Liu Qiong; Wu Zhiqiang; Huang Zuoxin; Zhang Feng; Zhao Yi

(Research Institute of Petroleum Processing, SINOPEC, Beijing 10083)

Abstract: Diesel engine technology innovation causes excessive soot accumulated in engine oil. Due to its detrimental effect on lubricant and diesel engine, improving the dispersibility of engine oil to restrain soot aggregation efficiently is the key technique for formulations. In this study, the aggregation of soot and interaction between dispersant and soot were investigated by molecular dynamic simulation. It was found that the molecular interaction between the dispersant and the soot aggregation system had a significant influence on disrupting the soot aggregation. Bis-PIBSI was more beneficial to having more interaction sites with soot molecules, while the mono-PIBSI with a high proportion of polar groups had stronger interaction with soot molecules. According to the simulation result, suggestions for use of additives were proposed.Carbon black dispersancy test was exploited to verify the dispersion effect of different dispersants on carbon black. The results indicate that mono-PIBSI and bis-PIBSI added at suitable mixture ratio to lubricant could perform good dispersion ability.

Key words: engine oil; soot; aggregation; dispersant; dispersancy; molecular interaction

1 Introduction

Heavy duty diesel engines continuously adopt new engine technology to reach an increasingly tighter NOx emission limits, which would lead to a significant increase of soot in lubricating oil. Soot is the incomplete combustion product of hydrocarbon fuel in heavy duty diesel engines transported to the crankcase oil by blowby. The high levels of soot in crankcase oil can produce a range of hazards, such as serious engine abrasive wear and increased viscosity of lubricants. Therefore, in order to maintain good performance of diesel engine lubricants with soot loading, the American Petroleum Institute (API)continues to issue new specifications and develop new engine oil categories with more stringent standards to evaluate the soot dispersancy performance of engine oil.In order to improve soot-handling capabilities and meet the tougher standards of API, oil formulation developers are devoted to solving the problem of high soot levels dispersancy to develop the high dispersancy oil. In lubricating oil, base oil and additives such as viscosity index improver and ashless dispersant both can affect the aggregation of soot[1-4]. One of the main methods to decrease soot accumulation and reduce the soot induced damage is to add a proper amount of dispersants into engine oil to increase the soot dispersion ability.Dispersants consisting of polar head and long alkyl chain can provide steric or electrostatic barriers to maintain soot particle dispersion. Different types of dispersants can slow down or stop the soot aggregation kinetics. The number of dispersant polar heads[4]and the structure of polar heads and alkyl chain[5]of dispersant may have a different impact on the stabilization of soot aggregation.Considerable researches[6-10]have been conducted to investigate the in fluence of dispersants on the aggregation of soot by using laboratory experiments and analytical methods. However, few studies have focused on the mechanisms of soot aggregation and the influence of dispersants on this aggregation behavior in a molecular level, and fewer reasonable suggestions are given based on the nature of soot.

In this work, a combination of computational simulation and experiment was conducted to study the effect of dispersant on soot and offer help to select a proper lubrication formulation. In the first part, we made an attempt to use molecular dynamics simulation to model the soot aggregation behavior and study the in fluence of dispersants on soot aggregation system to get a deeper understanding of the interaction between dispersant and soot. Based on the results of the first part, in the second half of the paper we tried to design a more reasonable dispersant combination to realize better soot dispersancy.Its performance was tested by carbon black dispersancy test.

2 Computational Details and Experimental

2.1 Computational details and models

2.1.1 Modeling

Based on previous studies[11-13]and characteristics of soot[14], the soot molecule has been modeled by large carbonaceous clusters, the edge carbon atoms of which are replaced by oxygen atoms and saturated by hydrogen atoms. The cluster contains 92 carbon atoms and 10 oxygen atoms plus 32 hydrogen atoms (see Figure 1).

Figure 1 The soot conformationThe molecules is shown as stick models with C (gray), O (red), and H (white).

Polyisobutylene succinimide (PIBSI) series are the main lubricant dispersant products with a high market share in engine oil accounting for more than 80%, among which the commonly used components are mono-PIBSI, bis-PIBSI and high molecular weight PIBSI. The selected dispersant structures are shown in Figure 2, and the models are built according to Table 1.

Figure 2 The chemical structure of dispersants

Table 1 Structural characteristics of dispersant models

2.1.2 Computational methods

The establishment and processing of the models were performed by using Material Studio software programs.The molecular dynamics simulations were carried out by Forcite plus module in the force field of COMPASS,condensed-phase optimized molecular potentials for atomistic simulations, which was the firstab intioforcefield that enables accurate and simultaneous prediction of gas-phase properties and condensedphase properties for a broad range of molecules and polymers[15]. The selected thermodynamic ensembles were NPT and NVT ensemble. The used temperature control method was the Nose-Hoover thermostat. The allowed maximum energy difference between successive steps of a production simulation was 50000.0 kcal/mol. The longrange electrostatic interactions were calculated by the Ewald method[16]and the van der Waals interactions were dealt with the atom based method.

First of all, the molecular mechanics and molecular dynamics simulation achieves the optimum structures of soot and dispersant. Optimized structures of soot and dispersant were utilized to construct the soot aggregation model and the dispersant and soot mixture model shown in Figure 3 by Amorphous cell module.Ten soot molecules are constructed in a box to study soot aggregation (Figure 3(a)), and then one dispersant is added to ten soot aggregation systems to investigate the influence of dispersant on the soot aggregation system shown in Figure 3(b).

Figure 3 (a) The initial conformation of soot aggregation system; (b) The initial conformation of dispersant and soot aggregation system(The dispersant molecule is highlighted in yellow.)

A temperature of 298 K (25 °C) and a time step of 2 ns was chosen for all dynamic simulations. The MD simulations of 500 ps at 298 K were implemented by the NPT ensemble and the remaining 1500 ps was carried out by the NVT ensemble. After the system was fully optimized and balanced, the intermolecular interactions were calculated according to Equation (1).

where Einteris the interaction energy between dispersant and soot; Etotalis the total energy of all molecules; Esootis the total energy of the soot molecules; Edispersantis the energy of the dispersant.

2.2 Materials

Two kinds of ashless dispersants were commercial polyisobutylene succinimide (PIBSI) produced by the Wuxi South Petroleum Additives Co., Ltd., The structure of selected ashless dispersants is shown in Figure 2(a)and (b). The same dosage of ashless dispersants was added into a fully formulated API CH-4 diesel-engine oil as the test oil. The formulated oil using the API Group Ⅱbase stocks contained all normal additives with the same dosage and composition of other additives, such as the overbased detergent, and the antiwear and antioxidant additives. The formulated oil has a kinematic viscosity of 15.67 mm2/s at 100 ℃ and a density of 868.7 kg/m3at 20 ℃, to comply with the quality requirements for SAE 15W-40 grade.

2.3 Carbon black dispersancy test

Carbon black is often chosen as a soot surrogate in lubricant in term of both kinds of particles exhibiting similar aggregate size and predicting the likely performance of lubricanting oils[8,17-19]. In this test, the Printex U carbon black produced by the Evonik Degussa Company was chosen as the soot substitute. The carbon black has a surface area of 254 m2/g, with its radius of elementary particles equating to 25 nm. 3% of carbon black was added to the test oil and was then agitated for five minutes into the suspension by an ultrasonic dispersion machine (Model:VCX 130; SONIC &MATERIALS INC.). After being dispersed, the kinematic viscosity of mixture at 100 ℃ was measured.

3 Results and Discussion

3.1 Soot aggregation behavior

Since soot aggregation leads to particle size increase and deposition that can cause adverse effect on lubricant performance, it is essential to study the soot accumulation and make great efforts to minimize the aggregation. So the optimized soot molecules are used to build a soot aggregation system to investigate the soot aggregation behavior firstly. It can be seen that 10 soot molecules are randomly dispersed in Figure 4(a), and with the increase in simulation time, the soot molecules tend to be closer in Figure 4(b). After 1500 ps, the soot system upon achieving adequate equilibrium almost tends to accumulate in a parallel con figuration shown in Figure 4(c).

Figure 4 The conformation of aggregation model with increase in simulation time

The details of soot aggregation are presented in Figure 5. Soot molecules tend to be horizontally stacked as a result of existing π-π strong interaction[20-21]among soot molecules with an aromatic ring structure, in which the surface stacking distance is about 0.287 nm, and the calculated interaction energy among soot molecules is-479.4kJ/mol, suggesting that the soot tends to agglomerate spontaneously by the attractive forces of π-π interactions.Hence under this attraction, the accumulation of soot would continue to accumulate until it builds up to form deposits.

Figure 5 The parallel accumulation structure of soot

3.2 Interaction between dispersant and soot system

Ashless dispersants are always used to stabilize the soot particles and protect soot-laden lubricants from excessive viscosity increase. The effect of different dispersants on soot aggregation system was investigated through establishing a soot aggregation and dispersant interaction simulation system with a mixing ratio of 1:10. Dispersant can destroy the parallel accumulation of soot molecules and make them distribute in a more scattered way (Figure 6). It can be seen that the dispersant interacts with multiple soot molecules when they are in close contact,so the parallel structure is disrupted by the interaction between the dispersant and the soot.

Figure 6 The conformation of interaction between dispersant and soot

Dispersants with different structures all have positive effects on soot dispersion (Figure 7 and Figure 8). In Figure 7,the interaction energy between D0 and soot aggregation system is close to -300 kJ/mol, the absolute value of which is less than the absolute energy of soot accumulation. It can be inferred that D0 may have little influence on soot aggregation. The interaction energy between HPD1 and soot aggregation system is around -1200 kJ/mol, which is the strongest interaction between dispersant and soot aggregation because of its large molecular weight. With the increase of alky1 chain length, the absolute value of intermolecular interaction energy between dispersant and soot aggregation system dramatically increases. It indicates that the interaction strength between dispersant and soot aggregation system markedly increases with the increase of dispersant molecule action sites. Compared with mono-PIBSI and bis-PIBSI with the same PIB length, the absolute value of interaction energy between double succinimide and soot aggregation system is significantly greater than that of single succinimide. There may be two ways causing this results. On the one hand, the increase of polar groups may increase the interaction intensity; on the other hand, the increase of alkyl chains leading to the increase of adsorption sites may also be an important reason.

Figure 7 The interaction energy between dispersant and soot molecules

As shown in Figure 8, the interaction between the dispersant and the soot aggregation system has a significant influence on the soot aggregation system,the interaction intensity among soot aggregation system allways decreases with the addition of dispersant. And the effect of bis-PIBSI on interaction energy of soot aggregation system is remarkably greater than that of mono-PIBSI. It can be indicated that bis-PIBSI with more polar groups and long alkyl chains had a preferable ability to disrupt soot aggregation system. In addition,the interaction among soot aggregation could hardly be affected by the alky chain length.

Figure 8 The interaction energy of soot aggregation system with dispersant

In order to further analyze the effect of polar group on interaction between dispersant and soot, the mixture system of dispersant and soot was established by mixing dispersant and soot at a weight ratio of 6:3, which was close to that adopted by the actual system. Based on the results shown in Figures 3 and 4, only the dispersants with long alkyl chain were chosen. It can be seen from Figure 9 that the interaction intensity between PD0 molecules and soot molecules is slightly higher than that between PD1 and soot,and is much higher than that between HPD1 and soot. The possible reason can be inferred that when the dispersant and soot molecule were mixed at the same ratio, the effect of molecular weight was weakened because the number of interaction sites was basically the same, and the mono-PIBSI with a high proportion of polar groups had stronger interaction with soot molecules than the bis-PIBSI.

Figure 9 The interaction energy between dispersants and soot molecules

3.3 Carbon black dispersion test

The computational results (see 3.1, 3.2) suggested that the larger the molecular weight was, the more advantageous it was to disperse the soot aggregation and enhance the interaction with soot molecules. Mono-PIBSI and bis-PIBSI could both reduce the soot aggregation to some extent in different ways. Bis-PIBSI was more bene ficial to possessing more interaction sites with soot molecules, and mono-PIBSI with a high proportion of polar groups had a stronger interaction with soot molecules. The different effect of PIBSI on soot dispersion indicated that rational utilization of different advantage may produce better dispersion effect. According to the simulation results, the carbon black dispersion test was established to verify the dispersion effect of dispersant on soot and an optimum proportion of dispersant compounding was determined.The dispersancy characteristics of oil sample with smaller viscosity growth rate are better.

The effect of dispersant on carbon black dispersion is listed in Table 2. The reference oil had passed the ASTM engine oil test in the Mack T-8E engine and had shown excellent dispersion performance. The viscosity growth rate of reference oil added to carbon black was set as the reference. The relative value of viscosity growth rate in the test oil was calculated by the formula, viz.: Relative value of viscosity growth rate of test oil = (viscosity growth rate of test oil)/(viscosity growth rate of reference oil). To compare the relative value of viscosity growth rate, the dispersion of oil sample with smaller relative value of viscosity growth rate was better. It can be found from Table 2 that the viscosity growth rate of OD3 oil is smaller than that of OD2 and OD3, and was close to the reference oil. The OD3 oil was an optimized combination of mono-PIBSI and bis-PIBSI. The data suggested that when mono-PIBSI and bis-PIBSI at a suitable mixture ratio would achieve a good dispersion ability.

Table 2 Effect of dispersant on carbon black dispersion test

4 Conclusions

Through a combination of computational and experimental method, the performance of dispersant in terms of soot dispersion was studied. The effect of different chain length and structural dispersants on soot dispersion has been studied by using the molecular dynamic simulations. It was found that soot molecules tend to be spontaneously and horizontally aggregated as a result of existing π-π strong interaction. The interaction between dispersant and soot aggregation system has a significant influence on the soot aggregation system, and the effect increases with the increase of alkyl chain length. Bis-PIBSI was more beneficial to having more interaction sites with soot molecules, and mono-PIBSI with a high proportion of polar groups had stronger interaction with soot molecules. On the basis of the simulation results,the ability of long-chain mono-PIBSI, bis-PIBSI and some of their combinations to reduce viscosity increase caused by soot aggregation has been studied in the carbon black dispersion test. The dispersant combination at a certain proportion exhibited good soot dispersancy. It must be noticed that this study was only conducted on a simple system without base oil for the convenience of calculation. Further research may be explored to determine the effects of dispersants on soot in the presence of base oil in order to get a deeper understanding of the interaction mechanism.

Acknowledgements: We gratefully acknowledge the support from the China Petrochemical Corporation funding (Sinopec Group, No.117022) on this work.