter, but the authors used a 0.2 µm filter without explanation. Furthermore, they filtrated all extracts at 0.2 µm, except for DCB extracts, for which they applied a Whitman filter with approx. 2 µm pore size. Not only is the inconsistency of the filter pore size between the different extracts incomprehensible, but also that it does not exclude the presence of particulate species in the DCB extracts - dissolved species, and only those are at stake, are defined via a 0.45 μm filter. These are not the only examples where studies/reviews are cited but the statements in them are not considered.

More specifically and minor concerns:
line 41: “Soil OC carbon” Please correct. And SOM would be correct here, not SOC, which is in most cases only a measurand.

line 48: Similarly, SOM is associated with minerals, not SOM. Please check and correct the entire manuscript accordingly.

lines 61-63: “Iron and Al oxides […] are formed in soils from the weathering and dissolution of primary minerals (Chadwick and Chorover, 2001).” Not exclusively, but Fe oxides also in redox cycles. This has been known before 2001.

lines 108-111: It is good to mention these studies on drawbacks of extraction procedures, but the findings in these studies must be considered in planning and conducting the study, and interpretation of the results. For instance, these studies show issues of phase separation, and selectivity and completeness of PP extraction, which were not considered.

line 113: Extractions cannot estimate SOM stabilisation, as they do not provide an estimate of the deceleration of SOM decomposition; they can roughly estimate the contents of stabilising phases.

lines 113-115: The extractions used here are unable to differentiate the SOM associated with Fe or Al oxides either.

line 121-122: “the important role of poorly crystalline Fe/Al oxides and/or short range order (SRO) Fe/Al oxides“ here is even a combination of incomprehensible phrases; please check my previous comment. Also, as the extractions have no proper mineralogical meaning, correlations mentioned before between extracted amounts and SOC can only indicate but never prove the importance of certain groups of minerals. Correlation is not causation.

line 124-125: DCB-extractable Fe alone is no measure of crystalline Fe, as mentioned in the cited references, but DCB releases Fe from SRO Fe oxides, too.

line 127-128: When mentioning PP extraction here, it would be good to mention the known drawbacks.

line 136: The selectivity of PP extraction is unknown, and some of the studies cited before indicate poor selectivity. This problematic issue is not discussed at all.

line 142: The most common methods are not necessarily the most suitable ones. Why not using citrate instead of the highly problematic PP? The studies/reviews the authors cited indicate its better suitability.

lines 151-155: The highest content does not prove the most important role in stabilisation. Stabilisation refers to processes that decelerate SOM decomposition. What if a less abundant compound decelerates SOM decomposition more effectively that the most abundant compound? There are no data on the decomposition/stabilisation rates given, and they cannot be derived from extraction data.

lines 160-161: I can understand the decision to sample the soils by depth intervals, but I am not convinced, according to the available information, that actually exclusively top- or subsoils were studied. The thickness of an A horizon may exceed 20 cm so that the deeper sample is mixed topsoil and subsoil, or it may be less than 20 cm so that the upper sample is a mixed sample. I cannot believe that the A horizon(s) was/were always exactly 20 cm thick.

Table 1: Apart from the implied difference between SRO and poorly crystalline minerals, the information is missing that OX extracts metals from organic forms. DCB does not generally dissolve metal oxides, but only those with a reducible metal cation, thus not Al oxides. These aspects are mentioned in several of the cited studies. Is it necessary to use the phrase ‘free’ oxides (even already in quotation marks) without definition? The phrase has no real mineralogical meaning then.

lines 207-208: OX extracts metals from organic association, too (see my comments before). Schwertmann (1964) exclusively referred to poorly crystalline Fe oxides, not “SRO and poorly crystalline minerals” like the authors claimed.

lines 215-216: “Dithionite-citrate-bicarbonate extraction was used to extract pedogenic Fe and Al phases in soils (Mehra and Jackson, 1960).” No. See my comments before. The reference is from 1958.

lines 239-240: “Inorganic HH extraction targets SRO and/or amorphous Fe and Al oxides (Chao and Zhou, 1983; Ross et al., 1985).” None of the studies mentioned studied the dissolution of Al oxides. Thus, the suitability of the method is unknown.

lines 406-407: OX-soluble Fe oxides are not necessarily non-crystalline. The authors even mentioned later in the text the dissolution of e.g., goethite. The whole text needs a very substantial revision regarding uniform interpretation of studies and findings derived from them. It is not only here that there are contradictions. It seems like piecemeal.

section 4.1: Please shorten. Everything is known, and the study using common, but partially not selective methods cannot provide new mechanistic findings.

lines 519-521: It is not possible to estimate the extent of re-adsorption with the applied methods either.

lines 524-525: “The DCB procedure is known to extract pedogenic or free Fe oxides (and Al substituted in the structure of goethite and hematite) as well as poorly crystalline and SRO Fe/Al oxides” The first part is correct regarding Al, the latter is not.

line 533-535: “In the oxalate extraction procedure, Fe present in non-crystalline, poorly ordered and short-range ordered (such as ferrihydrite) was dissolved and a fraction in the organic complexes may also be dissolved.” Why was extraction of metals in organic complexes not mentioned and included in the interpretation before? Ferrihydrite is not poorly ordered?

lines 538-539: “The small ratio of Fe_OX:Fe_DCB, referred to as active Fe ratio, indicate the degree of crystallinity of Fe oxides in soils”. Another example of the selective interpretation of the entire literature: If OX does not only dissolve oxide Fe (lines 533-535), it is not necessarily a good indicator of oxide crystallinity.

lines 543-545: “The significant correlation between Fe_DCB and Fe_OX (Table 4), may be due to some overlap in the phases extracted by the two procedures“ “Some overlap” may be between 0 and 100%. Anyway, this is known from several cited studied.

lines 546-548: “In addition to SRO Fe oxides, oxalate is known to dissolve crystalline Fe oxides, such as magnetite, maghemite, lepidocrocite, and poorly crystalline and nano-crystalline goethite and hematite” exemplifies selective interpretation again and see my previous comment on lines 406-407.

lines 549-551: “The close relationship between Fe_DCB and Fe_OX may also suggest that both Fe forms were function of similar pedogenic processes (Rezapour et al., 2015)” See my comments before, and this knowledge was gained long before 2015.

The entire section 4.2 is too extensive, only repeats existing knowledge, but does not apply this to one's own data in a critical discussion, particularly regarding PP extraction.

line 581: “Crystalline Al oxides are generally not dissolved during the DCB extraction” Correct, but several times the authors implied dissolution of Al oxides by DCB. Amorphous Al oxides are not dissolved by DCB either.

section 4.3: The authors do not consider that desorption of anionic OM adsorbed on any mineral surface may be induced by the alkaline pH during PP extraction. The contribution of this fraction of the extracted SOM cannot be quantified so that it is not possible to attribute the SOM released during PP extraction to SOM formerly associated with metals. This is particularly important with the acidic soils under study.

lines 649-650: The large differences between Fe extracted by PP at pH 7.5 and pH 10 (factor of 2.88) clearly shows the inappropriateness of the method - what is the "correct" value for Fe in organic association? Or could this value be even higher? Or does it have to be assumed to be lower because, depending on the pH, non-targets are also dissolved?

lines 688: “The C_DCB is expected to be C associated with secondary crystalline Fe oxides including SRO phases.” The pH of the DCB solution is 8 (Table 1) so that pH-induced desorption of SOM from an unknown variety of minerals surfaces occurs to unknown extent as mentioned before. DCB does not dissolve Al SRO phases completely.

line 690: “Since DCB extracts both SRO and crystalline Fe and Al oxides” According to line 581 and several papers the authors cited, neither SRO nor crystalline Al oxides are extracted by DCB.

lines 692-694: The calculations do not make sense, as the SOM released during extractions cannot be attributed to certain groups of minerals because of the known drawbacks and the pH-induced SOM release mentioned before.

line 705: The incorporation of Al in goethite and hematite, substituting Fe, is mentioned here for the fourth time.

section 4.4: Attributing metal:C ratios to mechanisms is very questionable given the known drawbacks of the methods used. For instance, pH-induced SOM release remains unknown, and so is the “correct” concentration of PP-extracted Fe and Al, depending on pH. The pH dependency of Fe and Al extraction by pyrophosphate is even reciprocal (Figure S3). In addition, the completeness of metal extraction by PP is unknown, as mentioned in cited studies.

section 5: The section does not provide conclusions, but a summary of the findings. At this point at the latest, the obvious problems with the application of the extractions should have been mentioned.

Due to my above-mentioned objections, a substantial revision of the manuscript is necessary.
Citation: https://doi.org/10.5194/egusphere-2026-2843-RC1
- AC1: 'Reply on RC1', Bright Amenkhienan, 26 Oct 2026
  
  The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2026/egusphere-2026-2843/egusphere-2026-2843-AC1-supplement.pdf
  
  Citation: https://doi.org/10.5194/egusphere-2026-2843-AC1
RC2: 'Comment on egusphere-2026-2843', Anonymous Referee #2, 27 Oct 2026

Congratulations to the authors for their efforts in conducting this research and preparing the manuscript. The manuscript addresses an important topic in soil science by investigating the interactions between soil organic carbon (SOC) and mineral phases, particularly Fe- and Al-bearing phases. However, several aspects of the methodology and data interpretation would benefit from further clarification to strengthen the robustness of the findings. My comments and suggestions are provided below.
Lines 13-14: The manuscript's aim explicitly concerns Fe, Al, Mn and Si, and the hypothesis proposes roles for all four. But the principal OC conclusions are essentially about Fe and Al. What does the study actually establish about Mn- or Si-associated OC? If their contribution cannot be resolved using these extractants, should Mn and Si be presented as supporting geochemical observations rather than as equivalent components of the central research question?
Line 15: The abstract explicitly frames the study as examining Fe-rich and P-deficient soils, but phosphorus is not reported among the measured soil properties. The Methods report pH, EC, texture, CEC, TC/TN and total Fe, Al, Mn and Si—but no P measurement. Please provide the relevant P measurements and threshold/reference, or remove the P-deficient framing from the objective and abstract.
Lines 21-22: The Methods indicate that pyrophosphate, ammonium oxalate, and DCB extractions were performed separately on different soil aliquots rather than sequentially on the same soil sample. Therefore, the statement in the Abstract that “SOC was extracted in the sequence: CPP > C-DCB > C-OX” is potentially misleading and should be revised. Please clarify that these represent three independent extraction procedures.
Lines 33-34: the abstract stated that "The significant amount of OC associated with OX extraction also indicated that a substantial portion of the SOC was associated with SRO and poorly crystalline Fe oxides", but in the conclusion it's mentioned that "Poorly crystalline and SRO alumino-silicate minerals were absent in the soils." Please clarify this statement.
Lines 222-223: The DCB extraction was repeated for some samples when brown or red coloration persisted in the soil residue. Since the objective of DCB extraction is to remove pedogenic Fe oxides as completely as practicable, please clarify how repeated extraction was accounted for when interpreting the relationships between DCB-extractable Fe and soil organic carbon, particularly if residual SOC was measured after extraction?
Lines 264-265: Some samples reach pH 7.6, but the manuscript simply states that TC represents OC because “most” soils are acidic, and carbonate was tested qualitatively with HCl in soils near neutral pH. Because TC serves as the denominator for the reported percentages and C:(Fe+Al) ratios, even modest inorganic C could affect the quantitative interpretation. Please provide evidence that inorganic C was negligible across the dataset.
Line 341-342: Al extracted by PP, OX, and DCB ranged only between 2-4% of the total Al in samples. What are the possible Al forms in soils?
Line 393. Was the procedure conducted in sequence (i.e. PP first, and followed by DCB, and OX), or is it the sequence with higher to lower organic C C-PP > C-DCB > C-OX? Perhaps the authors can clarify the sentence.
Lines 408-409: The interpretation appears to assume that the OC released by OX represents a subset of the OC released by DCB, such that CDCB - COX can be interpreted as OC associated exclusively with crystalline Fe/Al oxides. However, it is also mentioned somewhere that OX and DCB are independent extractions with different reaction chemistries and partially overlapping. How do the authors address this issue? Or perhaps, I suggest interpreting these values as operationally defined extractable-C fractions rather than assigning the 13% difference quantitatively to crystalline Fe/Al-associated OC.
Lines 630-632: This result highlights that the amount of carbon extracted by a reagent and the amount of mineral elements extracted by that same reagent do not necessarily follow the same pattern. Pyrophosphate can disperse organic matter and disrupt associations between organic molecules and soil particles. It can also extract organic matter associated with metal ions, particularly through complexation with Fe and Al. Other factors, such as the nature of the organic matter and its associations with soil minerals, may also be important. Consequently, the amount of OC extracted does not have to be proportional to the amount of Fe or Al extracted.
Lines 754-755: The authors stated that "Crystalline Fe oxides (i.e., goethite and hematite) were dominant…” and that SRO aluminosilicate minerals were absent." But elsewhere the authors acknowledge that DCB and OX are non-specific: DCB does not quantitatively extract all crystalline Fe phases, while OX may dissolve some crystalline Fe minerals." Can goethite and hematite dominance, and the absence of SRO aluminosilicates, be concluded from selective dissolution alone? Unless supported by independent mineralogical measurements, these statements may need to be reformulated as inferences based on operational extraction ratios rather than direct mineral identification.

Citation: https://doi.org/10.5194/egusphere-2026-2843-RC2

Bright E. Amenkhienan, Feike A. Dijkstra, Charles Warren, and Balwant Singh

Supplement

https://doi.org/10.5194/egusphere-2026-2843-supplement

Bright E. Amenkhienan, Feike A. Dijkstra, Charles Warren, and Balwant Singh

Viewed

Total article views: 443 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	Supplement	BibTeX	EndNote
259	129	55	443	77	52	56

HTML: 259
PDF: 129
XML: 55
Total: 443
Supplement: 77
BibTeX: 52
EndNote: 56

Views and downloads (calculated since 25 Aug 2026)

Month	HTML	PDF	XML	Total
Aug 2026	34	17	5	56
Sep 2026	126	52	36	214
Oct 2026	99	60	14	173

Cumulative views and downloads (calculated since 25 Aug 2026)

Month	HTML	PDF	XML	Total
Aug 2026	34	17	5	56
Sep 2026	126	52	36	214
Oct 2026	99	60	14	173

Viewed (geographical distribution)

Total article views: 440 (including HTML, PDF, and XML) Thereof 440 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 27 Oct 2026

Short summary

The smallest fraction of soil minerals is rich in iron (Fe) and aluminium (Al) are known to preserve and store organic carbon (OC) in soils. We subjected topsoil and subsoil to three separate chemical extractions, i.e. sodium pyrophosphate (PP), ammonium oxalate and dithionite. Iron was dissolved in the sequence- Fe_DCB>Fe_OX>Fe_PP, and Al was: Al_DCB>Al_OX>Al_PP. A large fraction (62%) of OC extracted by PP indicates that most of the SOC was present in organic-Fe/Al complexes in the studied soils.


Total:	0
HTML:	0
PDF:	0
XML:	0