Introduction

The majority of the Earth’s terrestrial carbon is stored in the soil, the top of which is primarily exposed to sunlight1,2,3,4 and intense microbial activity5,6. Climate scenarios forecast the loss of 55 ± 50 Pg of C from upper soil horizons by 2050, and each °C increase in temperature would reduce the total soil organic carbon (SOC) stocked by 30 ± 30 to 203 ± 161 Pg of C5. Furthermore, the worldwide increasing land use and land cover changes, e.g., by deforestation and drought7,8 favor SOC losses by enhanced soil respiration6,9, erosion8, and temperature-induced degradation7,10. These effects are expected to reduce soil C storage11,12 and C sequestration/stability13,14, thereby affecting land degradation15,16. Generally, biodegradation processes are mainly responsible for SOC losses with emissions of CO2, COS, and H2S10,17,18, whereas sunlight-exposed SOC may produce CO219 and NOx emissions1,3. These phenomena involve various soil biogeochemical processes, including SOC transformations20,21, soil oxygenic/anoxygenic photosynthesis22,23, soil respiration accelerated by temperature increase6,9, dissolution of redox and non-redox metals from minerals4,13,14,24,25,26,27, and As release28,29. Specifically, transformation and mineralization processes occur for SOC and humic substances (HS) associated with clay, and Fe-like minerals13,24,25,30 at the mineral-microbes interfaces, via reactive oxygen species generated from mineral-induced electron transfer14, Fe-mediated processes in humid soils25, and/or reductive dissolution of clay minerals14,31. These processes highlight the absence of unique potential mechanisms of SOC mineralization. Based on this research gap, the key scientific question emerges: is SOM involved in the integrated mechanism of mineralization of HS-associated organo-minerals and their subsequent sequestration or neoformation, along with the export of various dissolved and gaseous components into surface water, groundwater, and/or the atmosphere?

Moreover, mineral sequestration of SOC22,23 may occur via subsequent sequestration/uptake of carbon (C)32,33,34,35, sulfur (S)36,37, nitrogen (N)1,38,39, and various metals, including Hg4,24,25,30,40,41,42, which are mostly responsible for the neoformation of organo-mineral components24,25,30,32,43,44 that are involved in SOC stabilization and accumulation13,45. Such transformations and sequestration of soil organo-mineral components occur simultaneously in the soil matrix, but any proved experimental evidence of the various responsible processes and factors that drive these phenomena, as well as their mechanisms, remain mostly uncertain. Based on this research gap, another key scientific question emerges: do raw soils under sunlight and microbial/dark treatments experience field conditions that help us understand the sequestration of C, S, N, and various metals, as well as the subsequent transformation processes of soil organic matter (SOM)?

Many researchers have employed various techniques to extract soil HS as individual fractions, i.e., humic acids (HA), fulvic acids (FA) and protein-like substances (PLS), as well as to assess the solubility of soil HS components using different extraction times ranging from 10 min to 24 h. For extracting HA and FA, the extraction methods used include the IHSS method, which involves 0.1 M HCl extraction followed by 0.1 M NaOH extraction46, a modified IHSS method47,48, or Na4P2O7 mixed NaOH extraction49, and the Nagoya University method, which involves alkaline extraction using Na2SO4 mixed with 0.1 M NaOH50,51. The solubility characteristic methods includes various extraction solutions such as pure water52,53,54, alkaline solutions54,55, acid solutions using 2.5 M and 13 M H2SO4 at 105°C56, saline solutions57, hot water (up to 100 °C) extraction58,59,60, various solid to liquid ratios61, sequential extraction62, and pH variation61. Among these extraction methods, water extracts at room temperature (or 25 °C) represent the labile components of SOM, while alkaline extracts represent the organo-mineral components of SOM influenced by environmental conditions and factors in soils and sands20,54,63. All other methods have been considered irrelevant to the environmental conditions and factors of the soil20,54,63. In particular, the use of individual water and alkaline soil extracts with inconsistent timescales remains a research gap, highlighting the need to develop a systematic, time-dependent extraction method that is highly efficient not only yielding soil components but also in reducing extraction time54. Moreover, solely relying on solid-phase analysis13,14,64,65,66 does not accurately represent or estimate the exact contributions of soil labile and organo-mineral fractions. Additionally, this approach fails to account for the continuous leaching or export of soil components through water and/or rainwater runoff from terrestrial soils into surrounding surface water and groundwater ecosystems. Based on this research gap, another key scientific question emerges: how can soils (treated and untreated) be analyzed to gain a better understanding of their relevant components that remain in labile (water-extractable) and organo-mineral (alkaline extracts) forms, while also maximizing the efficiency of soil extracts with a shorter extraction time21?

Within the framework of these research gaps on studying SOM dynamics under existing ambient conditions14,20,54, we propose three hypotheses. First, the photosynthetic uptake/sequestration of C, N, S, and various metals from soils or the atmosphere drives the growth and survival of soil microorganisms, which in turn leads to the generation of extracellular polymeric substances (EPS) and sulfur-mediated sequestration-redox processes (S2− ⇌ Sᵒ and Sᵒ ⇌ SO42− ⇌ S2−/S22− ⇌ SO42‒). Second, the degradation of soil components, including EPS, driven by photosynthetically mediated S sequestration-redox processes, is coherently involved in organo-mineral neoformation, which persists in soils to maintain steady-state SOC levels, as well as the continuous export, leaching, or emission of degradative labile soil components from terrestrial soils into surrounding environments. Third, specific time-dependent water extraction (targeting labile components), followed by alkaline extraction (targeting organo-mineral components), for any soil can respectively represent the biogeochemical lability and organo-mineral stability of soil components under various soil conditions and factors.

To implement these three hypotheses, the main objectives of this work are to: (i) develop a comprehensive methodological procedure to separately extract soil labile and organo-mineral components; (ii) analyze the contents of C, S, N, P, and their respective nutrient or ending derivatives, as well as seven redox (Fe, Mn, Cu, Hg, Ni, As, and U) and four non-redox metals (Al, Ca, Mg, and Sr), soil total nitrogen (STN), soil total sulfur (STS), SOC, and soil inorganic carbon (SIC), as well as their stable isotopes in both the original soil and the soil residue after extraction, after 150 consecutive days of exposure to selected sunlight, dark, and control conditions, in order to achieve a robust understanding of the integrated transformation-sequestration of soil mineral components; and (iii) investigate the continuous export of dissolved-phase components potentially important for ecosystem functions and processes and also their subsequent sequestration for SOC stabilization, analyze the soil samples for their dissolved organic carbon (DOC), dissolved inorganic carbon (DIC), SO42−, N-nutrients, P-nutrients, and dissolved SiO32− contents, plus the fluorescence properties of HS in both water extracts (labile state, LS) and alkali-extracts (complexed state, CS). To achieve the objectives outlined above, various experiments are carried out in both field and laboratory settings in a wide temperature range (10–42 °C and 25 °C, respectively) to assess the potential effects that would be induced by global warming.

Results

Carbon transformation

Soil total carbon (STC) fluctuates, either increasing or decreasing under sunlight, dark, and control samples over various timescales, showing the highest increase in sunlight (by 0.43%) and the lowest decrease in dark conditions (by 10.8%; Fig. 1A). This increase in STC primarily results from a rise in soil inorganic carbon (SIC) in the sunlight and control treatments (Fig. S1A), suggesting that inorganic carbon sequestration occurs through photosynthetic uptake of CO2 or DIC17,33,34,35. Moreover, the SOC decrease follows the order: dark > control > sunlight (Fig. 1B), which correspondingly reflects the DOCLS+CS mineralization in the same order under dark and control conditions (Fig. 1C, D). In particular, the highest level of mineral-associated DOCCS mineralization observed during the incubation period (by 10.0-35.2%), combined with a relatively low decrease in DOCLS (by 9.3–15.5%), results in the highest level of SOC (by 6.4–10.8%) mineralization in the dark samples (Fig. 1C, D). This finding suggests the occurrence of microbial SOC transformation20,21 via heterotrophic soil respiration6,9,10,25. Such changes would produce DICLS+CS or CO2, COS or H2S11,17,18, and/or carbonate minerals by silicate weathering67. These phenomena can be induced in the dark by various microbial communities that are highly active in the day-night experimental temperature range (10–42 °C)66,68. Similarly, SOC or DOC transformation further supports the extended and/or partial mineralization of labile and complexed state forms of protein-like substances or fulvic acids, and the labile form of humic acids, as shown by fluorescence data21 (Figs. S2-S3). The occurrence of carbonate mineral sequestration67 is supported by the individual SIC increase and significant DICLS+CS decrease, together with stable isotope of STC (ẟ13C-STC) enrichment (Fig. 1F and Fig. S1A-C). Remarkably, the gradual depletion of ẟ13C-SOC and the fluctuations of ẟ13C-DOC in both labile or complexed state forms in the dark (Fig. 1E, G,H) suggest that microbial SOC sequestration would occur via lighter DICLS+CS/CO2 uptake, after inorganic carbon production from SOC + DOCLS+CS mineralization17,34,35. This phenomenon often goes undetected when monitoring the SOC alone, because of the overwhelming role of mineralization. Similarly, in the control at constant temperature, a moderate level of SOC mineralization and/or respiration (2.3-6.0%) would account for relatively low microbial respiration by temperature-dependent specific microbial activities at the experimental temperature (25 °C)6,66,68. This phenomenon results in relatively low mineral-associated DOCCS mineralization by 15.0% over the 0-30-75-150 day periods, while simultaneously causing a significant decrease by 28.7% in labile DOCLS over the 0-30-75 day periods, which is reversed under dark conditions. Furthermore, the observed depletion of ẟ13C-SOC from ‒27.15‰ to ‒27.41‰ during the incubation period in the control samples (Fig. 1E) suggests that some degree of SOC sequestration, approximately a 0.4% increase, would occur in parallel (Fig. 1B). This happens because biologically generated SOC is coherently enriched in 12C, due to the lighter DIC and/or CO2 uptake caused by the SOC + DOCLS+CS mineralization. The highest degree of ẟ13C depletion (Figs. 1E) occurs under sunlight by photosynthetic uptake of lighter DIC or CO2. Such changes are further highlighted by the high ẟ13C-STC enrichment over 0–75 days, which remains primarily stable afterwards (Fig. 1F). Ideally, ẟ13C-STC enrichment would arise primarily from the combined effects of lighter SOC + DOCLS+CS mineralization and carbonate sequestration, with SIC increasing over individual timescale under all conditions (Fig. 1FSA). This finding agrees with the enrichment of ẟ13C-SIC and ẟ13C-SOC in soil depth-profiles elsewhere33. For instance, C respiration and sequestration processes can be induced by bacterial and fungal communities6,11,66,68 during both oxygenic and anoxygenic photosynthesis either through lichen-associated soil fungi plus cyanobacterial symbioses22,23,38,69 and/or non-symbiotic pathways70,71,72. Soil photosynthetic processes are further supported by the results from 100 soil samples collected across China, which demonstrated that lighter or depleted δ13C-SOC values (-27.4‰ to -22.5‰) correlated with higher levels of soil components, including Hg species, while enriched δ13C-SOC (-22.5‰ to -7.1‰) corresponded to lower concentrations of soil components and Hg fractions42.

Fig. 1
figure 1

Photoinduced and microbial changes of: soil total carbon (STC, A), soil organic carbon (SOC, B), dissolved organic carbon (DOC) concentration in labile state (DOCLS, C) and complexed state (DOCCS, D), stable isotope of soil organic carbon (δ13CSOC, E), stable isotope of soil total carbon (δ13CSTC, F), δ13C-DOCLS (G), δ13C-DOCCS (H), soil total sulfur (STS, I), SO42‒LS (J), SO42‒CS (K) and soil organic sulfur (SOS) + sulfides/pyrites (Sn2 (L) in extracts from soils subjected to sunlight, dark and control conditions measured at time 0 (original soil) and after 30, 75 and 150 consecutive experimental days. Error bars indicate the standard deviation among three replicates. The average values labeled with different letters (a, b, ab, c, d) at various time period (0, 30, 75, and 150 days) indicate significant differences (p < 0.05) among the three treated samples (sunlight, dark, and control at constant temperature), as determined by a one-way ANOVA.

Full size image

The SOC content varies only slightly under sunlight, showing an initial decrease (0–75 days) followed by a slight increase (+ 0.8% at 75–150 days; Fig. 1B), whereas SIC shows an initial increase and then decrease (Fig. S1A). Such changes would be associated with, respectively, a gradual increase in DOCLS and fluctuations in DOCCS plus DICLS+CS (Fig. 1D-D and Fig. S1B-C)7,8,14, and carbonate mineral neoformation by subsequent DICLS+CS uptake67. This finding is further supported by the net enrichment of ẟ13C-STC and ẟ13C-DOC in both labile and complexed state forms, due to the lighter12C export by SOC photorespiration. Notably, the gradual δ13C-SOC depletion from ‒27.15‰ to ‒27.99‰ during the incubation period (Fig. 1E) would likely arise from the uptake of depleted CO2 and/or DIC resulting from the photodegradation of DOCLS and DOCCS, which is further supported by the enrichment of the stable isotope values of δ13C-DOCLS and δ13C-DOCCS (Fig. 1G-H). This phenomenon consequently leads to continuous SOC photo-sequestration, which would offset photoinduced respiration, especially during the second half of the experiment. The substantial photoinduced C sequestration might be favored by the increased air moisture derived from intense precipitations (~ 240 mm), which occurred during the last 75 days of the experiment23,73. Remarkably, the simultaneous increase of photoinduced DOCLS and δ13C-DOCLS+CS (Fig. 1C, G,H) suggests a continuous breakdown of SOC into relatively small, water-soluble compounds. Simultaneously, an increase in the fluorescence signals of humic acids and fulvic acids is observed in the water extracts (Figs. S1, S2 and Supplementary Table S1). Similar phenomena might be predicted on a global scale7,8,17 in deforested lands, where sunlight-induced SOC transformation into DOM/DOCLS in the top soil might accelerate soil erosion15, decrease soil carbon storage11,12, and cause land degradation16 as global issues.

Most importantly, highly depleted δ13C-DOCLS values in three treatments (-29.00‰ to -28.68‰: Fig. 1D), when compared to the δ13C-DOCCS values (-26.94‰ to -26.37‰: Fig. 1D), likely indicate that the lighter DOCLS would be exported into the surrounding surface and groundwater, while the highly enriched DOCCS would remain in organo-minerals, contributing to C preservation and stabilization. In particular, this enriched carbon in DOCCS may be kinetically favorable for the neoformation of organo-minerals. Furthermore, it has been reported that the highly depleted carbon in DOCLS typically shows the higher molecular sizes and weights compared to DOCCS34. This finding suggests that the relatively low molecular weights and sizes of organic substances, along with their enriched C content, may promote the neoformation of organo-minerals due to their lower steric hindrance.

Sulfur transformation

The net soil total sulfur (STS) increase with fluctuations (1.2–41.0%) under all conditions (sunlight, dark, and control at constant temperature, Fig. 1I) is the consequence of the fluctuation of soil organic sulfur plus sulfides/pyrites (SOS + Sn2−) and of the different trends of SO42−, depending on the occurrence of the LS or CS forms (Fig. 1I-L). These changes suggest that the occurrence of sulfide-dependent photosynthetic S sequestration36 is taking place and is induced by, for instance, lichen-associated cyanobacterial-fungal symbioses through COS uptake74 via irreversible redox reactions (e.g., COS + H2O → 2H2S + CO2 → CH2O + 2Sᵒ + H2O)23,37,75,76. Importantly, soils simultaneously uptake and produce COS and CO2 or DIC, which occurs through the photochemical and microbial degradation of SOM76,2,6,77,78,79,80,81 as well as redox reactions in oxygen-limited soils82. However, carbonic anhydrase activity in microbial communities for soil COS uptake is characterized by an enzyme-catalysed rate of COS hydrolysis, which ranges from 0.002 to 0.02 s- 1 and is significantly faster on the shortest timescale compared to the enzyme-catalysed rate of CO2-H2O isotopic exchange (0.01 to 0.75 s- 1)83. The rapid carbonic anhydrase activity in soil microorganisms is primarily responsible for COS uptake rather than CO2.

Simultaneously, fluctuations in SOS + Sn2− and SO42−CS suggest the occurrence of sequestration-oxidation-reduction processes31,36,37 consisting in Sᵒ oxidation to SO42− by soil fungi and bacteria31, microbial SO42− reduction to sulfides (Sn2−)31, and S2−/Sn2− oxidation to SO42− by Fe3+ and/or O231. Specifically, the eight electrons released from the reactions S2‒ → SO + 2e‒ and SO → SO42‒ + 6e–31 would generate H2O2 and •OH radicals via O2•− intermediacy43,84, which would contribute to mineralize the SOC or HS-bound clay and Fe minerals14,30. Overall, the simultaneous occurrence of the redox reactions S2− ⇌ Sᵒ and Sᵒ ⇌ SO42− ⇌ S2−/S22− ⇌ SO42‒ would induce the concurrent sequestration of S and C, along with photochemical and microbial mineralization of SOC and HS-bound organo-minerals. Such S-mediated redox processes are further supported by the results from 100 soil samples collected across China, which showed that high soil S and SO42‒ (both labile and mineral-bound) contents correlated well with low concentrations of various soil components, including Hg species in both solid and liquid phases across all soil types, and vice versa42. The detailed quantitative changes in SOS + Sn2− and SO42−CS are discussed in the supplementary note 1.

Differently, the linear SO42−LS increase over time under all experimental conditions would primarily arise from the corresponding photo-microbial dissolution of SO42‒-containing minerals (e.g., clay)24,27. Remarkably, the higher SO42−LS production in the dark, compared to sunlight and control at constant temperature, would attribute to enhanced dark/microbial mineral dissolution by intense microbial activities resulting from the wide-ranging day-night temperature variation (10–42 °C). The significant correlation of SO42−LS with CaLS, MgLS, and SrLS (Fig. S4A-C; Table S2) suggests the formation of soluble sulfate forms that can be continuously exported into surface waters and groundwater.

Nitrogen transformation

The overall increase of soil total nitrogen (STN) under all conditions (Fig. 2A) likely reflects the increase of the different nitrogen (N) forms, including dissolved organic nitrogen (DON), NO3−-N, and NH4+-N (Fig. 2) in LS samples, while the same forms fluctuate in CS samples, except for NH4+-N under sunlight. Simultaneously, stable isotopes of STN (ẟ15N-STN) fluctuates and is depleted under sunlight exposure. An increase in STN would derive from fresh N sequestration38,65 by lichen-associated cyanobacterial-fungal symbioses85. Furthermore, the combination of STN increase and ẟ15N-STN depletion under sunlight might suggest a continuous photoinduced uptake of lighter ẟ14N-N2O (from − 1.8‰ to -2.2‰)/N2 (0‰)1,86. These processes are similar to the microbial depletion of ẟ15N-STN occurring by biological uptake of lighter ẟ14N-N2O or N2. Conversely, dark/microbial ẟ15N-STN enrichment would probably arise from lighter N-losses in non-photosynthetic microbial processes17.

Fig. 2
figure 2

Photoinduced (sunlight) and microbial (dark and control at constant temperature) changes of soil total nitrogen (STN, A); stable isotope of soil total nitrogen (δ15NSTN, B); dissolved organic nitrogen in labile state (LS) (DONLS, C); DON in complexed state (CS) (DONCS, D); NO3−-N in LS (NO3−-NLS, E); NO3−-N in CS (NO3−-NCS, F); NH4+-N in LS (NH4+-NLS, G); and NH4+-N in CS (NH4+-NCS, H) in extracts from soils subjected to sunlight, dark and control conditions measured at time 0 (original soil) and after 30, 75 and 150 consecutive experimental days. Error bars indicate the standard deviation among three replicates. The average values labeled with different letters (a, b, ab, c, d) at various time period (0, 30, 75, and 150 days) indicate significant differences (p < 0.05) among the three treated samples (sunlight, dark, and control at constant temperature), as determined by a one-way ANOVA.

Full size image

Denitrification is the ending step of DON mineralization in both LS and CS forms under photoinduced and microbial conditions1,17, which associates to NH4+-NLS+CS and NO3−-NLS+CS production (Fig. 2). An increase of DON in both LS and CS forms has been reported at soil-microbe interfaces65, due to continuous N uptake by Geobacter microbes1, whereas its decrease would be related to the production of NH4+-NLS+CS followed by photochemical and microbial nitrification-denitrification. Simultaneously, a strong correlation exists between DOCLS-S and NH3-NCS-S (Fig. S4D; Table S2). A decrease in NO3‒-N is occasionally observed, which might involve N2O emissions with ẟ15N depletion17. Remarkably, the significantly higher export of NO3-NLS in the dark, compared to control and sunlight conditions, would arise from microbial mineralization of DON-bound DOC in both LS and CS forms (Figs. 1C and D and 2E). Furthermore, DON mineralization in both LS and CS forms mostly associates with amino functionalities in protein-like substances, resulting in complete degradation in LS forms and partial degradation in CS forms (Figs. S2, S3). Overall, the continuous N uptake is a key factor for subsequent DON mineralization, which leads to the export of dissolved N-nutrients into surrounding environments.

Metals transformations

FeLS levels decrease under sunlight over the periods of 0–30 days and 75–150 days (by 70.2% and 26.2%, respectively), but levels increase by 202% over a period of 30–75 days (Fig. 3). In contrast, FeCS levels decrease by 61.3% over the first 0–30 days and then gradually increase by 63.7–86.0% over the 30-50-150 day periods. A similar increasing and decreasing trend in both FeLS and FeCS levels is observed under both dark and control conditions across various time intervals (Fig. 3). These trends of increasing and decreasing FeLS+CS levels at specific time intervals under sunlight, dark, and control conditions (Fig. 3) suggest a dual behavior of Fe. A first process would entail the dissolution of HS-bound clay and Fe minerals14, which concurrently promotes Fe redox properties by increasing the lability and solubility of Fe species25. The other process would involve photochemical and microbial neoformation of secondary minerals, e.g. DOM or HS-bound clay and Fe14,24,30. The latter components might be formed by organo-mineral complexation of HS functionalities with Fe d-orbitals via π–d interactions30. These processes would subsequently lead to SOC stabilization and sequestration by formation of soil organo-mineral complexes13,14,40. Similar dual roles of Fe have been observed in the transformation of C and N in previous studies43,87,88,89. In principle, a strong Fe3+ complexation behavior is typically due to stabilization of the π–d electron bonding system in Fe–DOM complexes through the donation of electrons from HS functionalities into the Fe d-orbitals30 (Zhang et al. 2023) by the lower energy of the high-spin state, which has one electron in each of the five d-orbitals of 23Fe3+ ([Ar]183d54s0), due to its greater electron-nucleus attraction90. This property allows Fe3+ ions to occur to a higher extent than other metals in larger molecular DOM fractions30,91. Simultaneously, dark mineralization occurring at 10–42 °C due to various microbial activities66,68 would lead to the highest SOC loss in dark conditions (by 10.8% decrease) compared to control samples (by 6.0% decrease) and to a decrease of organo-mineral stability, thereby enhancing surface soil erosion15,16.

Fig. 3
figure 3

Changes of FeLS+CS, MnLS+CS, CuLS+CS, HgLS+CS, NiLS+CS and AsLS+CS in water-extracts (labile state, LS) and alkali-extracts (complexed state, CS) from soils subjected to sunlight, dark and control conditions measured at time 0 (original soil) and after 30, 75 and 150 consecutive experimental days. Error bars indicate the standard deviation among three replicates. The average values labeled with different letters (a, b, ab, c, d) at various time period (0, 30, 75, and 150 days) indicate significant differences (p < 0.05) among the three treated samples (sunlight, dark, and control at constant temperature), as determined by a one-way ANOVA.

Full size image

Furthermore, the trends (increase or decrease) of metals, both redox (MnLS+CS, CuLS+CS, HgLS+CS, NiLS+CS, AsCS, and UCS−S) and non-redox ones (AlCS, and MgCS) in specific timescales under sunlight (Fig. 3) suggests the occurrence of phenomena similar to those hypothesized for Fe, namely photoinduced and microbial change of redox properties together with increased lability or solubility and/or sequestration or neoformation of minerals and organo-minerals (see also supplementary Text 1–2)13,14,24,25,30,40,43. The latter phenomenon (neoformation of water-insoluble Fe minerals that include other metals as well) might also account for the observed ULS transformation24, which would also involve formation of U-phosphate minerals92 (see also supplementary Text 2). Such U-bound mineral formation could be attributed to a significant decrease in ULS over the first 0–30 days (by 39.0%) in dark samples, and over 30-75-150 days in both sunlight and control samples (by 16.0–40% and 8.4–42.0%, respectively). Additionally, a significant gradual increase in UCS over the 0-30-75-150 day periods (44.0-66.0-84.7%) is observed in sunlight samples. In essence, U mineral neoformation would occur when U forms complexes with Fe oxyhydroxides through the hexavalent U species (UO22+) (Fortier and Hayton 2010)93,94, presumably via surface complexation or co-precipitation. Therefore, ULS would be relatively less labile in dark and control conditions due to its strong complexation behavior in the form of water-insoluble U-complexes such as uraninite-UO2/U-phosphate minerals, whereas UCS shows a high photoinduced lability, but relatively low microbial lability in dark and control conditions.

Most importantly, AlCS levels decrease over a period of 75–150 days under sunlight, dark, and control conditions (by 7.2%, 15.3%, and 9.6%, respectively), with the simultaneous decrease of FeLS+CS, MnLS, CuCS, NiCS, MgCS, AsLS+CS, and SiO32−CS, which would cause by the photochemical (sunlight samples) and microbial (dark and control samples) neoformation of Fe secondary silicate minerals (Figs. 3 and 4)24,95,96. These occur via the degradative by-products of HS components, as well as photosynthetically-derived SOC that generates HS components, which are detected under both sunlight and control conditions. These processes might be due to the high AlCS occurrence that primarily promotes SOC stabilization and storage13,97,98.

Fig. 4
figure 4

Photoinduced and microbial changes of: total Hg, ULS+CS, AlLS+CS, CaLS+CS, MgLS+CS and SrLS+CS in water-extracts (LS) and alkali-extracts (CS) from soils subjected to sunlight, dark and control conditions measured at time 0 (original soil) and after 30, 75 and 150 consecutive experimental days. Error bars indicate the standard deviations among three replicates. The average values labeled with different letters (a, b, ab, c, d) at various time period (0, 30, 75, and 150 days) indicate significant differences (p < 0.05) among the three treated samples (sunlight, dark, and control at constant temperature), as determined by a one-way ANOVA.

Full size image

Remarkably, among the redox metals, MnLS levels undergo the most significant increase by 720–890% over the 0-30-75 day periods under sunlight (Fig. 3). This increase coincides with a rise in MnCS levels (by 54.0-71.8%) over the 30-75-150 day periods. This phenomenon might be attributed to thermodynamic instability of Mn(III) due to its unfilled d-orbitals (22Mn3+ = [Ar]183d44s0), which is stable only in specific soluble organic complexes or minerals; and often disproportionates into soluble Mn(II) and insoluble Mn(IV)12,99 (Tebo et al., 2004; Li et al. 2024; Wang et al. 2022) whereas Mn(II) is electronically more stable featuring five half-filled d-orbitals (23Mn2+ = [Ar]183d54s0). Moreover, the microbial (dark and control at constant temperature) MnLS+CS content is primarily regulated by SOC respiration and mineralization via S sequestration-redox reactions along with Mn2+ oxidation into Mn(III, IV), with neoformation of Mn minerals99,100. This process subsequently leads to lower MnLS+CS levels under both dark and control conditions, with the most significant decrease occurring over the first 0–30 days (by 38.5%) in dark conditions and over a period of 30–75 days (by 34.3%) in control samples. However, the simultaneous reductive dissolution of MnO2 via redox processes (Sᵒ ⇌ SO42‒ ⇌ S2‒/S22‒ ⇌ SO42‒) leads to increased Mn2 + LS+CS levels.

Similar trends as those observed in MnLS+CS and FeCS are also evident in CuLS+CS and NiLS+CS (Figs. 3 and 4), which leads to a continual export of Cu + Ni by photodissolution14,96 and to the neoformation of HS-bound clay and Fe minerals13,14,24. A significant increase in CuLS levels is observed over the 0-30-75-150 day periods in sunlight samples, with increases of 156%, 20.4%, and 23.1%, respectively, and over the first 0–30 days (by 25.3%) in dark samples. Furthermore, CuCS levels significantly increase over 30–75 days (by 6.7%) in sunlight samples and over the first 0–30 days (by 13.0%) in control samples, while a significant decrease in CuCS levels is observed over the 30-75-150 day periods, with decreases of 15.1% and 8.1%, respectively, in dark samples. Ideally, the high CuLS+CS increase would be caused by photoreduction of Cu2+ to Cu+ (Cu2+ + e− Δ Cu+), i.e., a photoredox cycling101 involving sulfur sequestration-redox cycle. Differently, Ni2+ undergoes one-electron redox processes to generate the Ni+ ion that is highly reactive due to its outer-shell, one-electron unfilled d-orbital (27Ni+= [Ar]183d94s0). In this respect, Ni+ is similar to Cu+ and this determines the involvement of Ni+ in a dinuclear Ni-bonding system (Ni-Ni) or Ni-organic compounds102. This hypothesis is further supported by a similar increasing trend in NiLS over the 0-30-75-150 day periods, with increases of 64.4%, 9.1%, and 14.0%, respectively, in sunlight samples, and a significant decrease in NiCS over the first 0–30 days in dark samples.

Remarkably, the high occurrence of HgCS (approximately, 25% of HgT) and its subsequent transformation/emission under all conditions leads to a decrease of HgT. For instance, up to 15% and 10% of HgT loss is observed under sunlight and dark conditions, respectively (Figs. 3 and 4). Such photochemical and/or microbial HgT losses would presumably occur by reduction of Hg2 + CS and Hg+ CS to highly volatile Hgᵒ103. Moreover, HgLS levels decrease over the 30-75-150 day periods (by 52.6% and 12.0%, respectively) in dark samples, while HgCS levels decrease over the first 0–30 days (by 46.0%) in dark conditions and over a period of 75–150 days in sunlight and control samples (by 21.0% and 22.8%, respectively).Such HgLS+CS decrease under all conditions likely derives from HgS precipitation26,104 triggered by the reduction of SO42−CS to S2− (Fig. 1). Simultaneously, HgLS levels increase (by 137%) over the first 0–30 days in dark samples and over the 0-30-75-150 day periods (by 139%, 11.2%, and 55.6%, respectively) in sunlight samples. This finding suggests that DOCLS+CS containing Hg-DOM complexes would undergo mineralization, leading to a high export of HgLS and volatilization of Hg as Hgᵒ, thereby decreasing HgT over the first 0–30 days. Similarly, HgCS levels increase under dark conditions over a period of 30–75 days (by 62.7%) and over the 0-30-75 day periods (by 6.7% and 9.6%, respectively) in the sunlight samples, which would be ascribed either to the formation of Hg-thiol complexes with degradative HS functionalities105 or to HgS precipitation by SO42−CS − D75 increase (3036%) via Sᵒ Δ SO42‒ ⇌ S2‒ ⇌ Sᵒ (Fig. 1). Additionally, an increase of HgCS would also suggest its sequestration by photosynthetically-derived SOC and/or HS functionalities, via Hg-thiols complexes (Fig. 4)26,44, where newly-derived SOC and HS components are typically generated through photosynthesis in soil and sand environments34,35. The sequestration of HgCS via Hg-thiol complexes over a period of 30–75 days is accompanied by a simultaneous increase in SOS + Sn2− (Fig. 1). Such processes result in a decrease in HgLS levels over 30–150 days and simultaneously involve the mineralization of DOCCS under dark conditions (Figs. 3 and 4), which results in the reduction of DOM-Hg in HgCS to Hgᵒ54, thereby decreasing HgT.

The consistently higher export of AsLS is observed in the dark over the 0-30-75-150 day periods (increase by 33.8%, 24.2%, and 14.2%, respectively) compared to sunlight samples (increase by 25.5%, decrease by 13.1%, and increase by 12.2%, respectively) and control samples (increase by 17.4% and 18.0%, and decrease by 6.2%, respectively) (Figs. 3 and 4). This phenomenon is attributed to the intense dissolution and respiration of HS-bound clay and Fe minerals, which are induced by the wide day-night temperature variation at TD10−42ᵒC and are facilitated by diverse fungal-bacterial communities that are highly active under these conditions66,68. This process involves high SOC and/or DOCLS+CS mineralization and absence of photosynthetically-derived SOC, which is further supported by a significant correlation between AsLS and SOC, or DOCLS, or DOCCS (Fig. S4E-F). This phenomenon might be a key worldwide signature for the increasingly high As export from soil due to mineral dissolution, a consequence of elevated temperatures induced by climate warming28,29. Differently, a decrease in AsLS, as previously mentioned, and AsCS over the 30-75-150 day periods in sunlight samples (by 34.4% and 5.7%, respectively) and dark samples (by 19.0%and 7.6%, respectively), as well as over the first 0–30 days (by 24.4%) in control samples would be associated with parallel decreases of FeCS, MnCS, AlCS, and MgCS (Figs. 3). This phenomenon would imply the possible involvement of these elements in the neoformation of As-containing secondary minerals via either degradative HS byproducts or photosynthetically-derived SOC and HS components (Figs. 1 and 3)43,106. Ideally, electron transfer from the S sequestration-redox cycle would facilitate the oxidation of arsenite (AsO33‒) to arsenate (AsO43‒)107, leading to As sequestration or immobilization through the neoformation of secondary minerals involving AlCS, FeLS, MnLS, and/or MgCS24,40,106,108. The outcome of these processes result in a decrease of these metals, including AsLS+CS, over the 30-75-150 day periods (Figs. 3 and 4). Remarkably, the higher electronic stability of AsO43‒ (28As5+ = [Ar]183d104s04p0) compared to AsO33‒ (30As3+ = [Ar]183d104s24p0) would result in the formation of more stable arsenate complexes with various Fe-oxides under alkaline conditions109,110. Ideally, the pH-dependent, negatively-charged arsenite/H2AsO3‒ plus arsenate/H2AsO4‒ would form strong complexes with the d-orbitals of Fe-minerals or oxides, and also with other metals106,108,111, via π–d electron bonding systems similar to HS functionalities (e.g., COO‒)30,41. This would result in the simultaneous association of AsO43‒, AsO33‒, and HS components with various oxides or hydroxides of Fe and other metals30,40,41,106,110,112.

Gradual increases in CaLS, MgLS, and SrLS over the 0-30-75-150 day periods under sunlight conditions (up to 47.8%, 15.3%, and 53.7%, respectively), dark conditions (up to 47.0%, 34.5%, and 45.0%, respectively), and control conditions (up to 19.6%, 17.1%, and 21.0%, respectively) are observed, with the sole exception of a MgLS decrease by 7.0% over a period of 30–75 days in dark samples. These findings would arise from lability and water-solubility of these cations that originate from the dissolution of the corresponding sulfate and clay minerals upon photochemical and microbial reactions27,96,113,114,115. In particular, Mg partially derives from the dissolution of chromite [(Fe, Mg)Cr(III)2O4]116. Differently, CaCS and SrCS show a similar behavior, with slightly decreasing-increasing trends over the 0-30-75-150 day periods under sunlight, dark, and control conditions. Their relatively low input and/or transformation might be due both to their low occurrence in clay and silicate minerals together with the predominant occurrence of Si, Al, and Fe96,113,115, and to the strong immobilization of Sr in clay minerals117,118,119. However, MgCS levels decrease over the first 0–30 days, both in sunlight and in dark (by 64.4% and 46.0%, respectively), over a period of 30–75 in control at constant temperature (by 50.6%) and over a period of 75–150 days in sunlight (by 26.8%) samples, while MgCS levels increase over the first 0–30 days in control at constant temperature (by 31.8%), over a period of 30–75 days in sunlight (by 157%) and over the 30-75-150 day periods in dark (by 39.0% and 27.8%, respectively) (Fig. 4). These phenomena presumably arise from the immobilized nature of Mg bound to various silicate interlayers120 and also from its strong involvement in the neoformation or sequestration of clay minerals24,30. Therefore, the MgCS decrease under individual treatments would accompany the corresponding decrease of AlLS+CS, FeLS+CS, MnLS+CS, CuCS, and NiLS+CS, due to formation of clay minerals through co-precipitation of degradative HS byproducts or photosynthetically-derived SOC14,24,34,96,121,. Lastly, the CaLS and MgLS contents are 1.8- and 8.2-fold higher than, respectively, CaCS and MgCS, and nearly equivalent to SrLS with SrCS, which implies that environmental factors (e.g. sunlight and microbes) dissolve primarily various Ca, Mg, and Sr minerals whereas these elements occur preferentially in LS rather than CS-forms. In turn, this issue affects the lability and water-solubility of these three essential cations, constantly leaching them from soil to the surrounding ecosystems. The individual metal behaviors are discussed in the supplementary note 4.

Discussion

Transformation and sequestration of soil organo-mineral components

The STS increase might be ascribed to lichen-associated cyanobacterial-fungal symbioses74, as well as to non-symbiotic pathways70,71,72. In the former case, photosynthetic COS uptake for symbiotic growth22,23,37 would occur at the microbe-mineral interface14,65. These processes would either increase or decrease the levels of SO42−CS + SOS + Sn2− (Figs. 1J, K,L), via sequential S2− ⇌ Sᵒ and/or Sᵒ ⇌ SO42− ⇌ S2−/S22− ⇌ SO42‒ transformations31,36. The S2− ⇌ Sᵒ plus Sᵒ ⇌ SO42− conversion generates eight electrons, whereas high-valence labile redox metals coherently accept electrons to produce reduced low-valence forms, which can generate reactive oxygen species via Fenton/photo-Fenton reactions. The latter process would then induce SOC or DOCLS+CS mineralization in minerals14. In return, the produced high-valence metals would cause neoformation or precipitation of minerals13,24,25,40,103. Such transformations involve the simultaneous dissolution and/or neoformation of HS-bound clay, and Fe minerals, through complexation/co-precipitation of HS or photosynthetically-derived SOC and HS13,14,20,24,30. These processes might also determine the export of water-soluble LS-forms (e.g. DOM, nutrients, and trace elements) to the surrounding ecosystems, which is depicted in a conceptual model (Fig. 5).

Fig. 5
figure 5

A conceptual model of the simultaneous photosynthetic sulfur/COS acquisition or uptake via S2− ⇌ Sᵒ and the subsequent oxidation-reduction via Sᵒ ⇌ SO42− ⇌ S2−/S22− ⇌ SO42‒ by fungi and cyanobacterial symbioses, showing the net S sequestration-redox in the corresponding transformation of soil organo-mineral components at mineral-microbe interfaces. These processes lead to dissolved-phase LS exports of dissolved organic matter (DOM)/humic substances (HS), nutrients, SO42−, COS, cations and anions together with the subsequent sequestration/neoformation of minerals by uptake of various components of labile state (LS) and complexed states (CS) forms.

Full size image

DICLS contents under sunlight appear to increase simultaneously with a net increase of DOM, HS fluorescence, nutrients, and redox/non-redox metals (Figs. 1, 2 and 3 and Figs. S3-S5), suggesting future increases of DIC in soils and waters under changing climatic conditions, which is a phenomenon similar to seawater carbonate dissolution. The export of various LS-components, particularly DOC, HA and FA, nutrients, redox and non-redox metals, and DIC, as well as CO2, COS or H2S emissions (Figs. 1 and 3 and Figs. S2-S5)11,17,18 might occur in deforested land ecosystems due to enhanced exposure of soil to sunlight. Moreover, the initial DICLS decreasing rate under dark conditions over the first 0–30 days was 3.46 ± 1.15 mg kg− 1day− 1, which was higher than the rate observed for the subsequent 30-75-150 day periods (0.65 ± 0.56 and 1.14 ± 0.40 mg kg− 1day− 1, respectively). Similarly, the initial DICCS decreasing rate over the first 0–30 days was 4.74 ± 0.84 mg kg− 1day− 1, which was higher than for the rate observed for the subsequent 75–150 days (1.54 ± 0.16 mg kg− 1day− 1) (Fig. S1). In contrast, the relatively low decreasing rates of DICLS under control conditions over the 0-30-75 day periods (then compared to the dark samples) were 2.92 ± 0.73 and 0.49 ± 0.42 mg kg− 1day− 1, respectively. These phenomena under dark and control conditions would suggest the occurrence of two processes: (a) DIC uptake by photosynthetically-derived SOC22,23,34,35,37, and/or silicate mineral weathering67,122, both of which would be associated with the neoformation of clay, Fe or carbonate minerals; and (b) atmospheric emissions of DIC as CO280,11,122.

Additionally, high soil respiration under dark conditions (10.0–42ᵒC) is primarily responsible for high As export (Fig. 3)14,28 and high SOC/HS mineralization. This process leads to a decrease in SOC sequestration and its organo-mineral stability, thereby enhancing surface soil erosion12,13. The increasing temperatures associated with climate warming, particularly under dark conditions, significantly elevate the risks of As intrusion in both surface water and groundwater. Moreover, lichen-associated cyanobacterial-fungal symbioses would lead to neoformation of SOC- or HS-bound clay, Fe, and silicate minerals14,24,25,30,40,43. Such mineral formation would cause the sequestration of metals (e.g. Fe) and various anions, including nutrients (Fig. 2 and Fig. S5), while Hg decrease would result from the reduction of DOM-Hg complexes to volatile Hgᵒ103,105, thereby increasing the risks of Hg exports into the surface water and groundwater, as well as its emissions into the atmosphere. Most importantly, the high AlCS presence (approximately 70 times higher than AlLS and 25 times higher than FeCS in untreated soil) suggests the fundamental role of clay or silicate minerals and/or organo-mineral complexes in the stabilization or storage of soil C97,123.

Remarkably, microbial SOC mineralization, phototrophic carbon fixation, S redox processes, mineral neoformation, microbial biomass, and respiration processes have been discussed here based on the observed increases and decreases in solid-phase and liquid-phase soil components, along with references from previous studies, because direct evidence from this study is lacking. These soil biogeochemical processes should be emphasized in future research.

Conclusions

The assessment of photochemical and microbial interactions at 10–42 °C and 25 °C provides useful information on the simultaneous transformations of soil biogenic and non-biogenic elements derived from mineral dissolution and neoformation. In particular, mineral dissolution can lead to the release of water-soluble species, including DOM, DIC, HS components, nutrients, and both redox and non-redox metals/elements, such as carcinogenic As and Hg, into the surrounding environments-namely surface water, groundwater, and the atmosphere. The key findings about the various features of mineralization-sequestration of clay, and Fe mineral components are summarized below. First, among redox metals, Fe always plays a dual role, i.e., organo-mineral complexation13,24,43 due to its five half-filled Fe d-orbitals30 and redox behavior14,25.

Second, the reduction of DOM-Hg in HgCS to volatile Hgᵒ103,124 would lead to a net Hg decrease under sunlight, dark and control conditions over the 0-30-75-150 day periods (a decrease by 9.7–14.8%, 4.98–10.2%, and 3.52–10.4%, respectively), which would indicate the continuous Hgᵒ emission from degraded and dry-affected lands. This process poses significant human health risks, either directly or through microbial food webs.

Third, the consistently higher AsLS export in dark conditions over the 0-30-75-150 day periods (6.6–55.2% and 14.0-46.1%, respectively) compared to sunlight and control conditions (Fig. 3) would be due to a continuous and high dissolution or respiration of HS-bound clay and Fe minerals induced by the wide day-night temperature variation at TD10−42ᵒC and operated by diverse fungal-bacterial communities that are highly active in these conditions66,68,125. This process involves high SOC and DOCLS+CS mineralization and absence of photosynthetically-derived SOC, which might be key signatures for the increasingly high As export from soil caused by mineral dissolution worldwide as a consequence of increased temperatures28,29,40 due to climate warming. This process also presents significant risks to human health, either directly or through microbial food webs.

Fourth, the high AlCS presence, approximately 70-fold higher than AlLS and 25-fold higher than FeCS in untreated soil, would suggest the fundamental role of clay or silicate minerals and/or organo-mineral complexes in stabilization and storage of soil carbon14,97,126,127.

Fifth, photorespiration appears to increase DICLS concentration at the same time as it causes a net increase of DOM, HS fluorescence, nutrients, redox and non-redox elements (Figs. 1, 2, 3 and 4 and Figs. S3-S5). These findings suggest a potential increase in the dissolution behavior of DIC in soils and waters under changing climatic conditions, a phenomenon akin to seawater carbonate dissolution.

Sixth, the photosynthetic uptake of C17,31,32,33,34, N1,2,37,38,134, S35,36, and various other elements4,24,25,30,40,41,42 is a key process driving increases in SOC levels through photosynthetically-derived EPS or HS components produced by soil microorganism21,22,34,35,42,43,73,133. This process, i.e., increases in SOC, counterbalances simultaneous degradative losses of EPS and other soil components through photosynthetically driven, sulfur-mediated sequestration-redox processes14,31,134,135. Increasing SOC persists in soils through organo-mineral neoformation involving EPS or HS components, which help maintain steady-state SOC levels, while degradative loses of labile soil components may continuously be exported, leached, or emitted from terrestrial soils into surrounding environments.

Seven, fluorescence studies of HS components and other soil components54 have confirmed that the extraction time used for two sequential water extractions (6 h + 1 h), followed by two sequential extractions (3 h + 3 h), is essential for the effective dissolution of both water- and alkali-extractable soil components. Reducing the time may result in incomplete dissolution, while extending it unnecessarily might partially alter discrete or functionally distinct components. Using this extraction procedure, the total SOC values ranged from 1.41 to 5.89% in water extracts and from 5.83 to 9.52% in alkali extracts from both raw and treated soils. These results demonstrate a substantial yield of SOC, which is critical for understanding the biogeochemical characteristics, functions, and influencing factors of various soils and sands30,34,35,42. This study, along with recent publications30,34,35,42, highlights the effectiveness and potential of this extraction technique.

Finally, the results of the experiments conducted on soil samples within the day-night experimental temperature range (10–42 °C) under both sunlight and dark conditions, in comparison with the control at constant temperature (25 °C), are reasonably expected to enhance our understanding of certain aspects of ongoing climate warming processes. In particular, sunlight-induced exports of redox metals and DOM are expected to affect soil worldwide by increasing erosion and will also degrade the soil and water ecosystem and their services, which remain a crucial issue of immense global ecological and economic significance.

Materials and methods

Samples

Approximately 5 kg soil, classified as haplic luvisol, free of plant litter and debris, were collected in black airtight polypropylene bags from the A-horizon (0–30 cm) of the Panshan deciduous forest floor, Tianjin, China (N40°00′990″, E117°15′804″). The forest site is located very close to a stream between two mountains and is covered mostly by a deciduous man-made forest densely populated by Pine trees21. The collected soil in the sampling site is undisturbed and remains under natural conditions, as reported elsewhere21. Soil samples were then 2-mm sieved and mixed homogeneously, to generate a sample that is spatially representative at the field scale.

Experimental protocols

Three experimental protocols, i.e., sunlight, dark, and control at constant temperature were conducted on three aliquots of about 50 g each of the homogeneously mixed 2-mm sieved soil during the warm season, from May 5th to October 1st, 2018, for 150 days. In particular, the ‘sunlight’ experiment aimed at evaluating the photorespiration and/or photodegradation with photosynthetically-derived SOC sequestration of soil organo-mineral components, and was conducted on soil samples placed in 500-mL quartz bottles. Notably, the quartz bottle measures exactly 19.0 cm in length and 7.5 cm in width, with a mouth opening of ~ 4.3 cm and a neck length of approximately 4.0 cm. Each bottle’s mouth was covered by parafilm with small holes to allow for oxygen exchange, and the bottles were placed under natural sunlight conditions on the roof of a building of Tianjin University, Tianjin, China. Quartz bottles were used to avoid the blockage of ultraviolet (UV) wavelengths by glass and were placed on a strong support at opposite angles of 45 degrees, with the mouth positioned downward to avoid the effect of any precipitation or contamination, taking advantage of their long-necked designed.

The ‘dark’ experiment, which aimed at evaluating the microbial respiration and/or biodegradation in the same system, was conducted on soil samples put in 500-mL glass bottles and placed in the same site indicated above, with the mouth loosely bound to allow for limited gas exchange and covered with aluminum foil to avoid sunlight exposure. In particular, the 500 mL amber bottles were specifically designed in consultation with company representatives and ordered for this experimental analysis. These bottles feature a brown color and a wide mouth, while the cap-type stoppers are made of light-resistant black plastic. Notably, the glass bottle measures exactly 14.0 cm in length and 8.0 cm in width, with a mouth opening of ~ 4.3 cm and a neck length of approximately 2.0 cm. Aluminum foil does not cover the caps/mouth of the bottles and is loosely bound to allow for adequate air circulation; therefore, heat retention does not occur in the experimental bottles. In these conditions the soil samples were subject to the effects of day-night temperature differences that can activate different fungi-bacteria, which are primarily responsible for respiration influenced by climatic change128. Finally, the ‘control at constant temperature’ experiment aimed at assessing the differences in microbial responses at a fixed temperature compared to day-night cycles, and was conducted in the laboratory. The soil samples were placed in identical 500-mL brown amber glass bottles and kept in a thermostat maintained at a constant temperature of 25ᵒC. The plastic caps were loosely sealed to facilitate air exchange with the soil samples in the thermostat. All ‘control at constant temperature’ bottles were removed from the thermostat, and the caps were completely opened for a moment to allow for adequate air circulation during the middle stage of each experimental interval before being placed back in the thermostat. Three replicate samples under sunlight, dark, and control conditions are applied for each time point: 0 days, 30 days, 75 days, and 150 days of incubation.

In summary, a ‘dark’ sample or experiment can be defined as a soil that has been subjected to subsequent microbial transformation (mineralization and sequestration) of soil organic matter (SOM), and has experienced fluctuations in temperature between day and night. These conditions can be compared to natural soil environments influenced by dark and microbial activities. In contrast, a ‘control at constant temperature’ sample or experiment refers to a soil that has also been subjected to subsequent microbial transformation (mineralization and sequestration), but is maintained at a constant temperature of 25ᵒC under controlled microbial and dark conditions in a thermostat, which allows for a comparison of its transformations with those of dark samples.

A ‘sunlight’ sample or experiment can be defined as a soil that has been subjected to subsequent photochemical transformation (mineralization and sequestration) of SOM, while being exposed to sunlight during the daytime. These conditions can be compared to natural soil environments affected by daytime sunlight exposure on deforested land, as well as on various surface soils. The objective of this ‘sunlight’ experiment is to evaluate the photorespiration and/or photodegradation in relation to photosynthetically-derived SOC sequestration at the soil surface.

During the ‘sunlight’ and ‘dark’ experiments, the highest air temperature recorded was 41.8 °C in July 2018, based on diurnal temperature data collected near the experimental site136. The lowest experimental temperature, 10 °C, was obtained from monthly meteorological data recorded in Tianjin in September. Therefore, the air temperature ranged from 10.0 to 41.8 ᵒC with significant monthly variations from May to September, with June, July, and August as the hottest months on average137 (Source: www.tianqi.com). The solar intensity during the experimental period was very variable as well, ranging from 231.2 Wh m−2 hr−1 in September to 262.4 Wh m−2 hr−1 in June127. Precipitation was also unevenly distributed, with the highest values in July (232 mm) and August (~ 190 mm) and the lowest in September (~ 8 mm)128. Note that soil samples were not directly exposed to rain, but humid air could come in contact with the studied soils. Water was not added to the samples because the experiments were focused on the impact of dry conditions under sunlight and temperature variations on the surface soil. However, an increased humidity during the rainy seasons was contributed in the experimental protocols. To prevent the direct reaching of rainwater to sunlight samples, the corresponding bottles are specifically configured as previously described. Notably, the direct moisture influence on the experimental soils are mitigated by using parafilm with small holes to cover the mouth of each bottle. This design also facilitates air exchange within the experimental bottles, as previously mentioned. In contrast, dark and control bottles are loosely sealed with plastic caps as mentioned previously, allowing air moisture to remain in a saturation state within the treated soil samples.

The pH of raw forest soil is measured at 7.20 ± 0.03 and it typically increases across all three treatments, varying from 7.29 ± 0.07 to 7.38 ± 0.07 in sunlight conditions, from 7.25 ± 0.09 to 7.32 ± 0.02 in dark conditions, and from 7.28 ± 0.12 to 7.40 ± 0.03 in control conditions (Fig. S1D). Similarly, the electric conductivity (EC) of raw forest soil is measured at 105 µS/cm, and it typically increases across all three treatments, showing the highest increase of 51.3% in the dark, 29.4% in sunlight, and 22.8% in control samples (Fig. S1E).

Three soil sample replicates were collected consecutively after 30, 75, and 150 days, oven-dried at 40 °C for 24 h, and then pestle-ground at a < 0.2 mm size in a mortar for further analyses.

Extraction protocol

The organo-mineral components were isolated from soil samples by extraction with water followed by a 0.1 M NaOH solution. The water extract (We) was obtained by a number of subsequent steps (Fig. S6). In the first step, ultrapure water (18.2 MΩ·cm, Mill-Q, Millipore) was added to the ground and 0.2-mm-sieved soil sample at a soil/water ratio of 1:10, the mixture was vortexed for 1 min in closed 500-mL brown bottles and then shaken for 6 h at 25 °C. The mixture was then centrifuged for 20 min at 4000 rpm using a Thermo Fisher Scientific SORVALL ST 16 centrifuge to remove suspended solids. The supernatant solution was then filtered through a 0.45-µm membrane filter (GF/F type, Shanghai Xin Ya Purification Equipment Co. Ltd, China), whereas the remaining solid residue was extracted again with fresh ultrapure water for 1 h, and the procedure described above applied again to obtain another supernatant solution that was mixed with the previous one and stored in freezer at − 20 °C until further processing.

The alkaline extract (Ae) under N2 was obtained by adding a 0.1 M NaOH solution to the corresponding We soil residue at a soil residue/alkaline solution ratio of 1:10, shaking the mixture for 3 h at 25 °C, and then centrifuging it as described above. The supernatant solution was filtered through a 0.45-µm membrane filter (polytetrafluoroethylene membrane, PTFE, Shanghai Xin Ya Purification Equipment Co. Ltd, China), whereas the remaining solid residue was extracted again with a fresh alkaline solution for 3 h and, after applying again the above procedure, a further supernatant solution was obtained that was mixed with the previous one and stored in freezer at − 20 °C until further processing. The We and Ae samples provide the water-soluble labile state (LS) and water-insoluble complexed state (CS) fractions of soil organo-mineral components20,21. They are expected to provide relevant information on the biogeochemical phenomena, mechanisms, and origin of soil organo-mineral components, DOM, HS, nutrients, and metals. Based on the extraction yield and timing, the contributions of We (water extracts) and Ae (alkali extracts) to the total SOC in raw soils were approximately 1.86% and 8.06%, respectively. These contributions were relatively higher for sunlight-exposed samples, ranging from 3.63% to 5.89% for water extracts and 8.20% to 9.52% for alkali extracts, and relatively lower for microbial (dark and control at constant temperature) samples, ranging from 1.41% to 1.80% and 5.83% to 8.79%, respectively.

Sample analyses

Soil total carbon (STC), soil organic carbon (SOC), N, and S contents were measured by a Vario EL Cube Elemental Analyzer (Elementar VarioEL III, Germany) as reported previously21, on approximately 20 mg aliquots of dried, ground, and homogenized soil sample contained in a clean, carbon-free, pre-combusted tin boat placed on an autosampler rack assembly. In the case of SOC measurement, the dried, ground, and homogenized soil samples were pretreated with 1 M HCl in a vial to remove carbonates52. Sulfanilamide was used as standard after every 10 measurements.

The stable isotopes13C and15N were analyzed on 2.0 mg and 30.0 mg, respectively, of dried soil samples contained in tin boats placed on the autosampler rack assembly of the Elemental Analyser (Flash 2000 HT) interfaced with Stable Carbon Isotope Ratio Mass Spectrometer (MAT 253 Plus, ThermoFisher, USA). The IAEA-600 Caffeine was used as the13C standard, whereas for15N measurements, both IAEA-600 Caffeine and USGS 40 plus 41 L-Glutamic Acid were used as standards.

The dissolved organic carbon (DOC) content in soil solutions was measured in triplicate using an Aurora combustion total organic carbon (TOC) analyzer (OI Analytical Aurora, Model 1030, USA), according to a procedure reported previously21. Soil particle sizes were determined by the hydrometer method using a Master sizer 3000 (Malvern), with soil classification based on the WRB Texture Classes, i.e., sand (63–2000 μm), silt (2–63 μm), and clay (< 2 μm). The pH and electric conductivity (EC) were determined in suspensions at a soil/water ratio of 1:2.5 using a pH meter (Multi 3630 IDS, Germany).

The nutrients NO3–, NH4+, total N (TN), PO43−, total phosphorus (TP), and Si were measured in the soil solutions colorimetrically using an automated continuous flow analyzer (Skalar San++ System, Skalar Analytical B.V., The Netherlands). The SO42− concentrations were measured using a Thermo Fisher SCIENTIFIC DIONEX ICS 5000 + USA ion chromatograph in controlled environment (23ᵒC and 40% humidity). Soil organic sulfur plus sulfides/pyrites (SOS + Sn2−) were estimated from the equation: SOS + Sn2− = STS ‒ SLS ‒ SCS, where STS is the soil total sulfur, while SLS and SCS are the S content in SO42−LS and SO42−CS, respectively. The S content in SO42−LS and SO42−CS was calculated as: A × (SMW/BMW), where A is the SO42− concentration, SMW is the atomic weight of S, and BMW is the molecular weight of SO42−.

The trace metals/elements (Ca, Mg, Fe, Al, and Si) were analyzed on digested soil samples by inductively coupled plasma optical emission spectroscopy (ICP-OES, Agilent 5110) and the other metals (Mn, Cu, Ni, As, Sr, and U) by inductively coupled plasma mass spectrometry (ICP-MS, Agilent 7900). Briefly, approximately 0.25 g of soil were treated with a solution of 3 mL ultrapure 40% HCl and 9 mL ultrapure 80% HNO3 in a Teflon vessel and then digested in a Microwave Digestion System (MARS 6, CEM, USA), according to the EPA method 3051 A. Digestion was continued until a few drops of solution remained in the digested soil residue which was then transferred to a Teflon bottle and finally diluted to 50 mL with 3% HNO3. The extracted solution was then filtered using a syringe filter (0.45 μm, PTFE, Pall Corporation, USA) and stored in 50 mL polypropylene bottles. Blank (only acid) and standard samples were prepared using a similar procedure. The standard sample provided by Agilent Technologies was used to make the calibration of the standard solution. Quality assurance and quality control procedures were ensured by using standard reference materials supplied by Agilent Technologies. Owing to the high sensitivity of ICP-MS, ultra-high-purity grade HNO3 was used as the blank. One standard was measured with an interval of every ten samples and a solvent blank set with an interval of each batch sample. Average recoveries of standard reference metals ranged from 95% to 105% with a test precision within 5%. The entire system was controlled by the MassHunter Workstation 4.4 software (Agilent Technologies).

The Hg concentrations in We and Ae samples were measured by a Cold Vapor Atomic Fluorescence Spectroscopy (CVAFS) Analyzer (Tekran 2600, Tekran Instruments Corporation, Knowville, TN, USA) according to the EPA method 1631E. The samples were processed in a dark cold room at 4 °C and oxidized with 40% (v/v) of 2HNO3/1HCl with 1% (v/v) solution of 0.2 M BrCl139. The total Hg in soil was measured on soil samples ground in an agate mortar and sieved at a size < 0.075 mm, with precautions taken to avoid cross-contamination of samples, using a Lumex RA915F Mercury Analyzer equipped with a PYRO-915 + unit and Zeeman Atomic Absorption Spectrometry139. In this method, Hg in the soil samples was released as Hg(0) vapor during pyrolysis at approximately 750 °C and subsequently measured using Zeeman Atomic Absorption Spectrometry. Prior to analysis, the instrument was calibrated by certified reference materials (CRMs) in the range of 20 to 50 ng of Hg. The values of CRMs were within ± 5% of their certified values.

Absorption spectra of solutions were recorded from 200 to 800 nm at 1-nm interval using a quartz cell of 1-cm path length by a UV-VIS spectrophotometer (UV-2401PC, Shimadzu). Fluorescence excitation-emission matrix (EEM) spectra were acquired by a fluorescence spectrophotometer F-7000, Hitachi (Japan) using a procedure reported previously21. Ultrapure water (18.2 MΩ cm) was used as the blank and measured every ten samples, to check for the performance of the instrument and ensure data quality. The pure water EEM data were subtracted from the corresponding sample EEM spectra. A 10-µgL− 1 quinine sulphate (QS) solution in 0.01 N H2SO4 was used for fluorescence normalization. The fluorescence intensities of all samples were calibrated using the intensity of the QS (1 µg L− 1 = 1 QS unit, QSU) peak at Ex/Em = 350/450 nm21. To avoid inner-filter effects and fluorescence quenching, the sample solutions were diluted based on the initially measured DOC concentration and then EEM was measured. The fluorescence intensity of each peak was rechecked and corrected using a common absorbance-based approach129. Preprocessed EEM data were analyzed by the parallel factor (PARAFAC) model using the N-way Toolbox for MATLAB130 as described elsewhere131. First, Rayleigh and Raman peaks, as well as an ultrapure water blank, were subtracted from the measured EEM spectra using a home-made Excel program. Second, in order to avoid the mixing of fluorescent components of different soil samples that could produce artifacts131, PARAFAC analysis was performed individually on each soil sample. Finally, non-negative constraints were applied to the PARAFAC model. The detailed procedure of using PARAFAC modeling of EEM spectra was reported previously21. All statistical data analysis was performed by multivariate analysis and descriptive statistics using the Origin 2018 version 9.5 (https://www.originlab.com).

Definitions

LS

Labile state. Labile state refers to the labile forms of water-soluble soil components found in water-extracts obtained by a specific time-dependent extraction process using ultrapure water from soil and are influenced by existing environmental conditions such as moisture, temperature, redox properties, mineral matrix, and microbial activities in the soils. These LS soil components are typically leaching from soils into the surrounding surface and groundwater via water infiltration or rainwater runoff.

CS

Complexed state. Complexed state refers to the organo-mineral forms of soil components that remain complexed with soil minerals and are insoluble in water. These organo-mineral components can only be dissolved and extracted using an alkali solution (O.1 M NaOH) through a specific time-dependent extraction method applied to the soil residue left after water extraction. CS soil components typically remain in the soils during water infiltration or rainwater runoff. These fractions, however, are operationally defined and may not correspond exactly to discrete or functionally distinct biogeochemical pools in natural soils.