*3.1. Source*

Almost all Fe-containing minerals are potential sources of Fe for red-bed chromogenic substances. During oxidative weathering, Fe is released from these minerals, mainly forming Fe-bearing clay minerals and Fe oxides/hydroxides. However, under reducing conditions, Fe carbonates, sulfides, and phosphates may also be formed [33,51]. During the formation of red beds, the alteration of unstable iron-rich particles is an important source of Fe for hematite [24,25,52–54].

The commonly altered mineral grains are iron–magnesium silicates (e.g., hornblende, pyroxene, and biotite), volcanic rock fragments and glass, magnetite, and ilmenite [45]. These grains formed at high temperature or high pressure during magmatism and metamorphism became unstable at surface and near-surface conditions in the sedimentary basin. During grain alteration, under reducing conditions, released iron goes into solution as Fe+2 and precipitates as iron oxide or oxyhydroxide under oxidizing conditions [55]. Walker et al. (1967, 1978) [22,56] elaborated on the formation mechanism of red beds during burial diagenesis and pointed out that the alteration of the interior of the iron–magnesium silicate layers by oxygen-containing groundwater during the burial process is the key. They showed that the hydrolysis of iron-bearing detrital minerals follows the Goldich dissolution reaction series and is governed by the Gibbs free energy of the specific reaction [22,56].

Mucke (1994) indicated that alteration after diagenesis can also occur through the oxidation of pyrite and siderite [57].

3O<sup>2</sup> + 4FeS<sup>2</sup> → Fe2O<sup>3</sup> (hematite) + 8S; E = −789 kJ/mol.

$$\text{CO}\_2 + 4\text{FeCO}\_3 \rightarrow 2\text{Fe}\_2\text{O}\_3 \text{ (hematite)} + 4\text{CO}\_2\text{; E} = -346 \text{ kJ/mol.}$$

Dissolution and displacement are the main processes leading to grain alteration in red beds. Partial dissolution of unstable minerals such as amphibole and pyroxene may proceed inward from the grain edges, while minerals such as plagioclase may preferentially disrupt the grain interior [32]. The different morphologies of partially dissolved minerals are an abiotic result of selective dissolution along lattice planes. Xiao et al. (2018), suggested that Fe and Ti-bearing silicate minerals such as biotite gradually release various elements during low-temperature alteration, with water-soluble elements such as K and Na being carried away with the transport of pore water, while water-insoluble iron–titanium oxides are deposited in situ around these altered minerals [49]. Clay mineral replacement of clastic particles is also a common alteration phenomenon in red beds, where the replaced clay forms in situ and shows random sheet crystal orientation [23].

Apart from the inheritance of parent minerals such as ferromagnesian silicates, clastic and clay minerals mechanically permeated by groundwater can also provide a direct source of iron for staining [7,54,58,59]. The mixing of hematite with clay minerals in the form of very fine-grained crystals is an important feature of red beds, and the relative proportions of hematite and clay vary considerably. In some cases, iron-bearing films may consist only of granular crystals of hematite that have precipitated in situ. In other cases, a thicker iron-containing clay film may be present. These Fe-bearing clays may have originated from surface soils or, more likely, from loose sediments formed in the adjacent floodplain [60]. An important feature of mechanical infiltration is that clay is not red when deposited, but reddens over time after contact with oxygenated groundwater [25]. In Cenozoic alluvium, the usual sediment color is yellow or red orange, due to iron oxyhydroxide, of which goethite and ferrihydrite are the main chromogenic minerals. Thus, the reddening of permeable clays represents an early stage of red material formation in red beds.

In addition, Blodgett et al. (1993) suggested that recirculation of red material may be the source of hematite in some red beds [32]. Such red beds would be produced by the erosion and redeposition of pre-existing red beds. These often have irregular color zoning, which is related to an unconformity in the weathering profile. Color boundaries cross lithological contacts and show more intense reddening near unconformities. Secondary red beds formed in this way are representative of late diagenesis, associated with previously deposited uplift, erosion, and surface weathering, requiring similar primary and diagenetic conditions [61].

## *3.2. Hematite Formation*

There are two main potential mineralogical pathways for the formation of hematite during diagenesis: ferrihydrite transformation and ferric oxyhydroxide dehydroxylation. Ferrihydrite is the most common precursor for the occurrence of iron oxides in soils and sediments, and the transformation from ferrihydrite to hematite in red beds is an important process [62]. The conversion of ferrihydrite to hematite appears to start with the internal rearrangement and dehydroxylation of ferrihydrite aggregates, which are converted to hematite over time [63]. This transition may occur faster than the new formation of ferrihydrite and may occur too quickly to be detected in most geological settings [33]. Jiang et al. (2018), proposed a new model for the transformation of ferrihydrite to hematite in soils and sediments, which can be divided into five stages according to the chronological order of transformation: (I) magnetically ordered (or core-shell structured) superparamagnetic (SP) ferrihydrite formation, (II) rapid hematite formation from SP ferrihydrite, (III) relative hematite stabilization, (IV) maghemite nanoparticle neoformation, and (V) completion of the transformation from maghemite to hematite [64]. This model provides a new framework for interpreting the formation of iron oxides and their paleoenvironmental significance.

The transformation of goethite to hematite has long been considered the important pathway for red beds [65,66]. Weibel et al. (1999) observed pseudo-crystals of goethite transforming to hematite in red sandstone samples from the Triassic Skagerak Formation in Denmark. Withing increasing burial depth (550 m to 2500 m), the ground temperature increased from 47 ◦C to >105 ◦C [67]. Berner (1969) pointed out that goethite and iron hydroxides are generally unstable relative to hematite, and goethite is easily dehydroxylated under water shortage or high-temperature conditions (2FeOOH (goethite) → Fe2O<sup>3</sup> (hematite) + H2O). Experiments show that, at 250 ◦C, the Gibbs free energy (G) of goethitehematite reaction is 2.76 kJ/mol [68]. Langmuir (1971) concluded that G becomes more and more negative as particle size decreases. The smaller the mineral particle, the easier it is for goethite to undergo dehydroxylation to form hematite [69]. Dehydroxylation of finegrained goethite may occur quickly, especially during deposition [7]. Thus, it appears that clastic iron oxyhydroxides (such as goethite) spontaneously dehydrate into red hematite over time. This process helps to explain not only the gradual reddening of the alluvium but also why desert dunes of older geological ages are redder than the newer ones.

In addition, other processes leading to the formation of diagenetic hematite may include the pseudo-crystalline substitution of hematite for minerals such as magnetite [7,70] and microbial interactions [44,71]. It should be noted that usually the formation of hematite in red beds may not be caused by a single mechanism of action but is the result of a combination of mechanisms. On long geological time scales, the origin of hematite in some older red beds may even be more complex because of the possibility of dissolution and reprecipitation of hematite during one or more earlier tectonic or crustal fluid events [43,45].

#### **4. Reduction and Leaching**

The local development of light bands of white, gray, or green color in many reddominated stratigraphic sequences is a typical feature of continental red beds. Many researchers refer to such light bands as secondary reduction zones [7]. Geologists have been observing secondary reduction zones for almost as long as the color has been discussed in

the literature [32,72–75]. According to the mineralogical characteristics and geochemical properties of secondary reduction zones in red beds, two categories have been identified: leaching zones and reduction zones. Rocks in the leaching zone tend to be white, as the iron content is depleted relative to the surrounding red host rock, due to leaching of soluble iron [7,76]. In contrast, rocks in the reduction zone are mainly gray or green and show a slight deficit in iron content relative to the peripheral red host rock. There, Fe2+ /Fe3+ values can be significantly higher and can be attributed to the formation of sulfides, carbonates, or clay minerals (mainly chlorite) [74]. Some red beds are characterized by both leaching and reduction, which are manifested by the coexistence of multiple-colored sediments in such secondary reduction zones.

According to scale of development and morphological characteristics, secondary reduction zones can exhibit three common forms in field outcrops: reduction spots, reduction strips, and reduction areas.

#### *4.1. Reduction Spots*

Light-colored zones having a circular or elliptical shape can be referred to as reduction spots or reduction points. Spatially, these reduction spots are a series of irregularly shaped reduction spheres or ellipsoidal spheres, which are early reduction features formed by diffusion of reduced material from the core of reduction spots to the surrounding area. Most reduction spheres in red beds are near-perfect spheres showing clear boundaries with the outer red host rock (Figure 3a,d,e). Some of the ellipsoidal spheres have long axes extending parallel to the stratigraphy (Figure 3e,f), mostly between 1 and 3 cm in diameter [32,59]. Their colors range from white to off-white or light green to blue-green. Reduction spots can occur in both impermeable argillaceous limestone and pore-rich conglomerates; they are like the surrounding red host rock in terms of sediment composition, particle size, roundness, porosity, and sedimentary structure [9]. In terms of chemical composition, the red host rock has a higher total iron content and a lower Fe2+/Fe3+ value than the reduction spot, but the Fe2+ content in the reduction spot is not necessarily higher than the Fe3+ value. Hematite is usually present in the matrix and cement of the red beds as microcrystalline particles or mixed with clay minerals, but is rarely present in reduction spots [7,59,77]. The formation of reduction spots is usually attributed to the in situ reduction of Fe3+ to Fe2+ from the dissolution of iron oxides. Some researchers have suggested that reduction spots form after the crystallization of pigment hematite, i.e., the sediment is red before the formation of reduction spots, and the hematite is converted to the corresponding reduction products by reduction [78,79]. Other researchers have suggested that the area where the reduction spots form never become red, i.e., the spots formed before the surrounding host rocks become pigmented, preventing the precipitation of hematite and preserving the original color of the sediment or reducing it further [7,25,55,59,77,80].

In some red beds, dark cores of black, brown, or red colors are also observed in the center of the reduced spots (Figure 3b). The dark core is often enriched with metallic elements such as Cu, V, U, and Ni and nonmetallic elements such as S and As [74,79,81,82]. Some researchers have suggested that the higher concentration of metallic elements such as copper in the reduction spots may be due to adsorption by organic matter [76–78]. More researchers have suggested that the formation of reduction spots containing dark cores is better attributed to the activity of bacteria or microorganisms that use organic matter as an energy source and reducing agent, allowing the dissolution of metallic elements and their precipitation in the cores of the reduction spots through bacterial-mediated reduction [74,75,81,83–85].

We note that carbonatite-bearing rock chip conglomerates also commonly appear in the center of the reduction spot in some conglomerate red beds. The formation of such reduction spots is closely related to the composition of the rock chips (Figure 3c,f).

**Figure 3.** (**a**) A large perfectly round white reduction spot in reddish-brown sandstone; (**b**) many reduction spots of different sizes in reddish-brown sandstone (note red core exists at center of a white reduction spot); (**c**) reduction ring around carbonate fragments in red sandstone; (**d**) sub-circular gray-white reduced spot in the red bed; (**e**) sub-elliptical light green reduction spot in red conglomerate (note its long axis is parallel to bedding); (**f**) gravelly debris in center of sub-elliptical reduction spot and crossed by late-forming fractures. Reduction spot photo from Danxia Mountain, Guangdong, China. **Figure 3.** (**a**) A large perfectly round white reduction spot in reddish-brown sandstone; (**b**) many reduction spots of different sizes in reddish-brown sandstone (note red core exists at center of a white reduction spot); (**c**) reduction ring around carbonate fragments in red sandstone; (**d**) subcircular gray-white reduced spot in the red bed; (**e**) sub-elliptical light green reduction spot in red conglomerate (note its long axis is parallel to bedding); (**f**) gravelly debris in center of sub-elliptical reduction spot and crossed by late-forming fractures. Reduction spot photo from Danxia Mountain, Guangdong, China.

#### In some red beds, dark cores of black, brown, or red colors are also observed in the *4.2. Reduction Strips*

center of the reduced spots (Figure 3b). The dark core is often enriched with metallic elements such as Cu, V, U, and Ni and nonmetallic elements such as S and As [74,79,81,82]. Some researchers have suggested that the higher concentration of metallic elements such as copper in the reduction spots may be due to adsorption by organic matter [76–78]. More researchers have suggested that the formation of reduction spots containing dark cores is better attributed to the activity of bacteria or microorganisms that use organic matter as an energy source and reducing agent, allowing the dissolution of metallic elements and their precipitation in the cores of the reduction spots through bacterial-mediated reduction [74,75,81,83–85]. We note that carbonatite-bearing rock chip conglomerates also commonly appear in the center of the reduction spot in some conglomerate red beds. The formation of such reduction spots is closely related to the composition of the rock chips (Figure 3c,f). *4.2. Reduction Strips* Secondary reduction zones distributed parallel to bedding planes, along with unconformity contacts and structural deformation zones or fractures, usually appear as strips or irregular bands in field outcrops. these are collectively referred to as reduction strips in this paper (Figure 4b,d). Compared with the surrounding red host rocks, the main composition of sediments in the reduction strip is similar, but with greater grain size, porosity, Secondary reduction zones distributed parallel to bedding planes, along with unconformity contacts and structural deformation zones or fractures, usually appear as strips or irregular bands in field outcrops. these are collectively referred to as reduction strips in this paper (Figure 4b,d). Compared with the surrounding red host rocks, the main composition of sediments in the reduction strip is similar, but with greater grain size, porosity, and permeability [59]. Therefore, the reduction strip is often developed on the sides of the relatively coarse-grained laminated units or tectonic fissures. The colors of the reduction strip are mainly white, gray, or off-white, followed by light pink, dark gray, green, etc. [42,54,86,87]. Reduction fluids use the pores or fissures in the reduction strips as channels and leach the free iron oxide from the sediments by reduction [88,89]. According to the chemical composition, reducing fluids can be divided into the following categories: (1) hydrocarbons (hydrocarbon organics associated with hydrocarbons) [90,91]; (2) organic acids (associated with surface facies formation) [54]; (3) acidic reduction fluids such as CO<sup>2</sup> and H2S (associated with volcanism and hydrothermal action) [92,93]. In addition, reduction fluids from reservoirs occasionally contain high H2S contents (>5 vol.%) as a result of thermochemical sulfate reduction of methane with capping anhydrite [94]. Most secondary reduction zones in red beds may not be a result of a single reduction, but include a variety of reducing fluids with multiple stages of reduction and leaching, and even the participation of bacteria or other microbial activities [54,87,95].

and permeability [59]. Therefore, the reduction strip is often developed on the sides of the relatively coarse-grained laminated units or tectonic fissures. The colors of the reduction strip are mainly white, gray, or off-white, followed by light pink, dark gray, green, etc. [42,54,86,87]. Reduction fluids use the pores or fissures in the reduction strips as channels and leach the free iron oxide from the sediments by reduction [88,89]. According to the chemical composition, reducing fluids can be divided into the following categories: (1) hydrocarbons (hydrocarbon organics associated with hydrocarbons) [90,91]; (2) organic

the participation of bacteria or other microbial activities [54,87,95].

**Figure 4.** (**a**) Coal measures within the green facies; (**b**) green facies and relicts of red facies in outcrop showing bleaching of red sandstone to green; (**c**) gray sandstones and conglomerates separated by red mudstone interbeds; (**d**) reduction strips in the red bed are distributed along the bedding plane. (**a**,**b**) are reprinted with permission from Ref. [15]. 2019, Elsevier; (**c**) is reprinted with permission from Ref. [91]. 2019, Elsevier; (**d**) is from Danxia Mountain, Guangdong, China. **Figure 4.** (**a**) Coal measures within the green facies; (**b**) green facies and relicts of red facies in outcrop showing bleaching of red sandstone to green; (**c**) gray sandstones and conglomerates separated by red mudstone interbeds; (**d**) reduction strips in the red bed are distributed along the bedding plane. (**a**,**b**) are reprinted with permission from Ref. [15]. 2019, Elsevier; (**c**) is reprinted with permission from Ref. [91]. 2019, Elsevier; (**d**) is from Danxia Mountain, Guangdong, China.

acids (associated with surface facies formation) [54]; (3) acidic reduction fluids such as CO<sup>2</sup> and H2S (associated with volcanism and hydrothermal action) [92,93]. In addition, reduction fluids from reservoirs occasionally contain high H2S contents (>5 vol.%) as a result of thermochemical sulfate reduction of methane with capping anhydrite [94]. Most secondary reduction zones in red beds may not be a result of a single reduction, but include a variety of reducing fluids with multiple stages of reduction and leaching, and even

#### *4.3. Reduction Areas 4.3. Reduction Areas*

Reduction areas have a similar genesis mechanism to reduction strips but require an adequate source of reduction fluids to enable the reduction of larger-scale zones and even entire stratigraphic units (Figure 4a,c). Reduction areas must involve the transport and storage of large-scale reduction fluids. The formation of large reductive areas is primarily related to hydrocarbon reservoirs. The reductive fluids are mainly hydrocarbons, followed by CO2. Most Jurassic and Cretaceous red beds in the Colorado Plateau of the United States developed abundant reducing areas with hydrocarbons as the main reducing fluid. These are closely related to nearby oil and gas reservoirs in space [89,96–98]. However, in other areas such as the red sandstone reduction zones near Green River, Utah, the reduction fluids were dominated by CO2. Fe and other trace metal elements in the matrix and cement were dissolved in low-pH reduction fluids and eventually reprecipitated near the redox front with fluid transport [93,99,100]. In addition, because the creation of reduction areas usually consumes large amounts of fluids, when they carry dissolved metals such as Cu, U, V, Co, and Au, creating enrich mineralization near redox fronts, geologists often locate copper, uranium, and other sedimentary minerals near the Reduction areas have a similar genesis mechanism to reduction strips but require an adequate source of reduction fluids to enable the reduction of larger-scale zones and even entire stratigraphic units (Figure 4a,c). Reduction areas must involve the transport and storage of large-scale reduction fluids. The formation of large reductive areas is primarily related to hydrocarbon reservoirs. The reductive fluids are mainly hydrocarbons, followed by CO2. Most Jurassic and Cretaceous red beds in the Colorado Plateau of the United States developed abundant reducing areas with hydrocarbons as the main reducing fluid. These are closely related to nearby oil and gas reservoirs in space [89,96–98]. However, in other areas such as the red sandstone reduction zones near Green River, Utah, the reduction fluids were dominated by CO2. Fe and other trace metal elements in the matrix and cement were dissolved in low-pH reduction fluids and eventually reprecipitated near the redox front with fluid transport [93,99,100]. In addition, because the creation of reduction areas usually consumes large amounts of fluids, when they carry dissolved metals such as Cu, U, V, Co, and Au, creating enrich mineralization near redox fronts, geologists often locate copper, uranium, and other sedimentary minerals near the reduction areas [7,15,42,82,101–104].

#### reduction areas [7,15,42,82,101–104]. **5. Conclusions and Future Work**

Since the Mesoproterozoic Great Oxidation Event (GOE), the free oxygen content of the atmosphere has increased sharply, and most of the Earth's surface is generally an oxidizing environment. All iron oxides or oxyhydroxides have high tinting strength and hiding power. Even if their content is less than 1% iron oxides, it is enough to stain the sediment. Therefore, the color of most iron-bearing continental sediments is reddish or yellowish. It is certain that hematite has stronger coloring and covering power than other iron oxides and oxyhydroxides; hence, the color of the red beds is almost dependent on microfine hematite. Thus, the color evolution of the red beds is essentially the evolution of hematite. In the evolution of these red beds, hematite is mainly formed during diagenesis, and iron originates from iron-rich mineral particles such as iron-bearing silicates, sulfides, and oxides, which are released into pore solutions through alterations, and then precipitated under suitable geochemical conditions forming hematite or precursors. Hematite can also be formed by dehydroxylation of iron oxyhydroxides such as goethite, an important pathway for hematite formation during red-bed burial and diagenesis, although the intermediate product of the dehydration process remains controversial [39,105–107]. Even though the free energies of the different surfaces of hematite crystals are roughly similar and their relative stability is easily modified by environmental conditions, crystal morphology usually shows great variability [63,108]. Therefore, authigenic hematite in red beds exhibit various shapes such as spherical, rod, flake, and lath, likely caused by different diagenetic environments. The reductive leaching phenomenon of red beds is often accompanied by the interaction between reducing fluids and iron oxides. Thus, the secondary reduction zone not only indicates the redox state of iron ions but also allows the color evolution history of the red beds to be probed by the redistribution traces of iron phases.

According to the progress and existing problems of previous research on the color origin of continental red beds, the authors believe that more in-depth and detailed research related to red beds in the future needs to consider the following topics:


**Author Contributions:** Conceptualization, W.H. (Wang He) and Z.Y.; methodology, W.H. (Wang He); validation, W.H. (Wang He), Z.Y. and H.D.; investigation, W.H. (Wang He); resources, H.L.; writing—original draft preparation, W.H. (Wang He); writing—review and editing, W.H. (Weisheng Hou), Z.Y. and K.Z.; visualization, H.D. and J.H.; supervision, Z.Y., W.H. (Weisheng Hou) and K.Z. All authors have read and agreed to the published version of the manuscript.

**Funding:** This study was financially supported by the Guangdong provincial Natural Science Foundation (Grant 2021A1515011472) and the Guangdong Special Fund for National Park Construction (Grant 2021GJGY026).

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**

