Characteristics of an Altered Diabase Dike in a Coal Seam: A Case Study from the Datong Coalfield, Shanxi, China

Carboniferous–Permian coal seams in the Datong coalfield are intruded by Mesozoic magmatic dikes and sills. Our observations indicate that these dikes and sills have been seriously altered and have lost their original characteristics. Although this phenomenon has been mentioned in some studies, there is a lack of detailed research on the alteration mechanism. To fill this research gap, six dike samples were collected in this study to analyze their alteration characteristics and mechanisms. Petrographic analysis, scanning electron microscopy, and X-ray diffraction were used to determine the mineralogy of the altered igneous samples. The results suggest that the mineral alteration is associated with severe carbonation and clavization. The carbonation of mafic minerals is due to the release of CO2 generated from the coking of high volatile bituminous coals that may have occurred during the intrusive event. However, labradorite did not react with CO2. The water activity in the coal seam was enhanced by the dikes, and mafic minerals were converted into mixtures of chlorite and smectite group minerals. The water activity is especially high at the edges of the dike. Albite and clay minerals were formed due to the albitization of labradorite. Based on these effects, the dike can be divided into weak alteration, carbonation, and clavization bands. The alteration processes in the coal seam are described in detail in this work, and potential formation mechanisms are suggested.


Introduction
Igneous intrusions in coal seams have been reported for many coalfields, for example, coalfields in the United States [1][2][3], Australia [4][5][6], China [7,8], and India [9][10][11]. A large body of literature is available regarding the contact metamorphism of coal. However, the alteration characteristics of intrusive bodies in coal seams have rarely been discussed. Kisch and Taylor reported the alteration of an andesite sill in the Collinsville coalfield [4] and that of a basanite sill in Baralaba [5], Australia. The intrusive bodies turned white after entering the coal seam and were thus named "White traps." Petrological and mineralogical analyses of the altered intrusions suggested that the porphyry in the intrusive rocks consists of plagioclase, while the matrix contains kaolinite, chert, siderite, ankerite, and calcite. Plagioclase crystals near the coal/rock contact zone were altered to clay minerals, and mafic minerals were converted into carbonate minerals. These changes may be related to the release of CO 2 from the coal. Ward et al.observed dike zoning while studying the impact of a diabase dike on a coal seam in Belmont, New South Wales, Australia [6]. Pyroxene was altered to biotite and chlorite, while plagioclase did not undergo alteration in the central part of the dike (fresh, dark green in color). Dolomite, siderite, chlorite, and montmorillonite were the dominant minerals at the edges of the dike (altered, pale brown in color). Chen reported changes in the color of the upper part of the sill near the coal seam in the Huainan coal mine, China [8], and suggested that this phenomenon is a result of changes in the lithology of the upper sill. The migration of silicate minerals from the lower part of the sill to the upper part may have resulted in the formation of light-colored granodiorite. Although color changes associated with igneous intrusions in coal seams have been reported in the above-mentioned studies, alteration mechanisms have rarely been discussed. The aim of the current study was to elucidate the alteration mechanisms in igneous dikes in the Datong coalfield.
During the exploration and mining of Carboniferous-Permian coal seams in the Datong coalfield, basic-ultrabasic lamprophyre, carbonatite (229 ± 11 Ma) [12][13][14][15], and diabase (115-140 Ma) [16,17] were encountered. These intrusions occur mainly as dikes and sills. The most notable feature of these intrusions is that they mainly are light gray or pale yellow (Figure 14-17 in the Supplementary Material), in contrast to typical lamprophyre or diabase ( Figure 18 in the Supplementary Material). As mentioned previously, this phenomenon was also reported in Queensland, New South Wales, Australia, and Huainan, China. Our aim was to determine the responsible alteration mechanisms using several analytical techniques.

Geologic Setting
The Datong coalfield is located in the northern part of the Shanxi Province, China, and is~50 km long (NE-SW) and 30 km wide (NW-SE), covering a total area of 1828 km 2 . The coalfield is bounded by the Qingciyao fault to the northeast, Kouquan fault to the south and southeast, and Hongtaoshan anticline to the southwest. The northwest is the predicted denudation boundary of the coal seam [16][17][18][19][20][21] ( Figure 1). The base of the Datong coalfield consists of an Archaean gneiss. Cambrian strata mainly consist of mudstones, limestones, and dolomites and have an angular unconformity with the Archaean strata. The Ordovician strata include limestones, dolomites, and dolomitic limestones and are conformable with the lower Cambrian strata. Due to Caledonian tectonic movement, the Upper Ordovician, Silurian, Devonian, and Lower Carboniferous strata are absent. Coal in the Datong Basin is mainly found in the Taiyuan and Shanxi formations in the Upper Carboniferous-Lower Permian system. Due to collisions between the North China Craton and Siberian plates, intense magmatism occurred during the Triassic period [22]. The Datong coalfield, located at the northern margin of the North China Craton, is intruded by basic-ultrabasic lamprophyres and carbonatites [12]. The Yanshan tectonic movement during the Late Jurassic to Early Cretaceous periods also affected the Datong coalfield; the coal seam is intruded by diabase. Furthermore, magmatic activity is mainly concentrated in faulted areas in the central and northern regions of the coalfield, where dikes and sills occur.

Samples and Methods
3.1. Samples. Rock samples (TS2-TS5), mixed coal-rock samples (TS1 and TS6), natural coke samples (TS7-TS9), and normal coal samples (TS10 and TS11) were collected from a diabase dike and from the surrounding area ( Figure 2). Samples TS10 and TS11 were collected 10 and 20 m away from the coal/dike contact zone, respectively. The mineral morphologies of TS2-TS5 were semiquantitatively analyzed using a TESCAN-MIRA3 cold-field emission scanning electron microscope (SEM) with an energy-dispersive spectrometer (EDS, Oxford-X-MaxN20, Si (Li) detector). The spectra were analyzed using the AZtecEnergy software (Oxford Instruments). The mineral composition of samples TS2-TS5 was determined with an X-ray diffractometer (XRD; Rigaku SmartLab X) using Ni-filtered Cu-Kα radiation at 18 kW and a scintillation detector. The XRD patterns were recorded in a 2θ range of 3°-60°with a step size of 0.02°. The initial data, High-Score Plus software, and ICDD PDF database were used to identify the mineral composition.

Field Observations and Macroscopic
Characteristics of the Samples. The average width of the dike is 4.6 m; it gradually widens upward. The dike exhibits notable color changes on the macroscopic scale. The micrographs can be roughly divided into three bands (Figure 2), which from the edge to the middle of the dike are: pale yellow, light green, and dark green.
(1) The pale yellow band is located 50-70 cm from the edge of the dike. The hand specimen is pale yellow, and the mineral particle size is less than 0.2 mm, which is almost indistinguishable to the naked eye (Figures 3(a) and 3(b)). The minerals were  TS1 TS2  TS3  TS4  TS5  TS6  TS7  TS8  TS9 145°C oal Natural coke  3 Geofluids completely clavizated and reacted with dilute hydrochloric acid. Notably, a large amount of gray white and translucent carbonate minerals can be observed in the coal/rock contact zone. The coal that solidified after liquefaction fills the fissures of the magmatic rock ( Figure 4) (2) The light green band is 0.3-0.6 m wide. The hand specimen is light green and fine grained. The mineral particle size is 0.3-1.5 mm, with an average size of 0.5 mm (Figure 3(c)). The plagioclase crystals are gray white and exhibit a subvitreous luster. Most of the pyroxene was altered. The unaltered pyroxene is dark green and has a vitreous luster.
The olivine crystals have diameters of 1-1.5 mm and were completely altered  Geofluids major phenocrysts include augite (~28-43 vol%, average of 31 vol%), plagioclase (~27-46 vol%, average of 37 vol%), and olivine (~3-8 vol%). The plagioclase crystals in the phenocrysts are 1-1.5 mm in size and euhedral. The extinction angles of the Carlsbad and albite twins indicate that the plagioclase is labradorite. Augite is the major mafic mineral. It is subhedral, pale green, and fills the gaps between the plagioclase crystals. Almost 30%-40% of the augite was converted into brown-green mixtures of chlorite and smectite. Olivine phenocrysts are 0.5-2 mm in size, euhedral, and colorless. The cracks and margins of the olivines were altered to serpentine. The matrix is fine and microcrystalline; only plagioclase and magnetite can be identified in the matrix. The samples taken from the light green band exhibit a porphyritic texture. The major phenocrysts in these samples are augite (~25-42 vol%, average of 33 vol%), plagioclase (~27-41 vol%, average of 35 vol%), and olivine (~3-8 vol%). The major minerals in the light green area (TS5) are augite (~37-44 vol%, average of 42 vol%) and plagioclase (~43-55 vol%, average of 50 vol%). The minor minerals include olivine (~3-8 vol%), magnetite (3 vol%), and carbonate (2 vol%). The labradorite crystals in the phenocrysts are 0.8-1.5 mm in size and euhedral. The augite in sample TS5 was altered to a greater extent than that in sample TS3; 70% of the augite was converted to mixtures of chlorite and smectite and carbonate. The euhedral olivine phenocrysts are 1.5-2 mm in diameter and were replaced by carbonate minerals (maintaining the olivine crystal shape; the cracks and margins of the olivine were altered to serpentine; Figures 5(e) and 5(f)). Note that the sample contains many carbonate spherules (0.4-1 mm; Figures 5(g) and 5(h)). The matrix varies from fine grained to microcrystalline and plagioclase, magnetite, and carbonate minerals can be identified.
All minerals in the pale yellow band underwent severe alteration (Figures 5(i)-5(l)). Plagioclase in the phenocrysts and matrix was replaced by fine clay mineral particles and retained its crystalline morphology. Augite in this band was altered to a greater extent than that in sample TS5; all augites in this band were altered to dark brown carbonate and mixtures of chlorite and smectite. The euhedral olivine phenocrysts are 1.5-3 mm in diameter and were also altered to dark brown carbonate minerals and mixtures of chlorite and smectite (maintaining the olivine crystal shape).
The point counting method was used to evaluate the felsic mineral content of sample TS3. According to the QAP diagram of the IUGS [30], the lithology of this dike can be classified as diabase ( Figure 6). The thin section indicates a porphyritic texture and euhedral plagioclase and subhedral augite. The olivine concentration is~5%, indicating an olivine diabase porphyrite dike. A similar texture can be observed in the altered samples TS2 and TS5, which indicates that the entire dike has the same lithology, but different bands experienced different alteration processes.
Based on the SEM-EDS analysis of sample TS3 (Figures 7(a) and 7(b)), the mineral composition at point c was altered; the main elements at this point are Mg, Al, Fe, and Si (Figure 7(c)). The mineral composition indicates that it is augite, and its surface was gradually altered into fine schistose mixtures of chlorite and smectite (Figure 7(a)). Almost no alteration can be observed at point d, and the EDS analysis indicates the lack of Al (Figure 7(d)), suggesting that the mineral at this point is olivine. The mineral texture at point f is smooth, and the long columnar morphology was not altered. Based on the EDS analysis results, the ratio of silicon to aluminum is~2 : 1 (Figure 7(e)). The mineral contains sodium and calcium (Figure 7(f)), suggesting the inclusion of labradorite. Fluorapatite can be found as an accessory mineral at point f.
The degree of mineral alteration in sample TS5 is much higher than that of sample TS3 (Figures 8(a) and  8(b)). The well-preserved mineral is labradorite, which exhibits almost no alteration (point d). Augite was almost completely altered (points c and e). At a similar magnification, it was observed that augite in sample TS5 was altered to schistose mixtures of chlorite and smectite (5-25 μm in size; points c and e). At the same time, many 10-15 μm rhombic Ca-bearing Mg-siderite crystals (point f) can be observed.
Sample TS2 was completely altered. The original mineral is absent and a large number of 2-5 μm clay fraction grains, which are common alteration products of labradorite, can be observed (Figure 9(a)). Based on the EDS analysis results, the ratio of silicon to aluminum in the clastic minerals at point c is~1 : 1 (Figure 9(d)), indicating that these are kaolin group clay minerals (1 : 1 clay mineral field). In sample TS6, kaolin group minerals (Figure 9(b)) and many Ca-bearing Mg-siderite crystals (8-20 μm) can be observed (Figure 9(c)).

4.3.
Mineralogy of the Samples of Different Bands. The alteration characteristics of the samples determined using macroscopic and microscopic methods correspond to the results of the XRD analysis ( Figure 10). The XRD spectrum of sample TS3 is simple and characteristic peaks are notable, which suggests that the sample underwent very little 5 Geofluids alteration. Based on the intense reflections at~3.18 Å and weaker reflections at~3.22 and~3.73 Å, the most prominent mineral is labradorite. The other major mineral is augite, with an intense reflection at~2.98 Å and weaker reflections at~2.93 Å and~2.53 Å. The spectrum corresponding to sample TS5 shows that the main minerals are labradorite and augite. However, unlike sample TS3, the characteristic siderite peaks, that is, an intense peak at 2.79 Å and weaker reflections at~3.58 Å and~2.14 Å, are also present. In addition, the peak intensity of augite decreased. The spectrum corresponding to sample TS2 is very different from the spectra of the other two bands. This sample contains albite (intense peak at~3.23 Å and weaker peaks at~4.11 Å and~3.74 Å), siderite, and clay minerals. The intense reflections at~7.1 Å and~3.56 Å are representative of the kaolin group clay minerals. In general, the mixtures of chlorite and smectite group minerals yield intense peaks at~12 Å and~14 Å and can be observed in weathered magmatic rocks.

Discussion
Based on the results, we infer that the minerals in the samples from the middle to the edge of the dike experienced different   6 Geofluids levels and types of alteration, which resulted in different colors. The mineral alteration of this dike is discussed in this study from the perspective of the alteration of mafic minerals and plagioclase.

Dike Carbonation.
Carbonation occurs at a distance of 1-1.5 m from the edge of the dike. Optical microscopy, SEM, and XRD data show that the pale yellow and light green bands contain many carbonate minerals. Based on Kisch [5], the carbonate production may be attributed to carbon dioxide generation from the coking of the coal seams at high temperature. Based on the normal coal VRr and quality data (

Geofluids
When the coal was thermally influenced by magma, the volatile content decreased and the carbon content significantly increased, indicating that the coal experienced devolatilization [1,3,5,32,33]. The normal coal samples collected in this study are high volatile bituminous coal. When the coal rank increases rapidly, the release of volatile gas by devolatilization can be divided into two stages. In stage I (VRr < 1:2%), a large quantity of CO 2 and a small amount of methane are released. In stage II (VRr > 1:2%), methane is produced [5,34,35]. The overall process based on which the coal rank rises can be summarized using the following equation [36]: During this process, the gases are thought to migrate out of the rapidly heated coal because of the decreased gas adsorption ability of the metamorphosed coal. The excess gas is pushed into the surrounding lithologies [37,38]. Mafic minerals have good CO 2 absorption abilities and have been widely used for the absorption of artificial CO 2 from the atmosphere [39,40]. The conditions during magma intrusion in coal seams may be similar to those during artificial CO 2 injection. The chemical formulas of the mafic minerals can be regarded as a combination of MgO, FeO, CaO, and NaO+Al 2 O 3 . In the presence of CO 2 , these compounds may undergo carbonation depending on the environmental conditions [41]. The temperature of the diabase intrusion is~1100°C-1200°C, while the depth of the buried coal seam is~750-1000 m and the hydrostatic pressure is~75-100 bar [10]. According to the spontaneous reaction conditions for minerals and CO 2 summarized by Marini [41], the conditions for the spontaneous reaction of pyroxene and olivine with CO 2 are met when the magma temperature is 350°C-400°C ( Figure 11). Therefore, more carbonate minerals are formed at this stage. The corresponding reactions are as follows: Unlike the mafic minerals, labradorite does not undergo carbonation. Compared with the spontaneous reaction of plagioclase with CO 2 , the temperature required for the reaction between labradorite and CO 2 is similar to the   9 Geofluids temperature required for the reactions of albite and anorthite with CO 2 (150°C-300°C) but lower than that of the reaction between pyroxene and olivine with CO 2 (Figure 12). Because the mafic minerals are not fully carbonated, they can absorb most of the produced CO 2 and plagioclase does not undergo carbonation.

Dike Clavization.
Clavization primarily occurs at a distance of 0.5-0.7 m from the edge of the dike. In this band, the color of the sample turned pale yellow and the rock hardness decreased significantly. Microscopically, labradorite disappeared and the albite and kaolinite contents increased. In addition, the mafic minerals were completely altered to siderite and mixtures of chlorite and smectite group minerals.
According to Boles [42,43], the alteration of plagioclase into clay may be attributed to the activity of pore water rich in Na + , which leads to the albitization of plagioclase. In this process, Ca 2+ in labradorite is replaced by Na + to form albite, while redundant Al 3+ forms clay minerals. When the magmatic rock intrudes in the coal, it cuts through the stratum and connects the water in the coal seam, roof, and floor, which would induce a localized convection cell and then resulting in an enhanced groundwater activity at the edges of the dike. Furthermore, high concentrations of Na + in the coal seam, roof, and floor water [44] may lead to a similar albitization process. The albitization reaction can be written as follows [42]: 10 Geofluids During the albitization of plagioclase, mafic minerals are also altered to chlorite. Several new alteration minerals, such as sericite, epidote, and iron oxides, are formed during this process [45]. Němec suggested that albitization represents the intense alteration of mafic minerals [46]. Therefore, compared with plagioclase, mafic minerals are more prone to alteration to mixtures of the chlorite and smectite group minerals in the presence of groundwater. Some of the mafic minerals in the middle of the dike were altered to chlorite and smectite group minerals, which means that the underground water influenced the entire dike. The groundwater activity was the strongest at 0.5-0.7 m from the edge of the dike, resulting in the albitization of labradorite and chloritization and smectitization of mafic minerals.

Dike
Zoning. Based on the observation of carbonation and clavization in the dike, the dike first underwent carbonation during the condensation process and mafic minerals were altered to carbonate minerals (major carbonate mineral: Ca-bearing Mg-siderite) at the edge of the dike with a width of 1-1.5 m. Subsequently, the groundwater activity affected the entire dike, causing different levels of chloritization and smectitization of the mafic minerals. At the edge of the dike with a width of 0.5-0.7 m, intense groundwater activity led to the albitization of labradorite. Therefore, the dike can be divided into three bands, that is, a weak alteration band (weak chloritization and smectitization), carbonation band (carbonation, chloritization, and smectitization), and clavization band (albitization as domination, chloritization, smectitization, and carbonation), which correspond to the dark green, light green, and pale yellow bands in the macroscopic images, respectively. The different alteration processes that occur in the three bands are illustrated in the schematic diagram in Figure 13.

Conclusions
In this study, we employed several analytical techniques, including SEM, optical microscopy, and XRD, to evaluate the mineral alteration of intrusions in coal seams in the Datong coalfield, Shanxi, China. Our main conclusions are as follows: (1) After magma intrudes the coal seam, in addition to differences in the degree alteration in the coal seams on both sides, the dike itself experiences an alteration zoning effect (2) When magmatic rock intruded into the low-rank coal, the coal first released a large quantity of CO 2 . The mafic minerals in the intrusion reacted with the CO 2 , altering the mafic minerals into carbonate minerals (Ca-bearing Mg-siderite, among others), that is, forming a carbonate band. Labradorite finds it more difficult to participate in the carbonation reaction when compared to mafic minerals and hence it remains mostly unaltered 11 Geofluids alteration. A weak alteration band and a clavization band formed due to the groundwater activity This study sheds light on the alteration mechanisms, and the findings support the mineralogical analysis and prediction of different coal seams. Understanding the mechanisms of alteration in igneous intrusions in coal seams will help to predict and analyze the mineralogical characteristics of coal seams.

Data Availability
The mineralogical, petrographic, and coal quality data used to support the findings of this study are included in the article. Robert Finkelman, Dr. David French, and Professor Shaoqing Wang for their help with the coal analyses; Professor Anchao Zhou for his help with the mineralogical analyses;  Figure 14: (a) a narrow diabase dike (60 cm wide, yellow) in the Xiaoyu coal mine, Datong coalfield. Photomicrographs of (b) PPL and (c) XPL of this dike. Plagioclase crystals were completely altered to clay minerals, while augite and olivine were altered to carbonate minerals and chlorite. Figure 15: (a) a diabase dike (pale gray) in the Wajinwan coal mine, Datong coalfield. Photomicrographs of (b) PPL and (c) XPL of this dike. Plagioclase crystals were completely altered to clay minerals, while augite and olivine were altered to carbonate minerals and chlorite. Figure 16: a diabase dike (pale yellow) in the Jinzhuang coal mine, Datong coalfield. Figure 17: a diabase sill (pale gray) in the Dongzhouyao coal mine, Datong coalfield. Figure 18: a lamprophyre dike (dark green to dark gray) in Emaokou, Datong coalfield. Photomicrographs of (b) PPL and (c) XPL of this dike. For lamprophyric texture, the minerals were not seriously altered. (Supplementary Materials)