The Occurrence and Distribution of Nitrogen in Coal of Different Ranks and Densities

: Nitrogen is one of the significant pollutants emitted from coal combustion, and the study of its distribution and occurrence is very important for the efficient and clean utilization of coal resources. Four kinds of coal with different metamorphic ranks from major coal-producing provinces of China were studied. A gravity float-and-sink experiment was applied to obtain coal samples with different densities from Shanxi Province. The microscope optical method, Kjeldahl method, and X-ray photoelectron spectroscopy (XPS) were used to assess the occurrence, form, and distribution of nitrogen in the coal. The results show that the nitrogen content was about 0.47%–1.85%, and the maximum nitrogen content was positively correlated with the rank of coal, but the difference was not obvious. In the low-rank coal, the nitrogen content was mainly related to vitrinite and inertinite, while in the middle–high-rank coal, the nitrogen content was mainly related to inertinite and minerals. Pyrrolic (N-5) and pyridinic (N-6) were the main forms of nitrogen in the low-rank coal. The contents of N-6 and N-5 decreased with increases in the coal density, but the contents of quaternary N-Q1 and quaternary N-Q2 increased. N-Q2 mainly comes from fixed ammonia nitrogen in minerals, and vitrinite and liptinite contain more N-6 and less N-Q1 than inertinite. This research provides valuable evaluation guidance for the efficient utilization of coal.


Introduction
Nitrogen is the main constituent of the Earth's atmosphere, but its provenance in the Earth remains uncertain [1,2].The relative contribution of primordial nitrogen inherited during carbonization is unclear.Most scientists believe that nitrogen in coal mainly comes from the proteins, amino acids, porphyrins, alkaloids, etc., of plants trapped during the process of peat accumulation [3,4].These are combined with organic matter and also exist in clay minerals, such as in ammonium illite in the form of fixed ammonium [5][6][7].Previous studies have shown that the nitrogen content in Chinese coal is 0.4% to 1.6%, and the main coal accumulation periods in China were the middle Jurassic, Cenozoic Paleogene, Neogene, and Paleozoic Carboniferous-Permian, according to their low to high nitrogen contents [8][9][10].The combustion of coal produces a lot of harmful volatile substances containing nitrogen, such as NO x , N 2 O, NH 3 , and HCN, causing heavy air pollution.Therefore, research addressing nitrogen in coal is very important, especially in China with its huge coal consumption of over four billion tons per year [11][12][13].The types of organic nitrogen in coal exist as pyrrolic nitrogen, pyridinic nitrogen, quaternary, amine, protonated or oxidized heterocyclic nitrogen, and pyridinic N-oxide complexes [6,[14][15][16][17][18][19].In immature coal, pyrrolic nitrogen is the main type, followed by pyridines, pyridones, and aromatic amines [20][21][22][23]; from low-grade metamorphism to the meta-anthracite stage, edge-located nitrogen (such as pyridinic and pyrrolic nitrogen) is gradually eliminated, and some nitrogen atoms are introduced into the condensation ring to form a more stable structure of quaternary nitrogen (N-C 3 ) [19].
The different thermal stabilities of nitrogen species have been confirmed by many pyrolysis experiments, which have shown that unstable nitrogen species are converted into stable species.The protonated pyridine nitrogen is transformed into pyridine nitrogen, and pyrrolic nitrogen is converted into pyridine or quaternary nitrogen during pyrolysis [6,24,25].Under the combustion conditions, quaternary nitrogen tends to be converted into pyrrolic and pyridinic nitrogen [24].Hydrothermal treatments change pyrrolic nitrogen into pyridinic or quaternary nitrogen [25,26].The results of the thermal extraction of coal samples using methylnaphthalene and light cycle oil showed that the nitrogen in HyperCoal was mainly composed of pyrrolic nitrogen, with lower amounts of pyridinic nitrogen and no quaternary nitrogen, which also suggests that the stability of quaternary nitrogen is high [27].
Usually, coal needs to undergo a washing and flotation process before it is utilized as fuel, a coking material, or a chemical material.Washing and flotation are necessary because they can improve the calorific value and decrease the negative effects of the coal on the environment, as well as increase the ratio of active components (such as vitrinite and liptinite) to inert components (such as minerals), which has a significant influence on the metallurgy [24,[28][29][30].Flotation could substantially decrease the content of inorganic sulfur due to the high density of pyrite [31].Potentially harmful trace elements can also be segregated from the coal, due to their different occurrence states; for example, chromium is mainly combined with clay minerals, and nickel exists in both iron-bearing sulfide and organic matter [32,33].
Although there have been some studies on nitrogen in coal, few researchers have focused on the study of nitrogen content and its occurrence modes in coal samples of different densities, and there is not enough valuable guidance for the processing and washing of fine minerals.Therefore, four coals from different coalfields in Shanxi Province of China were collected, and the nitrogen content and density fractions of the samples of different species were investigated.This research is helpful for deeply understanding the occurrence mechanism of nitrogen in coal and provides a valuable reference for the clean use of coal.

Geological Setting
Shanxi Province is an important coal production base in North China.There are six coal fields in the province.Except for the Datong-Ningwu area in the north, which comprises Carboniferous Permian and Jurassic coal fields, they are all Carboniferous Permian coal fields.All the coal fields share a unified basin with the Carboniferous Permian giant coal field in North China, that is, the weathering surface of the early Paleozoic Ordovician and its lower strata.Devonian, Silurian, and Early Carboniferous strata are absent from the region [34].
The four coal seam samples in this study were collected from the four major coal fields from the north to south of Shanxi Province (Figure 1).The Datong coal field is mainly composed of Carboniferous Permian, Jurassic, and Lower Cretaceous strata, and the Xishan, Huoxi, and Qinshui coal fields are mainly composed of Carboniferous Permian and Triassic strata.The Paleogene and Neogene strata are exposed at different levels in the four fields.

Materials and Methods
The No. 8 coal from the Xiaoyu mine of the Datong coalfield (XY08) had an Ro,max of 0.82% (Table 1), the No. 8 coal from the Tunlan mine of the Xishan coalfield (TL08) had an Ro,max of 1.10%, the No. 10 coal from the Lingshi mine of the Huoxi coalfield (LS10) had an Ro,max of 1.32%, and the No. 15 coal from the Sihe mine of the Qinshui coalfield (SH15) had an Ro,max of 1.77% (Figure 1 and Table 1).All coal samples were collected from underground via the channel sampling method.These four coal seams are all from the middle and lower part of the Taiyuan Formation of Carboniferous-Permian coal-bearing strata, and belong to the same coal seam, which has regional comparability.

Materials and Methods
The No. 8 coal from the Xiaoyu mine of the Datong coalfield (XY08) had an R o,max of 0.82% (Table 1), the No. 8 coal from the Tunlan mine of the Xishan coalfield (TL08) had an R o,max of 1.10%, the No. 10 coal from the Lingshi mine of the Huoxi coalfield (LS10) had an R o,max of 1.32%, and the No. 15 coal from the Sihe mine of the Qinshui coalfield (SH15) had an R o,max of 1.77% (Figure 1 and Table 1).All coal samples were collected from underground via the channel sampling method.These four coal seams are all from the middle and lower part of the Taiyuan Formation of Carboniferous-Permian coal-bearing strata, and belong to the same coal seam, which has regional comparability.
Coal petrography analyses were performed using a Leitz Orthoplan microscope (Leica Microsystems, Wetzlar, Germany) equipped with a high-resolution digital camera and counter.The polished section of the oil-immersed coal was observed using reflected polarization, and the volume fractions of organic macerals and minerals were calculated using a counting point method.On one side of the polished section, the composition was identified at the intersection of the cross of the visual field with a fixed horizontal and vertical shifting length.Cements, cavities, cracks, and unidentified tiny particles were recorded as invalid measuring points, and no fewer than 500 testing effective points were recorded for each polished section.Then, the percentage of statistical points of each composition out of the total effective statistical points was calculated as the content of the macerals.Note: The superscript "&" indicates that the density fraction is a mixture of 1.6 g/cm 3 -1.7 g/cm 3 and 1.7 g/cm 3 -1.8g/cm 3 ; vol.%, volume percentage; R o,max , maximum vitrinite reflectance.
The maximum reflectance of the vitrinite was tested with a Berek prism illuminator and a polarizer located at 45 degrees.The polished section of coal was moved slightly to obtain a suitable area for vitrinite testing where there were no cracks, polishing sag, mineral inclusions, or other maceral fragments.The testing area was kept away from the boundary of the macerals and not influenced by the bulge.There was no high-reflectivity material, such as pyrite, within 10 µm of the outer edge of the test area.Then, light was projected onto a photoelectric converter, and the maximum reflectance was recorded as the object platform slowly rotated 360 degrees.
Coal samples (0.2 g) of less than 0.2 mm (air-dry basis) were set into a 50 mL Kelvin bottle, and 2 g of catalyst (a mixture of anhydrous sodium sulfate, mercuric sulfate, and selenium powder) and 5 mL of concentrated sulfuric acid were added.The mixture was heated to about 350 • C until it was transparent and the floating black particles disappeared completely.Then, the solution was transferred into a 250 mL Kelvin bottle and diluted to 100 mL.Subsequently, it was mixed with 25 mL of lye (37 g of sodium hydroxide +3 g of sodium sulfide dissolved in 1000 mL of distilled water) and distilled.The distilled NH 3 was absorbed using 20 mL of boric acid solution (30 g/L) with an acid-base indicator, and the color of boric acid solution changed from purple to green.Then, the sulfuric acid solution (0.025 mol/L) was used to titrate the boric acid solution until the color changed from green to steel gray.The content of nitrogen in the sample was calculated according to the amount of sulfuric acid solution, and 0.2 g of sucrose was used to replace the sample for a blank control experiment.
where N ad is the mass percentage of nitrogen in the sample (air-dry basis), %; C is the concentration of sulfuric acid solution, mol/L; M is the quality of the sample, g; V 1 is the dosage of the sulfuric acid solution in the coal sample test, mL; V 2 is the dosage of sulfuric acid solution in the blank test, mL; and 0.014 is the molar mass of nitrogen, g/mmol.
All of the DDF coal samples were tested using the XPS method with the ESCALAB 250 system (Thermo Fisher Scientific Technology (China) Co., Ltd., Beijing, China) equipped with an anode Al X-ray source and a hemispherical analyzer.The samples were crushed to 75 µm in an agate mortar and then pressed at 8 t/cm 2 before XPS analysis.The transmission energy values of the survey scan and the narrow scan were 150 eV and 60 eV, respectively.The binding energy was determined to a precision of ± 0.1 eV and calibrated taking C1s (285 eV at the maximum) as a standard.Version 4.1 of XPSPEAK software was used for the spectra curve resolution process, and the envelopes were curve-fitted with mixed Gaussian-Lorentzian functions (30% Lorentzian and 70% Gaussian).

Coal Petrology and Coal Quality
The contents of vitrinite, inertinite, liptinite, and mineral were between 6.2% and 89.36%, 5.5% and 75%, 0.0% and 5.3%, and 1.5% and 69.3%, respectively, in all DDF samples (Table 1).This showed that vitrinite, inertinite, and mineral were the dominant components of the coal, while the contents of liptinite were low in all four kinds of coal (Figure 2).The Ad yield (dry basis) of the DDF samples of the four coals had a positive correlation with the density, and the XY08 coal samples had a higher Ad yield than the TL08 coal, LS10 coal, or SH15 coal (Table 2).Except for the LS10 coal, the Odaf content (dry, ash-free oxygen) had an increasing trend with the increase in the density level and sharply increased to 35.17%, 54.79%, and 21.06%, respectively, in the XY08 coal, TL08 coal, and SH15 coal at a density of >1.8 g/cm 3 .
The Vdaf yields (dry ash-free basis) were between 16.09% and 36.05%,18.70% and Generally, with increases in density, the vitrinite content had a downward trend, the inertinite content showed an increasing-decreasing trend, and the mineral content had an upward trend.The mineral content increased sharply and reached 69.30%, 54.21%, 53.35%, 39.00% in the XY08, TL08, LS10, and SH15 samples at the density level of >1.8 g/cm 3 , while the contents of vitrinite, inertinite, and liptinite tended to sharply decrease.This reflects that the mineral contents were dominant in the high-density DDF samples.The 1.4 g/cm 3 -1.5 g/cm 3 samples of the XY08 coal and the 1.5 g/cm 3 -1.6 g/cm 3 samples of the TL08 coal, LS10 coal, and SH15 coal had the lowest mineral contents compared with the other DDF samples.With the increase in the coal's rank, the maximum value of the mineral content decreased in the DDF samples of >1.8 g/cm 3 (Figure 2d).
Vitrinite mainly appeared at a density level of <1.5 g/cm 3 , and its relative content in the XY08 coal was less than in the TL08 coal, LS10 coal, and SH15 coal (Figure 2a).The vitrinite content in the XY08 coal and SH15 coal had better negative correlation with the density level, while the vitrinite contents in the TL08 coal and LS10 coal suddenly increased at the >1.8 g/cm 3 density level compared with the 1.7 g/cm 3 -1.8g/cm 3 level.
The inertinite contents in the DDF samples of the LS10 coal were obviously lower than in the XY08 coal, TL08 coal, and SH15 coal.The DDF samples of 1.5 g/cm 3 -1.8g/cm 3 density level comprised high values of inertinite content, greater than 50% in the XY08 coal, TL08 coal, and SH15 coal (Figure 2b).
The relative contents of liptinite were between 0.2% and 5.3% in the DDF samples of the XY08 coal, obviously higher those that in the TL08 coal, LS10 coal, and SH15 coal (Figure 2c).
The A d yield (dry basis) of the DDF samples of the four coals had a positive correlation with the density, and the XY08 coal samples had a higher A d yield than the TL08 coal, LS10 coal, or SH15 coal (Table 2).Except for the LS10 coal, the O daf content (dry, ash-free oxygen) had an increasing trend with the increase in the density level and sharply increased to 35.17%, 54.79%, and 21.06%, respectively, in the XY08 coal, TL08 coal, and SH15 coal at a density of >1.8 g/cm 3 .
The V daf yields (dry ash-free basis) were between 16.09% and 36.05%,18.70% and 25.90%, 15.56% and 22.56%, and 5.59% and 40.16% in the DDF samples of the XY08 coal, TL08 coal, LS10 coal, and SH15 coal, respectively.At the density level of <1.6 g/cm 3 , the V daf yield in the DDF samples of low-rank coal was higher than that in the higher-rank coal.This change was consistent with the degree of coal metamorphism.The V daf yield was obviously higher in the DDF samples of SH156 coal (Table 2); this was partly due to the thermal decomposition of the carbonate minerals (CaO comprised 13.18%).
The FC d contents (dry basis) were between 9.92% and 57.97%, 51.57% and 74.50%, 30.36% and 78.15%, and 28.10% and 87.85%, respectively, in the DDF samples of the XY08 coal, TL08 coal, LS10 coal, and SH15 coal.With the increase in the density level, the FC d contents in the DDF samples of the four coals generally decreased.
According to the results of the ultimate analysis, the C daf also had a decreasing trend when the density of the DDF samples increased, which demonstrates that the organic matter was lower in the relatively high-density samples.The C daf value was between 33.86% and 81.35%, 22.66% and 88.76%, 62.87% and 89.86%, and 73.64% and 93.52% in the DDF samples of the XY08 coal, TL08 coal, LS10 coal, and SH15 coal, respectively (Table 2).This shows that the carbon content was higher in the higher-rank coal than it was in the low-rank coal.
The S t,daf had a sharp increase at a density of >1.8 g/cm 3 , with values of 23.02%, 15.69%, and 31.97%,respectively, in the XY08 coal, TL08 coal, and LS10 coal.The DDF samples of the SH15 coal had a relatively stable S t,daf value, which was obvious lower than that in the XY08 coal, TL08 coal, and LS10 coal (Table 2).Note: M, moisture; V, volatile; FC, fixed carbon; A, ash; S t , total sulfur; d, dry basis; daf, dry ash-free basis; the superscript "*" indicates that the nitrogen content was obtained using the Kjeldahl method; wt.%, weight percentage.

Nitrogen Content
The Kjeldahl nitrogen contents (N daf %) in the DDF samples of the XY08 coal, TL08 coal, LS10 coal, and SH15 coal were 0.85%-1.35%,0.47%-1.47%,1.09%-1.49%,and 1.17%-1.85%,respectively (Table 2).There were sharp decreases in the nitrogen contents in the DDF samples of the XY08 coal and TL08 coal with the increase in density (Figure 3a).In contrast, the nitrogen contents in the DDF samples of the SH15 coal increased sharply, and the nitrogen contents in the DDF samples of the LS10 coal increased first and then decreased.Taking 1.5 g/cm 3 -1.6 g/cm 3 as the boundary, when the density was higher than 1.6 g/cm 3 , the nitrogen contents in the SH15 DDF samples were higher than those in the DDF samples of the XY08 coal, TL08 coal, and LS10 coal; when the density was lower than 1.5 g/cm 3 , the nitrogen contents in the SH15 DDF samples were less than those in the DDF samples of the XY08 coal, TL08 coal, and LS10 coal (Figure 3a).
Variations in the nitrogen content with the carbon content are often used to characterize the evolution behavior of nitrogen during coal formation [38][39][40].The relationships between the N daf % and C daf % in all the DDF samples were established (Figure 3b).They showed that with the increase in the C daf %, the distribution of the N daf % changed along a parabola trend, and the N daf % reached its peak value when the C daf % was near 85 wt.%.Some studies have shown a similar relationship between the N daf % and C daf % in coal [19,41,42].It can also be seen that the DDF samples may have had very different N daf % when the C daf % was similar; near the C daf % of 85 wt.%, the N daf % values in the SH154 and SH155 coals were significantly higher than in the other DDF samples (Figure 3b).The differences in the maceral characteristics were compared with the DDF samples whose C daf % was near Minerals 2024, 14, 549 8 of 14 85 wt.%, and it was found that SH154 and SH155 had obviously low vitrinite contents (10.8% and 14%, mineral matter basis) and high inertinite contents (71.1% and 67.4%, mineral matter basis).In addition, the DDF sample of SH156 had a low C daf % value of 73.64%, which was less than 85%; however, its N daf % value reached the highest content of 1.85% and it had the lowest vitrinite content (6.2%, mineral matter basis) among all DDF samples.This means that in higher-rank coal, the content of N daf % was more closely correlated with the mineral content.The SH156 had the highest mineral content of 53.04%.Variations in the nitrogen content with the carbon content are often used to char terize the evolution behavior of nitrogen during coal formation [38][39][40].The relationsh between the Ndaf% and Cdaf% in all the DDF samples were established (Figure 3b).Th showed that with the increase in the Cdaf%, the distribution of the Ndaf% changed alon parabola trend, and the Ndaf% reached its peak value when the Cdaf% was near 85 wt Some studies have shown a similar relationship between the Ndaf% and Cdaf% in c [19,41,42].It can also be seen that the DDF samples may have had very different Nd when the Cdaf% was similar; near the Cdaf% of 85 wt.%, the Ndaf% values in the SH154 a SH155 coals were significantly higher than in the other DDF samples (Figure 3b).The d ferences in the maceral characteristics were compared with the DDF samples whose Cd was near 85 wt.%, and it was found that SH154 and SH155 had obviously low vitrin contents (10.8% and 14%, mineral matter basis) and high inertinite contents (71.1% a 67.4%, mineral matter basis).In addition, the DDF sample of SH156 had a low Cdaf% va of 73.64%, which was less than 85%; however, its Ndaf% value reached the highest cont of 1.85% and it had the lowest vitrinite content (6.2%, mineral matter basis) among all D samples.This means that in higher-rank coal, the content of Ndaf% was more closely c related with the mineral content.The SH156 had the highest mineral content of 53.04% The differences in macerals may have affected the nitrogen contents in the DDF sa ples of coal.There were strong positive correlations between the Ndaf% and the vitrin inertinite, and liptinite contents in the XY08 coal, with r = 0.68, 0.67, and 0.80, respective and there were weak positive correlations between the Ndaf% and the macerals in the TL coal (Figure 4a-c).Strong negative correlations between the nitrogen content and the m eral content appeared in the XY08 and TL08 coal samples, with r = −0.99 and −0.98, resp tively (Figure 4d).This indicates that nitrogen has closer relationships with organic m erals, and minerals make less contribution to the nitrogen in lower-rank coal.Several stu ies have shown that the nitrogen content in vitrinite is higher than that in inertinite [4 45], and the content of liptinite in coal is very low (no more than 5.5%) (Table 1).Therefo it is thought that macerals, especially those in vitrinite, make a significant contribution The differences in macerals may have affected the nitrogen contents in the DDF samples of coal.There were strong positive correlations between the N daf % and the vitrinite, inertinite, and liptinite contents in the XY08 coal, with r = 0.68, 0.67, and 0.80, respectively, and there were weak positive correlations between the N daf % and the macerals in the TL08 coal (Figure 4a-c).Strong negative correlations between the nitrogen content and the mineral content appeared in the XY08 and TL08 coal samples, with r = −0.99 and −0.98, respectively (Figure 4d).This indicates that nitrogen has closer relationships with organic macerals, and minerals make less contribution to the nitrogen in lower-rank coal.Several studies have shown that the nitrogen content in vitrinite is higher than that in inertinite [43][44][45], and the content of liptinite in coal is very low (no more than 5.5%) (Table 1).Therefore, it is thought that macerals, especially those in vitrinite, make a significant contribution to the content of nitrogen in lower-rank coal.
For the LS10 coal, only the inertinite had a high correlation coefficient of 0.69 (Figure 4b).The nitrogen content in the SH15 coal had a strong negative correlation with the vitrinite content (Figure 4a) and strong positive correlations with the inertinite content (Figure 4b) and the mineral content (Figure 4d).This showed that the inertinite and minerals made an important contribution to the nitrogen occurrence in SH15 coal.
These results indicate that the macerals had a certain influence on the N daf % in the coals of different ranks.In the low-and middle-rank coal, the nitrogen content was the most correlated with the macerals; in the low-rank coal, it was correlated with the vitrinite; in the middle-low-rank coal, it was correlated with the inertinite, and in the middle-high-rank coal, it was most correlated with the inertinite and minerals.Although the overall nitrogen content slowly increased, this also meant that the occurrence state of N was related to the coal rank, and the inorganic nitrogen content may have increased.
content slowly increased, this also meant that the occurrence state of N was related to the coal rank, and the inorganic nitrogen content may have increased.

Nitrogen Occurrence State of DDF Samples
The N1s XPS spectra of the XY08 coal DDF samples were analyzed (Figure 5).A total of five sub-peaks were determined in the curve-fitting process of the N1s XPS spectra.The binding energies of the sub-peaks were fixed at 398.8 eV (N-6), 400.6 eV (N-5), 401.3 eV (N-Q1), 402.7 eV (N-Q2), and 403.6 eV (N-X).The full width at the half-maximum of the sub-peaks was fixed at 1.4 eV.Sub-peak N-6 is related to pyridinic nitrogen, in which one nitrogen atom replaces one carbon atom in the aromatic ring and connects with two carbon atoms.Sub-peak N-5 is related to three kinds of organic nitrogen: (i) the pyrrolic nitrogen structure, in which a nitrogen atom replaces a carbon atom in a non-aromatic ring and is bonded to a hydrogen atom and two carbon atoms; (ii) pyrrolic or pyridinic nitrogen with an oxygen substituent; (iii) amine or pyridone.It was not possible to distinguish between pyrrolic, amide, amines, and pyridone using XPS, and hereafter, N-5 is collectively referred to as pyrrolic nitrogen.Sub-peak N-Q1 is mainly related to quaternary nitrogen (named N-C3), in which a nitrogen atom with a positive charge or a neutral sp 3 coordination is bonded with three carbon atoms in a condensation partial aromatic system.Sub-peak N-Q2 is related to fixed ammonia in minerals, and sub-peak N-X is mainly related to pyridinic N-oxide complexes.Meanwhile, the N-C3 type may also contribute to N-Q2 and N-X [6,19].

Nitrogen Occurrence State of DDF Samples
The N1s XPS spectra of the XY08 coal DDF samples were analyzed (Figure 5).A total of five sub-peaks were determined in the curve-fitting process of the N1s XPS spectra.The binding energies of the sub-peaks were fixed at 398.8 eV (N-6), 400.6 eV (N-5), 401.3 eV (N-Q1), 402.7 eV (N-Q2), and 403.6 eV (N-X).The full width at the half-maximum of the sub-peaks was fixed at 1.4 eV.Sub-peak N-6 is related to pyridinic nitrogen, in which one nitrogen atom replaces one carbon atom in the aromatic ring and connects with two carbon atoms.Sub-peak N-5 is related to three kinds of organic nitrogen: (i) the pyrrolic nitrogen structure, in which a nitrogen atom replaces a carbon atom in a non-aromatic ring and is bonded to a hydrogen atom and two carbon atoms; (ii) pyrrolic or pyridinic nitrogen with an oxygen substituent; (iii) amine or pyridone.It was not possible to distinguish between pyrrolic, amide, amines, and pyridone using XPS, and hereafter, N-5 is collectively referred to as pyrrolic nitrogen.Sub-peak N-Q1 is mainly related to quaternary nitrogen (named N-C3), in which a nitrogen atom with a positive charge or a neutral sp 3 coordination is bonded with three carbon atoms in a condensation partial aromatic system.Sub-peak N-Q2 is related to fixed ammonia in minerals, and sub-peak N-X is mainly related to pyridinic N-oxide complexes.Meanwhile, the N-C3 type may also contribute to N-Q2 and N-X [6,19].

Sub-Peaks N-X: Oxidized Nitrogen
The sub-peaks of N-X and N-Q2 with a binding energy of >402 eV remain unresolved and have three possible contributions, including pyridinic N-oxide complexes, quaternary nitrogen, and ammonium fixed in clay minerals [6,19].The weak N-X peak in residues from low-temperature ashing of ammonium illite-containing coal and the presence of a N-X peak in kerogen without silicate minerals demonstrated that the N-X may be not related to the NH4 + in clay minerals, and it may correspond to the pyridinic N-oxide complexes rather than quaternary nitrogen [19].The N-X sub-peak could also result from oxidation during preparation and storage [19,46].The content of N-X in the DDF samples of the XY08 coal did not show a significant change; it presents an abnormally low point in Figure 6, which may be due to the complexity of the N-X source and the heterogeneity of the coal.Since the content of N-X was relatively low, it is not discussed further in this work.

Sub-Peak N-Q2: Contributing to Inorganic Nitrogen or Organic Nitrogen
A strong N-Q2 sub-peak was found in the low-temperature ashing of ammonium illite-containing coal, which confirmed the significant contribution of NH4 + in clay minerals to the N-Q2 peak [19].Figure 7a shows the abundance changes in the N-Q2 sub-peak with ash yield in the DDF samples of the XY08 coal.It can be seen that although the relative content of N-Q2 had a period of decline, it experienced an overall upward trend with the increase in ash yield.The abundance of N-Q2 had a strong positive correlation with the ash yield (r = 0.82).This demonstrates that the inorganic ammonium nitrogen in the

Sub-Peaks N-X: Oxidized Nitrogen
The sub-peaks of N-X and N-Q2 with a binding energy of >402 eV remain unresolved and have three possible contributions, including pyridinic N-oxide complexes, quaternary nitrogen, and ammonium fixed in clay minerals [6,19].The weak N-X peak in residues from low-temperature ashing of ammonium illite-containing coal and the presence of a N-X peak in kerogen without silicate minerals demonstrated that the N-X may be not related to the NH 4 + in clay minerals, and it may correspond to the pyridinic N-oxide complexes rather than quaternary nitrogen [19].The N-X sub-peak could also result from oxidation during preparation and storage [19,46].The content of N-X in the DDF samples of the XY08 coal did not show a significant change; it presents an abnormally low point in Figure 6, which may be due to the complexity of the N-X source and the heterogeneity of the coal.Since the content of N-X was relatively low, it is not discussed further in this work.The sub-peaks of N-X and N-Q2 with a binding energy of >402 eV remain unresolved and have three possible contributions, including pyridinic N-oxide complexes, quaternary nitrogen, and ammonium fixed in clay minerals [6,19].The weak N-X peak in residues from low-temperature ashing of ammonium illite-containing coal and the presence of a N-X peak in kerogen without silicate minerals demonstrated that the N-X may be not related to the NH4 + in clay minerals, and it may correspond to the pyridinic N-oxide complexes rather than quaternary nitrogen [19].The N-X sub-peak could also result from oxidation during preparation and storage [19,46].The content of N-X in the DDF samples of the XY08 coal did not show a significant change; it presents an abnormally low point in Figure 6, which may be due to the complexity of the N-X source and the heterogeneity of the coal.Since the content of N-X was relatively low, it is not discussed further in this work.A strong N-Q2 sub-peak was found in the low-temperature ashing of ammonium illite-containing coal, which confirmed the significant contribution of NH4 + in clay minerals to the N-Q2 peak [19].Figure 7a shows the abundance changes in the N-Q2 sub-peak with ash yield in the DDF samples of the XY08 coal.It can be seen that although the relative content of N-Q2 had a period of decline, it experienced an overall upward trend with the increase in ash yield.The abundance of N-Q2 had a strong positive correlation with the ash yield (r = 0.82).This demonstrates that the inorganic ammonium nitrogen in the  + in clay minerals to the N-Q2 peak [19].Figure 7a shows the abundance changes in the N-Q2 sub-peak with ash yield in the DDF samples of the XY08 coal.It can be seen that although the relative content of N-Q2 had a period of decline, it experienced an overall upward trend with the increase in ash yield.The abundance of N-Q2 had a strong positive correlation with the ash yield (r = 0.82).This demonstrates that the inorganic ammonium nitrogen in the minerals in the DDF samples of the XY08 coal made a significant contribution to the content of N-Q2.minerals in the DDF samples of the XY08 coal made a significant contribution to the content of N-Q2.The N-Q2 sub-peak in the pure kerogen samples without minerals indicates that organic nitrogen may also contribute to N-Q2 and is thought to represent organic quaternary nitrogen (N-C3), because when nitrogen replaces carbon in a condensed aromatic system, the charge transfer and hybridization produce a higher binding energy [19].Figure 7b-d show the abundance changes in N-Q2 with vitrinite, inertinite, and liptinite in the DDF samples of the XY08 coal.With the increases in the vitrinite and liptinite contents (vol.%,mineral matter basis), there was an overall downward trend in the content of N-Q2, and there were strong negative correlations between the abundance of N-Q2 and the contents of vitrinite (r = −0.85)and liptinite (r = −0.88).The inertinite content (vol.%,mineral matter basis) showed a weak negative correlation with the abundance of N-Q2.Therefore, it is inferred that the contribution of organic nitrogen N-C3 to the abundance of the N-Q2 peak may have been very weak in the XY08 coal, and the N-Q2 peak mainly originated from the fixed ammonium in the minerals.The sub-peaks N-6, N-5, and N-Q1 are distinctly from the contribution of organic nitrogen, therefore the relative percentage content of N-6, N-5, and N-Q1 in the DDF samples of the XY08 coal was normalized to a hundred percent, and the organic macerals (vitrinite, inertinite, and liptinite) were also normalized to the mineral-free basis.Then, the relationships between N-6, N-5, and N-Q1 with the organic macerals were assessed (Figure 8).The N-Q2 sub-peak in the pure kerogen samples without minerals indicates that organic nitrogen may also contribute to N-Q2 and is thought to represent organic quaternary nitrogen (N-C3), because when nitrogen replaces carbon in a condensed aromatic system, the charge transfer and hybridization produce a higher binding energy [19].Figure 7b-d show the abundance changes in N-Q2 with vitrinite, inertinite, and liptinite in the DDF samples of the XY08 coal.With the increases in the vitrinite and liptinite contents (vol.%,mineral matter basis), there was an overall downward trend in the content of N-Q2, and there were strong negative correlations between the abundance of N-Q2 and the contents of vitrinite (r = −0.85)and liptinite (r = −0.88).The inertinite content (vol.%,mineral matter basis) showed a weak negative correlation with the abundance of N-Q2.Therefore, it is inferred that the contribution of organic nitrogen N-C3 to the abundance of the N-Q2 peak may have been very weak in the XY08 coal, and the N-Q2 peak mainly originated from the fixed ammonium in the minerals.The sub-peaks N-6, N-5, and N-Q1 are distinctly from the contribution of organic nitrogen, therefore the relative percentage content of N-6, N-5, and N-Q1 in the DDF samples of the XY08 coal was normalized to a hundred percent, and the organic macerals (vitrinite, inertinite, and liptinite) were also normalized to the mineral-free basis.Then, the relationships between N-6, N-5, and N-Q1 with the organic macerals were assessed (Figure 8).minerals in the DDF samples of the XY08 coal made a significant contribution to the content of N-Q2.The N-Q2 sub-peak in the pure kerogen samples without minerals indicates that organic nitrogen may also contribute to N-Q2 and is thought to represent organic quaternary nitrogen (N-C3), because when nitrogen replaces carbon in a condensed aromatic system, the charge transfer and hybridization produce a higher binding energy [19].Figure 7b-d show the abundance changes in N-Q2 with vitrinite, inertinite, and liptinite in the DDF samples of the XY08 coal.With the increases in the vitrinite and liptinite contents (vol.%,mineral matter basis), there was an overall downward trend in the content of N-Q2, and there were strong negative correlations between the abundance of N-Q2 and the contents of vitrinite (r = −0.85)and liptinite (r = −0.88).The inertinite content (vol.%,mineral matter basis) showed a weak negative correlation with the abundance of N-Q2.Therefore, it is inferred that the contribution of organic nitrogen N-C3 to the abundance of the N-Q2 peak may have been very weak in the XY08 coal, and the N-Q2 peak mainly originated from the fixed ammonium in the minerals.The sub-peaks N-6, N-5, and N-Q1 are distinctly from the contribution of organic nitrogen, therefore the relative percentage content of N-6, N-5, and N-Q1 in the DDF samples of the XY08 coal was normalized to a hundred percent, and the organic macerals (vitrinite, inertinite, and liptinite) were also normalized to the mineral-free basis.Then, the relationships between N-6, N-5, and N-Q1 with the organic macerals were assessed (Figure 8).N-5 abundance maintained a basically stable level with the increase in vitrinite, inertinite, and liptinite (mineral-free basis), while the N-6 and N-Q1 abundances changed significantly, indicating that the organic macerals may have had an important effect on the abundances of N-6 and N-Q1 in the DDF samples of the XY08 coal. Figure 8a,c show that with the increase in the vitrinite and liptinite contents (mineral-free basis), the abundance of N-6 exhibited an increasing trend, and the abundance of N-Q1 decreased.Figure 8b shows that with the increase in the inertinite content, the abundance of N-6 decreased and the abundance of N-Q1 increased.Therefore, there may be more N-6 and less N-Q1 in vitrinite and liptinite, and less N-6 and more N-Q1 in inertinite.According to the relative contents of the three organic macerals in the coal, the vitrinite and inertinite accounted for the main contents, so priority was given to analyzing the effects of vitrinite and inertinite on the changes in N-6 and N-Q1 in the DDF samples.
As a result of coal-mining activities, the nitrogen cycle in the environment has been significantly altered and nitrogen pollution has increased [47,48].Especially, the change of the occurrence forms of N in char during ammonia-coal co-pyrolysis affects the release of N and the formation of NO x [49].Therefore, if coal washing can be carried out according to the content and occurrence form of nitrogen, it will reduce the proportion of nitrogen in raw coal, which is beneficial to reducing the environmental pollution caused by coal combustion.

Conclusions
In this study, the occurrence and distribution of nitrogen in coal of different ranks and densities were analyzed by coal petrology, geochemistry, and the XPS method.The results showed that there were obvious differences in the contents of N daf in the coals of different ranks and densities.
(1) The nitrogen content was about 0.47%-1.85%,and the maximum nitrogen content was positively correlated with the rank of the coal, but the difference was not obvious.
The N daf % reached a maximum value near 85 wt.% of carbon content.(2) In the low-rank coal, the nitrogen content was mainly related to vitrinite and inertinite, while in the middle-high-rank coal, the nitrogen content was mainly related to inertinite and minerals.The higher rank coal had higher nitrogen content when the density was >1.6 g/cm 3 , while it had lower nitrogen content when the density was <1.5 g/cm 3 .(3) Pyrrolic (N-5) and pyridinic (N-6) were the main forms of nitrogen in the low-rank coal.The contents of N-6 and N-5 decreased with increases in the coal density, but the contents of quaternary N-Q1 and quaternary N-Q2 increased.In addition, the vitrinite and liptinite may have contained more N-6 and less N-Q1, while the inertinite may have contained less N-6 and more N-Q1.N-Q2 mainly came from fixed ammonia nitrogen in minerals.The nitrogen tended to form a more stable structure in the high-density-level samples.

Figure 1 .
Figure 1.The location of the coal sample collection.

Figure 1 .
Figure 1.The location of the coal sample collection.

Figure 3 .
Figure 3. (a) Characteristics of nitrogen contents (N daf %) in DDF samples of XY08 coal, TL08 coal, LS10 coal, and SH15 coal; (b) relationship between N daf % and C daf % in samples with different density fractions (DDF).

Figure 4 .
Figure 4.The correlations between the nitrogen content and the contents of vitrinite (a), inertinite (b), liptinite (c), and minerals (d) in the DDF samples.

Figure 4 .
Figure 4.The correlations between the nitrogen content and the contents of vitrinite (a), inertinite (b), liptinite (c), and minerals (d) in the DDF samples.

Figure 5 .
Figure 5. XPS spectra of N1s of DDF samples of XY08 coal.

Figure 6 .
Figure 6.Change trends of the nitrogen species contents in the DDF samples of the XY08 coal.

Figure 5 .
Figure 5. XPS spectra of N1s of DDF samples of XY08 coal.

Figure 6 .
Figure 6.Change trends of the nitrogen species contents in the DDF samples of the XY08 coal.

Figure 6 .
Figure 6.Change trends of the nitrogen species contents in the DDF samples of the XY08 coal.

4. 3 . 2 .
Sub-Peak N-Q2: Contributing to Inorganic Nitrogen or Organic Nitrogen A strong N-Q2 sub-peak was found in the low-temperature ashing of ammonium illite-containing coal, which confirmed the significant contribution of NH 4

Figure 7 .
Figure 7.The relationships betwee the N-Q2 abundance and the ash yield (a), vitrinite (b), inertinite (c), and liptinite (d) contents of the DDF samples of the XY08 coal.

Figure 7 .
Figure 7.The relationships betwee the N-Q2 abundance and the ash yield (a), vitrinite (b), inertinite (c), and liptinite (d) contents of the DDF samples of the XY08 coal.

Figure 7 .
Figure 7.The relationships betwee the N-Q2 abundance and the ash yield (a), vitrinite (b), inertinite (c), and liptinite (d) contents of the DDF samples of the XY08 coal.

Table 1 .
Maceral compositions of coal with different density fractions.

Table 1 .
Maceral compositions of coal with different density fractions.

Table 2 .
Proximate and ultimate analyses of coal with different density fractions.