Density Function Study of the Interaction of a Surface Modifier with the Oxidized Coal Surface Model.

A density function approach has been used to screen an appropriate surface modifier for oxidized coal to enhance its hydrophobicity in a flotation process. Two oxidized coal surface models, coal-COOH and coal-COONa, based on the substitution of 10-fused benzene rings with COOH and COONa functional groups, have been constructed to mimic the surface hydrophilic sites at acidic and alkaline pHs, respectively. A nonpolar molecule and five polar candidate molecules with different functional groups have been examined on each oxidized coal model surface. Our present study indicates that octane is ineffective toward increasing the surface hydrophobicity for both coal-COOH and coal-COONa models due to its preferential adsorption on hydrophobic aromatic sheet, although it can spontaneously bind to the coal model surfaces at 298 K. Unlike octane, 4-pentylpyridine will present the preferred hydrophobic conformation on both models. However, its adsorption process is favorable energetically only on the coal-COOH model. The optimized geometries of all four oxygen-containing molecules (1-methoxyheptane, 1-octanol, octanal, and octanoic acid) show that directional hydrogen bonds will be formed between their oxygenated groups and the COOH group of coal-COOH model. This results in the protrusion of the hydrocarbon chain toward the water phase, which is beneficial for increasing coal surface hydrophobicity. However, the calculated Gibbs free energies suggest that octanoic acid is the best candidate. The adsorption of all four oxygen-containing molecules on the coal-COONa model is a spontaneous process. However, only sodium octanoate can be regarded as the effective surface modifier according to its optimized adsorption conformation at alkaline pH.

Introduction

During the coal oxidation process, the content of inherently hydrophobic areas is decreased, whereas the content of oxygen-containing functional groups on the coal surface, such as hydroxyl, carbonyl, and carboxyl groups, is increased. These groups easily form hydrogen bonds with water molecules, reducing the natural hydrophobicity of the coal surface.[1]

For industries that utilize fine coal processing, froth flotation is a widely used cleaning technology. The flotation behavior of coal after oxidation has been widely investigated.[2,3] It was found that the adsorption of collector droplets at the coal–water interface is strongly affected by the coal surface properties.[4,5] Oily collectors such as kerosene and diesel oil have long been used in coal flotation. However, the oxidized coal is difficult to float with these oily collectors.[6,7]

It is evident that the surface hydrophobicity of oxidized coals can be improved using a polar molecule or a blend of a polar molecule and a nonpolar oil. Harris et al.[8] reported that a much lower dosage of polyethoxylated nonylphenol was required to achieve the same flotation performance for oxidized coals compared to that of dodecane. Jia et al.[6] proposed that tetrahydrofurfuryl esters have the capability of restoring the floatability of oxidized coal through the formation of hydrogen bond with oxygen-containing functional groups on the oxidized coal surface. Jena et al.[9] found that a mixture of a long-chain fatty acid and kerosene could promote flotation of oxidized high-ash Indian coal. Qu et al.[10] concluded that the floatability of a Shendong low-rank coal can be enhanced by the addition of 2-ethyl hexanol into the collector. Xia et al.[11] showed that biodiesel, which contains much more abundant fatty acids, was more effective collector than diesel. Chang et al.[12] demonstrated that a composite collector of diesel and Triton X-100 could lead to a significant increase in the flotation of oxidized coal compared with diesel alone. In the above studies, the primary aim of the addition of polar molecules is to render the oxidized coal surface hydrophobicity through the interaction of the polar functional groups and the oxidized sites on the coal surface. Hereafter, these polar molecules will be referred to as coal surface modifiers.

In fact, due to the structural differences, the adsorption strength of different types of oxygen-containing groups on the coal surface should vary from one to another. To optimize the effect of surface hydrophobicity, it is desirable to precisely match a suitable surface modifier to every type of surface functional group. What is the appropriate surface modifier for a particular type of oxygen-containing group on the coal surface? How does the surface modifier interact with it on the coal surface? A molecular-level understanding about the adsorption behavior of a surface modifier on the coal surface can provide answers to these two fundamental questions. However, because of the complexity of the coal surface structure,[13] the microscopic characterization of reagent–coal surface interaction is very difficult. Until now, few reports in the literature about these fundamental questions have been found.[14] The screening of coal surface modifiers is still a conventional trial-and-error method.

The purpose of our study was to gain a general understanding about how a range of organic molecules with various functional groups interact with the oxidized coal models using a quantum chemical approach. For studying the oxygenated group on the oxidized coal surface, we focus on a carboxyl group attached to a simple coal model. The interactions between the surface modifier and other oxygenated groups incorporated into coal, such as phenolic and alcoholic hydroxyl groups and carbonyl groups, are currently under investigation. This work may throw light on the screening and design of new chemical reagents for coal flotation.

Quantum Chemistry Calculations

To understand the adsorption mechanism of surface modifiers on the coal surface at the molecular level, it is necessary to establish a suitable model to simulate the oxidized coal surface. The structure of coal is inherently complex and varies widely, depending on the origin, history, and rank of the particular coal examined. The molecular substructure of a typical bituminous coal has, on average, three or four fused aromatic rings with pendant short chains.[15] However, ring condensation in anthracite is believed to be between 10 and 100.[16] For anthracite, few aliphatic chains and heteroatom-containing functional groups survive in its molecule. In this study, a 10-fused benzene ring model was chosen to represent the basic hydrophobic coal surface substructure, which may be regarded as a simplified anthracite model. Aromatic carboxylic acid is believed to be the major form of the carboxylic acid group in an oxidized coal matrix.[17] Therefore, to model the local structure of the COOH site on the oxidized coal surface, we place one COOH group on the edge of this cluster. In this manner, a coal–COOH surface model was constructed, as presented in Figure a. In alkaline coal flotation pulp, the COOH group dissociates from the COO– group. The ion pair of COO– and Na+ could be formed when Na+ ions are abundant (for example, the addition of NaOH, a basic pH adjuster). The Na+ ion may form a salt bridge between COO– and the polar molecule. On the basis of this assumption, a coal–COONa surface model (Figure b) was also constructed.

Figure 1

(a) Coal–COOH and (b) coal–COONa models used to model the oxidized coal surface. Gray, red, purple, and white balls represent C, O, Na, and H atoms, respectively. Their electrostatic potentials (ESPs) and representative extreme points (energies in kcal/mol) on a 0.001 au molecular surface are also shown.

Six candidate model molecules were examined in this study. As we focus on the understanding about how different functional groups interact with the oxidized coal surface, four polar oxygenated molecules, including 1-methoxyheptane (ether group), 1-octanol (hydroxyl group), octanal (carbonyl group), and octanoic acid (carboxylic acid group, sodium octanoate at alkaline pH), were used as the model modifiers. Pyridine is a good solvent for the swelling and extraction of coal[18] due to its strong adsorption and penetration ability. Besides, it has also been proved that pyridine can form hydrogen bond with the oxidized coal surface.[19] Thus, a molecule that contains a pyridine group may be a promising candidate for a surface modifier for oxidized coal. In this study, we also evaluated the interactions of a model molecule, 4-pentylpyridne, and the coal models. It should be noted that the side chain length of all of these molecules is approximately equal. The effect of side groups on adsorption behavior will be considered in the further study. Finally, for comparison, the binding effect of an octane molecule was also investigated, which was used as a model of nonpolar collector.

All of the quantum chemistry calculations were carried out using the Gaussian 09 package.[20] A meta-GGA M06-2X functional,[21] in conjunction with the 6-311G(d) basis set, was used to optimize geometries. The M06-2X functional has shown promising performance for studying main group thermochemistry and noncovalent interactions.[22] Following the recommendations of the developers of the M06-2X functional,[21] the basis set superposition error correction has not been applied. All of the geometries were optimized using the Berny algorithm. Single-point energy calculations were performed at the M06-2X/6-311++G(d,p) level. The effect of water as a solvent has been modeled by means of the polarizable continuum model (PCM).[23] In the framework of PCM, the solvent is represented as a structureless continuum characterized mainly by its dielectric permittivity. The molecular electrostatic potential on the van der Waals surface (electron density = 0.001 au) was analyzed by the Multiwfn program.[24]

To find the most stable adsorption geometry for a particular molecule, calculations were performed for various starting geometries, where the molecule was placed on the coal model surface to promote possible interactions between them via geometry optimization. For every modifier/coal model system, only the most stable adsorption geometry was reported.

The adsorption Gibbs free energy (ΔGads) of the candidate molecule in the aqueous phase was calculated at room temperature (298 K) and 1 atm pressure, corresponding to the actual application conditions for coal flotation. This was determined fromwhere Gsurface, Gmolecule, and Gmolecule–surface represent the total Gibbs free energies of a clean oxidized coal surface, a candidate molecule, and a system with the candidate molecule adsorbed on the coal surface, respectively. A negative adsorption Gibbs free energy determined in this way corresponds to a spontaneous process, and a larger negative value indicates stronger interaction with the surface.

To obtain the Gibbs free energies and confirm the nature of the stationary points, vibrational frequency calculations have been performed for all of the models. The thermodynamic analysis yielded zero point energies and thermal corrections to the electronic energies due to translational, electronic, and vibrational motions, which were used to calculate Gibbs free energies. It was confirmed that geometries corresponding to stable structures do not have imaginary vibration frequencies.

Results and Discussion

Geometrical Structures

In this article, we have chosen two models, coal–COOH (Figure a) and coal–COONa (Figure b), to represent the oxidized coal fragments in acidic and alkaline coal pulps, respectively. The aromatic sheet in these two models is used to represent the hydrophobic region on the oxidized coal surface, whereas the COOH and COONa groups are treated as the hydrophilic sites. It could be found that the atoms of COOH and COONa groups and the aromatic sheet are coplanar in these two models.

Figures and 3 show the obtained adsorption geometries for six candidate molecules on the coal–COOH and coal–COONa models, respectively. For the coal–COOH model, the most stable structure for an octane molecule adsorbed on the coal–COOH is shown in Figure a. This straight-chain alkane is in direct contact with the fused ring basal plane in a parallel orientation. However, the hydrophilic COOH group cannot be covered by this nonpolar chain.

Figure 2

Predicted adsorption free energy (ΔGads) and geometry for (a) octane, (b) 1-methoxyheptane, (c) 1-octanol, (d) octanal, (e) octanoic acid, and (f) 4-pentylpyridine on the oxidized coal–COOH model surface. Gray, red, blue, and white balls represent C, O, N, and H atoms, respectively.

Figure 3

Predicted adsorption free energy (ΔGads) and geometry for (a) octane, (b) 1-methoxyheptane, (c) 1-octanol, (d) octanal, (e) sodium octanoate, and (f) 4-pentylpyridine on the oxidized coal–COONa model surface. Gray, red, purple, blue, and white balls represent C, O, Na, N, and H atoms, respectively.

The preferred adsorption configurations of four oxygen-containing polar molecules, 1-methoxyheptane (Figure b), 1-octane (Figure c), octanal (Figure d), and octanoic acid (Figure e), were also determined. From Figure , it can be found that all of these molecules can form hydrogen bonds with COOH groups present on the model surface. The hydrogen bond exhibits strong orientation preference.[25] It has been reported that a stable conventional hydrogen bond angle has a range of 140–180°.[26] Interactions with hydrogen bond angle in the range of 120–140° have reduced stabilization energies, and angles below 120° are generally unlikely to correspond to significant interactions. As shown in Figure , all of these four oxygen-containing polar molecules can form stable hydrogen bonds with the COOH groups present on the coal surface. However, unlike octane, their hydrophobic chains cannot lie flat on the basal plane of fused rings. This can be understood in terms of two competing energy contributions, hydrogen bond and dispersion interaction. The stable hydrogen bonds between oxygen-containing groups in surface modifiers and COOH group in the coal–COOH model cause steric hindrance between the hydrophobic chains and fused aromatic rings. For 4-pentylpyridine, the obtained geometry shows that it acts as a hydrogen bond acceptor molecule. Its alkyl chain also stays away from the basal plane of nonpolar aromatic fused rings. For all of the adsorption systems, no obvious deviation was found from planarity for the geometry of the coal–COOH model.

Next, the adsorption configurations of six candidate molecules on the coal–COONa model were investigated, which is relevant to the alkaline coal flotation pulp. In this situation, octane was still adsorbed on the hydrophobic aromatic sheet region. The atom positions of the COONa group are nearly unchanged after the adsorption of octane. This suggests that the interaction between the COONa group and octane is very weak. In addition to octane, 1-menthoxyheptane, 1-octanol, and octanal also lie flat on the basal plane of fused aromatic rings (Figure b–d). 1-Menthoxyheptane, 1-octanol, and octanal attach to a Na site of the model surface through their oxygen atoms in oxygen-containing groups. As the positively charged Na+ ion acts as an electrophile, it will seek electron density on other atoms. This leads to the association of Na+ and oxygen atoms in the candidate molecules. Obviously, the ion-pair association strength of Na+ and COO– of the coal–COONa model is weaker than that of the covalent carboxylic OH bond of the coal–COOH model. Therefore, the position of Na+ is more flexible and less directional than that of hydrogen in the COOH group, which can be easily observed from Figure . As a result, the hydrocarbon chains of 1-menthoxyheptane, 1-octanol, and octanal can adsorb flat on the basal plane of fused aromatic rings, which maximize the dispersion interaction. Under the conditions relevant for coal flotation in an alkaline pH environment, similar to the case of the coal–COOH model, we supposed that octanoic acid is deprotonated to form sodium octanoate. Unlike octane as well as 1-menthoxyheptane, 1-octanol, and octanal, the obtained preferred adsorption geometries of sodium octanoate and 4-pentylpyridine show that their hydrophobic chains still stick out from the coal model.

Evidently, for all of the studied candidate model molecules, two distinctly different kinds of adsorption modes on the oxidized coal surface can be identified. In model I, the molecule adheres to the hydrophilic site of the coal surface through its polar group while its hydrophobic end is projected away from the coal surface. In model II, the candidate molecule binds to the hydrophobic basal surface of coal in a parallel orientation.

Electrostatic Potential on Molecular Surfaces

The adsorption of surface modifiers will modify the surface characteristics, which may influence the interaction between the coal surface and water (i.e., the hydrophobicity of coal surface). The electrostatic interactions, including hydrogen bonds, play dominant roles in the interactions of the coal surface and water.[27] The electrostatic features of the coal surface can be predicted well by its electrostatic potential (ESP) map computed on a 0.001 au molecular surface because this map can reflect the specific feature of a molecule and is also appropriate for studying noncovalent intermolecular interactions.[28] Therefore, the ESP map of unmodified and modified coal on the 0.001 au molecular surface is used to evaluate the effect of coal surface modification.

First, the ESP maps of coal–COOH and coal–COONa models were investigated. Figure a,b shows the calculated ESP surfaces and representative extreme points of two oxidized coal models. It can be seen that the hydrogen and double bond oxygen atoms of the carboxyl acid group of the coal–COOH model have much more positive (51.14 kcal/mol) and negative (−30.58 kcal/mol) surface potentials, respectively, than do those of the fused aromatic rings. This suggests that these regions can easily adsorb polar molecules, like water. For the coal–COONa model (Figure b), the two oxygen atoms of the carboxylate functional group (COO–) and the Na+ ion have strong negative (−35.55 and −31.65 kcal/mol) and positive (134.72 kcal/mol) surface potentials, respectively, which will result in the electrostatic association of the COO– group and the Na+ counterion in aqueous solution. In addition, the Na+ in the coal–COONa model has stronger ESP compared to that of hydrogen of the carboxyl acid group in the coal–COOH model, indicating that the local polarity and hence hydrophilicity of the coal–COONa model should be stronger than that of the coal–COOH model.

Next, the ESP maps of octane/coal–COOH and octane/coal–COONa model systems were investigated (Figures a and S1e), which offered an opportunity to evaluate the modification effect of coal surface with the oily collector due to the fact that the properties of this nonpolar molecule are similar to those of the conventional oily collectors. Compared with those in the coal–COOH model, it can be found that the ESP features of COOH group in the octane/coal–COOH model remain unchanged (Figure a). Therefore, the hydrophobicity of the coal–COOH model will not improve significantly after the adsorption of octane. Similarly, the hydrophobicity of the coal–COONa model also cannot be improved according to its ESP features (Figure S1e).

Figure 4

Electrostatic potentials and representative extreme points (energies in kcal/mol) of (a) octane/coal–COOH, (b) octanoic acid/coal–COOH, and (c) 1-octanol/coal–COONa model systems on a 0.001 au molecular surface. Gray, red, purple, and white balls represent C, O, Na, and H atoms, respectively.

Corresponding to their adsorption modes, two kinds of ESP maps of coal surface could be found after the adsorption of the candidate polar molecules. The first kind of ESP map applies to the adsorption of 1-methoxyheptane (Figure S1a), 1-octanol (Figure S1b), octanal (Figure S1c), octanoic acid (Figure b), and 4-pentylpyridne (Figure S1d) on the coal–COOH model as well as that of sodium octanoate (Figure S1h) and 4-pentylpyridne (Figure S1i) on the coal–COONa model (adsorption model I), which would improve the hydrophobicity of oxidized coal surface. The octanoic acid/coal–COOH system (Figure b) is a representative case. From Figure b, it can be found that the strong negative ESP region is around the two COOH groups in the octanoic acid/coal–COOH system. After adsorption of octanoic acid, the ESP maximum (51.14 kcal/mol) around the hydrogen of the carboxylic acid group in the coal–COOH model (Figure b) disappeared. At the same time, the adsorption of octanoic acid also weakened the ESP value of the double bond oxygen of the COOH group of the coal–COOH model (from −30.58 to −17.27 kcal/mol), but it increased the ESP value of singly bonded oxygen of the studied coal model (from −14.43 to −18.65 kcal/mol). In addition, two new ESP minimum values emerged, around the oxygen atoms of carboxylic group of octanoic acid. This should be attributed to the formation of hydrogen bond. However, according to the present ESP results, it is difficult to compare the slight difference in local hydrophilicity between the hydrogen-bonded COOH groups in the octanoic acid/coal–COOH system and the isolated COOH group in the original coal–COOH model. However, this should not be a key factor to determine the hydrophobicity of a modified coal surface. According to the optimized geometry of the octanoic acid/coal–COOH system, the nonpolar hydrocarbon chain of octanoic acid protruded into water solution. Meanwhile, the associated two COOH groups are buried in the interior of octanoic acid/coal–COOH system. In the flotation process, if an oil droplet or an air bubble collides with this region, it will have to interact with the exposed hydrocarbon chain. Thus, it is reasonable to expect that the electrostatic potential of the hydrocarbon chain will determine the surface hydrophobicity of modified coal. Obviously, the ESP of the alkane chain of octanoic acid is very weak, just like that of a fused aromatic sheet. Therefore, the hydrophilic site on the coal surface is essentially converted into a hydrophobic site. In this manner, the hydrophobic chain of octanoic acid could provide a good hydrophobic anchor point for the oil droplet or the air bubble, which is beneficial to coal flotation. This is applied to the adsorption of 1-methoxyheptane (Figure S1a), 1-octanol (Figure S1b), octanal (Figure S1c), and 4-pentylpyridne (Figure S1d) on the coal–COOH model as well as that of sodium octanoate (Figure S1h) and 4-pentylpyridne (Figure S1i) on the coal–COONa model.

The ESP maps of adsorption of 1-methoxyheptane (Figure S1f), 1-octanol (Figure c), and octanal (Figure S1g) on the coal–COONa model (adsorption model II) indicate that this kind of adsorption mode cannot improve the hydrophobicity of the oxidized coal surface. The 1-octanol/coal–COONa system is a typical case. Figure c shows the ESP map and several extremes around the polar region of the 1-octanol/coal–COONa system. It can be found that Na+ in this system has a smaller ESP maximum value (116.62 kcal/mol) as compared to that in the original unabsorbed coal–COONa model (Figure b, 134.72 kcal/mol). However, a new maximum region around the hydroxyl hydrogen of 1-octanol arises after its adsorption. In addition, the negative ESP around two oxygen atoms of the carboxylate group became stronger. Overall, no significant weakening of polarity around the hydrophilic region of the 1-octanol/coal–COONa system can be observed at all. More importantly, the polar hydrophilic region in the 1-octanol/coal–COONa system will be exposed to the water phase. As a result, the flat adsorption geometry of polar 1-octanol in model II cannot obviously improve the hydrophobicity of coal in the flotation process. This analysis can also be applied to 1-methoxyheptane/coal–COONa and octanal/coal–COONa systems.

Gibbs Free Energies

It should be noted that all of the geometries presented above were optimized at 0 K without consideration of the influence of temperature. The Gibbs free energy accounts for the enthalpy contribution as well as the entropic penalty, which is a predictor of the spontaneity of adsorption at a certain temperature. Thus, the Gibbs free energy computed at 298 K and 1 atm was used to determine the adsorption possibility of the candidate molecules.

Figures and 3 present the adsorption Gibbs free energy (ΔGads) for octane, 1-methoxyheptane, 1-octanol, octanal, octanoic acid/sodium octanoate, and 4-pentylpyridine in the aqueous phase at 298 K and 1 atm pressure. It is found that the candidate organic molecules bind to the coal–COOH model in the orderThis indicates that among all of the above-mentioned molecules the highest binding affinity toward the coal–COOH model is seen for octane, and the negative ΔGads value (−0.49 kcal/mol) encourages its adsorption on the coal–COOH model. However, its adsorption cannot shield the COOH group present on the coal–COOH model according to the ESP results. Thus, it is reasonable to suppose that the adsorption of nonpolar oil cannot ineffectively improve the surface hydrophobicity of oxidized coal, which is in agreement with the common experimental results.[6]

For octanoic acid that contains a COOH group, the negative ΔGads (−0.19 kcal/mol) indicates that its adsorption on the coal–COOH model is spontaneous. Thus, our calculations imply that octanoic acid should be the qualified surface modifier for oxidized coal considering its adsorption geometry discussed above, which is consistent with the experimental results.[29] With respect to the other oxygen-containing molecules, the ΔGads values for those with ether, hydroxyl, and carbonyl groups are positive (2.50 kcal/mol for 1-methoxyheptane, 3.97 kcal/mol for 1-octanol, and 3.99 kcal/mol for octanal). This indicates that the adsorption of those surface modifiers that contain ether, hydroxyl, and carbonyl groups on the oal–COOH model may not be completed spontaneously at 298 K. Besides, 4-pentylpyridine that possesses a nitrogen-containing group also can bind to the coal–COOH model although its binding is not as strong as that of octanoic acid. On the basis of its adsorption conformations and ΔGads value, 4-pentylpyridine should also be effective toward shielding the COOH group on the coal surface.

Interestingly, despite the formation of hydrogen bond between their oxygen-containing groups and the carboxyl acid group on the coal–COOH model, the adsorption of all five polar candidate molecules is weaker than that of nonpolar octane. This should be ascribed to the competing result of hydrogen bond and dispersion interaction, as discussed above. The oriental hydrogen bond between these oxygen-containing molecules and carboxyl acid group incorporated into the coal model surface hinders the dispersion interaction between their hydrophobic chains and the aromatic sheet, resulting in their lower adsorption abilities.

Under the alkaline conditions, when the COONa group is present on the coal surface, the adsorption strength of six selected model molecules on the coal–COONa model is very differentThe ΔGads values for all four oxygen-containing molecules are negative (Figure ), meaning that they can bind to the coal–COONa model spontaneously at 298 K. Unfortunately, the adsorption configurations of 1-octanol, 1-methoxyheptane, and octanal reveal that they cannot improve the hydrophobicity of the coal–COONa model, as discussed in Section . Therefore, the only effective surface modifier for the coal–COONa model is sodium octanoate. For 4-pentylpyridne, the positive ΔGads values discourage their adsorption in alkaline pH conditions. Like the octane/coal–COOH system in the acidic conditions, the adsorption of octane on the coal–COONa model is spontaneous. However, its adsorption cannot improve the hydrophobicity of the oxidized coal surface, as discussed in Section .

Conclusions

The density function method has been used to screen the appropriate surface modifier for oxidized coal flotation. It was found that the effective modification of the COOH and COO– groups on the coal surface by a polar molecule depends not only on the free energy preferences of adsorption but also on the corresponding adsorption conformation.

Without any polar groups in the molecule, octane is ineffective toward shielding the neutral and ionized carboxyl groups due to its preferential adsorption on the hydrophobic region of the coal surface despite the favorable ΔGads. When the COOH group prevails on the oxidized coal surface, all five selected polar model molecules can form hydrogen-bonded complexes with the coal model. The strong orientation preference of hydrogen bond hinders the binding of the hydrocarbon chain of the candidate polar molecules on the coal hydrophobic region. The resulting protrusion of the hydrocarbon chain will improve the hydrophobicity of the oxidized coal surface. However, the calculated Gibbs free energies of adsorption evidence that only octanoic acid and 4-pentypyridine may be able to adsorb on the coal–COOH model in the aqueous phase at 298 K and the adsorption of octanoic acid is energetically more favorable than that of 4-pentypyridine. These results suggest that the molecule that contains a carboxyl group should be the preferred surface modifier to impart hydrophobicity toward the carboxyl site on the coal surface at acidic pH. Under pH conditions where COONa groups exist on the oxidized coal surface, the optimized geometries of surface modifier/coal–COONa model systems show that the COONa group will attract the polar groups of the candidate molecules. However, the orientation of this interaction is less constrained than that of hydrogen bond, resulting in the binding of hydrophobic end of 1-methoxyheptane, 1-octanol, and octanal to the hydrophobic region of the coal model. Their adsorption geometry cannot obviously improve the hydrophobicity of coal. From the calculated ΔGads results, we propose that sodium octanoate is the only qualified surface modifier.