Molecular representation of coal-derived asphaltene based on high resolution mass spectrometry

Yonghong Zhu; Chongpeng Du; Huaan Zheng; Feili Wang; Feng Tian; Xu Liu; Dong Li

doi:10.1016/j.arabjc.2021.103531

View/Download PDF

Buy Reprints

PDF

Translate this page into:

Original article

01 2021

:15;

103531

doi:

10.1016/j.arabjc.2021.103531

Molecular representation of coal-derived asphaltene based on high resolution mass spectrometry

Yonghong Zhu^{^a,^b}, Chongpeng Du^{^a,^c}, Huaan Zheng^{^a}, Feili Wang^{^a}, Feng Tian^{^a}, Xu Liu^{^d}, Dong Li^{^a,⁎}

a School of Chemical Engineering, Northwest University, Chemical Engineering Research Center of the Ministry of Education for Advanced Use Technology of Shanbei Energy, Shaanxi Research Center of Engineering Technology for Clean Coal Conversion, Xi’an 710069, Shaanxi, China

b Hydrocarbon High-Efficiency Utilization Technology Research Center, Yanchang Petroleum Co. Ltd., Xi’an 710075, Shaanxi, China

c Chemical Engineering Department, Yulin Vocational Technical College, Yulin 719000, Shaanxi, China

d China Qiyuan Engineering Co. Ltd., Xi’an 710018, Shaanxi, China

⁎Corresponding author at: School of Chemical Engineering, Northwest University, Taibai North Road, Beilin District, Xi’an 710075, Shaanxi, China. lidong@nwu.edu.cn (Dong Li)

Received: 2021-8-29, Accepted: 2021-10-27,

Disclaimer:
This article was originally published by Elsevier and was migrated to Scientific Scholar after the change of Publisher.

Peer review under responsibility of King Saud University.

Abstract

The asphaltene separated by solubility in small molecular alkanes and toluene is the most structurally diverse and complex components in heavy oil, such as vacuum residue and coal tar. The coal-derived asphaltene is always regard as a succession of maltene fraction from small molecules to large molecules, and also a continuum of island- and archipelago-type structures, which is difficult to be identified accurately through current characterization methods. This limits the further study of molecular dynamics and reaction dynamics simulation of asphaltenes. In this work, a representation model of molecular composition and structure for coal-derived asphaltene is developed mainly based on Fourier transform ion cyclotron resonance mass spectrometry (FT-ICR MS) coupled with collision induced dissociation (CID) and traditional methods of nuclear magnetic resonance spectroscopy (¹³C NMR), Fourier transform infrared spectroscopy (FT-IR), X-ray photoelectron spectroscopy (XPS). Island- and archipelago-type structures are considered qualitatively in the representation of asphaltene. The asphaltene molecules are systematic assembled using stochastic algorithms and optimized by simulated annealing algorithm according to the group contribution method. The bulk properties for simulating asphaltenes are in good agreement with the experimental results giving acceptable predictions for the composition and structure of the asphaltenes. Moreover, the representative average structure asphaltene molecules are obtained using the developed molecular similarity function, which could be applied in the further study of molecular aggregation simulation and reaction kinetics simulation.

Keywords

Coal-derived asphaltene

CID FT-ICR MS

Molecular composition

Average asphaltene structure

Show Related Articles from PubMed

Nomenclature

CID: Collision induced dissociation
IRMPD: Infrared multiphoton dissociation
PDF: Probability density function
DBE: Double-bond equivalences
PAHs: Polycyclic aromatic hydrocarbons
CNs: Carbon numbers
HTSD: High temperature simulated distillation
α: The shape parameter in gamma function
β: The scale parameter in gamma function
F_MS: Molecular similarity function
F_AR+NR: Aromatic and naphthenic ring function
F_HG: Heteroatom group function
F_ASC: Aliphtic side chain function
F_HG-N: N-containing group function
F_HG-O: O-containing group function
F_HG-S: S-containing group function
F_ASC-n: The average alkyl substituent number function
F_ASC-l: The average substituted aliphtic chain length function
NNG: The number of N-containing groups
NOG: The number of O-containing groups
NSG: The number of S-containing groups

1 Introduction

With the rapid development of modern coal chemical industry, the output of coal tar as the key product of coal pyrolysis technology has achieved more than fifteen million tons at recent years (Sun et al., 2019). How to achieve efficient and clean utilization of coal tar, a high-quality secondary processing fossil energy, is particularly important and urgent for countries where coal is the main form of energy, especially for China (Feng et al., 2018). However, the heavy components in coal tar always result in many thorny issues, such as the catalyst poisoning, pipeline blockage, reactor coking and low conversion rate, during the hydrogenation, catalytic cracking and delayed coking processes (Zhu et al., 2021). Asphaltene is with no doubt the most complex fraction, with high content of heteroatoms and polycyclic aromatic hydrocarbons (PAHs), of coal tar in terms of chemical structure and composition (Shao et al., 2018, Pei et al., 2017, Sun et al., 2015 ). In order to solve the problems of low conversion and serious coking in heavy oil processing technology, and promote the further study of reaction kinetics of heavy oil system and molecular simulation dynamics of asphaltene, it is necessary to have a clear understanding of the molecular composition and structure of coal tar asphaltenes (coal-derived asphaltene), including the confirmation for average molecular structure and the establishment of the molecular library of asphaltenes consistent with the macroscopic physical and chemical properties (Sheremata et al., 2004).

However, the coal-derived asphaltene is difficult to analyze comprehensively in the molecular level with traditional characteristic methods including Fourier transform infrared spectroscopy (FT-IR), X-ray photoelectron spectroscopy (XPS), X-ray diffraction (XRD), nuclear magnetic resonance spectroscopy (¹³C NMR, ¹H NMR), thermogravimetric analysis (TG), pyrolysis gas chromatography-mass spectrometer (Py-GCMS) (Sun et al., 2015, Alvarez-Majmutov et al., 2019) and even with modern advanced detected technology of Fourier transform ion cyclotron resonance mass spectrometry (FT-ICR MS), atomic force microscopy (AFM) and scanning tunnelling microscopy (STM) (Schuler et al., 2015, Schuler et al., 2020, Schuler et al., 2017, Zheng et al., 2020, Cho et al., 2011, Glattke et al., 2020, Chacón-Patiño et al., 2020a ). It is importantly noted that a great quantity of excellent work has been done by Schuler et al. (2015, 2020) and Zhang et al. (2018), who innovatively obtained the asphaltene molecular skeleton images of petroleum and coal-derived samples from the advanced AFM and STM technologies at recent years. Up to now, the limitations of this method is that this technology may improve the possibility of selective, nonrepresentational extraction from the bulk, or more likely, only aim at quasi-planar components for selective visualization due to the complexity of natural asphaltene materials and the surface nature of imaging technology (Scott et al., 2021). The current results haven’t fully defined the “mean molecular composition” of native asphaltenes. Totally, each measurement technology could only reveal the limited aspects of heavy components structure (Zhang et al., 2014). Therefore, it is necessary to incorporate various characterization methods to obtain complete fundamentally composition and structure information of asphaltene. Nevertheless, the current characterization methods are still inadequate for asphaltene mixtures due to the large variety of chemical structures, leading to the polydispersity of composition information is entirely lost (De Oliveira et al., 2004).

Molecular reconstruction modeling is an efficient method for expanding the boundaries of asphaltene analysis. This type of model enables simulation of heavy oil composition from model hypotheses, analytical characterization data and chemical knowledge by systematically generating a set of computational mixtures of representative hydrocarbon molecules (Feng et al., 2019, Boek et al., 2009). Neurock et al. (1990, 1994) established the Monte Carlo algorithm for generating model molecules of heavy oil fractions from detail physical and chemical properties obtained by various characterization methods with the concept of attributes and probability density function (PDF). Sheremata et al. (2004) proposed a quantitative molecular representation (QMR) method to determine the average molecular structure of asphaltenes. In this method, a series of asphaltene structure fragments were determined and a large number of possible asphaltene molecular structures are formed according to the defined chemical rules. Boek et al. (2009) and Headen et al. (2009) improved the QMR method by extending the connection mode of structural fragments from alkyl chain to chemical bond. Frigerio and Molinari (2011) also used QMR method to study the average molecular structure of six kinds of asphaltenes through expanding the structure fragment library of asphaltenes.

Quann and Jaffe modeled the light, middle and heavy petroleum fractions by developing the method of structure oriented lumping (SOL) (Quann and Jaffe, 1992, Quann and Jaffe, 1996, Jaffe et al., 2005 ). The structure blocks of molecules in mixtures were described by vector of different elements. Two algorithms were developed based on stochastic reconstruction (SR) and reconstruction by entropy maximization (REM), respectively, by generating a group of model molecules from overall petroleum, which has been applied to light cycle oil, vacuum gas oil and vacuum residues (Hudebine and Verstraete, 2004, Verstraete et al., 2004, Verstraete et al., 2010 ). Zhang et al. (2014) proposed the structural representation of petroleum residue molecules through sampling with PDF and a juxtaposition of structure attributes according to FT-ICR MS and knowledge of light oil.

Although there are still many uncertainties of composition concentrations affected by the signal responses of analytes, which are highly dependent on the ionization efficiency of different molecular structure, the FT-ICR MS coupled with collision induced dissociation (CID) equipped with positive electrospray ionization (+ESI), negative electrospray ionization (-ESI) or positive atmospheric pressure photoionization (+APPI) sources can still provide specific qualitative composition information in terms of double bound equivalent (DBE), carbon numbers (CNs), distribution of aromatic ring, basic and non-basic functional groups and chemical formula, greatly improving the accuracy of molecular representation model (Hsu, 2012, Pereira et al., 2014, Pan et al., 2012, Ballard et al., 2020, Shi et al., 2010, Gaspar et al., 2012, Thomas et al., 2021 ). Niles et al. (2020) studied an unaltered Illinois coal No. 6 asphaltene using an extrography separation method by (+) APPI FT-ICR MS coupled with infrared multiphoton dissociation (IRMPD) demonstrating that coal asphaltene contains archipelago and island compounds with low aromatic polyfunctional oxygenated species similar to asphaltenes in solubility. Multifunctional archipelago compounds are rich in the most polarizable parts of coal and petroleum asphaltenes, but their ionization ability is poor.

In present study, the coal-derived asphaltene is prepared by an improved separating method proposed in previous study (Zhu et al., 2021). A composition representation model for coal-derived asphaltene is firstly developed applying stochastic reconstruction and simulated annealing optimization algorithms based on the fundamental experimental information. The fundamental composition and structure information (island and archipelago) of coal-derived asphaltene are innovatively obtained by CID FT-ICR MS with (±) ESI and (+) APPI ionization sources, solid state ¹³C NMR, XPS, FT-IR, etc. The bulk properties for simulating asphaltene (about 100,000 molecules) are coincident with the experimental results. Importantly, the representative average structure molecules are determined using the developed molecular similarity function, which is also verified by the molecular chemical fingerprint depend on the classical Morgen fingerprint algorithm.

2 Experimental section

2.1

2.1 Materials

The middle and low temperature coal tar for this study is obtained from the coal pyrolysis process in Shenmu industrial park, Shaanxi Province, China. The asphaltenes are obtained based on an improved separation method published in our previous paper including two-stage asphaltene extraction process (Zhu et al., 2021). 3 g coal tar sample is mixed with 150 mL heptane, refluxed for 2 h, and then filtered into heptane soluble component and heptane insoluble component. The heptane insoluble components with filter are extracted by Soxhlet method with 150 mL toluene. The soluble part of toluene is evaporated and dried in vacuum to obtain the unpurified asphaltene. Then, the unpurified asphaltene is crushed using an agate mortar and extracted by Soxhlet apparatus for 24 h with 200 mL heptane as the washing solvent. Finally, the maltene and purified asphaltene are obtained under vacuum evaporation and drying.

2.2

2.2 Sample characterization

Element composition of C, H, N, S was determined by the VarioEL III elemental analyzer. The molecular weight is obtained from GPC on HP110 analyzer. The solvent is tetrahydrofuran with a flow vrate of tetrahydrofuran solvent at 1 mL·min⁻¹. The distillation range is detected by high temperature simulated distillation (HTSD) based on Agilent 7890B-GC referring to ASTM D7169 standard. X-ray photoelectron spectroscopy is operated on An ThermoFisher ESCALAB 250Xi with an Al kα source. ¹³C NMR is performed on Bruker AVANCE III 600 at a resonance frequency of 150.92 MHz using TMS as the external reference for ¹³C spectrum chemical shifts. The asphaltene is characterized by Bruker SolariX XR FT-ICR MS equipped with a 9.4 T superconducting magnet. The ionization sources are positive electrospray ionization (+ESI), negative electrospray ionization (-ESI) and positive atmospheric pressure photoionization (+APPI). For the positive-ion ESI mode, the samples are dissolved in toluene to 10 mg/mL and then diluted to 0.15 mg/mL with toluene/methanol (3:7, v:v), then 10 μL of formic acid was added in the sample solution prior to injection. The operating conditions for Positive -ion ESI formation are a −4.3 kV emitter voltage, a −4.8 kV capillary column introduction voltage, and 320 V capillary column end voltage. For the negative-ion ESI mode, the samples are also dissolved in toluene to 10 mg/mL and then diluted to 0.15 mg/mL with toluene/methanol (3:7, v:v), then 12 μL of ammonium hydroxide (28 wt%) was added in the sample solution prior to injection. The operating conditions for negative-ion ESI formation are a 3.8 kV emitter voltage, a 4.3 kV capillary column introduction voltage, and −320 V capillary column end voltage. For the positive-ion APPI mode, the samples are dissolved in toluene to 10 mg/mL and then diluted to 0.25 mg/mL with toluene. The operating conditions for Positive -ion APPI formation are a −4.0 kV emitter voltage, a −4.5 kV capillary column introduction voltage, and 320 V capillary column end voltage.

In the CID experiments, the collision voltage is −45 eV. In the quadruple isolation experiment, due to the low signal intensity of isolation and CID FT-ICR MS, the cumulative time is increased to 1.0 s. The isolation window width is 3 Da. Double-bond equivalences (DBE), representing the number of rings plus the number of double bonds for a given molecular formula of C_cH_hN_nO_oS_s, is calculated by the following equation:

(1)

DBE = c-h/2 + n/2 + 1 where c and h are unlimited, 0 ≤ n ≤ 3, 0 ≤ o ≤ 10, and 0 ≤ s ≤ 3.

3 Modeling approach

3.1

3.1 Structure attributes

Probability density functions (PDFs). PDFs are applied to describe the quantitative aspects of the attribute distributions (Zhang et al., 2014). The definition of a PDF is shown in Eq (2), where f(x) is the PDF of variable x with the range of 0 to 1.

(2)

\sum f (x) = 1 (0 \leq x ≪ 1)

Each structural attribute can also be represented by a PDF, such as gamma and histograms functions used in this model. The functional form of histogram distribution is described in Eq (3), where X_i is the PDF value of each structure attribute:

(3)

f (x) = X_{i}

The gamma function is regard as the best PDF to represent the distribution for most of the heavy oil structural attributes (Verstraete et al., 2010, Pyl et al., 2011). The expression form is defined in Eq (4):

(4)

f (x) = \frac{{(x - γ)}^{α - 1} e^{- (x - γ) / β}}{Γ (α) β^{α}}

Where α and β are the shape and scale parameters, respectively.

The gamma functions describe a continuous distribution containing a wild range of possible values, as is the case for the number of aromatic rings, naphthenic rings, S, N, O atoms, alkyl side chains and length of per alkyl side chain, while the histogram is suitable for the description of structural attributes with a narrow range of possible values, such as the type of cores, basic nitrogen groups and oxygen groups (De Oliveira et al., 2004). In order to reconstruct the coal-derived asphaltene molecules, 23 parameters of 12 PDFs are generated in the model depicted in Table 1.

Table 1 Probability distribution functions (PDFs) applied in the molecular reconstruction model of coal-derived asphaltene.

Structure attribute	Values	PDF type	No. of parameters
No. of aromatic cores^a	0, 1	Histogram	1
No. of aromatic rings	1–10	Gamma	2
No. of naphthenic rings	≥0	Gamma	2
No. of sulfur atoms	0–1	Gamma	2
No. of basic nitrogen atoms	0–3	Gamma	2
Type of basic nitrogen groups^b	0 or 1	Histogram	1
No. of non-basic nitrogen atoms	0–3	Gamma	2
No. of oxygen atoms	0–10	Gamma	2
Type of oxygen groups^c	0, 1, 2 or 3	Histogram	3
No. of alkyl side chains	≥0	Gamma	2
Length of per alkyl side chain	≥1	Gamma	2
Type of inter-core linkages^d	0, 1 or 2	Histogram	2

a Type of cores: 0, mono-aromatic cores; 1, multi-aromatic cores.

b Type of basic nitrogen groups: 0, pyridine functional group; 1, amino group;

c Type of oxygen groups: 0, hydroxyl functional group; 1, furan group; 2, carbonyl functional group; 3, carboxyl functional group;

d Type of inter-core linkages: 0, two PAHs connected by a single bond; 1, two PAHs connected by –CH₂–; 2, two PAHs connected by –O–.

Definition of islands and archipelago structures. At present, the structures of petroleum asphaltenes are divided into island- and archipelago-type. Many studies have discussed the proportion and structure features of the island and archipelago molecules, mainly focusing on the asphaltenes derived from vacuum residue or other heavy petroleum fraction (Juyal et al., 2013, Chacón-Patiño et al., 2017, Mckenna et al., 2010, Podgorski et al., 2013, Chacón-Patiño et al., 2018a, 2018b, McKenna et al., 2019 ). Petroleum asphaltenes are the continuum of island and archipelagic structures. The dominant structure of island or archipelagic depends on different asphaltene samples. Fig. 1 shows the fragmentation FT-ICR mass spectra and iso-abundance map of carbon numbers (CNs) vs DBE for precursor and fragment ions at m/z of 365–367 Da for coal-derived asphaltene. A boundary line (blue dotted) of DBE is defined to obtain qualitative conclusions of the distributions of island- and archipelago-type molecules by Chacón-Patiño et al. (2018a, 2018b) and McKenna et al. (2019). The boundary value of DBE is calculated by the weighted average of DBE values of the precursor ions minus the weighted standard deviation. The fragment ions below this boundary are attributed to the archipelago structures generated by the loss of DBE, and correspondingly, fragment ions above this boundary line represent island structures. The DBE value of the boundary line are calculated by the weighted average values of DBE for the precursor ions minus the weighted standard deviation. As the iso-abundance map shows, the proportion of fragment ions above the DBE boundary is 80%, indicating the dominant structure type for coal-derived asphaltene is island based on the definition of boundary line. From the fragmentation mass spectra, it’s obviously that the intensity of archipelago fragment ions mark in purple dotted line are weak. However, it is worth noting that the monomeric (non-aggregated) asphaltene are preferentially observed by FT-ICR MS (Chacón-Patiño et al., 2017). Compared with island structures, the samples enriched in archipelagic structures tend to strongly aggregate and form larger nano aggregates, which reduces the ion signal during the ionization process resulting in the difficulties of asphaltene molecule detecting (Chacón-Patiño et al., 2020b, Scott et al., 2021). Therefore, the conclusion, that the dominant structure type for coal-derived asphaltene is island, should be mainly qualitative and semi quantitative.

Left: fragmentation FT-ICR mass spectra. Right: iso-abundance map of CNs vs DBE for precursor and fragment ions at m/z of 365–367 Da for coal-derived asphaltene.

The reactions of dealkylation side chains evidently occurred during CID, which is supported by the island-derived fragment ions marked in green circle with a n*CH₂ (14*n Da) decrease of m/z from precursor ions to the left. A series of fragment ions with a difference of 14 Da are generated by the β- or γ- bonds cleavage of alkyl side chains during CID experiment. As depict in Fig. 1, the value of n is less than 7 for most of the fragment ions meaning that the range of total loss of alkyl side chain carbon atoms is nearly 0–7. The intensity of the island-type structure peak group is in the form of quasi normal distribution. This finding is consistent with the previous research results that the coal-derived asphaltenes have few and short alkyl side chains with average alkyl substituent number and substituted alkyl chain length of 4.4 and 2.9, respectively (Zhu et al., 2021). Schuler et al. (2015) innovatively study petroleum and coal-derived asphaltene structures by combining atomic resolution imaging using AFM method and molecular orbital imaging using STM method. The molecular skeleton images of asphaltenes are firstly obtained providing direct evidence for many conventional cognitions. For example, the island- and archipelago-type coal-derive asphaltenes with cata- and peri-condensed fragments and peripheral alkane chains are directly detected in that study. The island-type is regarded as the dominant asphaltene molecular architecture, while the archipelago-type molecules are present with several distinct PAHs connected by a single bond. According to the CID results, the proportion of island- to archipelago-type molecules in this model is determined for about 80%: 20%. Meanwhile, the types of inter-core linkages are divided into three classes including a single bond, –CH₂–, or –O– groups connecting two PAHs (Schuler et al., 2015). The maximum number of cores is set to two for the coal-derived asphaltene molecule in this study.

The bond cleavage paths in CID process are mainly divided into two directions of homolytic and heterolytic bond cleavages [27]. The relative abundances of radical fragments (M^+•) and protonated fragments ([M+H]⁺) for different class compounds in asphaltene before and after CID process by positive-ion ESI FT-ICR mass spectra is shown in Fig. S1. The relative abundance of N₁ species increase significantly after CID, which is mainly generated from the poly-N heteroatomic compounds. The N₁ and O_x species also has an increment during the CID process. Besides, there are some hydrocarbons (HC) generated through homolysis modes of precursors ions. It can be found that there are still some multiple heteroatom species, such as N₁O₂, N₁O₃ and N₁O₃, survived after CID.

PDFs obtained from FT-ICR MS. In this model, the initial PDF parameters are estimated from the results of (+) ESI, (−) ESI and (+) APPI FT-ICR MS through relating the relative abundance and DBE of aromatic hydrocarbon, basic nitrogen, non-basic nitrogen, oxygen and sulfur atoms. The iso-abundance maps of DBE versus CNs of asphaltenes in (+) ESI, (−) ESI, (+) APPI modes of FT-ICR MS are shown in Fig. 2. As Fig. 2 shows, the average DBE and CNs increase from (−) ESI to (+) ESI to (+) APPI data, respectively. It indicates that different ion modes have different ability to identify asphaltene molecules and the weakly polar compounds detected by positive APPI and basic compounds detected by positive ESI perform higher average CNs and aromaticity for the coal tar asphaltene (Pan et al., 2012). There are some asphaltene molecules with low carbon numbers and DBE values detected by (+) ESI, (−) ESI and (+) APPI FT-ICR MS although the purified process has been operated in the sample preparation, illustrating the asphaltenes obtained in experiment method are not only the continuum of small molecules to macromolecules, which generally mixes in a small amount of aromatic and resin molecules, but also island and archipelagic molecules as a function of increasing molecular weight and polarity (Chacón-Patiño et al., 2018a, 2018b). These “atypical” structures are classified as asphaltenes according to their functionality and high polarity of ultra-high heteroatom content rather than high aromaticity (Ballard et al., 2020). It should be pointed out that these “atypical” compounds are mainly exist in species with multiple nitrogen atoms. Besides, the compounds with high DBE are mainly the NO species with multiple heteroatoms.

The iso-abundance maps of CNs versus DBE of asphaltenes based on (+) ESI, (−) ESI, (+) APPI FT-ICR MS. Red and green lines represent the average DBE and CNs, respectively.

The PDFs and cumulative distribution functions (CDFs) are shown in Fig. 3a-e. Specifically, the PDF of aromatic rings is obtained from hydrocarbons class distributions from (+) APPI FT-ICR MS by estimating the maximum number of aromatic rings based on the relative abundance and its DBE (Fig. 3a). In the case of nitrogen-containing group, the PDFs are obtained by normalizing the relative abundance of N₁, N₂ and N₃ species detected form (+) ESI (basic nitrogen) and (−) ESI (non-basic nitrogen) FT-ICR MS. The PDF for number of basic nitrogen atoms consisting of pyridinic ring and amino group is obtained by (+) ESI FT-ICR MS (Fig. 3b), while the PDF for the number of non-basic nitrogen atoms designated as pyrrole ring is obtained by (−) ESI FT-ICR MS (Fig. 3c). With the same manner of nitrogen PDF, the PDF for number of oxygen atoms (O₀-O₁₀) is obtained by (−) ESI FT-ICR MS (Fig. 3d) and PDF for number of sulfur atoms (S₀-S₃) is obtained by (+) APPI FT-ICR MS (Fig. 3e).

Probability density functions (PDFs) and cumulative distribution functions (CDFs) for different structure attribute based on (+) ESI, (−) ESI, (+) APPI FT-ICR MS.

Roles of the conventional analysis method in this model. Although FT-ICR MS can provide very meaningful qualitative analysis data for the PDFs of vital structure attributes, it is necessary to incorporate the information obtained from conventional analysis methods, which have been published in our previous paper (Zhu et al., 2021). In this model, the types of nitrogen include the basic (46.95 wt%) and non-basic (53.05 wt%) nitrogen based on the XPS results. The basic nitrogen groups are quantitatively divided into pyridinic ring (36.13 wt%) and amino group (10.82 wt%). The types of oxygen groups are attributed to hydroxyl (50.16 wt%), furan ring (42.95 wt%), carbonyl (4.02 wt%) and carboxyl (2.87 wt%) groups from the qualitative and quantitative analysis of FT-IR and ¹³C NMR methods. In view of the detecting limitation for the effective information of aliphatic structures, the gamma parameters of the PDFs for naphthenic ring numbers, alkyl side chain numbers and length of per alkyl side chain are initialized randomly referring to the average information got by solid state ¹³C NMR method and the rough range of the carbon atoms loss of alkyl side chain obtained through the CID technology discussed above.

3.2

3.2 Core attributes library

Multi-aromatic core with a few naphthenic rings is the key structure attribute for the asphaltene molecules. It is very important to build the initial aromatic core attribute library accurately according to the analysis data. At present, the accurate identification of the aromatic core of macromolecules is still facing a great challenge. The CID technology provides a feasible way to obtain useful core structure information by the analysis of molecular ionized fragments, although it could not get a complete quantitative result of aromatic core structure attribute. Fig. 4 shows the FT-ICR MS and the carbon number distribution of coal-derived asphaltene before and after CID. The aliphatic side chain and bridge between aromatic cores will be broken under optimized ionization voltage, while the aromatic core structure is reserved due to its high bond energy (Zhang et al., 2014). There is remarkable reduction for the mass and carbon number distributions after CID led by the dealkylation reactions.

FT-ICR MS and the carbon number distribution of coal-derived asphaltene before and after CID.

The iso-abundance maps of CNs versus DBE of N₁, N₂ and N₁O₁ species based on FT-ICR MS is shown in Fig. 5. The average DBE values of N₁, N₂ and N₁O₁ species before CID are higher than that after CID for coal-derived asphaltene in this study, which is different from the petroleum asphaltene (Chacón-Patiño et al., 2018a, 2018b). This maybe because that the fragmentation of island compounds will generate new unsaturated bonds in view of the CID heterolysis mechanism leading to the increase of DBE. Also, the dehydrogenation of PAH cores during CID will also make an increase in DBE of 1 by loss of H₂ (Chacón-Patiño et al., 2017). The significant increase in DBE proves the absolute dominance of island-type structures in coal tar asphaltene, while the archipelago type structures occupy a certain important proportion in petroleum asphaltene (Chacón-Patiño et al., 2018a, 2018b). The average CNs of N₁, N₂ and N₁O₁ species all decreased after CID mainly caused by the cleavage of aliphatic side chains. The fragment molecules obtained by CID tend to shift to left approaching to the planar limit, which defines the composition space of asphaltene corresponding to the DBE as an indicator of aromatic rings and carbon numbers (Hsu, 2012, Marshall, 2011). It is used to interpret fossil fuel molecule structure through defining lines connecting maximum DBE values at given CNs (Jameel et al., 2019, Giraldo-Davila et al., 2016). The CNs and DBE value per molecule are restricted by following relationship: 0.9·CNs > DBE. The oil mixture with rich aromatic structure produces high line slope. Under the given CNs, the aromaticity of PAH cannot exceed a maximum value which increases linearly with the increase of CNs (Cho et al., 2011). In this model, the aromatic cores are inferred and predefined manually to fit to the PAH boundary as far as possible according to the qualitative analysis on the slope of boundary of fragment ions from CID FT-ICR MS results (Zhang et al., 2014).

The iso-abundance maps of DBE versus CNs of N1, N2 and N1O1 species based on FT-ICR MS. Red, green, orange lines represent the average DBE, carbon numbers, planar limit, respectively.

Below are the considerations that are taken into account to set the aromatic core attributes:

The maximum fused aromatic ring is up to 10, as obtained by APPI FT-ICR MS data.
The maximum naphthenic ring is up to 3, which are initialized randomly referring to the average information got by solid state ¹³C NMR method. Besides, the six-membered ring (cyclohexane) is the only form for naphthenic ring (Zhang et al., 2014).
There are up to 4 oxygen heterocycles of furan type per aromatic core determined by the qualitative and quantitative analysis of FT-IR and ¹³C NMR data. Phenols are not regard as the oxygen heterocycles in this model and hydroxyl groups are attributed to the branched chains of aromatic core.
Each aromatic core allows to have at most 3 nitrogen heterocycles containing the pyridine and pyrrole rings based on the XPS and FT-ICR MS results.
Each sulfur-containing core has 0–3 thiophene rings according to the FT-ICR MS results.

3.3

3.3 Computer simulation

The molecular representation model is embedded in an optimization loop to determine the PDF parameters that sample out the asphaltene molecular structure attributes based on Monte Carlo algorithm to make the properties of simulated molecular library accordant with the experimental data. The algorithm of asphaltene molecular reconstruction is shown in Fig. 6.

Schematic of asphaltene molecule reconstruction.

The objective function (Eq (5)) is solved by the simulated annealing algorithm using a least-squares comparison of the experimental and simulated properties of coal-derived asphaltene. For the Monte Carlo algorithm, the PDF parameters are adjusted on basis of initial values to minimize the deviations between experimental and simulated properties. The fragments sampling process by PDF functions is shown in Fig. 7.

Fragments sampling process by PDF functions.

Fig. 8 exhibits an example of assembly process for the mono-aromatic core molecule. Initially, the fragments of core structure and lateral branch including heteroatom-based lateral branch and alkyl side chain are sampled out by the PDFs. The complete core structure is then selected from the core library based on the above fragment information. After that, all the lateral branch fragments are assembled to the core structures one by one through the connection sites, which is generated based on the machine learning language. The final version molecule is dealt with the group contribution method to obtain the predictive physical and chemical properties.

The properties of density (D), element composition of C, H, S, N, O, distillation fraction yields of each boiling point (BP) range and C, O, N atom types, are given preference to be as the optimization objectives.

(5)

O b j e c t i v e f u n c t i o n (F) = m i n i m i z e \{\begin{matrix} {(\frac{D_{\exp -} D_{cal}}{D_{\exp} \cdot φ_{1}})}^{2} \\ + \sum_{i = 1}^{Element #} {(\frac{W_{Element i, e x p -} W_{Element i, c a l}}{W_{Element i, e x p} \cdot φ_{3}})}^{2} \\ + \sum_{i = 1}^{C - t y p e #} {(\frac{{Mol}_{C - t y p e i, e x p -} {Mol}_{C - t y p e i, c a l}}{{M o l}_{C - t y p e i, e x p} \cdot φ_{4}})}^{2} \\ + \sum_{i = 1}^{N - t y p e #} {(\frac{{Mol}_{N - t y p e i, e x p -} {Mol}_{N - t y p e i, c a l}}{{Mol}_{N - t y p e i, e x p} \cdot φ_{5}})}^{2} \\ + \sum_{i = 1}^{O - t y p e #} {(\frac{{Mol}_{O - t y p e i, e x p -} {Mol}_{O - t y p e i, c a l}}{{Mol}_{O - t y p e i, e x p} \cdot φ_{6}})}^{2} \\ + \sum_{i = 1}^{Fraction #} {(\frac{W_{Fraction i, e x p -} W_{Fraction i, c a l}}{W_{Fraction i, e x p} \cdot φ_{7}})}^{2} \end{matrix}\}

The compositions and structures of coal-derived asphaltene molecules are much diverse and complex resulting in big difficulties in the detecting of some vital properties using experiment method. It’s necessary to use the simulating model to estimate the chemical and physical properties of a molecule starting from the structure (Feng et al., 2019). In this model, the properties of density and boiling point of the molecules are calculated by the group contribution method proposed by Hukkerikar et al. (2012) and Gani et al. (2005) that sums the number of the occurrences of each contribution group by an empirical function from the first-, second- or third-order predefined group library. Feng et al. (2019) and Cai et al. (2018) confirm that the group contribution method is reliable to predict the properties of heavy oil fractions which contain the molecules with large-size and high-carbon number. Besides, the structural units considered in the optimization function are C-types (CH₃, CH₂, CH in aliphatic chain, protonated aromatic C, aromatic bridgehead C, aromatic branched C, C bonded to O), N-types (pyridine, pyrrole, amino), O-types (hydroxyl, carbonyl, carboxyl, furan). It is notable that the complete distillation distribution is difficult to obtained depending on HTSD for coal-derived asphaltene caused by the difficult evaporation of high boiling point molecule and strong aggregation function of polar molecules leading to weak solubility. Therefore, only the distillation fraction yields of boiling point range for 300–400, 400–500, 500–600 °C are selected to be the fraction types as the optimization objectives, which can ensure the accuracy of the experimental data to the most extent.

In consideration of the errors generated from the model and analytical methods, there should be a balance among the contributions of the function terms. The weight factor (φ) is applied by Zhang et al. (2014) to eliminate the errors between the simulated results and experimental analysis data. Generally, the value of weight factor is enlarged to lower its contributions if the uncertainty is great. In this model, weight factor is composed of three levels of high (8), medium (2) and low (0.5) to balance the contribution values referring to the method of Zhang et al. (2014). Among these properties, the element composition and C-types, calculated directly from experimental analysis with high accuracy, are attributed to the weight factor of low level. The weight factor of the density is assigned to the high level for the high uncertainty in either the experimental measurement or the prediction by group contribution method. The weight factors of left properties for C-, N-, O- types and distillation fraction yields are set to the medium level.

4 Results and discussions

The molecular representation model is obtained with about 100,000 asphaltene molecules in total being reconstructed after sampling and optimization according to above approach. The bulk properties and molecular compositions of the molecular library are carried out confirmatory and extrapolative analysis with experimental data. Combined with our characterization results (Zhu et al., 2021) and similarity algorithm, a small set of the representative average structure molecules of coal-derived asphaltene are determined from the molecular library.

4.1

4.1 Bulk properties and molecular compositions

The experimental bulk properties and the predicted values of coal-derived asphaltenes are listed in Table 2. It can be observed that most of the prediction values agree reasonably well with the experimental data, such as the element compositions, H/C, O/C and carbon types and HTSD results. It is noted that only a partial distillation profile covering up to 50 wt% of the data is available limited to the evaporation difficulty and aggregation function of asphaltene molecules.

Table 2 Experimental and simulated bulk properties of coal-derived asphaltene.

Property	Unit	Experimental (Zhu et al., 2021)	Simulated
C	wt%	80.61	81.05
H	wt%	5.64	5.81
S	wt%	0.29	0.32
N	wt%	1.92	1.84
O	wt%	11.55	10.98
H/C	at/at	0.84	0.86
O/C	at/at	0.11	0.10
Molecular weight		630	578

¹³C NMR
Aliphatic CH₃	mol%	14.15	10.04
Aliphatic CH₂	mol%	6.43	8.36
Aliphatic CH	mol%	4.23	4.87
Aromatic protonated C	mol%	22.73	21.29
Aromatic bridgehead C	mol%	19.35	24.53
Aromatic branched C	mol%	10.36	9.30
C linked to O^a	mol%	22.75	21.61

HTSD
0.5 wt% BP	°C	315	326
10 wt% BP	°C	346	371
30 wt% BP	°C	459	473
50 wt% BP	°C	582	597

a C linked to O group includes the carbonyl carbo, carboxyl carbon, aliphatic carbon bonded to oxygen and aromatic carbon bonded to oxygen (Zhu et al., 2021).

The predicted average molecular weight is slightly smaller than the experimental value obtained from the GPC method. In consideration of the influence of asphaltene self-aggregation on the vapor pressure osmometry (VPO) and gel permeation chromatography (GPC) methods, which may lead to high average molecular weight, such error is acceptable in the simulation model (McKenna et al., 2013). The average molecular weight is calculated based on the monomeric (non-aggregated) entities from the molecule library in this work (Alvarez-Majmutov et al., 2019). Fig. 9 exhibits the molecular weight data of predictive asphaltenes which presents the normal distribution.

The molecular weight distribution of predictive asphaltenes.

For C types, the predictive errors of aliphatic CH₂, aliphatic CH, aromatic protonated C, aromatic branched C, C linked to O are small enough, while the errors of aliphatic CH₃ and aromatic bridgehead C (cata- and peri- condensed carbons) are a little bit bigger. But in the whole, the predictive C-type distribution is close to the ¹³C NMR results. The distributions for carbon types of aliphatic CH, CH₂, CH₃, C linked to O, aromatic protonated C, aromatic bridgehead C, aromatic branched C in asphaltene obtained from the molecular representation are displayed in Fig. 10.

Distributions for carbon types of aliphatic CH3, CH2, CH, C bonded to O, aromatic protonated C, aromatic bridgehead C, aromatic branched C in asphaltene obtained from the molecular representation.

As shown in Fig. 10, the molar fractions of the number of aliphatic CH₃, CH₂, CH in the asphaltene molecules decrease from 0 to 7, 0 to 9⁺ and 0 to 4, respectively, indicating that the less and shorter side chain distribution are the vital features for coal-derived asphaltenes, which is coincident with the research conclusion of Wu et al. (2014) and our previous findings (Zhu et al., 2021). This also proves that the molecular representation model is liable and accurate. The molar fractions of the number of aromatic protonated C increase nearly from 0 to 9+ indicating there are many C in the aromatic ring are unsubstituted. The molar fractions of the number of aromatic bridgehead C, aromatic branched C, C linked to O are close to the normal distributions peaking at the number of 10, 2, 4, respectively. It has to be said that the number of aromatic bridgehead C in each molecule must be even (0, 2, 4…0.20) following its occurrence characteristics in the aromatic lamellar structure.

The distributions for aromatic, aliphatic, total carbon numbers in asphaltenes obtained from the molecular representation are described in Fig. 11. The range of the aromatic, aliphatic and total carbon numbers are nearly 0–25, 10–30 and 15–55, respectively. The shape of the total carbon number distribution is similar with that of the molecular weight depicted in Fig. 11.

Distributions for aromatic, aliphatic, total carbon numbers in asphaltene obtained from the molecular representation.

In order to obtain the detail composition information, the iso-abundance map of DBE versus CNs of predictive N_1-3 and O_1-3 species with the predictive structure images of the representative asphaltene molecules are shown in Fig. 12. The iso-abundance maps of CNs versus DBE of predicted N₁ and O₁ species are similar to the experimental data of FT-ICR MS in Fig. 2. There's a little difference that the predicted carbon distribution is slightly wider and the species with high carbon numbers are more plentiful. It is important to explain that although the FT-ICR MS method can provide the qualitative comparison data, the carbon and DBE distribution are not absolutely accurate in view of the asphaltene aggregation effect emerging at asphaltene concentration low to 50 µg·mL⁻¹, which reduces the ion signal during the ionization process resulting in the difficulties of asphaltene molecule detecting, especially for the high carbon number species (Chacón-Patiño et al., 2017). Therefore, the carbon numbers in our model are not limited strictly according to the FT-ICR MS results. Instead, the optimization objectives in Eq. (5) play crucial roles during composition determination of asphaltene molecular library.

The iso-abundance map of CNs versus DBE of predictive N1-3 and O1-3 species with the predictive structure images of the representative asphaltene molecules.

Fig. 13 exhibits the iso-abundance map of CNs versus DBE of predictive N₁ (Top: pyridine-type, pyrrole-type, amino-type) and C_xH_yO₁ (Bottom: hydroxyl-type, furan-type, carbonyl-type) species with the detail distribution of carbon numbers, respectively, for the representative asphaltene molecules. It is found that the ranges of carbon number and DBE for pyridine-, pyrrole- and amino-type N₁ species are close to each other and so is to hydroxyl-, furan- and carbonyl-type O₁ species. Furthermore, the iso-abundance map of CNs versus DBE of predictive HC, N₁O₁ and N₁O₁S₁ species with predictive structure images of the representative asphaltene molecules are depicted in Fig. 14.

The iso-abundance map of CNs versus DBE of predictive N1 (Top: pyridine-type, pyrrole-type, amino-type) and O1 (Bottom: hydroxyl-type, furan-type, carbonyl-type) species with the detail distribution of carbon numbers, respectively, for the representative asphaltene molecules.

The iso-abundance map of CNs versus DBE of predictive HC, N1O1 and N1O1S1 species with predictive structure images of the representative asphaltene molecules.

4.2

4.2 Determination of representative average structure molecules

As analyzed above, this model is capable of giving very detailed structural information of coal-derived asphaltene molecules, which could be applied in molecular-based directional separation or kinetic computational models. In order to explore the inducement of asphaltene aggregation and the effect of asphaltene on oil-water interface, the average structure molecules are always necessary to be determined in the molecule dynamics simulation work (Sedghi et al., 2013, Liu et al., 2015, Wang and Ferguson, 2016 ). In consideration of the deviation errors of asphaltene molecule weight by GPC method discussed above, the experimental data from ¹³C NMR spectrum is dealt again based on the molecule weight got from representation model. Table 3 shows the structural parameters of coal-derived asphaltene calculated by GPC and molecular representation methods based on ¹³C NMR spectrum, respectively (Zhu et al., 2021). The main structural parameters are reasonably corrected by a small margin. Besides, the average share of N-, O-, S-containing heteroatom groups per asphaltene molecule got by quantitative and qualitative analysis from FT-IR, XPS and ¹³C NMR spectrum are summarized in Table S1.

Table 3 Structural parameters of coal-derived asphaltene calculated by GPC and molecular representation methods based on ¹³C NMR spectrum.

Description	Parameter	GPC (Zhu et al., 2021)	Molecular representation
Average molecular formula	–	C_42.3H_35.5O_4.5N_0.9S_0.1	C_38.8H_32.6O_4.2N_0.8S_0.1
Molecular weight	M	630	578
Aromatic ring number	R_A	5.9	5.4
Naphthenic ring number	R_N	2.3	2.1
Total ring number	R_T	8.2	7.5
Alkyl substituent number	n	4.4	4.0
Substituted alkyl chain length	L	2.9	2.9

The molecular similarity function (F_MS) is defined reasonably to selected out the representative asphaltene molecules rely on the average structural and composition data described in Table 3 and Table S1. The detailed selected process is depicted in Fig. 15. High F_MS means the selected molecule is close to the average molecular structure and composition. F_MS includes three parts of aromatic and naphthenic rings (F_AR+NR), heteroatom groups (F_HG), aliphtic side chains (F_ASC). F_HG are composed of N-containing groups (F_HG-N), O-containing groups (F_HG-O), S-containing groups (F_HG-S). F_ASC are composed of the average alkyl substituent number (F_ASC-n) and average substituted aliphtic chain length (F_ASC-l). A series weighting factors of 0.5 (first level, W₁), 0.2 (second level, W₂), 0.1 (third level, W₃) are set to obtain the reasonable F_MS. NNG, NOG, NSG express the number of N-containing, O-containing and S-containing groups, respectively. The big weighting factor is chosen when the values of the parameter from the modelling molecules are close to the average values listed in Table 3 and Table S1. The molecules with the parameters exceeding the boundary set in Fig. 15 are ignored in this method.

The selected process of asphaltene molecules based on the molecular similarity function.

Fig. 16 provides a calculating example of molecular similarity with the average experimental data on the basis of the group-finding algorithm (Cho and Hansen, 2019). Finally, 60 asphaltene molecules are picked out from the calculated molecule library with molecular weight range of 500–650 g·mol⁻¹. The represented coal-derived asphaltene molecules are shown in Fig. S2.

The calculating process of molecular similarity with the average experimental data for the coal-derive asphaltene.

In order to verify the rationality of the molecular similarity function based on the experimental data above, the similarity calculation through the molecular chemical fingerprint is applied to the preferred asphaltene molecules in Fig. S2, which is developed by the similarity property principle that similar molecules have similar biological and chemical activities (Gortari et al., 2017). The similarity map, based on the Morgen fingerprint algorithm, is used to visualize the similarity contribution of atoms in two molecules (Riniker and Landrum, 2013). The greener the color, the higher the similarity on the map. The asphaltene molecules of A1, A6 are regard as the reference molecules, while A2-A5, A7-A10 are selected to be the target molecules, respectively. The similarity maps between the reference and target molecules are shown in Fig. 17. As Fig. 17 expresses, the target molecules perform high similarities with the reference molecules indicating the molecular similarity function based on the experimental data proposed in this paper are feasible to determine the average structure asphaltene molecules from the simulation library.

The similarity maps between the target and reference molecules.

5 Conclusions

In this study, a novel compositional model of coal-derived asphaltene is developed based on the advanced high resolution mass spectrum and traditional representation methods. All the attribute groups including cores, alkyl side chains, heteroatom functional group and inter-core linkages are assembled depending on the PDFs. The data of FT-ICR MS coupled with CID are fully exploited to obtain the distribution information of basic nitrogen, non-basic nitrogen, oxygen, sulfur atoms, aromatic core and the qualitative data of island- and archipelago-type structures for coal-derived asphaltene. The experimental core structure attributes and heteroatom functional group information are gained form ¹³C NMR, FT-IR, XPS methods. The modelling asphaltene molecules are systematic assembled through stochastic algorithms and then optimized using simulated annealing algorithm. The simulation asphaltene library composed of about 100,000 molecules covers the common species in practical asphaltene. The bulk properties for simulating asphaltenes are coincident with the experimental results. Finally, sixty representative average structure asphaltene molecules are obtained using the developed molecular similarity function. In the follow-up study, we will research on the aggregation behavior of asphaltene macromolecules in heavy oil, oil-water interface and other environments using the representative average structure asphaltene molecules got in this study.

Acknowledgement

We gratefully acknowledge the financial support of the National Natural Science Foundation of China (21978237), Scientific and Technological Research Projects of Yulin City (CXY-2020-017-01), Natural Science Basic Research Program of Shaanxi (2021JLM-19), Natural Science Basic Research Program of Shaanxi (2019JLP-03).

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

References

Alvarez-Majmutov A., Gieleciak R., Chen J., . Modeling the molecular composition of vacuum residue from oil sand bitumen. Fuel. 2019;241:744-752.
[Google Scholar]
Ballard D.A., Chacón M., Qiao P., Roberts K.J., Harbottle D., . Molecular characterization of strongly and weakly interfacially active asphaltenes by high-resolution mass spectrometry. Energy Fuels. 2020;34:13966-13976.
[Google Scholar]
Boek E.S., Yakovlev D.S., Headen T.F., . Quantitative molecular representation of asphaltenes and molecular dynamics simulation of their aggregation. Energy Fuels. 2009;23:1209-1219.
[Google Scholar]
Cai G.Q., Liu Z.F., Zhang L.Z., . Quantitative structure property relationship model for hydrocarbon liquid viscosity prediction. Energy Fuels. 2018;32:3290-3298.
[Google Scholar]
Chacón-Patiño M.L., Rowland S.M., Rodgers R.P., . Advances in asphaltene petroleomics. Part 1: Asphaltenes are composed of abundant island and archipelago structural motifs. Energy Fuels. 2017;31:13509-13518.
[Google Scholar]
Chacón-Patiño M.L., Rowland S.M., Rodgers R.P., . Advances in asphaltene petroleomics. Part 2: Selective separation method that reveals fractions enriched in island and archipelago structural motifs by mass spectrometry. Energy Fuels. 2018;32:314-328.
[Google Scholar]
Chacón-Patiño M.L., Rowland S.M., Rodgers R.P., . Advances in asphaltene petroleomics. Part 3. Dominance of island or archipelago structural motif is sample dependent. Energy Fuels. 2018;32:9106-9120.
[Google Scholar]
Chacón-Patiño M.L., Smith D.F., Hendrickson C.L., Marshall A.G., Rodgers R.P., . Advances in asphaltene petroleomics. part 4. compositional trends of solubility subfractions reveal that polyfunctional oxygen-containing compounds drive asphaltene chemistry. Energy Fuels. 2020;34:3013-3030.
[Google Scholar]
Chacón-Patiño M.L., Moulian R., Barrère-Mangote C., Putman J.C., Weisbrod C.R., Blakney G.T., Bouyssiere B., Rodgers R.P., Giusti P., . Compositional trends for total vanadium content and vanadyl porphyrins in gel permeation chromatography fractions reveal correlations between asphaltene aggregation and ion production efficiency in atmospheric pressure photoionization. Energy Fuels. 2020;34:16158-16172.
[Google Scholar]
Cho K., Hansen P.E., . Structure identification of novel compounds using simple IR, ¹H, and ¹³C-NMR spectroscopy and computational tools. Bull. Korean Chem. Soc.. 2019;41:78-83.
[Google Scholar]
Cho Y., Kim Y.H., Kim S., . Planar limit-assisted structural interpretation of saturates / aromatics / resins / asphaltenes fractionated crude oil compounds observed by Fourier transform ion cyclotron resonance mass spectrometry. Anal. Chem.. 2011;83:6068-6073.
[Google Scholar]
De Oliveira L.P., Vazquez A.T., Verstraete J.J., Kolb M., . Molecular reconstruction of petroleum fractions: application to vacuum residues from different origins. Energy Fuels. 2004;27:3622-3641.
[Google Scholar]
Feng S., Cui C., Li K., Zhang L., Xu C., . Molecular composition modelling of petroleum fractions based on a hybrid structural unit and bond-electron matrix (SU-BEM) framework. Chem. Eng. Sci.. 2019;201:145-156.
[Google Scholar]
Feng X., Li D., Chen J., Niu M., Liu X., Chan L.L., Li W., . Kinetic parameter estimation and simulation of trickle-bed reactor for hydrodesulfurization of whole fraction low-temperature coal tar. Fuel. 2018;230:113-125.
[Google Scholar]
Frigerio F., Molinari D., . A multiscale approach to the simulation of asphaltenes. Comput. Theor. Chem.. 2011;975:76-82.
[Google Scholar]
Gani R., Harper P.M., Hostrup M., . Automatic creation of missing groups through connectivity index for pure-component property prediction. Ind. Eng. Chem. Res.. 2005;44:7262-7269.
[Google Scholar]
Gaspar A., Zellermann E., Lababidi S., Reece J., Schrader W., . Characterization of saturates, aromatics, resins, and asphaltenes heavy crude oil fractions by atmospheric pressure laser ionization Fourier transform ion cyclotron resonance mass spectrometry. Energy Fuels. 2012;26:3481-3487.
[Google Scholar]
Giraldo-Davila D., Chacon-Patino M.L., Orrego-Ruiz J.A., Blanco-Tirado C., Combariza M.Y., . Improving compositional space accessibility in (+) APPI FT-ICR mass spectrometric analysis of crude oils by extrography and column chromatography fractionation. Fuel. 2016;185:45-58.
[Google Scholar]
Glattke T.J., Chacón-Patiño M.L., Marshall A.G., Rodgers R.P., . Molecular characterization of photochemically produced asphaltenes via photooxidation of deasphalted crude oils. Energy Fuels. 2020;34:14419-14428.
[Google Scholar]
Gortari F.D., García-Jacas C., Martinez-Mayorga K., Medina-Franco J.L., . Database fingerprint (DFP): an approach to represent molecular databases. J. Cheminf.. 2017;9:9.
[Google Scholar]
Headen T.F., Boek E.S., Skipper N.T., . Evidence for asphaltene nanoaggregation in toluene and heptane from molecular dynamics simulations. Energy Fuels. 2009;23:1220-1229.
[Google Scholar]
Hsu C.S., . Compositional space boundaries for organic compounds. Anal. Chem.. 2012;84:3410-3416.
[Google Scholar]
Hudebine D., Verstraete J.J., . Molecular reconstruction of LCO gasoils from overall petroleum analyses. Chem. Eng. Sci.. 2004;59:4755-4763.
[Google Scholar]
Hukkerikar A.S., Sarup B., Kate A.T., Abildskov J., Sin G., Gani R., . Group-contribution+ (GC+) based estimation of properties of pure components: improved property estimation and uncertainty analysis. Fluid Phase Equilib.. 2012;321:25-43.
[Google Scholar]
Jaffe S.B., Freund H., Olmstead W.N., . Extension of structure-oriented lumping to vacuum residua. Ind. Eng. Chem. Res.. 2005;44:9840-9852.
[Google Scholar]
Jameel A.G.A., Khateeb A., Elbaz A.M., Emwas A.H., Zhang W., Roberts W.L., Sarathy S.M., . Characterization of deasphalted heavy fuel oil using APPI (+) FT-ICR mass spectrometry and NMR spectroscopy. Fuel. 2019;253:950-963.
[Google Scholar]
Juyal P., Mckenna A.M., Fan T., . Joint industrial case study for asphaltene deposition. Energy Fuels. 2013;27:1899-1908.
[Google Scholar]
Liu J., Zhao Y., Ren S., . Molecular dynamics simulation of self-aggregation of asphaltenes at an oil/water interface: formation and destruction of the asphaltene protective film. Energy Fuels. 2015;29:1233-1242.
[Google Scholar]
Marshall A.G., . Compositional boundaries for fossil hydrocarbons. Energy Fuels. 2011;25:2174-2178.
[Google Scholar]
Mckenna A.M., Blakney G.T., Xian F., Glaser P.B., Rodgers R.P., Marshall A.G., . Heavy petroleum composition. 2. Progression of the boduszynski model to the limit of distillation by ultrahigh-resolution FT-ICR mass spectrometry. Energy Fuels. 2010;24:2939-2946.
[Google Scholar]
McKenna A.M., Donald L.J., Fitzsimmons J.E., Juyal P., Spicer V., Standing K.G., . Heavy petroleum composition. 3. asphaltene aggregation. Energy Fuels. 2013;27:1246-1256.
[Google Scholar]
McKenna A.M., Chacón-Patiño M.L., Weisbrod C.R., Blakney G.T., Rodgers R.P., . Molecular-level characterization of asphaltenes isolated from distillation cuts. Energy Fuels. 2019;33:2018-2029.
[Google Scholar]
Neurock M., Libanati C., Nigam A., Klein M.T., . Monte carlo simulation of complex reaction systems: molecular structure and reactivity in modelling heavy oils. Chem. Eng. Sci.. 1990;45:2083-2088.
[Google Scholar]
Neurock M., Nigam A., Trauth D., Klein M.T., . Molecular representation of complex hydrocarbon feedstocks through efficient characterization and stochastic algorithms. Chem. Eng. Sci.. 1994;49:4153-4177.
[Google Scholar]
Niles S.F., Chacón-Patiño M.L., Smith D.F., Rodgers R.P., Marshall A.G., . Comprehensive compositional and structural comparison of coal and petroleum asphaltenes based on extrography fractionation coupled with Fourier transform ion cyclotron resonance MS and MS/MS analysis. Energy Fuels. 2020;34:1492-1505.
[Google Scholar]
Pan N., Cui D., Li R., Shi Q., Chung K.H., Long H., . Characterization of middle-temperature gasification coal tar. part 1: bulk properties and molecular compositions of distillates and basic fractions. Energy Fuels. 2012;26:5719-5728.
[Google Scholar]
Pei L., Li D., Liu X., Cui W., Shao R., Xue F., Li W., . Investigation on asphaltenes structures during low temperature coal tar hydrotreatment under various reaction temperatures. Energy Fuels. 2017;31:4705-4713.
[Google Scholar]
Pereira T., Vanini G., Tose L.V., Cardoso F., Fleming F.P., . FT-ICR MS analysis of asphaltenes: asphaltenes go in, fullerenes come out. Fuel. 2014;131:49-58.
[Google Scholar]
Podgorski D.C., Corilo Y.E., Nyadong L., Lobodin V.V., Bythell B.J., Robbins W.K., . Heavy petroleum composition. 5. Compositional and structural continuum of petroleum revealed. Energy Fuels. 2013;27:1268-1276.
[Google Scholar]
Pyl S.P., Hou Z., Van Geem K.M., Reyniers M.F., Marin G.B., Klein M.T., . Modeling the composition of crude oil fractions using constrained homologous series. Ind. Eng. Chem. Res.. 2011;50:10850-10858.
[Google Scholar]
Quann R.J., Jaffe S.B., . Structure-oriented lumping: describing the chemistry of complex hydrocarbon mixtures. Ind. Eng. Chem. Res.. 1992;31:2483-2497.
[Google Scholar]
Quann R.J., Jaffe S.B., . Building useful models of complex reaction systems in petroleum refining. Chem. Eng. Sci.. 1996;51:1615-1631.
[Google Scholar]
Riniker S., Landrum G.A., . Similarity maps - a visualization strategy for molecular fingerprints and machine-learning methods. J. Cheminf.. 2013;5:43.
[Google Scholar]
Schuler B., Meyer G., Peña D., Mullins O.C., Gross L., . Unraveling the molecular structures of asphaltenes by atomic force microscopy. J. Am. Chem. Soc.. 2015;137:9870-9876.
[Google Scholar]
Schuler B., Fatayer S., Meyer G., Rogel E., Moir M., Zhang Y.L., . Heavy oil based mixtures of different origins and treatments studied by atomic force microscopy. Energy Fuels. 2017;31:6856-6861.
[Google Scholar]
Schuler B., Zhang Y., Liu F., Pomerantz A.E., Andrews A.B., Gross L., . Overview of asphaltene nanostructures and thermodynamic applications. Energy Fuels. 2020;34:15082-15105.
[Google Scholar]
Scott D.E., Schulze M., Stryker J.M., Tykwinski R.R., . Deciphering structure and aggregation in asphaltenes: hypothesis-driven design and development of synthetic model compounds. Chem. Soc. Rev.. 2021;50:9202-9239.
[Google Scholar]
Sedghi M., Goual L., Welch W., Kubelka J., . Effect of asphaltene sructure on association and aggregation using molecular dynamics. J. Phys. Chem. B. 2013;117:5765-5776.
[Google Scholar]
Shao R., Shen Z., Li D., Sun Z., Pei L., Liu X., . Investigation on composition and structure of asphaltenes during low-temperature coal tar hydrotreatment under various reaction pressures. J. Anal. Appl. Pyrol.. 2018;136:44-52.
[Google Scholar]
Sheremata J.M., Gray M.R., Dettman H.D., Mccaffrey W.C., . Quantitative molecular representation and sequential optimization of athabasca asphaltenes. Energy Fuels. 2004;18:1377-1384.
[Google Scholar]
Shi Q., Hou D., Chung K.H., Xu C., Zhao S., Zhang Y., . Characterization of heteroatom compounds in a crude oil and its saturates, aromatics, resins, and asphaltenes (SARA) and non-basic nitrogen fractions analyzed by negative-ion electrospray ionization Fourier transform ion cyclotron resonance mass spectrometry. Energy Fuels. 2010;24:2545-2553.
[Google Scholar]
Sun Z.H., Li D., Ma H.X., Tian P.P., Li X.K., Li W.H., Zhu Y.H., . Characterization of asphaltene isolated from low-temperature coal tar. Fuel Process. Technol.. 2015;138:413-418.
[Google Scholar]
Sun M., Li Y., Sha S., Ma X., . The composition and structure of n-hexane insoluble-hot benzene soluble fraction and hot benzene insoluble fraction from low temperature coal tar. Fuel. 2019;262:116511
[Google Scholar]
Thomas M.J., Jones H.E., Lozano D., Gavard R., Barrow M.P., . Comprehensive analysis of multiple asphaltene fractions combining statistical analyses and novel visualization tools. Fuel. 2021;291:120132
[Google Scholar]
Verstraete J., Revellin N., Dulot H., Hudebine D., . Molecular reconstruction of vacuum gas oils. ACS Division Fuel Chem.. 2004;49:125-132.
[Google Scholar]
Verstraete J.J., Schnongs P., Dulot H., Hudebine D., . Molecular reconstruction of heavy petroleum residue fractions. Chem. Eng. Sci.. 2010;65:304-312.
[Google Scholar]
Wang J., Ferguson A.L., . Mesoscale simulation of asphaltene aggregation. J. Phys. Chem. B. 2016;120:8016-8035.
[Google Scholar]
Wu L.L., Deng W.A., Li C., Zhang H.Y., Wang X.J., . Analysis of asphaltene properties of coal tar heavy fraction and its effect on hydrocracking coking. Acta Fuel Chem.. 2014;42:938-944.
[Google Scholar]
Zhang L., Hou Z., Horton S.R., Klein M.T., Shi Q., Zhao S., Xu C., . Molecular representation of petroleum vacuum resid. Energy Fuels. 2014;28:1736-1749.
[Google Scholar]
Zhang Y., Schuler B., Fatayer S., Gross L., Harper M.R., Kushnerick J.D., . Understanding the effects of sample preparation on the chemical structures of petroleum imaged with noncontact atomic force microscopy. Ind. Eng. Chem. Res.. 2018;57:15935-15941.
[Google Scholar]
Zheng J., Li D., Fan X., Huang Y., Dan Y., . Characterization of heteroatom class species in asphaltenes from medium/low temperature coal tar. Energy Sources Part A. 2020;1:1-15.
[Google Scholar]
Zhu Y., Tian F., Liu Y., Cui L., Li D., . Comparison of the composition and structure for coal-derived and petroleum heavy subfraction by an improved separation method. Fuel. 2021;292:120362
[Google Scholar]

Appendix A

Supplementary material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.arabjc.2021.103531.

Appendix A

Supplementary material

The following are the Supplementary data to this article:

Supplementary data 1

Download

Introduction

Experimental section

Materials
Sample characterization

Modeling approach

Structure attributes
Core attributes library
Computer simulation

Results and discussions

Bulk properties and molecular compositions
Determination of representative average structure molecules

Conclusions

Fulltext Views
2,442

PDF downloads
1,315

View/Download PDF
Download Citations

BibTeX
RIS

Show Sections

[1] Alvarez-Majmutov A., Gieleciak R., Chen J., . Modeling the molecular composition of vacuum residue from oil sand bitumen. Fuel. 2019;241:744-752.
[Google Scholar]

[2] Ballard D.A., Chacón M., Qiao P., Roberts K.J., Harbottle D., . Molecular characterization of strongly and weakly interfacially active asphaltenes by high-resolution mass spectrometry. Energy Fuels. 2020;34:13966-13976.
[Google Scholar]

[3] Boek E.S., Yakovlev D.S., Headen T.F., . Quantitative molecular representation of asphaltenes and molecular dynamics simulation of their aggregation. Energy Fuels. 2009;23:1209-1219.
[Google Scholar]

[4] Cai G.Q., Liu Z.F., Zhang L.Z., . Quantitative structure property relationship model for hydrocarbon liquid viscosity prediction. Energy Fuels. 2018;32:3290-3298.
[Google Scholar]

[5] Chacón-Patiño M.L., Rowland S.M., Rodgers R.P., . Advances in asphaltene petroleomics. Part 1: Asphaltenes are composed of abundant island and archipelago structural motifs. Energy Fuels. 2017;31:13509-13518.
[Google Scholar]

[6] Chacón-Patiño M.L., Rowland S.M., Rodgers R.P., . Advances in asphaltene petroleomics. Part 2: Selective separation method that reveals fractions enriched in island and archipelago structural motifs by mass spectrometry. Energy Fuels. 2018;32:314-328.
[Google Scholar]

[7] Chacón-Patiño M.L., Rowland S.M., Rodgers R.P., . Advances in asphaltene petroleomics. Part 3. Dominance of island or archipelago structural motif is sample dependent. Energy Fuels. 2018;32:9106-9120.
[Google Scholar]

[8] Chacón-Patiño M.L., Smith D.F., Hendrickson C.L., Marshall A.G., Rodgers R.P., . Advances in asphaltene petroleomics. part 4. compositional trends of solubility subfractions reveal that polyfunctional oxygen-containing compounds drive asphaltene chemistry. Energy Fuels. 2020;34:3013-3030.
[Google Scholar]

[9] Chacón-Patiño M.L., Moulian R., Barrère-Mangote C., Putman J.C., Weisbrod C.R., Blakney G.T., Bouyssiere B., Rodgers R.P., Giusti P., . Compositional trends for total vanadium content and vanadyl porphyrins in gel permeation chromatography fractions reveal correlations between asphaltene aggregation and ion production efficiency in atmospheric pressure photoionization. Energy Fuels. 2020;34:16158-16172.
[Google Scholar]

[10] Cho K., Hansen P.E., . Structure identification of novel compounds using simple IR, ¹H, and ¹³C-NMR spectroscopy and computational tools. Bull. Korean Chem. Soc.. 2019;41:78-83.
[Google Scholar]

[11] Cho Y., Kim Y.H., Kim S., . Planar limit-assisted structural interpretation of saturates / aromatics / resins / asphaltenes fractionated crude oil compounds observed by Fourier transform ion cyclotron resonance mass spectrometry. Anal. Chem.. 2011;83:6068-6073.
[Google Scholar]

[12] De Oliveira L.P., Vazquez A.T., Verstraete J.J., Kolb M., . Molecular reconstruction of petroleum fractions: application to vacuum residues from different origins. Energy Fuels. 2004;27:3622-3641.
[Google Scholar]

[13] Feng S., Cui C., Li K., Zhang L., Xu C., . Molecular composition modelling of petroleum fractions based on a hybrid structural unit and bond-electron matrix (SU-BEM) framework. Chem. Eng. Sci.. 2019;201:145-156.
[Google Scholar]

[14] Feng X., Li D., Chen J., Niu M., Liu X., Chan L.L., Li W., . Kinetic parameter estimation and simulation of trickle-bed reactor for hydrodesulfurization of whole fraction low-temperature coal tar. Fuel. 2018;230:113-125.
[Google Scholar]

[15] Frigerio F., Molinari D., . A multiscale approach to the simulation of asphaltenes. Comput. Theor. Chem.. 2011;975:76-82.
[Google Scholar]

[16] Gani R., Harper P.M., Hostrup M., . Automatic creation of missing groups through connectivity index for pure-component property prediction. Ind. Eng. Chem. Res.. 2005;44:7262-7269.
[Google Scholar]

[17] Gaspar A., Zellermann E., Lababidi S., Reece J., Schrader W., . Characterization of saturates, aromatics, resins, and asphaltenes heavy crude oil fractions by atmospheric pressure laser ionization Fourier transform ion cyclotron resonance mass spectrometry. Energy Fuels. 2012;26:3481-3487.
[Google Scholar]

[18] Giraldo-Davila D., Chacon-Patino M.L., Orrego-Ruiz J.A., Blanco-Tirado C., Combariza M.Y., . Improving compositional space accessibility in (+) APPI FT-ICR mass spectrometric analysis of crude oils by extrography and column chromatography fractionation. Fuel. 2016;185:45-58.
[Google Scholar]

[19] Glattke T.J., Chacón-Patiño M.L., Marshall A.G., Rodgers R.P., . Molecular characterization of photochemically produced asphaltenes via photooxidation of deasphalted crude oils. Energy Fuels. 2020;34:14419-14428.
[Google Scholar]

[20] Gortari F.D., García-Jacas C., Martinez-Mayorga K., Medina-Franco J.L., . Database fingerprint (DFP): an approach to represent molecular databases. J. Cheminf.. 2017;9:9.
[Google Scholar]

[21] Headen T.F., Boek E.S., Skipper N.T., . Evidence for asphaltene nanoaggregation in toluene and heptane from molecular dynamics simulations. Energy Fuels. 2009;23:1220-1229.
[Google Scholar]

[22] Hsu C.S., . Compositional space boundaries for organic compounds. Anal. Chem.. 2012;84:3410-3416.
[Google Scholar]

[23] Hudebine D., Verstraete J.J., . Molecular reconstruction of LCO gasoils from overall petroleum analyses. Chem. Eng. Sci.. 2004;59:4755-4763.
[Google Scholar]

[24] Hukkerikar A.S., Sarup B., Kate A.T., Abildskov J., Sin G., Gani R., . Group-contribution+ (GC+) based estimation of properties of pure components: improved property estimation and uncertainty analysis. Fluid Phase Equilib.. 2012;321:25-43.
[Google Scholar]

[25] Jaffe S.B., Freund H., Olmstead W.N., . Extension of structure-oriented lumping to vacuum residua. Ind. Eng. Chem. Res.. 2005;44:9840-9852.
[Google Scholar]

[26] Jameel A.G.A., Khateeb A., Elbaz A.M., Emwas A.H., Zhang W., Roberts W.L., Sarathy S.M., . Characterization of deasphalted heavy fuel oil using APPI (+) FT-ICR mass spectrometry and NMR spectroscopy. Fuel. 2019;253:950-963.
[Google Scholar]

[27] Juyal P., Mckenna A.M., Fan T., . Joint industrial case study for asphaltene deposition. Energy Fuels. 2013;27:1899-1908.
[Google Scholar]

[28] Liu J., Zhao Y., Ren S., . Molecular dynamics simulation of self-aggregation of asphaltenes at an oil/water interface: formation and destruction of the asphaltene protective film. Energy Fuels. 2015;29:1233-1242.
[Google Scholar]

[29] Marshall A.G., . Compositional boundaries for fossil hydrocarbons. Energy Fuels. 2011;25:2174-2178.
[Google Scholar]

[30] Mckenna A.M., Blakney G.T., Xian F., Glaser P.B., Rodgers R.P., Marshall A.G., . Heavy petroleum composition. 2. Progression of the boduszynski model to the limit of distillation by ultrahigh-resolution FT-ICR mass spectrometry. Energy Fuels. 2010;24:2939-2946.
[Google Scholar]

[31] McKenna A.M., Donald L.J., Fitzsimmons J.E., Juyal P., Spicer V., Standing K.G., . Heavy petroleum composition. 3. asphaltene aggregation. Energy Fuels. 2013;27:1246-1256.
[Google Scholar]

[32] McKenna A.M., Chacón-Patiño M.L., Weisbrod C.R., Blakney G.T., Rodgers R.P., . Molecular-level characterization of asphaltenes isolated from distillation cuts. Energy Fuels. 2019;33:2018-2029.
[Google Scholar]

[33] Neurock M., Libanati C., Nigam A., Klein M.T., . Monte carlo simulation of complex reaction systems: molecular structure and reactivity in modelling heavy oils. Chem. Eng. Sci.. 1990;45:2083-2088.
[Google Scholar]

[34] Neurock M., Nigam A., Trauth D., Klein M.T., . Molecular representation of complex hydrocarbon feedstocks through efficient characterization and stochastic algorithms. Chem. Eng. Sci.. 1994;49:4153-4177.
[Google Scholar]

[35] Niles S.F., Chacón-Patiño M.L., Smith D.F., Rodgers R.P., Marshall A.G., . Comprehensive compositional and structural comparison of coal and petroleum asphaltenes based on extrography fractionation coupled with Fourier transform ion cyclotron resonance MS and MS/MS analysis. Energy Fuels. 2020;34:1492-1505.
[Google Scholar]

[36] Pan N., Cui D., Li R., Shi Q., Chung K.H., Long H., . Characterization of middle-temperature gasification coal tar. part 1: bulk properties and molecular compositions of distillates and basic fractions. Energy Fuels. 2012;26:5719-5728.
[Google Scholar]

[37] Pei L., Li D., Liu X., Cui W., Shao R., Xue F., Li W., . Investigation on asphaltenes structures during low temperature coal tar hydrotreatment under various reaction temperatures. Energy Fuels. 2017;31:4705-4713.
[Google Scholar]

[38] Pereira T., Vanini G., Tose L.V., Cardoso F., Fleming F.P., . FT-ICR MS analysis of asphaltenes: asphaltenes go in, fullerenes come out. Fuel. 2014;131:49-58.
[Google Scholar]

[39] Podgorski D.C., Corilo Y.E., Nyadong L., Lobodin V.V., Bythell B.J., Robbins W.K., . Heavy petroleum composition. 5. Compositional and structural continuum of petroleum revealed. Energy Fuels. 2013;27:1268-1276.
[Google Scholar]

[40] Pyl S.P., Hou Z., Van Geem K.M., Reyniers M.F., Marin G.B., Klein M.T., . Modeling the composition of crude oil fractions using constrained homologous series. Ind. Eng. Chem. Res.. 2011;50:10850-10858.
[Google Scholar]

[41] Quann R.J., Jaffe S.B., . Structure-oriented lumping: describing the chemistry of complex hydrocarbon mixtures. Ind. Eng. Chem. Res.. 1992;31:2483-2497.
[Google Scholar]

[42] Quann R.J., Jaffe S.B., . Building useful models of complex reaction systems in petroleum refining. Chem. Eng. Sci.. 1996;51:1615-1631.
[Google Scholar]

[43] Riniker S., Landrum G.A., . Similarity maps - a visualization strategy for molecular fingerprints and machine-learning methods. J. Cheminf.. 2013;5:43.
[Google Scholar]

[44] Schuler B., Meyer G., Peña D., Mullins O.C., Gross L., . Unraveling the molecular structures of asphaltenes by atomic force microscopy. J. Am. Chem. Soc.. 2015;137:9870-9876.
[Google Scholar]

[45] Schuler B., Fatayer S., Meyer G., Rogel E., Moir M., Zhang Y.L., . Heavy oil based mixtures of different origins and treatments studied by atomic force microscopy. Energy Fuels. 2017;31:6856-6861.
[Google Scholar]

[46] Schuler B., Zhang Y., Liu F., Pomerantz A.E., Andrews A.B., Gross L., . Overview of asphaltene nanostructures and thermodynamic applications. Energy Fuels. 2020;34:15082-15105.
[Google Scholar]

[47] Scott D.E., Schulze M., Stryker J.M., Tykwinski R.R., . Deciphering structure and aggregation in asphaltenes: hypothesis-driven design and development of synthetic model compounds. Chem. Soc. Rev.. 2021;50:9202-9239.
[Google Scholar]

[48] Sedghi M., Goual L., Welch W., Kubelka J., . Effect of asphaltene sructure on association and aggregation using molecular dynamics. J. Phys. Chem. B. 2013;117:5765-5776.
[Google Scholar]

[49] Shao R., Shen Z., Li D., Sun Z., Pei L., Liu X., . Investigation on composition and structure of asphaltenes during low-temperature coal tar hydrotreatment under various reaction pressures. J. Anal. Appl. Pyrol.. 2018;136:44-52.
[Google Scholar]

[50] Sheremata J.M., Gray M.R., Dettman H.D., Mccaffrey W.C., . Quantitative molecular representation and sequential optimization of athabasca asphaltenes. Energy Fuels. 2004;18:1377-1384.
[Google Scholar]

[51] Shi Q., Hou D., Chung K.H., Xu C., Zhao S., Zhang Y., . Characterization of heteroatom compounds in a crude oil and its saturates, aromatics, resins, and asphaltenes (SARA) and non-basic nitrogen fractions analyzed by negative-ion electrospray ionization Fourier transform ion cyclotron resonance mass spectrometry. Energy Fuels. 2010;24:2545-2553.
[Google Scholar]

[52] Sun Z.H., Li D., Ma H.X., Tian P.P., Li X.K., Li W.H., Zhu Y.H., . Characterization of asphaltene isolated from low-temperature coal tar. Fuel Process. Technol.. 2015;138:413-418.
[Google Scholar]

[53] Sun M., Li Y., Sha S., Ma X., . The composition and structure of n-hexane insoluble-hot benzene soluble fraction and hot benzene insoluble fraction from low temperature coal tar. Fuel. 2019;262:116511
[Google Scholar]

[54] Thomas M.J., Jones H.E., Lozano D., Gavard R., Barrow M.P., . Comprehensive analysis of multiple asphaltene fractions combining statistical analyses and novel visualization tools. Fuel. 2021;291:120132
[Google Scholar]

[55] Verstraete J., Revellin N., Dulot H., Hudebine D., . Molecular reconstruction of vacuum gas oils. ACS Division Fuel Chem.. 2004;49:125-132.
[Google Scholar]

[56] Verstraete J.J., Schnongs P., Dulot H., Hudebine D., . Molecular reconstruction of heavy petroleum residue fractions. Chem. Eng. Sci.. 2010;65:304-312.
[Google Scholar]

[57] Wang J., Ferguson A.L., . Mesoscale simulation of asphaltene aggregation. J. Phys. Chem. B. 2016;120:8016-8035.
[Google Scholar]

[58] Wu L.L., Deng W.A., Li C., Zhang H.Y., Wang X.J., . Analysis of asphaltene properties of coal tar heavy fraction and its effect on hydrocracking coking. Acta Fuel Chem.. 2014;42:938-944.
[Google Scholar]

[59] Zhang L., Hou Z., Horton S.R., Klein M.T., Shi Q., Zhao S., Xu C., . Molecular representation of petroleum vacuum resid. Energy Fuels. 2014;28:1736-1749.
[Google Scholar]

[60] Zhang Y., Schuler B., Fatayer S., Gross L., Harper M.R., Kushnerick J.D., . Understanding the effects of sample preparation on the chemical structures of petroleum imaged with noncontact atomic force microscopy. Ind. Eng. Chem. Res.. 2018;57:15935-15941.
[Google Scholar]

[61] Zheng J., Li D., Fan X., Huang Y., Dan Y., . Characterization of heteroatom class species in asphaltenes from medium/low temperature coal tar. Energy Sources Part A. 2020;1:1-15.
[Google Scholar]

[62] Zhu Y., Tian F., Liu Y., Cui L., Li D., . Comparison of the composition and structure for coal-derived and petroleum heavy subfraction by an improved separation method. Fuel. 2021;292:120362
[Google Scholar]

Molecular representation of coal-derived asphaltene based on high resolution mass spectrometry

Abstract

Keywords

Coal-derived asphaltene

CID FT-ICR MS

Molecular composition

Average asphaltene structure

Nomenclature

List of symbols

1 Introduction

2 Experimental section

2.1 Materials

2.2 Sample characterization

3 Modeling approach

3.1 Structure attributes

3.2 Core attributes library

3.3 Computer simulation

4 Results and discussions

4.1 Bulk properties and molecular compositions

4.2 Determination of representative average structure molecules

5 Conclusions

Acknowledgement

Declaration of Competing Interest

References

Appendix A

Supplementary material

Appendix A

Supplementary material

Supplementary data 1

Suggested read for related articles:

We value your privacy