Iron Oxyhydroxide Transformation in a Flooded Rice Paddy Field and the Effect of Adsorbed Phosphate

The mobility and bioavailability of phosphate in paddy soils are closely coupled to redox-driven Fe-mineral dynamics. However, the role of phosphate during Fe-mineral dissolution and transformations in soils remains unclear. Here, we investigated the transformations of ferrihydrite and lepidocrocite and the effects of phosphate pre-adsorbed to ferrihydrite during a 16-week field incubation in a flooded sandy rice paddy soil in Thailand. For the deployment of the synthetic Fe-minerals in the soil, the minerals were contained in mesh bags either in pure form or after mixing with soil material. In the latter case, the Fe-minerals were labeled with 57Fe to allow the tracing of minerals in the soil matrix with 57Fe Mössbauer spectroscopy. Porewater geochemical conditions were monitored, and changes in the Fe-mineral composition were analyzed using 57Fe Mössbauer spectroscopy and/or X-ray diffraction analysis. Reductive dissolution of ferrihydrite and lepidocrocite played a minor role in the pure mineral mesh bags, while in the 57Fe-mineral–soil mixes more than half of the minerals was dissolved. The pure ferrihydrite was transformed largely to goethite (82–85%), while ferrihydrite mixed with soil only resulted in 32% of all remaining 57Fe present as goethite after 16 weeks. In contrast, lepidocrocite was only transformed to 12% goethite when not mixed with soil, but 31% of all remaining 57Fe was found in goethite when it was mixed with soil. Adsorbed phosphate strongly hindered ferrihydrite transformation to other minerals, regardless of whether it was mixed with soil. Our results clearly demonstrate the influence of the complex soil matrix on Fe-mineral transformations in soils under field conditions and how phosphate can impact Fe oxyhydroxide dynamics under Fe reducing soil conditions.


S1. Soil characterization
The soil samples were taken during the dry season in February 2020 at the Ubon Ratchathani Rice Research Center (URRC) in Thailand.A soil profile of 2 m depth (Figure S1) was established at the sampling site for soil characterization.The soil was classified as a Hydragric Loamic Anthrosol on sandstone after the World Reference Base for Soil Resources. 1 The puddled horizon was 20 cm thick, with a dense and hard plough pan at 20 cm (anthraquic horizon).Beneath the plough pan, prominent orange-brown hydromorphic features were present (hydragric horizons).While hydromorphic features were sharp and small, and partly followed plant root channels in the horizon beneath the plough pan (20-30 cm), the features became larger and blurred at greater depths (30-130 cm).Below the groundwater table (>130 cm), the soil had a bleached color and only showed few hydromorphic features.
Table S1: Characterization of the experimental rice paddy soil (0-15 cm depth), with element concentrations measured in dried, sieved (<2 mm) and milled soil by X-ray fluorescence spectroscopy (XRF), an elemental analyzer (EA), or after total digestion with hydrofluoric acid (HF) and the texture determined on dried and sieved (<2 mm) soil.These values have been reported previously in Schulz et al.

Method details
For the preparation of Mössbauer samples for NA Fe minerals without soil, triplicate samples were combined and ~15 mg of mineral was suspended in 1.  S2.  S3.

S21
Mineral transformations in NA Fe-mineral mesh bags without soil  S7.

S8. Element contents in incubated NA Fe minerals
Figure S15: Element contents of phosphorus (P) and silicon (Si) in initial and incubated NA Fe ferrihydrite ( NA Fe-Fh), ferrihydrite-P ( NA Fe-FhP) and lepidocrocite ( NA Fe-Lp) samples without soil, determined after mineral dissolution in acid.The P/Fe ratio at 16 weeks was 0.01 for NA Fe-Fh and NA Fe-Lp, and 0.11 for NA Fe-FhP.The Si/Fe ratio at 16 weeks was 0.01 for NA Fe-Fh, NA Fe-Lp, and NA Fe-FhP.

Figure S1 :
Figure S1: Soil profile at the Ubon Ratchathani Rice Research Center, Thailand, February 2020.Images show the soil profile across the full depth (left; 0-180 cm) and the soil at the depth where samples were installed in June 2021 (right; 10-15 cm).

Figure S2 :
Figure S2: Depth profiles of solid element contents of carbon (C), nitrogen (N), iron (Fe), aluminum (Al) and silicon (Si) determined by X-ray fluorescence spectroscopy in samples from the soil profile established in the experimental rice paddy field during the sampling campaign in the dry season in February 2020 at the Ubon Ratchathani Rice Research Center, Thailand.

Figure S3 :
Figure S3: Mössbauer spectra collected at 77 K (left) and 5 K (right) from of the rice paddy soil sampled in February 2020 during the dry season.Respective spectral areas of fitted mineral fractions are given in pie charts.The colors and labels in the pie charts correspond to the colors of the fitting components in the Mössbauer spectra.Abbreviations: Fh = ferrihydrite, Gt = goethite, D = doublet, S = sextet.This data has been reported previously in Schulz et al.2

Figure S7 :
Figure S7: Photo of the experimental setup in at the Ubon Ratchathani Rice Research Center, Thailand, in June 2021, and an empty sample holder.

Figure S8 :
Figure S8: Soil redox potentials (Eh) on the day of the experimental setup (0 weeks) and at sampling days after 8, 12 and 16 weeks.The depth where mineral samples were installed is marked by the dashed line.Replicate measurements (Rep) are indicated by colors.

Figure S9 :
Figure S9: Aqueous element concentrations of magnesium (A), sodium (B), total sulfur (C) and calcium (D) in soil porewaters at the start and during the experiment.
Figure S10: Mössbauer spectra collected at 77 K of initial and reacted NA Fe minerals.Abbreviations: w = weeks, Gt = goethite, CF = collapsed feature.Fitting parameters are presented in TableS2.

Figure S14 :
Figure S14: Mineral fractions in initial and field-incubated NA Fe ferrihydrite (A), ferrihydrite-P (C) and lepidocrocite (E) with corresponding X-ray diffraction patterns of initial and 16 week-incubated (16 w) NA Fe-Fh (B), NA Fe-FhP (D) and NA Fe-Lp (F).Error bars show the standard error of experimental triplicates, errors <0.02 are smaller than symbols and are not shown.Main diffraction peaks are labeled with Fh = ferrihydrite, G = goethite, L = lepidocrocite, Q = quartz.The quartz peak in panel D has been cut in height to improve the visibility of smaller peaks in the diffractogram.Fit results and parameters are presented in TableS7.

Figure S17 :
Figure S17:Fractions of 57 Fe atoms from added57  Fe-labeled minerals in 57 Fe-mineral-soil mixes before and during the 16-week incubation in the field.The displayed fractions were calculated from Fe contents and 57 Fe fractions measured after aqua regia digestion of the mineral-soil mixes (FigureS16): Signal fraction_ 57 Feadded-mineral = 1-( 57 Fesoil/ 57 Femineral-soil-mix), where57  Fesoil is the content of 57 Fe in the soil in µmol g -1 , and 57 Femineral-soil-mix is the content of 57 Fe in the mineral-soil-mix samples soil in µmol g -1 , as calculated from the Fe content and 57 Fe fraction after aqua regia digestion of the samples.The difference between fractions shown for 57 Fe-Fh and 57 Fe-FhP at 16 weeks is significant according to the Wilcoxon signed-rank test (p = 0.03) performed in RStudio 4.1.2. 2
a Center shift; b Quadrupole splitting (for doublets); c Quadrupole shift (for sextets); d σ, standard deviation of QS, ε or H; e Hyperfine field; f Red.χ 2 , goodness of fit; * Indicates values that were fixed during the fitting process.