Multi-characterization-assisted construction of the molecular structure of high-volatile bituminous coal
Plos.org·July 20, 2026
AI Summary
Researchers used multiple analytical techniques to determine the molecular structure of a high-volatile bituminous coal sample from Inner Mongolia. Understanding coal's molecular composition helps establish structure-property relationships and improve the efficiency of coal resource utilization.
Elucidating coal molecular structures is critical for studying its structure–property relationships and advancing efficient coal resource utilization. In this study, a bituminous coal (JSM_C) was selected from the coal-rich Inner Mongolia region of China as the research object. Multi-characterization technologies, including elemental analysis, 13C solid-state nuclear magnetic resonance spectroscopy (13C NMR), Fourier transform infrared spectroscopy (FTIR), and X-ray photoelectron spectroscopy (XPS), were integrated to characterize the elemental composition, carbon skeleton, functional groups, and nitrogen species in JSM_C. The results showed that the molecular formula of JSM_C is C176H128O19N2. The carbon skeleton is centered on mono-/bi-/tricyclic aromatics connected by aliphatic or oxygen-linked chains. The oxygen-containing functional groups are mainly phenols and ethers, while nitrogen species exist in the form of pyridine and pyrrole heterocycles. The simulated spectra of NMR and FTIR are consistent with the experimental data, confirming the reliability of the constructed structure. This study achieved the construction of complex coal macromolecules from the specific mine through “characterization analysis-model construction-simulation verification,” providing a feasible paradigm for coal molecular structure research. It also lays a molecular foundation for investigating structure-performance relationships and introducing related algorithmic models in coal research.
Full story reconstructed from Plos.org. Formatting and media may differ from the original.
Citation: Feng X, Dou Y, An Y, Xiong Z, Wang Q, Shi T, et al. (2026) Multi-characterization-assisted construction of the molecular structure of high-volatile bituminous coal. PLoS One 21(7): e0354266. https://doi.org/10.1371/journal.pone.0354266
Editor: Javed Iqbal, University of Sahiwal, PAKISTAN
Received: April 16, 2026; Accepted: July 6, 2026; Published: July 20, 2026
Data Availability: All relevant data are within the paper.
Funding: This research was funded by CHN Energy Investment Group Technology Project (GJNY-24-32). The funder provided support in the form of salaries for authors Xiang Feng, Youquan Dou, Qingsong Wang and Tan Shi, but did not have any additional role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. The specific roles of these authors are articulated in the ‘author contributions’ section.
Competing interests: The authors declare that they have received funding from a commercial source: CHN Energy Investment Group Technology Project (GJNY-24-32). Authors Xiang Feng, Youquan Dou, Yuanyuan An, Zheyao Xiong, Qingsong Wang, Tan Shi, and Rudan Chen are employed by commercial companies. This commercial affiliation and funding do not alter our adherence to PLOS ONE policies on sharing data and materials.
Coal, as an important primary energy source and chemical feedstock, plays a critical role in achieving energy-structure transformation and carbon-neutrality goals [1,2]. Coal utilization performance is mainly determined by its skeletal structure and functional groups [3]. Although coals of the same rank generally possess similar macroscopic characteristics, significant microstructure heterogeneity still exists among coals from different production areas due to the influences of geological evolution histories, coal-forming pathways, and depositional environments [4–8]. This difference will be further amplified during conversion processes such as pyrolysis, gasification, and liquefaction, thereby affecting their application performance. For instance, Zhang et al.[8] reported that coking coal from Xinjiang contains more aliphatic side chains and highly reactive oxygen-containing functional groups in its macromolecular structure than North China coal of the same rank. Consequently, the resulting coke exhibits a lower coke strength after reaction (CSR), failing to meet the operational requirements of large and medium blast furnaces (above 1000 m³). Therefore, investigating the molecular structure of coal from specific mining areas holds great significance for the customized development and efficient utilization of coal resources.
Methodologies for investigating coal molecules have continually evolved, with traditional studies predominantly relying upon chemical analysis and basic spectroscopic characterization. Common characterizations, such as ultimate analysis, nuclear magnetic resonance (NMR), and Fourier-transform infrared spectroscopy (FTIR), are widely utilized to reveal the elemental composition, skeletal frameworks, and functional groups of coal molecules [4,9]. Subsequently, the rapid advancement of computational chemistry and molecular simulation technologies, coupled with traditional methods, has further enhanced the accuracy of constructing molecular models [1,10]. In recent years, the rapid iteration of artificial intelligence technologies has demonstrated substantial potential for applications within the coal industry [11,12]. For example, spectroscopic techniques combined with machine learning-based recognition models can enable rapid and accurate compositional analysis of different coal types [13–15]. However, for deep chemical reaction processes such as pyrolysis, which involve complex bond cleavage, existing end-to-end black-box models often fail to deliver satisfactory predictive performance [16]. The fundamental reason is that the coal pyrolysis behavior is governed by its highly complex macromolecular structure. Therefore, models must rely on sufficient structural data to establish an intrinsic structure–property relationship to improve both prediction accuracy and generalizability. Despite over 130 published average molecular models for coal, structural data for specific mining areas, such as the Balongtu mine area, remain scarce, which restricts the accurate assessment of coal applications and performance prediction in this region [15,17–19].
Based on the aforementioned, this study focuses on coal samples collected from the Balongtu mine area in Inner Mongolia, China. Situated within the “Energy Golden Triangle” region, this area boasts abundant coal resources, yet systematic investigations at the molecular level remain conspicuously absent. The coal sample was denoted as JSM_C, and systematically analyzed via multi-characterization techniques, including 13C NMR, FTIR, elemental analysis, and XPS. Subsequently, a three-dimensional (3D) structure of the coal molecule was constructed based on specific construction rules, and its energy was optimized using the molecular mechanics method. Finally, the 13C NMR and FTIR spectra of the 3D molecule were simulated to refine and validate the model against the experimental data. This spectral simulation ensured that the constructed 3D molecular structure was consistent with the measured constraints, establishing it as one of the representative average structural models for this coal type and laying a solid foundation for subsequent property investigations.
The coal samples used in this study were commercial coal, purchased from China Energy Trading Group Co., Ltd., and were assembled and dispatched from the Batuta Station. Therefore, no specific field site access or environmental permits were required for this work. The coal was originally collected from the Balongtu coal mine in Nalintaohai Town, Eerduosi (Ordos) City, Inner Mongolia Autonomous Region, China. The mine is located within the Shanxi–Shaanxi–Inner Mongolia energy triangle region. The coal sample was crushed, mixed, and divided in accordance with the Chinese National Standard GB/T 474‑2008 Method for preparation of coal sample, yielding coal particles with a nominal size of 3 mm. Subsamples were randomly selected multiple times for subsequent experimental use.
Proximate analysis of the coal samples was carried out according to the Chinese National Standard GB/T 212–2008 Proximate analysis of coal, to determine the contents of moisture, ash, volatile matter, and fixed carbon. Elemental analysis was performed following the relevant Chinese National Standards: GB/T 476–2008 Determination of carbon and hydrogen in coal, GB/T 19227–2008 Determination of nitrogen in coal, and GB/T 214–2007 Determination of total sulfur in coal.
The coal sample was thoroughly ground and sieved through a 400-mesh standard sieve. The sieved sample was uniformly mixed with KBr and pressed into thin pellets (thickness: ~ 0.2 mm) for FTIR testing. The spectral measurement was conducted over a wavenumber range of 400–4000 cm ⁻ ¹ with 48 scans.
The molecular structure was constructed using KingDraw software. Structural optimization was performed with the Forcite module (Dreiding force field) in Materials Studio, with the maximum iteration steps set to 60000 and energy convergence tolerance of 0.001 kcal/mol. The Fourier transform infrared (FTIR) vibrational spectrum was then simulated utilizing the DMol3 module. Additionally, the NMR spectra were simulated by the built-in function of Mnova.
Proximate analysis reveals that the air-dry basis ash content of JSM_C is 7 wt%. Its dry ash-free volatile matter and carbon content reach 37 wt% and 82 wt%, respectively. In accordance with Chinese National Standard GB/T 5751–2009 Chinese Classification of Coals, JSM_C is identified as high-quality bituminous coal characterized by low coalification degree and ultra-low ash yield [20–22]. Generally, sulfur, which has the lowest content and a high relative atomic mass, is chosen as the reference element to deduce the molecular formula. For JSM_C, the sulfur content is only 0.3 wt% (Table 1). Such a low sulfur content would result in an excessively large number of atoms in the deduced molecule, which would not only greatly increase the difficulty of structural construction but also contradict the information of a relatively low degree of coalification. Therefore, nitrogen was selected as the reference element, consistent with reports in the literature [23,24]. When the number of nitrogen atoms in a molecule is 1, 2, or 3, the corresponding molecular weights are 3889, 2593, and 1296, respectively. Li et al.[25] pointed out that the molecular weight of coal is approximately 2000–3000. In the study of Jing et al.[20], the molecular weight of high-volatile bituminous coal is about 2500. Hence, the number of nitrogen atoms was determined to be 2. Since the sulfur content is extremely low and the number of sulfur atoms is less than one in a single molecular structure, sulfur elements are temporarily excluded in the structural construction. Therefore, the molecular formula of JSM_C was confirmed to be C176H128O19N2, and the molecular weight was 2572 (without sulfur).
https://doi.org/10.1371/journal.pone.0354266.t001
Solid-state ¹³C NMR can directly identify the chemical environment of carbon atoms, so it is widely applied to characterize the carbon skeleton of coal macromolecules [20]. Fig 1a shows the 13C NMR spectrum of JSM_C, which is mainly split into three regions corresponding to aliphatic carbon peaks (0–90 ppm), aromatic carbon peaks (90–165 ppm), and carboxyl and carbonyl carbon peaks (165–220 ppm) [18,23,26,27]. The carbon types can also be classified based on SP2 and SP3 hybridization (Fig 1b) [28]. In the aliphatic carbon peak region, the chemical shifts from low to high correspond to methyl, methylene, methine, quaternary, and ether carbons, respectively. Of these, methine and quaternary carbons cannot be completely distinguished, while methylene carbon is dominant, accounting for 21.8% of total carbon atoms. The aromatic carbon region corresponds to protonated aromatic carbon, bridged carbon, alkylated aromatic carbon, and oxygen-bonded aromatic carbon. Among these, protonated aromatic carbon had the highest content, accounting for 30.6%. Carboxyl and carbonyl carbons accounted for relatively low proportions, at 1.0% and 1.9%, respectively. Detailed peak information is listed in S1 Table.
(a) Experimental and fitted curves of 13C NMR of JSM_C; (b) Schematic diagram of chemical shift ranges and structural parameters corresponding to carbon in different chemical environments.
https://doi.org/10.1371/journal.pone.0354266.g001
There are twelve structural parameters of the coal molecule derived from NMR spectroscopy (Fig 1b and Table 2). Eight parameters directly correspond to carbon structural units, namely fal* (methyl carbon), falH (methylene/methine/quaternary carbon), falO (ether carbon), faH (protonated aromatic carbon), faB (aromatic bridgehead carbon), faS (alkylated aromatic carbon), faO (oxygen-bonded aromatic carbon), and faC (ketone/carboxyl carbon). Furthermore, the remaining four carbon parameters, derived from the aforementioned eight, are faN (non-protonated aromatic carbon), fal (total sp3-hybridized carbon), fa* (total aromatic carbon), and fa (total sp2-hybridized carbon). In light of the molecular formula of JSM_C, the number of different carbon structure types in a single molecule was calculated (Table 2), which served as the basis for constructing the coal molecule.
https://doi.org/10.1371/journal.pone.0354266.t002
Generally, the aromatic rings of coal molecules are composed of stable six-membered rings [29]. During coal evolution, higher coalification degree corresponds to more six-membered rings and a larger proportion of aromatic carbon [4,30]. Six-membered rings are capable of forming clusters of different sizes, which are connected through aliphatic or oxygen-linked chains to form the coal molecule. Solum et al.[31] pointed out that the size of clusters is related to the ratio (χb) of bridgehead carbon (faB) to aromatic carbon (fa*). The χb value of JSM_C is 0.18, corresponding to an average of ~10 carbon atoms per cluster, indicating an average bicyclic configuration. In consequence, mono-/bi-/tricyclic aromatics are mainly adopted during structure construction, which is consistent with the conclusion of judging cluster size by fixed carbon content [28]. According to the χb value and structural parameters, a quantitative allocation of carbon clusters was performed, yielding 4 monocyclic, 3 bicyclic, 2 linear tricyclic, and 1 non-linear tricyclic clusters, respectively (Fig 2). The ratio of bridgehead carbon to aromatic carbon of 10 cluster units was 0.19, which was essentially consistent with the χb value of JSM_C (0.18).
https://doi.org/10.1371/journal.pone.0354266.g002
Fig 3 exhibits the FTIR spectra of JSM_C for characterizing the functional groups of the coal molecule. The FTIR spectrum (Fig 3a) is divided into three regions: 1000–1800 cm ⁻ ¹ (mainly oxygen-containing functional group vibrations), 2800–3000 cm ⁻ ¹ (mainly alkyl group vibrations), and 3300–3700 cm ⁻ ¹ (mainly hydroxyl group vibrations) [25,32–34]. Subsequently, peak fitting was performed rigorously (Figs 3b–d), and the fitting results are summarized in Table 3. The coefficients of determination (R²) for the peak fitting are 0.995 (1000–1800 cm ⁻ ¹), 0.998 (2800–3000 cm ⁻ ¹), and 0.976 (3300–3700 cm ⁻ ¹), respectively, indicating high-quality fitting results. The results reveal that oxygen-containing functional groups predominantly exist as phenols and aryl ethers, with alkyl ether content relatively low and acidic groups at the lowest level. For alkyl groups, methylene is the most predominant component, followed by methyl, while methine is the least abundant. And hydroxyl groups are mainly present as phenols and adsorbed free water. The analytical conclusions derived from these FTIR spectra are consistent with the analysis of NMR spectra.
https://doi.org/10.1371/journal.pone.0354266.t003
(a) FTIR spectrum of JSM_C; (b–d) Peak fitting results of FTIR spectra in (b) 1000–1800 cm ⁻ ¹, (c) 2800–3000 cm ⁻ ¹, and (d) 3300–3700 cm ⁻ ¹ regions. The black solid line represents the experimental curve, and the red one is the total fitted curve.
https://doi.org/10.1371/journal.pone.0354266.g003
XPS technology was used to characterize the elemental chemical environments in coal. In the wide-scan XPS spectrum (Fig 4a), the major peaks are attributed to carbon and oxygen, indicating the high abundance of these two elements. The spectrum calibration was achieved by calibrating the binding energy of the C-C bond to 284.8 eV [32,35]. As shown in Fig 4b, the C 1s narrow-scan spectrum was obtained and then subjected to peak fitting, with binding energies from low to high corresponding to C = C (283.4 eV), C-C (284.8 eV), C-O (286.1 eV), C = O (287.5 eV), and COOH (289.2 eV). Peak signal intensities for carboxylic acid groups were extremely low, with their corresponding abundance also low, which was consistent with NMR and FTIR results. For the nitrogen spectrum, there are four chemical states assigned to metal nitrides (397.1 eV), pyridine (398.8 eV), pyrrole (400.5 eV), and nitrogen oxides (402.5 eV), respectively [36,37]. Notably, pyridine (six-membered heterocycle) and pyrrole (five-membered heterocycle) are the predominant nitrogen-containing structures in JSM_C, in line with well-established insights into coal molecular studies [4]. Nitrogen oxides, meanwhile, predominantly originate from small molecules, while metal nitrides mainly arise from the combination of nitrogen and residual transition metals.
(a) XPS survey spectrum of JSM_C; (b–c) XPS spectra for (b) C 1s, and (c) N 1s. The black solid line represents the experimental curve, and the red one is the total fitted curve.
https://doi.org/10.1371/journal.pone.0354266.g004
In a word, XPS characterization reveals that nitrogen is predominantly present as pyridine (a six-membered heterocycle) and pyrrole (a five-membered heterocycle) in JSM_C.
The construction and optimization of the coal molecule of JSM_C were performed as follows:
Linking ten cluster units of different sizes via 2–3 atom-long aliphatic chains or oxygen-linked chains [4,27].
Simulating the NMR spectrum of the constructed molecule, comparing deviations between the simulated and experimental spectra, and then iteratively optimizing the molecular structure.
As shown in Figs 5a and 5b, the 2D and 3D molecular structures are presented, which have been energy-optimized using molecular mechanics. It should be noted that coal is a highly heterogeneous and complex mixture, and its measured characterization data reflect average properties. Accordingly, the proposed molecular model is an average structure representative of the coal characteristics. The constructed molecule has a formula of C176H128O19N2 and a molecular weight of 2572, consistent with the previous elemental analysis (Table 4). Actually, the determination of this final molecular structure mainly relied on the NMR experimental data as a reference. Through multiple rounds of iterative optimization, the simulated NMR spectrum was ultimately consistent with the experimental spectrum (Fig 5c), thereby ensuring structural accuracy. A comparison of the key structural parameters (fa, fa*, fal) derived from simulated NMR spectra with experimental values is presented in Table 4. The absolute deviations of these parameters are less than 2%, which confirms that the optimized structure is consistent with the constraints from experimental measurements. However, to further verify the structure’s reliability, the FTIR spectrum of the constructed molecule was also simulated and compared against the experimental FTIR data. As the constructed molecule is an average molecular structure that primarily represents the characteristic information of the coal, the FTIR simulation verification focuses on the consistency of vibrational frequencies (peak positions). The results in Fig 5d demonstrate that the main peak positions of the simulated FTIR spectrum are fundamentally consistent with the experimental spectrum, confirming that the primary functional group structures are in agreement with the measured constraints.
https://doi.org/10.1371/journal.pone.0354266.t004
(a) The optimized 2D molecular structure; (b) The 3D molecular structure after energy optimization; (c–d) Comparison of simulated and experimental (c) 13C NMR spectra, and (d) FTIR spectra.
https://doi.org/10.1371/journal.pone.0354266.g005
In summary, analyses of the simulated NMR and FTIR spectra confirm that the constructed 3D molecular structure possesses structural features consistent with experimental constraints, confirming its reliability.
In this study, the coal molecular structure of the sample JSM_C was analyzed and constructed. The molecular formula was determined to be C176H128O19N2 with a molecular weight of 2572. By integrating characterization techniques such as NMR, FTIR, and XPS, the carbon skeleton, functional groups, and chemical structure of nitrogen were confirmed, and a molecular structure model was successfully constructed. Then, the constructed structure was optimized multiple times to ensure consistency between the simulated NMR spectrum and the experimental spectrum, thus guaranteeing the correctness of the molecular structure optimization. Finally, the FTIR spectrum was simulated, and the results were found to align with the experimental spectrum, further validating the reliability of the molecular structure model. In this work, a complete process from “multi-characterization structural analysis-model construction-spectral verification” was implemented, offering a viable framework for molecular structure research on analogous coal types and establishing a molecular basis for future performance investigations and the integration of algorithmic models in practical application.
https://doi.org/10.1371/journal.pone.0354266.s001
4. Xie K-C. Structure and reactivity of coal. Beijing, China: Science Press; 2002.
5. Wiser WH. Conversion of bituminous coal to liquids and gases: chemistry and representative processes. Magnetic resonance: introduction, advanced topics and applications to fossil energy. Dordrech, Netherlands: Springer. 1984:325–50.
20. Elliott MA. Chemistry of coal utilization. Second supplementary volume ed. New York, USA: John Wiley & Sons. 1981.
21. Chinese classification of coals. Beijing, China: China Standard Press. 2009.
22. Classification for quality of coal. Beijing, China: China Standard Press. 2018.
35. Davies M. High resolution XPS of organic polymers: The Scienta ESCA300 database. Beamson G, Briggs D, editors. Chichester, UK: John Wiley. 1992.
