Iron Composition of a Typical Loess-Paleosol Sequence in Northeast China

Abstract

Iron isotope compositions, along with the partial extraction of iron in its various forms, can be utilized to investigate the complex interplay of iron migration and transformation with respect to iron isotope patterns. This study investigated the iron composition of a typical loess-paleosol sequence in Northeast China and aimed to understand the influence of iron migration and transformation of the typical loess-paleosol sequence on iron isotopes and environmental and climatic changes that occurred in the region over time by analyzing the distribution and characteristics of iron compositions in sedimentary layers. Samples were collected from Chaoyang in Northeast China, and the iron isotopic composition was analyzed using the multi-receiver inductively coupled plasma mass spectrometer (MC-ICP-MS). The findings revealed depth-dependent variations in the content of different iron forms, reflecting paleoclimatic shifts primarily through pedogenic transformation processes. Notably, iron migration within the section was observed to be limited. The variations in the reddening index and magnetic susceptibility of the loess-paleosol were primarily influenced by the presence of free iron (Fed), exhibiting a range of colors from yellow to red-yellow and red. The δ⁵⁶Fe values for loess and paleosols ranged from 0.097 ± 0.035‰ to 0.167 ± 0.010‰, with an average of 0.133 ± 0.024‰ and a coefficient of variation (CV) of 15.66% at the stratum scale. These values indicated a systematic enrichment of heavy iron isotopes and a significant negative correlation with the slightly fluctuating total iron content. Specifically, our analysis highlighted distinct differences in δ⁵⁶Fe values between paleosol (0.126 ± 0.024‰) and loess (0.146 ± 0.021‰). The δ⁵⁶Fe in Fed was negative, averaging −0.101 ± 0.022‰, while the δ⁵⁶Fe in silicate-bound iron was positive, averaging 0.156 ± 0.032‰. Intense pedogenesis, driven by warm and wet climates, facilitated iron transformations and migrations, resulting in the accumulation of light iron isotopes in the paleosols. These transformations and migrations were predominantly observed in microdomains characterized by iron depletions and concentrations, as reflected in the profile morphologies. However, the limited iron transformations and migrations did not result in significant Fe redistribution within the soil section, as evidenced by the limited variations in δ⁵⁶Fe with soil depth at the stratum scale. Sampling from the stratum or pedogenic horizon could potentially create the illusion of the minimal fractionation of iron isotopes within the sequence. Therefore, a detailed examination of the iron isotope composition in the micro-domains of the loess-paleosol sequence is crucial to elucidate the fractionation processes and mechanisms of iron isotopes during the formation of these sequences.

1. Introduction

The loess-paleosol sequence as superimposed loess and paleosol is an aeolian deposit of alternating seasonal monsoon circulations in an atmospheric environment [1]. Loess comprises mainly aeolian dust accumulations transported by cold-dry East Asian winter monsoons associated with the northern Siberian High. Paleosols are dust deposits affected strongly by pedogenesis under warm-wet East Asian summer monsoons derived from the South China Sea and the East China Sea [1,2,3]. Variations in the composition of loess, formed during distinct periods, were reported as a result of diverse formation conditions related to the original aeolian dust composition and weathering–pedogenic processes [1,2,3,4]. The formation process complicates the interpretation of its pedogenic evolution and further extraction accuracy of paleoclimatic information [4].

Iron, a relatively stable element, undergoes migration and transformation, which constitutes a crucial process in surficial soil pedogenesis [1,5]. The pedogenesis process mainly modified iron’s form in loess and had few influences on the total iron amount due to limited leaching. Even in the southern region of the China Loess Plateau (CLP) under a relatively humid climate, iron in most paleosols had very limited migration within the profile [1,6]. Excluding the influence of other material migration and deposition on iron content, the total iron in loess closely reflects its original aeolian dust content. A comprehensive study of iron compositions in the loess-paleosol sequence, along with the exploration of iron activation, migration, redistribution, and mineral transformation, will aid in understanding its pedogenesis and formation environment.

The common chemical weathering indices and methods cannot well reflect the essential characteristics of iron migration and transformation in the loess-paleosol sequence. Partial extractions of Fe in different forms were used in previous studies—mainly focused on iron oxides such as total Fe (Fet), free Fe (Fed), poorly crystalline Fe (Feo), silicate-bound Fe (Fer = Fet − Fed), crystalline Fe (Fec = Fed − Feo) and their ratios such as the iron freeness index (freeness = Fed × 100/Fet), iron activity index (Feact = Feo × 100/Fed), and iron crystallinity index (Fey = (Fed − Feo) × 100/Fed)—to reveal the characteristics of iron migration and transformation during pedogenesis [7,8]. However, the iron transformation between different forms often occurred simultaneously in a highly heterogeneous soil system, and extraction results varied with experimenters and analysis conditions [7,8]. Thus, the obtained iron oxide data and indexes could easily cause misunderstandings about soil iron transformations and migrations.

The stable iron isotope technique provides a novel tool for tracing iron cycles in the natural biogeochemical process and a new method for further research on pedogenesis [9,10,11]. Isotope fractionations refer to various isotopic atoms or molecules of an element in a system, which are distributed into various substances or phases in different proportions [12]. Various effects that induce isotope fractionation are called the isotope fractionation effect [12]. There are four stable isotopes of iron, including ⁵⁴Fe (5.84%), ⁵⁶Fe (91.76%), ⁵⁷Fe (2.12%), and ⁵⁸Fe (0.28%) [13]. Iron isotope compositions are usually expressed as relative amounts of δ⁵⁶Fe or δ⁵⁷Fe with respect to international iron isotope standard samples (IRMM-014) [14]. The δ⁵⁶Fe is expressed as δ⁵⁶Fe = [(⁵⁶Fe/⁵⁴Fe)sample/(⁵⁶Fe/⁵⁴Fe)IRMM-014-1] × 1000‰, and δ⁵⁷Fe can be calculated using the formula of δ⁵⁷Fe = 1.5 × δ⁵⁶Fe [11]. During pedogenesis, iron was sensitive to redox environments [15]. After reaching an equilibrium with different redox environments, iron comprises different valence states and compound forms [16], and isotope fractionations might also occur [17,18].

Compared to the iron geochemical composition analysis, high-accuracy iron isotope analyses can trace iron migration and identify the mechanism involved during pedogenesis, accurately reflect the redox and organic chelation changes in pedogenic environments [7,19,20,21], and trace soil evolution [7,10,19,20,22,23,24,25,26]. Iron isotope fractionation is induced by the complex interactive processes of silicate dissolution, precipitation, redox, and organic chelation, and its magnitude varies with global soil types [16,18,23,27,28] within a range of −0.62‰ to +0.72‰ [7,16,19,20,21,23,26]. The information on iron isotope fractionation changes and observed fractionation mechanisms in soil is still very limited [23,29].

Previous research reported that the δ⁵⁶Fe of loess-paleosol sequences from the CLP was between 0.06 ± 0.02‰ and 0.12 ± 0.02‰ for the Yimaguan section [30] in Gansu province and 0.07 ± 0.05‰ for the Luochuan section (GBW07454), and it ranged from 0.078 ± 0.051‰ to 0.102 ± 0.028‰ for the Xifeng section [31]. Although fractionation was reported as insignificant when compared to baseline continental igneous rocks, δ⁵⁶Fe varied across different regions. Compared to aeolian loess-paleosol sequences in the CLP, the typical loess-paleosol sequence in Northeast China, which is an ideal region for studying the response of East Asian monsoons relative to high- and low-latitude interactions outside the China Loess Plateau, had experienced stronger pedogenesis with an averaged CIA of 77.6 [32], which is greater than the 60.1 value that is typical of loess-paleosol sequences from the CLP [30]. Obvious iron migrations and transformations with uniformly distributed Fe-Mn neo-formations in strata were observed [4]. We see a potential for using the typical loess-paleosol sequence in Northeast China to explore the loess-paleosol sequence in iron migrations and transformations during pedogenesis. We hypothesize that iron isotope fractionations combined with the partial extraction of iron in different forms could be employed to trace the elementary mechanism of iron migration and transformation. Iron isotope behaviors during migration and transformation should first be addressed. The objective of the paper was to investigate the influence of the iron migration and transformation of the typical loess-paleosol sequence on iron isotopes. The long-term novel migration and transformation of iron species with different depths will be addressed during pedogenetic progress, which will help in better understanding loess-paleosol pedogenesis and provide insights into the paleo-environmental conditions that influenced the formation of the loess-paleosol sequence in the studied area.

2. Materials and Methods

2.1. The Study Area Descriptions

The typical loess-paleosol sequence of the Chaoyang section (41°33′9.6″ N, 120°30′20.8″ E) is located in the hilly area of western Liaoning province where it is adjacent to the middle reaches of the Daling River. This area is mainly influenced by the northern temperate continental monsoon climate (Figure 1). The southeast portion of the region is impacted by warm and humid oceanic air, while the northern region is influenced by dry and cold air originating from the Inner Mongolia Plateau, ultimately creating a semi-dry and semi-humid climate. The area’s average annual precipitation is 450–500 mm, and its average annual temperature is 9 °C. Vegetation is dominated by North China flora such as wattle and elm. The Chaoyang section is located in a well-closed intermountain basin and formed by aeolian dust deposition under Quaternary climate fluctuations, which has no pedogenic relationship with the underlying bedrock [4]. This section is located in a closed watershed and is little affected by external water flow. In addition, conditions for the section are well-preserved. No signs of man-made disturbances and accelerated erosion were detected.

2.2. Section Description and Sampling

The Chaoyang section is 19.85 m thick and comprises Holocene soil S0, five loess (L), and four paleosol (S) superimposed strata (Figure 1) according to the standard stratigraphic division method [2,33,34] and forty-two pedogenic horizons according to detailed morphological descriptions referring to “The field soil description and samplings” [35]. A total of forty-two soil samples were collected from the bottom to the top of the section according to the standard soil sampling method [36]. Then, 92 sub-samplings were collected at about 20 cm intervals from the section’s bottom to the surface. Then, all soil samples were air-dried, ground, and passed through 60- and 200-mesh nylon sieves for subsequent determinations. The composition of the upper part (UPP) and middle part (MIP) of the section above 228 cm was complicated by flowing water; the following study mainly focuses on the lower part of the section (LOP) below 228 cm, which comprises aeolian loess with uniform parent materials and experienced pedogenetic processes [4,37].

Based on morphologic data [4], the examined soil section was primarily composed of loess from a single source, except for the S0 layer. This layer contained a thin substratum of coarse limestone gravel at its base. Soil texture changed from loam to sandy loam above 195 cm, with a silt loam transition indicating local water-reworked materials. Below 228 cm, the soil was uniformly silt. Five loess and four paleosol strata were present beneath S0, exhibiting distinct color patterns. Clay and Mn-Fe coatings were observed throughout, suggesting significant illuviation. The paleosols displayed argillic horizons and greater clay content than the loess. The CaCO₃ content was stable in L1, indicating minor eluviation, while intense leaching had occurred in the lower section (LOP). The detailed descriptions of pedogenic horizons can be found in a previously published paper [4].

Despite LOP’s uniformity in field morphology, further micromorphological evidence [4] was summarized as follows. Significant variations in microfabric and matrix between UPP and LOP suggested different parent materials. UPP contained large, sharply margined quartz and feldspar particles, while LOP had finer, weathered quartz particles. The MIP, a transitional zone, exhibited characteristics of both UPP and LOP. LOP’s matrix remained consistent across various loess horizons, indicating a uniform parent material without vertical stratification. L5, a specific layer within LOP, demonstrated a simple structure with minimal pedogenesis, making it a useful reference for comparing overlying soil horizons. Clay and Fe-Mn coatings, more prevalent in paleosols than in L5, were observed in LOP. Field observations reveal no signs of water-induced stratification, suggesting that LOP was not a water-reworked loess deposit. Coarse, clay-depleted zones and distinct circular features filled with coarser silt were noted, particularly in paleosols. Weathering and pedogenesis resulted in various clay coatings and infillings, more prominent in paleosols and decreasing in frequency towards the L5 base.

The magnetic susceptibility of samples was measured using a Bartington magnetic susceptibility meter with an MS2B (Bartington, Oxford, UK) dual frequency sensor [4]. Magnetic susceptibility measured at low frequencies (0.47 kHz) was used in this study.

2.3. Laboratory Methods

2.3.1. The Partial Extraction of Iron in Different Forms

The content of free Fe (Fed), including crystalline Fe and poorly crystalline Fe (Feo) in samples, was extracted with a mixed dithionite–citrate–bicarbonate (DCB) solution at 85 °C [38]. The content of poorly crystalline Fe (oxhydryl) oxides (Feo) in samples was extracted with an acid ammonium oxalate solution (pH = 3.2) [39]. Fed and Feo were finally measured using an atomic absorption spectrophotometer (AAS) (Z2000, 2019, Hitachi, Ltd., Tokyo, Japan) at a 248.3 nm wavelength using air/acetylene flames.

Total soil Fe (Fet) was prepared using the sodium carbonate melting method [40] and determined using an inductively coupled plasma-atomic emission spectrometer (VISTA-MPX, 2006, Varian, Palo Alto, CA, USA). About 0.2 g of an air-dried soil sample was weighed (accuracy of ±0.0001), placed on phosphorus-free, quantitative filter paper, and then uniformly mixed with 1.5 g of pulverized Li₂CO₃-H₃BO₃ (1:2). Next, it was kneaded into a small ball and placed into a graphite crucible. The graphite crucible was placed in a muffle furnace at 950 °C. The process was gradually heated from 0 to 500 °C for 30 min, kept at 500 °C for 15 min, and then heated from 500 °C to 950 °C for 30 min. Then, the molten ball was removed from the graphite crucible by utilizing a graphite spoon and placed into a 250 mL beaker containing 200 mL of 4% nitric acid (HNO₃). The beaker was immediately placed in the ultrasonic washing tank for ultrasonic separation for 40 min. After the molten state was completely dissolved, the solution was filtered and moved to a 250 mL volumetric flask with 4% HNO₃ solution [40]. At the same time, a blank sample and a national standard soil sample (GBW (E) 070045) were added and analyzed for quality control. Finally, the content of total Fe was determined using the inductive coupled plasma atomic emission spectrometer (VISTA-MPX, 2006, Varian, Palo Alto, CA, USA). Elemental compositions had relative errors of <5% and the standard deviation of one randomly selected sample in triplicate was <3%.

Iron was used in different forms, and the indices used were as follows: total Fe (Fet), poorly crystalline Fe (Feo), free Fe (Fed), silicate-bound Fe (Fer = Fet − Fed), crystalline Fe (Fec = Fed − Feo), iron freeness index (freeness = Fed × 100/Fet), iron activity index (Feact = Feo × 100/Fed), and iron crystallinity index (Fey = (Fed − Feo) × 100/Fed).

2.3.2. The Iron Isotope Composition

Less than 0.149 mm of the soil sample containing about 50 μg Fe was digested using the mixed HNO₃-HCl-HClO₄ acids. The solution was evaporated to dryness and then dissolved in 6 N HCl; anionic resin (6 N HCl solution balance) was used to purify iron until the impurity content was negligible and the chemical process recovery rate was more than 99.5%. Finally, the iron purification solution was diluted with 2% HNO₃, and its iron isotope composition was determined in the medium-resolution mode of the multi-receiver inductively coupled plasma mass spectrometer (MC-ICP-MS) (Neptune plus, Thermo Fisher Scientific, Waltham, MA, USA). According to the international iron isotope standard IRMM-014, δ is used to represent the iron isotope composition with the unit of “‰”. The representation of iron isotopes is as follows:

$${\mathsf{\delta }}^{56}\mathrm{F}\mathrm{e}=\left(\frac{{\left(\raisebox{1ex}{${}^{56}\mathrm{F}\mathrm{e}$}\!\left/ \!\raisebox{-1ex}{${}^{54}\mathrm{F}\mathrm{e}$}\right.\right)}_{\mathrm{s}\mathrm{a}\mathrm{m}\mathrm{p}\mathrm{l}\mathrm{e}}}{{\left(\raisebox{1ex}{${}^{56}\mathrm{F}\mathrm{e}$}\!\left/ \!\raisebox{-1ex}{${}^{54}\mathrm{F}\mathrm{e}$}\right.\right)}_{\mathrm{I}\mathrm{R}\mathrm{M}\mathrm{M}-014}}-1\right)\times 1000$$

Here, IRMM-014 was provided by the Association for Reference Materials and Measurements of the European Commission [41]. δ⁵⁷Fe and δ⁵⁶Fe can be interconverted by the δ⁵⁷Fe = 1.475 × δ⁵⁶Fe formula [11].

2.3.3. The Redness Rating Index (RR)

The redness rating index can estimate the content of hematites in soil [42].

$$RR=\left(12.5-H\right)\times \raisebox{1ex}{$C$}\!\left/ \!\raisebox{-1ex}{$V$}\right.$$

Here, H, C, and V represent hue, chroma, and value, respectively.

2.4. The Mass Balance Model

The mass balance model was used to quantify relative elemental differentiations compared to a selected reference base in pedogenesis [4,43,44]. According to a principle [45], a layer with the weakest development close to the original parent material in the sediment—L5 in the Chaoyang section [4]—was selected as the reference base. The stable Ti, which mainly existed in non-clay and weathering-resistant minerals, was selected as a reference element to quantitatively calculate the pedogenesis [4,45,46,47].

The model’s formula is summarized as follows:

$${\tau }_{j,Ti}=100\times \left[\frac{{\left(\raisebox{1ex}{$C_j$}\!\left/ \!\raisebox{-1ex}{$C_{Ti}$}\right.\right)}_s}{{\left(\raisebox{1ex}{$C_j$}\!\left/ \!\raisebox{-1ex}{$C_{Ti}$}\right.\right)}_p}-1\right]$$

Here, C_j and C_Ti represent the concentration of an element j and Ti, respectively; s represents a weathered layer, and p represents the reference base [48]. The positive τ_j,Ti indicates that j increases and accumulates compared to the reference base and vice versa.

2.5. Data Processing

The SigmaPlot 12.5 software was used to plot figures, and statistical data analysis was performed using a generalized linear model SAS procedure (SAS Institute, Hong Kong, China, 2000). The variation in soil properties was evaluated using the coefficient of variation (CV). A CV of a property that was less than 15% was considered to comprise a small variation and exhibit a relatively uniform distribution [49].

3. Results

3.1. Partial Extraction of Iron Composition in the Loess-Paleosol Sequence

The distribution of iron content across different forms exhibited distinct patterns with depth, characterized by four peaks in paleosols and four troughs in loess layers (Figure 2). Specifically, the Fet content fluctuated between 42.98 and 62.41 g/kg, averaging 52.66 g/kg, and it exhibited a relatively uniform distribution with respect to depth, as evidenced by a coefficient of variation (CV) of 8.96%. The Fer, extracted from primary silicate, ranged from 18.91 to 37.97 g/kg, averaging 29.85 g/kg. This form of iron exhibited minimal variations with depth (CV = 11.98%) and accounted for a significant portion (60.38%) of Fet, highlighting its dominance in the iron composition and suggesting weak weathering intensity. The Fed, derived from primary silicate weathering, exhibited greater variability with depth (ranging from 16.43 to 31.50 g/kg and averaging at 22.81 g/kg), as reflected by a higher CV of 19.58%. The Fec content, which ranged from 15.79 to 30.75 g/kg with an average of 22.18 g/kg, also showed significant depth-related variations. This indicates significant alterations in iron crystallization during pedogenesis, with Fec dominating Fed by 96.80%. The Feo content, varying from 0.34 to 0.80 g/kg (average for 0.63 g/kg, CV = 18.82%), further contributed to the complexity of iron distributions. Notably, the CV for iron content was greatest in Fec, followed by Fed, Feo, and Fer, with Fet showing the least variation. Paleosols generally exhibited higher iron content across all forms compared to loess layers. The most prominent peaks were observed in paleosol layers S3, S4, and S1-2, while the lowest troughs occurred in loess layers L4 and L5. Significant positive correlations were observed between Fet and Fed (R² = 0.61, p < 0.01) (

Table 1. Correlation analysis between iron contents in different forms with magnetic susceptibility and the redness rating index (RR) in the loess-paleosol sequence.

Name	Magnetic Susceptibility	Redness Rating	δ56Fe	Fet	Fed	Feo	Fec	Fer
Magnetic susceptibility	1
Redness rating	0.45 **	1
δ56Fe	−0.06	−0.26	1
Fet	0.37 **	0.45 **	0.45 *	1
Fed	0.80 **	0.46 **	−0.06	0.49 **	1
Feo	0.34 **	0.45 **	0.004	0.25 **	0.40 **	1
Fec	0.80 **	0.45 **	−0.06	0.49 **	1.00 **	0.38 **	1
Fer	−0.10	−0.04	−0.01	0.20 **	−0.11	−0.02	−0.11	1

“*” and “**” represent p < 0.05 and p < 0.001, respectively; “−” indicates a negative correlation relationship.

), as well as between Fed and both Feo and Fec (R² = 0.37, p < 0.01). These findings highlighted significant depth-related changes in iron content and suggested that iron responded primarily through form transformations during pedogenesis. The relatively small changes in Fet indicated limited iron migration within the section.

Freeness ranged from 33.15% to 59.76%, with an average of 43.13%. Feact varied between 1.91% and 3.90%, averaging 2.80%. Fey ranged from 96.10% to 98.09%, with an average of 97.54%. This indicated that crystalline iron was the predominant form of free iron. The distribution of these iron content ratios with respect to depth further supported the conclusion of higher pedogenic intensity in paleosol layers S3, S4, and S1-2 compared to loess layers L4 and L5.

3.2. Iron Isotope Compositions within the Loess-Paleosol Sequence

Various amounts of Fe-Mn coatings were observed below 228 cm of the studied section, indicating the presence of significant oxidation-reduction processes and potential iron isotope fractionations. To delve deeper into the iron transformations and limited migrations, we determined the iron isotope composition of soils sampled at the stratum scale within the typical loess-paleosol sequence. Our findings revealed that δ⁵⁶Fe values ranged from 0.097 ± 0.035‰ to 0.167 ± 0.010‰, with an average of 0.133 ± 0.024‰ and a CV of 15.66% (Figure 2). This systematic enrichment of heavy iron isotopes was evident across the section.

When compared to continental igneous rock δ⁵⁶Fe (0 ± 0.05‰), as reported in previous research, our data suggested no obvious iron isotope fractionations. However, a notable difference was observed between the paleosol and loess layers. Paleosols generally appeared in the troughs of the δ⁵⁶Fe versus the depth curve, with δ⁵⁶Fe values ranging from 0.097 ± 0.035‰ to 0.149 ± 0.013‰. This range encompassed a spread of 0.052‰, an average of 0.126 ± 0.024‰, and a CV of 16.09%. In contrast, loess layers were positioned at the peaks of the δ⁵⁶Fe versus depth curve, exhibiting higher δ⁵⁶Fe values from 0.140 ± 0.011‰ to 0.167 ± 0.010‰, averaging 0.146 ± 0.021‰. These values were more uniformly distributed with depth (CV = 11.90%), with S3 and L2 exhibiting the lowest and highest δ⁵⁶Fe values, respectively.

Further comparisons between δ⁵⁶Fe and various forms of iron revealed interesting trends. The δ⁵⁶Fe curve showed a contrasting trend compared to Fet, Fed, Feo, Fec, freeness, and Fey but aligned closely with Fer and Feact. This trend significantly differed from partially extracted irons. Importantly, δ⁵⁶Fe had a statistically significant negative correlation with Fet (R² = 0.45, p < 0.05). These observations strongly suggest that intense pedogenesis drove iron transformations and migrations, resulting in the accumulation of light iron isotopes in paleosols during warm and wet climatic periods.

Additionally, we determined the iron isotope composition of DCB-extracted free iron [38] and its remaining silicate-bound iron residue. The δ⁵⁶Fe in free iron was negative, averaging at −0.101 ± 0.022‰, while the δ⁵⁶Fe in silicate-bound iron was positive, averaging at 0.156 ± 0.032‰. These findings suggest that light iron isotopes preferentially migrated away from primary silicate minerals during weathering and pedogenesis, resulting in the relative enrichment of heavy iron isotopes in the silicate-bound fraction.

3.3. Comparisons of Iron Compositions with Magnetic Susceptibility and RR in the Loess-Paleosol Sequence

3.3.1. Comparisons of Iron Composition with Magnetic Susceptibility

Magnetic susceptibility is a reliable proxy for assessing the intensity of the East Asian summer monsoon [50]. Peaks in the magnetic susceptibility curve aligned with warm and humid epochs characterized by a pronounced summer monsoon. Moreover, troughs corresponded to colder, drier periods dominated by a strong winter monsoon and elevated dust deposition. Our analysis demonstrated that the depth-dependent trends of Fet, Fed, Feo, and Fec contents exhibited a strong correlation with the magnetic susceptibility curve. The Fer content curve displayed an inverse correlation with magnetic susceptibility (Figure 3).

Statistical correlations further revealed a significant positive relationship between the contents of Fet, Fed, Feo, and Fec and magnetic susceptibility (p < 0.001), with determination coefficients of 0.37, 0.80, 0.34, and 0.80, respectively. Although Fer content demonstrated a negative correlation with magnetic susceptibility, this association was not statistically significant. These findings implicated that variations in soil magnetic susceptibility were predominantly influenced by certain iron forms, especially the pedogenic neo-formation iron (Fed).

3.3.2. Comparison of Iron Compositions with RR

RR serves as a valuable indicator of hematite contents derived from pedogenesis, exhibiting a pronounced linear correlation with hematite abundance and reflecting the intensity of paleosol development [39,42,51]. Our data showed that RR varied considerably with depth, ranging from 0.94 to 12.00 (CV = 66.73%) (Figure 3).

The peaks in the RR curve corresponded to the presence of paleosols, while the troughs concurred with the presence of loess layers. The depth profiles of Fet, Fed, and Feo contents and freeness were consistent with the RR curve trend. Statistical analysis revealed a significant positive correlation between the contents of Fet, Fed, Feo, and Fec and RR (p < 0.001), with determination coefficients of 0.45, 0.46, 0.45, and 0.45, respectively. In contrast, the Fer content demonstrated a negative correlation with RR, albeit with a relatively low determination coefficient of 0.04. Based on these observations, it is evident that iron content, particularly in the form of ferric oxide (Fed), exerts a considerable influence on the RR. Consequently, variations in Fed content with depth were reflected in color shifts and RR values, distinguishing loess from paleosols.

4. Discussion

4.1. Iron Transformations in the Loess-Paleosol Sequence

The formation of the loess-paleosol sequence involved complex iron transformations, with iron existing in both Fe²⁺ and Fe³⁺ valences and manifesting in various forms such as Fed, Feo, Fec, and Fer. This process is primarily driven by chemical and biological weathering [2,52]. Paleosol formation coincided with peaks in magnetic susceptibility, indicating warm and humid climatic conditions conducive to flourishing plant life and intense microbial activity [53,54]. Under these conditions, the weathering of primary silicate minerals released Fe²⁺, which was rapidly oxidized to Fe³⁺ in the presence of oxygen within the well-developed pores [4] between coarse silt particles. Given the soil’s pH range of 7.4 to 8.6, Fe³⁺ predominantly existed as amorphous ferric hydroxide and ferrihydrite. These compounds subsequently undergo dehydration, crystallization, and aging processes, partially transforming into goethite and hematite, which were then stored within the soil layer [38,53,54]. These iron forms collectively constituted the free iron oxides present in the soil.

Due to the limited mobility of iron in its varying forms and the proximity of these transformations, Fed, Feo, and Fec accumulated within the paleosols, exhibiting peaks in their respective curves with an increasing cumulative trend. Conversely, Fer appeared in the troughs of the curve, exhibiting a decreasing trend due to these transformations. The minimal organic carbon content, ranging between 0.03% and 0.07% in both loess and paleosols, was inadequate to significantly drive iron complex migration and evolution. Instead, soil temperature, moisture, and oxygen fugacity were the primary factors influencing iron migration and transformation in the soil [39,53], as evidenced by the diverse and abundant secondary iron oxides. Consequently, the elevated Fet content observed in the reddish paleosol layer of the section was primarily attributed to the relative enrichment of Fed through iron transformations, intense carbonate leaching, and even silification.

During loess formation, the cold and arid climate constrained plant growth and microbial activity, resulting in weaker chemical and biological weathering. Consequently, the loess layers contained lower concentrations of Fed, Feo, and Fec compared to the paleosols. The loess layers were characterized by decreasing troughs, whereas the paleosol layers exhibited increasing peaks. Notably, the paleosol layer had higher contents of Fed and Fec, indicating a more pronounced iron crystallization process compared to the loess layer.

4.2. Iron Isotope Fractionations in the Loess-Paleosol Sequence

During the formation of the loess-paleosol sequence, Fe²⁺ migrated along with the soil solution, accompanying partial reduction, migration, and subsequent transformation into various iron oxide minerals [31]. This intricate process was potentially associated with notable iron isotope fractionations, as indicated in previous studies [7,19,44].

The section exhibited relatively small variations in Fet content, with a CV of 8.96%. The peak of these variations was observed in the paleosol layer, whereas the loess layer corresponded to the trough. The average Fet content in the paleosol layer was significantly elevated, being 19.53% higher than that in the loess layer. Compared to the reference base L5, Fet content varied from −15.13% to 29.05% with respect to depth during pedogenesis (Figure 4). The paleosol layers had an average Fet content of 54.51 g/kg, which was higher than the 48.63 g/kg observed in the loess layers. These layers displayed cumulative Fet gains with variations ranging from −13.74% to 29.06%, while the loess layers showed cumulative Fet losses between −15.13% and 2.79%. These findings indicated that the variations in Fet content within the section were primarily driven by differences in the layer type (paleosol versus loess) resulting from pedogenesis rather than depositional processes.

The iron isotopes within the sequence exhibited a range from 0.097‰ to 0.167‰, indicating a systematic enrichment of heavy iron isotopes in comparison to continental igneous rocks. The elevated δ⁵⁶Fe value observed in the reference base L5 suggested that the aeolian dust had undergone a certain degree of development prior to deposition. Postdeposition, the weathering of primary iron-bearing silicate minerals liberated iron, resulting in iron isotope fractionations. Under the influence of atmospheric oxygen oxidation, ferric iron states attained thermodynamic stability. During the weathering process, Fe²⁺ was released and subsequently underwent oxidation, precipitating into disordered iron oxide minerals. As soil development progressed, these minerals underwent further transformations, evolving into more crystalline forms of goethite and hematite. These transformations likely occurred within specific microdomains, characterized by iron depletions and concentrations observed in profile morphologies. However, the limited extent of iron transformations and migrations did not result in a significant redistribution of Fe within the soil section, as evidenced by the limited variations in δ⁵⁶Fe with soil depth at the stratum scale. The absence of significant fractionations suggested that changes in loess-forming conditions did not induce notable iron migrations at this scale. Instead, the weathering of iron minerals resulted in mutual transformations of iron among different forms within the microdomain scale.

The Fet content in the studied section, ranging from 33.36 to 62.41 g/kg, significantly exceeded the global average for loess, which Taylor et al. [41] reported as 22.7 ± 17.2 g/kg). Despite this elevated Fet content, the δ⁵⁶Fe values within the sequence, ranging from 0.097‰ to 0.167‰, were comparable to the global average for loess, documented by Beard et al. [22] as 0.05 ± 0.04‰. The significantly higher Fet content in the paleosol compared to the loess indicated a relative enrichment of iron in the former and a depletion in the latter. While these changes in iron distribution did not cause significant iron isotope fractionations, the dust particles within the sequence underwent a certain degree of weathering, releasing iron from primary silicate minerals. However, the presence of calcium carbonate in the dust particles created a slightly alkaline environment during dust accumulation. Additionally, the open system nature of the loess, characterized by low precipitation and a dry climate, facilitated deposition primarily in coarse silt particles [55]. These particles, with large intergranular pores and excellent air permeability, typically maintained an oxidizing environment. Under these oxidizing and well-drained conditions, water primarily migrated longitudinally within the soil. Due to the short retention time of water flow within the large intergranular pores, the reduction and transformation process of iron was limited. Only a small amount of Fe²⁺ was reduced and dissolved, rapidly oxidizing into low-solubility Fe³⁺ [56]. This process resulted in heavy iron isotopes that preferentially entered the crystal lattice of iron oxides or iron hydrogen oxides, resulting in the enrichment of heavy iron isotopes in the soil [16,22,57,58].

The narrow range of δ⁵⁶Fe from 0.097 ± 0.035‰ to 0.167 ± 0.010‰ reflected limited iron migration within the sequence. This variation suggested subtle, unobvious fractionation processes during pedogenesis. Nevertheless, the formation of iron oxides and the transformation of iron between various forms resulted in a heavier iron isotope composition in the soil compared to continental igneous rocks (0 ± 0.05‰). These observations were consistent with previous studies [7,16,18,19,20,21,23]. The low solubility of secondary Fe³⁺ minerals within the section constrained the amount and distance of iron migration with longitudinal water flow. Consequently, there were no significant changes in iron isotope composition at the stratum scale. These findings suggested that the section underwent brief redox processes, resulting in limited iron migration and minimal alterations in iron isotope composition. However, the presence of abundant Fe-Mn coatings, especially in paleosols, raised intriguing questions. This phenomenon could potentially be attributed to two factors. Firstly, iron redistribution occurred with the subtle fractionation of iron isotopes. Secondly, the soil evolution process occurred over a short distance, resulting in localized iron isotope fractionation and migration. For instance, longitudinal iron migrations at the stratum scale were absent, or iron export from the soil was highly restricted.

4.3. Comparisons of Iron Isotope Composition between the Loess-Paleosol Sequences from Different Areas

Loess deposits are extensively distributed across Northeast China, where, similarly to the CLP region, they have been influenced by East Asian monsoon circulation throughout the Quaternary period. This monsoon-induced rainfall significantly impacted the formation of loess-paleosol sequences in the area [2,59]. Northeast China is situated within a temperate humid to semi-humid continental monsoon climate zone, whereas the Loess Plateau lies in the transitional zone between a temperate monsoon and a temperate continental climate. These distinct climatic conditions resulted in markedly different soil-forming environments.

Research has revealed that Quaternary loess and paleosol deposits in Northeast China along the Yangtze River and even on the islands of the East China Sea shared similarities with those of the CLP, all of which are classified as aeolian loess [2,4]. These deposits originated from high-latitude deserts and Gobi areas, exhibiting comparable provenance compositions [60]. However, variations in geographical locations and soil-forming environments across these regions led to variations in soil-forming intensity, grain size, pH, chemical weathering degree, and other physicochemical properties of the loess [4,31,34].

For example, the δ⁵⁶Fe values of loess-paleosol sequences showed significant regional variations. The Yimaguan section in Gansu province ranged from 0.06 ± 0.02‰ to 0.12 ± 0.02‰ [30], while the Xifeng section spanned from 0.078 ± 0.051‰ to 0.102 ± 0.028‰ [31]. In contrast, the Xuanzhou and Langxi sections in Anhui province varied from −0.035 ± 0.110‰ to 0.086 ± 0.052‰ and −0.057 ± 0.071‰ to 0.117 ± 0.048‰, respectively [31]. The Luochuan section had a δ⁵⁶Fe value of 0.07 ± 0.05‰ (GBW07454). Despite these regional variations, iron isotope values remained close to those of continental igneous rocks, indicating insignificant iron isotope fractionation. This suggested that long-distance physical transport, thorough dust mixing, and varying degrees of chemical weathering did not induce a significant stratum-scale fractionation of iron isotopes [20,21,30,31]. This observation could be attributed to the limited distance of iron migration within the loess-paleosol sequence, which is insufficient to cause notable iron isotope fractionation or the absence of isotope fractionation during iron migration.

The characteristics of iron isotope composition across different regions revealed the existence of fractionation in soils, albeit to a small degree. Previous studies typically collected samples from strata or pedogenic horizons, often combining samples from a specific area to prepare a representative sample for iron isotope analysis. However, this sampling methodology may obscure the micro-domain migration patterns of iron, including concentration and depletion changes caused by soil oxidation-reduction reactions, as well as the fractionation signals of iron isotopes. This could potentially lead to the misperception of the insignificant fractionation of iron isotopes within the sequence. Therefore, further investigation into the micro-domain iron isotope composition of loess-paleosol sequences is crucial to elucidate fractionation processes and mechanisms during their formation.

5. Conclusions

Based on an in-depth investigation of iron migration and transformation in a loess-paleosol sequence from Northeast China and utilizing iron isotope fractionation and partial iron extraction techniques, the study concludes the following:

(a)

Depth-dependent variations in iron forms are primarily influenced by pedogenic transformation processes.

(b)

Limited iron migration was observed within the studied loess-paleosol section.

(c)

The presence of free iron significantly influences the reddening index and magnetic susceptibility of the loess-paleosol sequence, resulting in a range of color variations.

(d)

Heavy iron isotope enrichment was systematically observed, exhibiting a significant negative correlation with the slightly fluctuating total iron content.

(e)

Distinct differences in δ⁵⁶Fe values were identified between the paleosol and loess layers, with intense pedogenesis driven by warm and wet climates, facilitating iron transformations and migrations, particularly in microdomains.

(f)

Despite these transformations and migrations, significant Fe redistribution within the soil section was not evident at the stratum scale.

Detailed examination of the iron isotope composition in the micro-domains of the loess-paleosol sequence is essential for elucidating the fractionation processes and mechanisms of iron isotopes during the formation of these sequences.