Reaction Molecular Dynamics Study on the Mechanism of Alkali Metal Sodium at the Initial Stage of Naphthalene Pyrolysis Evolution

Abstract

In order to study the role of metal sodium in the spray pyrolysis of biomass tar, this paper designs a sodium-containing naphthalene pyrolysis system (NSS) and a pure naphthalene pyrolysis system (PNS) using naphthalene as the carbon source and sodium chloride as the sodium metal donor for comparison. This enables an exploration of the effect of sodium on the initial nucleation of carbon fumes formed by naphthalene pyrolysis using reaction molecular dynamics (ReaxFF MD). The simulation results show that NSS undergoes pyrolysis reactions earlier and faster than PNS at the same temperature. Simulated at 3250 K temperature for 2 ns, the naphthalene pyrolysis consumption rate of the NSS was faster than that of the PNS, and the addition of sodium atoms during the condensation process provided more active sites and accelerated the condensation of macromolecular products. Moreover, Na⁺ and carbon rings form a Na⁺-π structure to promote the bending of graphite lamellae to facilitate the formation of carbon nuclei. Molecular dynamics simulations were used to simulate the formation of carbon nuclei during the initial stage of naphthalene pyrolysis, revealing that the mechanism of sodium salt catalyzed the acceleration of organic matter pyrolysis from a microscopic visualization perspective.

1. Introduction

Excessive CO₂ emissions have led to global climate change and have caused numerous disasters. Reducing CO₂ emissions is urgent. Carbon reduction strategies mainly include source control and terminal renewable utilization [1]. Reducing fuel combustion is the most effective means of source control. Coal, oil, and biomass combustion are the main pathways for CO₂ emissions. Combustion is also the lowest energy utilization method. The deep development and utilization of fuel can not only reduce CO₂ emissions, but also improve the comprehensive utilization value of fuel. CCSU is recognized as one of the most effective and promising carbon terminal renewable utilization strategies [2]. Carbon materials can be combined with various materials to obtain different structure materials and have shown good performance in CCSU [3].

Biomass has a wide source, large yield, and is renewable. Most biomass is used as a combustion raw material. However, the low heat generated by biomass combustion makes it difficult to scale up its application in the fields of heating and power generation. Biomass combustion causes a large amount of CO₂ emissions, and the combustion form reduces the utilization rate of biomass, which is not conducive to energy conservation and emission reduction. Biomass is the only renewable energy source that can be converted into gas or liquid fuel (biomass tar) [4]. How to improve the high value-added utilization of biomass resources and reduce the CO₂ emission caused by combustion will become one of the focus issues of human beings. In recent years, the in-depth development and utilization of biomass has achieved initial results [5]. Salid et al. [6] summarized the process of producing biomass fuel from biomass in terms of Pakistan’s future demand for biomass fuel and investigated the effect of homogenous/heterogeneous chemical reactions on the Casson cross nanofluid flow towards a porous wedge. They found that tetra nanoparticles amplify the heat transfer rate and thermal conductivity and production of bioethanol by up to 59.96%. By controlling the pyrolysis temperature, biomass can be converted into various liquid hydrocarbon compounds and be used as chemical raw materials and carbon materials [7]. This can improve the high value-added utilization of biomass resources and reduce CO₂ emissions caused by combustion. When the pyrolysis temperature is 800–1000 °C, it can be converted into aromatic compounds rich in naphthalene, methylnaphthalene, biphenyls, anthracene, and other 2~3 rings [8], such as 2~3 ring tar, which is a high-quality raw material for the industrial production of special materials. The rapid conversion of biomass tar into high-quality carbon materials will be one of the effective ways to improve its high value-added utilization.

There are many production methods for carbon materials, among which the spray pyrolysis (SP) method can obtain materials with different structures by regulating the factors of material ratio, pyrolysis material degree, and residence time. With a simple method and fast production speed, SP will become the mainstream method to produce functional carbon materials. SP is an important technology to synthesize powder materials and films from a solution using a one-pot method. Liquid raw materials are atomized into aerosols and enter the pyrolysis chamber. In the pyrolysis chamber, the aerosol undergoes a series of reactions, including solvent evaporation, solute diffusion, pyrolysis, and deposition, generating materials with different structures [9]. SP has been widely used to prepare various nanostructured materials, such as hollow balls [10,11], dense balls [12], egg yolk-shell structures [13], core-shell structures [14], thin films, and composite materials [15,16]. By regulating the ratio of Co(NO₃)₂·6H₂O and glucose, Yu et al. [17] prepared hollow porous carbon microspheres doped with Co₃O₄ with a wall thickness less than 50 nm using spray pyrolysis, which were applied to the positive electrode of a battery to obtain a lithium battery with high capacity (1297.1 mAhg⁻¹) and good stability. Carbon materials are widely used in CCSU and electrode materials, and carbon materials prepared by different methods can be predicted in advance using a conductor-like screening model for real solvents (COSMO-RS) for material performance [18]. In the SP process, the formation of each particle only occurs in one micron-sized droplet. This means that millions of droplets will have a series of rapid physical and chemical reactions simultaneously at the same time. This makes it extremely difficult to observe the transformation of each particle in both the temporal and spatial dimensions. Many articles have reported the application of SP in the preparation of various materials, few articles have focused on the mechanism of obtaining materials using the SP method.

The alkali metal Na can affect the pyrolysis reaction of coal, thereby altering the morphology and physicochemical properties of carbon black [19,20,21]. At present, some researchers have chosen to add sodium salts during the pyrolysis process of organic compounds to regulate the properties of carbon materials [22,23,24]. Using Na as an additive, it is expected that carbon black or other carbon materials with a special morphology and surface functional groups will be obtained, while the excellent water solubility of sodium salt makes it easy to remove it from the product, thereby obtaining high-purity products. Kim et al. [25] synthesized porous carbon microspheres using spray pyrolysis using sucrose as carbon source. NaCl was applied as a pore template and activating agent. The porosity of carbon microspheres was controlled by adjusting the ratio of sucrose to NaCl. The temperature of the furnace was 900 °C. The powder was washed with distilled water to remove the NaCl. Porous carbon microspheres were obtained after drying.

Reaction molecular dynamics (ReaxFF MD), a reaction implementation of molecular dynamics, was developed in 2001 by van Duin [26]. It is proposed that the bond energy, bond angle energy, torsion energy, lone pair electron term, hydrogen bond, Coulomb force, and Van der Waals force between multiple “pairs of atoms” can be calculated and synthesized. ReaxFF is a classical force field theory based on bond order, so the force field can judge the bonding and breaking of chemical bonds, thus, simulating chemical reactions [27]. In recent years, reactive molecular dynamics has been widely used in the fields of organic polymerization, combustion, and pyrolysis, such as the nucleation mechanism of small molecular polymerization, nanostructured carbon fullerenes, carbon nanotubes, and graphene; the cracking of hydrocarbon fuels; and the formation of soot at high temperatures, as it enables a visual study of the reaction process at the molecular level and has a profound impact on the clarification of pyrolysis reaction mechanism. It is a powerful means to study the microprocess [28,29].

In summary, the production of functional carbon materials by SP with tar as a raw material and sodium salt as an additive is a very promising method. However, in order to make SP better serve the production of carbon materials, it is important to obtain the reaction mechanism of the SP production of carbon materials, especially the role of metal salt additives in the pyrolysis process. There are no reports on this area yet. Therefore, we propose an idea to simulate the spray pyrolysis process with ReaxFF MD to study the spray pyrolysis mechanism. We used naphthalene instead of tar, added NaCl, built a nano aerosol model, and simulated the spray pyrolysis and carbonization process. We analyzed the trajectory data, carbonization rates, and products to study the reactions occurring in nucleation at the initial stage of carbonation and to explore the mechanism of generating carbon nanocarbon materials from sodium salt in the initial stage of spray pyrolysis. This provides a new idea for exploring the mechanism of the preparation of related materials by SP.

2. Method

The whole simulation process was divided into two parts. First, we established the periodic potential cell of the naphthalene molecule and sodium chloride model, and then we conducted the high-temperature pyrolysis simulation.

2.1. Model Establishment

The modeling method was mainly used to establish the periodic bounding cell of naphthalene and sodium chloride using Materials Studio 2019 software (MS). In MS, we selected the “Amorphous Cell” module to establish a simulation box, filled 162 naphthalene molecules and 62 sodium chloride molecules into the cell, and set the initial density to 0.005 g/cm³. The Forcite calculation module, Compass II force field, and medium accuracy were selected to perform energy minimization relaxation and annealing on the initial cell. First, we selected “Energy” under the “Setup” option for the energy minimization optimization. Then, for the annealing treatment, the Berendsen temperature control mode was selected; the time step was set to be 1 fs. Canonical (NVT) annealing took place between 300 and 1500 K in 100 ps. Second, under the isothermal isobaric ensemble (NPT), in the temperature range of 300–1000 K, we set the “Temperature” to 750 K and “Pressure” to 0.2 MPa and performed the simulated annealing for 100 ps. We repeated the above annealing process once, and finally obtained the NSS of the atomized state with a density of 0.114 g/cm³. We established a PNS model using the same method. See Figure 1 for both models.

2.2. Molecular Dynamics Simulation of Reaction

Reactive molecular dynamics were simulated with the ReaxFF force field in LAMMPS (3Aug2022-MPI) software. The LAMMPS simulation data were analyzed using the Python package.

2.2.1. Effect of Reaction Temperature on the Model Pyrolysis Process

In an actual production process, the temperature is generally between 1000 and 1900 K. In order to accelerate the simulation decomposition rate, increasing the temperature can shorten the simulation time. During the simulation, we mainly considered cases of pyrolysis at 2000, 2500, 3000, and 3500 K. Kowalik, Purse, and Orekhov successfully simulated the pyrolysis process of hydrocarbons at 2800 K, 3500 K, and 6000 K, providing us a reasonable basis between 2000 and 4000 K [30,31,32]. We recorded the trajectory information to examine the effect of temperature on the pyrolysis rate and process.

2.2.2. Selection and Setting of Simulation Process Parameters

Since the pyrolysis reaction takes place at high temperatures and the system is in the high energy region, the time step should be set as small as possible to avoid the loss of atoms during the test by using a larger time step. This structure can result in a non-physical simulation process to avoid simulation errors. We set the time step to 0.1 fs. The reasonable choice of the coefficient synthesis can better express the stiffness of the pyrolysis material during the carbonization process and avoid the problem of cumulative expansion [33]. We chose a canonical ensemble (NVT) for the simulation to ensure the rationality of the simulation results.

The simulation process was divided into four processes: velocity initialization and relaxation, temperature rise, constant temperature, and temperature drop processes. The specific parameters of each process were set as follows. (1) Relaxation process: relaxation was performed for the whole system under the LAMMPS operating environment. The initial temperature was set to 300 K. The NVE system synthesis and Berendsen thermostat were used to relax the model for 10 ps to obtain a more reasonable equilibrium state. (2) Heating process: in order to avoid large temperature oscillations in the thermostat, we used “NVE + Berendsen” for temperature control. We raised the temperature of the two models from 300 K to 2000, 2500, 3000, 3500, and 4000 K, with a 10.0 fs temperature damping constant and a 10 K/ps heating rate. The advantage of this temperature control is that when the temperature rises to a very high level, the degree of structural fragmentation is temperature-dependent and not affected by the heating rate. (3) Constant temperature pyrolysis process: the temperature was increased to a preset temperature and then thermo-thermolysis simulations were carried out using the NVT system and the Nosé–Hoover thermostat to control the temperature with a 10.0 fs temperature damping constant. We used the default Reaxff value for the product-atom-pair calculation (>0.3), and the product information was recorded as the average bond order per 100 bond order sample numbers once per 100 calculations. The duration of the entire thermostatic pyrolysis simulation was 2 ns. (4) Cooling process: after the isothermal pyrolysis was completed, the NVE ensemble combination “Berendsen” thermostat was used for the temperature control simulation. The temperature dropped from the isothermal pyrolysis temperature to 1000 K, with a 10.0 fs temperature damping constant and a 1 K/ps cooling rate. We recorded the output simulation trajectory and reaction process information, including the number of steps, temperature, healthy sequence, and energy.

2.2.3. Data Processing Methods

After the simulation, OVITO (Release 3.7.12) software was used to visualize the formation process of the products of the pyrolysis process, and external Python scripts were used to analyze the number of product species, the maximum molecules, and the carbonization process during the simulation, and to analyze the effect of the addition of NaCl on the naphthalene pyrolysis system.

3. Results and Discussion

3.1. Selection of Simulation Temperature

In order to obtain nanosecond time scales to accelerate the simulation reaction progress and reduce the simulation time cost, it is necessary to choose the appropriate temperature [34,35]. In this paper, simulations were carried out in the temperature range of 2000–4000 K. The number of molecules was plotted against the simulation time by counting the changes in the number of molecules at each temperature, and the number of Na⁺ and Cl⁻ molecules removed by NSS at the time of comparison; the results are shown in Figure 2. As can be seen from Figure 2a,b, the NSS and PNS molecular-number-variation curves fluctuate very little and almost do not respond under the 2000–2500 K simulation conditions. Under 3000 K simulation conditions, NSS is approximated as a straight line past the origin, while PNS can be seen as a straight line with an intercept on the x-axis, indicating that NSS undergoes pyrolysis reaction earlier than PNSS. Under 3500 K simulation conditions, NSS shows a convex curve and PNS shows a concave curve, indicating that the pyrolysis rate of NSS is higher than that of PNS. Under 4000 K simulation conditions, the molecular-number-change curves of NSS and PNS can be divided into two stages; namely, the rapid pyrolysis stage and the pyrolysis equilibrium stage, and it can be seen from the figure that NSS completes the first stage earlier.

From the analysis of the above results, it is clear that the number of product molecules fluctuates very little for both systems when simulations are performed at 2000–2500 K. The naphthalene molecules hardly decompose. The reaction at 3000 K is slow and time consuming for simulation; at 3500 K and 4000 K the number of reaction molecules increases rapidly, but the NSS pyrolysis is too fast, fragmentation is serious, and all carbon rings are opened (see Figure 3), which is not conducive to the formation of large volume carbon spheres [36,37]. For this reason, we conducted simulations of the two systems at 3250 K. The product analysis curves are shown in Figure 2. The change in the number of molecules of pyrolysis of the two systems at this temperature is linear, and the number of molecules increases steadily. Combined with the temperature selection method of Montgomery-Walsh et al. [38] for simulating the evolution mechanism of glass carbon nanostructures, we determined the simulation temperature to be 3250 K.

3.2. Comparison of the NSS and PNS Naphthalene Molecular Evolution Process

During the pyrolysis process, naphthalene is the carbon source provider of the condensation system, and the change in the number of naphthalene molecules reflects the extent to which the simulation reaction proceeds. Figure 4 shows the pyrolysis evolution of NSS and PNS naphthalene molecules, and it can be seen that the pyrolysis rate of NSS naphthalene molecules is significantly higher than that of PNS naphthalene molecules. Within 600 ps, the naphthalene consumption curve of PNS approximates a straight line, and the number of naphthalene molecules decreases at a more stable rate. The NSS naphthalene consumption curve approximates the concave curve of the hyperbola, the naphthalene molecules decrease rapidly at the beginning of pyrolysis, and the reduction rate decreases gradually at the later stage; at T = 25 ps, 33.95% of the naphthalene has been consumed, while for PNS the amount is only 1.85% (see Figure S1). NSS was 134 ps and 220 ps ahead of PNS when 50% and 90%, respectively, of the naphthalene molecules were consumed.

Compared with PNS, the addition of sodium chloride is equivalent to introducing heteroatom catalysts. Na and Cl are the most reactive metal and non-metallic groups, and more prone to substitution and C-H bond cleavage reactions with naphthalene molecules. Naphthalene molecules undergo rapid dehydrogenation and aromatic ring opening reactions. Quyn and Gao [39,40] studied the pyrolysis process of Victoria brown coal and NaCl, confirming the existence of interactions between them. They pointed out that, at a rapid heating rate, there is a strong interaction between Cl and the coal/carbon matrix; the retention of Cl in carbon was related to a large number of bond breaks in the pyrolytic coal/carbon matrix. There was an interaction between Na and nascent char, and free radicals were produced on the surface of the coal/char matrix when Na volatilized. At the simulated temperature, NaCl exists as a gaseous ion, and Na⁺ and Cl⁻ accelerate naphthalene pyrolysis at the early stage of pyrolysis mainly by promoting the breakage of the C-H bond of the aromatic ring. Cl⁻ activates the H on the naphthalene molecule, forming a C-H-Cl bond, accelerating C-H bond breakage and dehydrogenation, “escaping” from the naphthalene molecule with HCl gas, disrupting the cyclic energy balance of the naphthalene molecule, and providing acceleration for subsequent ring opening and cleavage. Na⁺ forms hydrogen bonds with hydrogen atoms on naphthalene molecules, and the H atom is activated, causing C-H to break and dissociate unstable HNa, which then decomposes into Na⁺ and H⁺. When Na⁺ and naphthalene ring bridge-bonding carbon into a Na⁺-π bond [19,41], the law of bridge-bond carbon bonding is changed, the bridge bond energy is reduced, and the C=C bridge bond breaks to form an unstable 10C ring, forming a five-membered and seven-membered ring complex. This is the main pathway for the formation of five carbon rings and seven-membered rings. However, there are no heteroelements in PNS, and we can only rely on high-temperature heat to force the aromatic ring C-H to break. Figures S1 and S2 allow for a clear comparison of the changes in reactants and products in the first 50 ps of the two pyrolysis systems. The typical reactions involving Na⁺ and Cl⁻ in the first 25 ps of the simulation are shown in

Table 1. Typical reactions involving Na⁺ and Cl⁻.

Reaction Equation	OVITO Visualization Capture Structure
${\mathrm{C}}_{10}{\mathrm{H}}_8+\mathrm{C}{\mathrm{l}}^{-}\stackrel{}{\to }{\mathrm{C}}_{10}{\mathrm{H}}_7^{-}+\mathrm{HCl}$
${\mathrm{C}}_{10}{\mathrm{H}}_8\stackrel{{\mathrm{Na}}^{+}}{\to }{\mathrm{C}}_{10}{\mathrm{H}}_7^{-}+{\mathrm{H}}^{+}$
${\mathrm{C}}_{10}{\mathrm{H}}_8\stackrel{{\mathrm{Na}}^{+}}{\to }{\mathrm{C}}_8{\mathrm{H}}_6^{2-}+{\mathrm{C}}_2{\mathrm{H}}_2$
${\mathrm{C}}_{10}{\mathrm{H}}_8+{\mathrm{Cl}}^{-}\stackrel{{\mathrm{Na}}^{+}}{\to }{\mathrm{C}}_{10}{\mathrm{H}}_6^{2-}+\mathrm{HCl}+{\mathrm{H}}^{+}$

Notes: C H Cl Na.

.

3.3. Evolution of the Pyrolytic Condensation of NSS and PNS

Pyrolysis is a semi-combustion process that undergoes reactions such as cracking, dehydrogenation, rearrangement, condensation, and carbonization, which occur under anoxic conditions. Among them, the condensation reaction is an important process for the growth of small molecules into large molecules and the nucleation of carbon fumes, and the molecular species and number within the system will be reduced when the condensation reaction occurs. The condensation process can be analyzed by statistically simulating the change of molecular species in the system to analyze the initial nucleation mechanism of carbon soot. Combining Figure 5, Figure 6, Figure 7, Figures S3 and S4 (because the pyrolysis visualization process is similar and only time differences exist, we only performed a visualization analysis for NSS pyrolysis), it is evident that this pyrolysis simulation mainly goes through the process of naphthalene molecule dehydrogenation and ring opening to form small molecules and long-chain hydrocarbon radicals, the free condensation of small molecules and long-chain hydrocarbons into multi-carbon rings, and macromolecular condensation-depolymerization-condensation growth into carbon soot particles. Firstly, the naphthalene molecule is rapidly pyrolyzed to produce small molecules and radicals, and the molecular species and number increase rapidly. In this process, naphthalene molecules are dehydrogenated and ring-opened to form aromatic radicals, C1–C4 small molecules, and chain radicals. Secondly, the simulation enters the process of small molecules and free condensation into rings. This process is the key period of core formation; small molecules and chain carbon radicals condense into long-chain carbon. Long-chain carbon continues to grow into “octopus-like” aliphatic carbon chains, which are condensed into carbon rings through cross-linking and the ring closure of the side chains. The number of 6C rings, as well as 5C and 7C rings, rapidly increase and condense into multiple C macromolecular graphite fragments to form the initial carbon smoke core. During this process, NSS and PNS polycondensation curves showed different patterns, see Figure 5. NSS fluctuates with a flat top dynamic equilibrium at t = 230~670 ps, and the number of molecular species does not increase significantly. The condensation curve of PNS fluctuates in a sharp peak shape in the range of t = 445–700 ps. During the Reaxff MD simulation, no medium was added and the system was in a vacuum state. Each small-molecule free radical has a very small volume compared to the simulated cell. Under a simulated density, the probability of fragmented small molecules colliding with each other is very low. There is a positive electric field around Na⁺. The “C” on small-molecule radicals formed by dehydrogenation or cracking are either negatively charged or positively charged. Within the range of electric field force, free radicals will undergo directional movement and arrangement. Directional movement causes negatively charged radicals to approach Na⁺, and positively charged radicals to move away from Na⁺. Directional arrangement results in negative charge C facing towards Na⁺ and positive charge C facing away from Na⁺. This directional movement and arrangement increase the probability of positive and negative radicals colliding and combining with each other during their movement, and the condensation between small molecules is strengthened. This is why the inflection point of NSS molecular species changes appears earlier than for PNS. The difference between the two curves intuitively indicates that the addition of Na⁺ affects the linear growth and cyclization process of small molecules. Finally, the macromolecular condensation-depolymerization-condensation growth phase is entered, and the condensation curve moves downward. In this process, the multiple rings evolve to more stable six-membered rings by rearrangement, the multi-C macromolecular fragments undergo polymerization, the number of C in the largest molecule increases rapidly, and the molecular species decreases rapidly. During this process, we found a special phenomenon whereby the curves of the two systems oscillate violently rather than smoothly. Through subsequent analysis, we found that, due to the instability of the polycarbon macromolecules formed through condensation, a repeated process of polymerization-depolymerization-polymerization occurred during polymer growth, continuously releasing configurational energy until a stable structure was achieved. This result is in agreement with the findings of Hang et al. [42], who found that bond breakage and the formation of frequent fluctuations occur during the condensation process. As can be seen from Figure 5 and

Table 2. Trigger and deadline schedule of each stage of NSS and PNS.

System	Stage 1 (ps)	Stage 2 (ps)	Stage 3 (ps)
NSS	0–230	230–670	670–2000
PNS	0–445	445–700	700–2000

, all processes of NSS occur earlier than those of PNS, indicating that the addition of sodium salt can catalyze the process of the simulated system. This is because the generation of active sites during high-temperature rapid pyrolysis decreases continuously with the enhancement of condensation, the H-C reaction activity of PNS is very low, and the generation of new active sites is difficult. Meanwhile, in NSS Na⁺ can interact with H-C and convert into effective carbon gasification sites, providing more active sites and promoting the condensation of the system into nucleation [39,40,43].

At the end of the simulation, the polymerization state of NSS and PNS polymers were compared, see Figure 8. NSS polymerizes into spherical particles, whereas PNS has a graphene-lamellae stacking morphology. The reason for this result is that the strong electropositivity of the Na⁺ surface (the electrostatic potential at the extreme point is 217.39 kcal/mol) can be mutually attracted to the graphene carbon ring center with a very small value of negative potential to form the Na⁺-π structure. The structure can increase the lamellar curvature and promote the bending of graphene to form carbon nanospheres.

3.4. Evolution of the Largest Molecule

During pyrolysis or combustion, the condensation of carbon atoms into carbon-containing macromolecules growth process is the main mechanism of soot formation. To study the formation of soot during the anoxic pyrolysis of naphthalene, we traversed the simulation trajectory with the number of bonded carbon atoms in a single chemical structure, plotted the evolution pattern of the chemical structure with the highest carbon number, and analyzed the soot formation process. Figure 9 shows the variation of the maximum molecular C number versus the simulation time during pyrolysis for both NSS and PNS systems.

The NSS curve shows that at 0–100 ps, two naphthalene molecules rapidly form about 20 molecules of C through the polymerization of Cl⁻ and Na⁺ on the benzene ring, with occasional occurrences of 30 molecules (Figure 9b). Based on Figure 9c, it is interesting to find that at 0–300 ps, the variation pattern of the C number of macromolecules is an integer multiple of 10, an integer multiple of naphthalene molecules. This indicates that during this simulation time, the large molecules are mainly formed by weak H-H bonds of naphthalene molecules, but this structure is in small quantities and unstable during the simulation process. In this time, mainly hydrocarbon bond breaking and ring opening to form small molecules are dominant, and the results are consistent with the analysis of the rapid increase in molecular numbers in Figure 5. At 300–450 ps, the curve resembles a low sloping straight line, and combined with the smooth vibration curve in Figure 5 at 230–670 ps, it suggests that the number of molecules breaking bonds and opening rings to form new molecules is close to equilibrium with the number of molecules consumed by the new core of C formation, and the number of molecular cleavage reaches the peak, while the number of C70–C80 condensates increases rapidly. This stage is an important stage of soot core formation.

At 450–670 ps the maximum molecular C number fluctuates significantly. From the curve observation, the maximum molecular C number at this stage is 2–4 times higher than the previous stage. The reason for the doubling is that the condensation-depolymerization reaction occurs in the C70–C80 condensate; the newly condensed macromolecular configuration may release configuration energy after repeated condensation depolymerization due to uneven energy distribution, completing the structural doubling. At 670 ps after the condensate polymerization, the slope of the NSS curve increases abruptly and the macromolecular C number increases rapidly, and, finally, the maximum molecular carbon number of NSS reaches 1377 at 2 ns. For the PNS curve, the increase of the maximum carbon number is not as rapid as that of NSS. The formation of the maximum carbon number molecule was slow at 0–800 ps and fluctuated more in the interval of 1050–1420 ps, and the maximum molecular carbon number was 1384 at 2 ns.

The comparison of the NSS and PNS maximum carbon number statistical curves indicates that the addition of sodium salts promotes the condensation of carbon atoms, as demonstrated by our previous results. Active sites are required for molecular condensation. Macromolecules merge into condensers by linking aliphatic side chains, which require active site support when merging into condensomes. The H atom on the macromolecule is mainly distributed on the edge C (Figure 10). When the moving Na⁺ passes through the edge of the macromolecule, it forms a hydrogen bond with H on the edge C through electrostatic force, activating C-H to undergo cleavage and dehydrogenation, providing an active site for the merger of macromolecules into condensates. The visualization snapshots in Figure 11 and Figure S5 show the dehydrogenation process of Na⁺ interacting with edge H-C. The presence of Na⁺ has a great influence on the nucleation process of graphite lamellae; Na⁺ is positively charged and forms a positive electrode, the carbonated dehydrogenation forms the center of the carbon ring with negative electric polarity, and both form a Na⁺-π structure through electric field force. At a high temperature, when Na⁺ moves rapidly through the surface of the graphite layer, the graphite layer bends towards Na⁺ under the action of the Na⁺-π structure, making the flat graphite layer form a spherical surface, providing conditions for the formation of a spherical core. A snapshot of the molecular dynamics simulation visualization of the Na⁺-π-structure interaction is shown in Figure 12. Na⁺ accelerates the formation of soot at the early stage of pyrolysis through the two mechanisms of action described above.

3.5. Evolution of the H/C Ratio of the Largest Molecule

The above comparison between NSS and PNS simulations indicates that the addition of sodium salts is beneficial for the graphitization of naphthalene pyrolysis. Our subsequent statistical analysis only focuses on NSS. Figure 13 shows the variation of the H to C ratio and quantity during pyrolysis. As seen in Figure 13, the change in H atoms can be divided into two stages; in the first stage the number of H atoms grows with the number of carbon atoms, but the H/C ratio decreases. The second stage is the macromolecular carbonization and dehydrogenation polymerization stage. In the first stage, the simulation is dominated by dehydrogenation and ring opening at the beginning, and the H/C ratio curve decreases rapidly. When entering the growth of small molecule condensation, the number of H atoms grows with the number of C atoms (Figure 13b), but the rate of H number growth is smaller than the rate of C atom number growth, resulting in a still decreasing trend of the H/C ratio with the increase of the molecular carbon number (Figure 13a). When the maximum C number molecule condensation-depolymerization-condensation tends to condense into larger molecules, the growth of the marginal H-atom number reaches a peak, and the hydrogen atom change enters the second stage. In the second stage the polycondensation precursors continue to dehydrogenate to provide active sites for condensation into larger molecules. The more active the sites, the more stable the polymer. To stabilize the polycondensation, rapid dehydrogenation is required to provide multiple active sites, leading to a lower H content of the polycondensation precursor fragment. When the macromolecule C number undergoes a drastic change, the H number does not increase exponentially, but slowly decreases, and the H/C ratio rapidly decreases. When a more stable polycondensation is formed and then carbonized and dehydrogenated, the C number changes slowly and the H atoms dissociate rapidly. This stage can be regarded as the pyrolysis system entering the rapid condensation graphitization stage.

3.6. Changing Pattern of Carbon Rings in NSS System

The anaerobic pyrolysis process is a carbonization process of organic matter, and the number of rings, especially the number of six-membered rings, is an important indicator of the graphitization occurring during the pyrolysis process. C5 and C7 rings are the main direct sources of C6 rings, and we only counted 5–7 meta-rings to analyze the naphthalene pyrolysis carbonation process. Figure 14, Figures S3 and S4 reveal the evolution of the rings during the pyrolysis of naphthalene. The change in the number of rings can be divided into three stages. In the early stages of pyrolysis, the number of 6C rings is provided by naphthalene molecules. When Na⁺ is close to the center of the naphthalene molecule aromatic ring, the Na⁺ electric field changes the naphthalene electric field distribution and accelerates the ring opening of the naphthalene molecule at a high temperature. As the naphthalene molecule keeps opening its rings, the number of unsaturated fatty rings, C1–C4 small molecules, and linear carbon continues to increase, the 6C ring curve rapidly decreases. This is the first stage. Entering the second stage, fatty rings, C1–C4 small molecules, and linear carbon rapidly increase the multiple carbon ring through condensation and rearrangement. Due to continuous dehydrogenation, the chain carbon begins to grow. Unsaturated long chains capture each other to form a disordered carbon skeleton, and carbon rings begin to form. The number of 5C and 7C rings increases rapidly. The increase in the number of 6C rings is less than the consumption of naphthalene decomposition, and still shows a decreasing trend. The 5C rings increasing provides enough positive curvature to the structure, forming an “octopus claw”-like macromolecular structure when condensed with long carbon chains. The “octopus claw” macromolecules undergo intramolecular condensation due to dehydrogenation, resulting in a rapid increase in carbon rings. The 6C loop curve begins to rise. During the ring formation process, Na⁺ complexes with free radicals to form transition-state bridge bonds and change the bond angle or distance of adjacent C by gathering or pushing them apart. When the transition-state bridge bond breaks, the rearrangement reaction of adjacent C accelerates the formation of the ring structure. The increase in the number of carbon rings is also the reason for the decrease in the number of molecular species of the system and the increase in the amount of macromolecular carbon, and this stage is the core stage of soot formation. In the third stage, the macromolecular carbon rings tend to stabilize. The 5C and 7C rings evolve towards the 6C rings, and their numbers begin to decrease and finally stabilize. The number of 6C rings continues to increase until reaching stability. At the beginning of this stage, the internal ring structure is complex and the structure is unstable. Figure S4 shows a representative process of carbon ring formation within macromolecules. With the progression of graphitization, the internal bonds of macromolecules break and produce new active sites. Chain-like active sites are prone to forming 5C and 7C rings. Meanwhile, 6C rings can be generated through the formation of the bridged bond between a 3C and a 6C ring within the 7C ring. It is also possible to convert the 6C ring by breaking the bridge bond between the 3C and 5C rings, and 6C can also be produced by closing the bond inside the large ring structure. This is the graphitic growth phase. The simulation results are consistent with Marcus and Han simulating the carbon ring conversion law in the carbon smoke formation process [30,36].

3.7. Detailed Visualization of the NSS Pyrolysis Process

Figure 15 shows the visualization of the NSS pyrolysis process, from which the changes in the state of matter throughout the pyrolysis process can be visualized very well.

Before 25 ps, the initial system reaction of the pyrolysis simulation is mild and mainly focuses on dehydrogenation. Although Figure 5 shows the logarithmic depletion of naphthalene molecules, this is due to the fact that only the hydrogen bonds on the benzene ring are broken and dehydrogenated under the catalytic action of Na⁺ and Cl⁻, and naphthalene molecules mainly underwent dehydrogenation reactions, with only a small amount of benzene ring opening. The products generated by the reaction are mainly C₁₀H₇, C₁₀H₆, and HCl, with 1~2 H₂ and C₂H₂. The amount of H⁺ produced during this time is less than Cl⁻; they react to form HCl, so less H₂ is produced. Dehydrogenation remains the main reaction within 100–200 ps, with an increase in the ring opening rate, yielding small amounts of straight chain hydrocarbon of C7–C8.

At t = 200–600 ps, the ring opening reaction is intense, and the production of 5C–7C rings is accelerated. A large number of “octopus claw”-like aliphatic chains are formed. These “octopus claws” “grab” each other, polymerize, and rearrange to form multi-rings, and form multi-carbon small graphite lamellae. The maximum number of C molecules increases from 80 to 350. At this point, the macromolecules are unstable and undergo condensation and depolymerization reactions at any time. The high variety of heterocycles within the molecule leads to structural instability.

At t = 600~1000 ps, the macromolecule repeatedly undergoes the condensation-depolymerization-condensation reaction, the unstable heterocyclic ring inside the molecule is transformed to the stable 5C~7C ring, and the 5C and 7C rings evolve to the 6C ring, gradually forming a stable macromolecular graphite layer. The maximum molecular C number grows from 350 to 1190.

At t = 1000–2000 ps, the macromolecules continue to carbonize to remove marginal hydrogen and deepen graphitization. The conversion of the 5C and 7C rings to the 6C ring within the molecule is gradually completed, and the maximum molecular C number grows to 1380. With time, the structure tends to stabilize the graphite flake layer under the action of the Na⁺-π structure, which gradually bends into spherical particles. The high dipole moment of Na⁺ forms the center of the positive charge collection [44], and the carbon ring center forms a negative charge collection. Under the action of the dipole moment, the graphite sheet bends and converges towards the center, undergoes continuous dehydrogenation, excessive rearrangement, and bending, ultimately forming smooth ellipsoidal particles. In the same simulation time, the result of PNS simulation is a multi-layered graphene stack structure. Purse et al. [31] simulated the complete pyrolysis and carbonization process (at 3500 K, in 2 ns) using phenol formaldehyde resin as a carbon source without a catalyst. The simulation result is a graphite layer structure. This is consistent with our PNS results. The PNS pyrolysis visualization process is shown in Figure S6.

4. Conclusions

Pyrolysis simulations of NSS and PNS models were performed by ReaxFF MD to compare and analyze the effect of NaCl on the pyrolysis process of naphthalene. At the end of the 2 ns simulation, good spherical nanoparticles could be observed in the NSS, revealing that the addition of NaCl has a catalytic effect on the naphthalene pyrolysis process and the initial stage of carbon nanosphere formation. Temperature has a significant impact on the simulation rate of pyrolysis, and it has been widely confirmed that high temperature can accelerate the simulation rate. In our study, excessive temperature leads to the severe fragmentation of molecules, which is not conducive to the condensation reaction. It is necessary to determine the appropriate simulation temperature based on different simulation systems. In a ReaxFF MD simulation, the pyrolysis mechanism of NSS and PNS is similar in nature. The simulation process went through three stages: dehydrogenation ring opening, linear chain and aliphatic ring growth, and macromolecular polymerization. There are significant differences in pyrolysis rate and structure.

• The first stage: naphthalene molecules undergo dehydrogenation and ring opening fragmentation. Na⁺ and Cl⁻ are prone to substitution and complexation reactions on naphthalene molecules and C-H fracture and aromatic ring opening are accelerated. The fragmentation of naphthalene molecules is faster in NSS than in PNS.
• The second stage: linear chain and fatty ring growth. Compared with PNS, there is a Na⁺-positive-electric-field effect in NSS, which increases the collision combination probability of heteroelectric radicals and promotes linear chain growth. Na⁺ forms a transition-state bridge bond and shortens the ring closing time.
• The third stage: Polymerization and carbonization of macromolecules. Na⁺ interacts with macromolecular edge C-H, shortens the generation time of the edge carbon active site, and accelerates macromolecular condensation polymerization. The Na⁺-π structure formed by electropositive Na⁺ and negative graphite layers improves the bending efficiency of graphite layers. This structure facilitates the formation of spherical particles.
• During a Reaxff MD simulation of 2 ns at 3500 K, NSS is a three-dimensional sphere, while PNS is a two-dimensional layer. It has been confirmed that sodium salts can promote the formation of early soot cores during organic matter pyrolysis.

In this work, the mechanism of the role of sodium salt in the formation of carbon nuclei at the early stage of pyrolysis was simulated by molecular dynamics. For the first time, sodium metal ions were introduced into the ReaxFF MD simulated organic pyrolysis system, and, from a microscopic visualization perspective, this demonstrates the catalytic acceleration of organic matter pyrolysis by sodium salts. The mechanism of action of metallic sodium in the pyrolysis of organic compounds has been explored. The mechanism of preparing carbon materials using spray pyrolysis was first discussed through ReaxFF MD.