Optimizing Experimental Design to Determine Friction and Wear Characteristics of Used Oil Contaminated with E20 Fuel ^†

Abstract

The aim of this paper is to develop an experiment design to determine the wear and friction parameters of an SAE 0W20-grade reference oil and an SAE 0W20-grade used oil contaminated with E20 fuel. Wear tests in a ball–disc arrangement were conducted at temperatures and loads established using Design of Experiments. A surface analysis of the wear scars was performed using a digital microscope, while the oils were investigated using a viscometer and a Fourier-transform infrared spectrometer. Design of Experiments was further employed to develop two models using R software. For the reference oil, the developed model was only suitable for screening, while the model for the E20 fuel contaminated used oil showed a 9.8% difference between the measured and estimated coefficients of friction.

1. Introduction

Advancements in automotive technology and the growing emphasis on environmental responsibility have highlighted the use of alternative fuels as a crucial opportunity to reduce carbon emissions and enhance the sustainability of mobility and transportation. The International Energy Agency projects that, although alternative fuels made up only 4% of global transport energy demand in 2020, this percentage is expected to rise to 14% by 2050. [1]. The most common alternative fuel is ethanol, which has a widespread use in pure or blended form with petrol in spark ignition engines due to its beneficial emission properties [2,3]. Global ethanol production has increased by an average of 3.8% over the past 10 years, peaking at 115 billion liters in 2019, which also underlines the importance of investigating ethanol as a relevant alternative fuel [4]. However, the use of these blends can alter the viscosity, tribofilm thickness [5], and chemical composition of the lubricant, which requires a comprehensive investigation of this phenomenon, as the fuel enters the lubricant system in small volume percentages.

Specifically, under cold-start conditions, alcohols can decompose, resulting in the formation of acetaldehyde in the case of ethanol [6]. In laboratory studies, engine oil was artificially altered with combustion products like acetaldehyde and acetic acid to examine their effects on oil properties. Besser et al. found that the anti-wear additive (ZDDP) content significantly decreased to 25% and 20% of the fresh oil concentration after exposure to acetaldehyde and acetic acid, respectively [7,8]. According to another study, the tribometer results, as reported by Lenauer et al., indicate that the wear rates and tribofilm thicknesses decrease when using acetaldehyde and acetic acid, which are considered the most reactive combustion products of bioethanol [8]. Design of Experiments is an important element in scientific research activities, as it reduces the time and cost of conducting experiments and allows the graphical representation of the results and the study of the interactions of variables [9]. Thus, in this study, the analyses of the effect of E20 fuel on the coefficients of friction and the wear-scar diameters were carried out using an experimental design methodology.

2. Materials and Methods

In this study, an SAE 0W20-grade reference oil (Shell Helix Ultra, Shell, London, UK) and an SAE 0W20-grade used oil contaminated with E20 (20% ethanol and 80% petrol) were investigated to determine the friction and wear characteristics of ethanol blended fuel. The used oil was taken from the engine dynamometer test of a 3-cylinder turbocharged spark ignition engine after 10 h running (~1000 km) to represent oil dilution with E20 under real conditions. Due to the low running time, the used oil had similar properties as the reference oil—except the fuel contamination—which allowed the observation of the effect of ethanol on lubricating oil.

2.1. Design of Experiments (DoE)

Throughout this research, a central composite circumscribed experimental design methodology (Figure 1) was used to define the tribometer test parameters. This method enabled the investigation of the friction characteristics of the oils and the development of a predictive model that can determine the expected minimal value of coefficient of friction (CoF). Temperature and load were selected as variables for the tribometer experiments, with central values (80 °C and 200 N) based on ISO 19291 [10] and practical experience. Considering that the experiments were not only run with uncontaminated fresh oil, a load of 200 N was applied instead of the 300 N required by the standard. Three measurements were carried out at the center point to verify reproducibility for each oil. One additional measurement was performed in the other eight points, as shown in the Figure 1, to gather information on the behavior of the system.

2.2. Friction and Wear Test

Also, for the determination of the measurement parameters, the friction and wear experiments were carried out using an Optimol SRV5 tribometer (Optimol Instruments Prüftechnik GmbH, München, Deutschland) in a ball–disk arrangement. Measurement parameters are defined based on ISO 19291 [10]. The test specimens were run in an oil bath to represent the hydrodynamic lubrication condition. The coefficients of friction from the experiments were averaged for five seconds of data halfway through the measurement, since the variance of the data during one measurement was smaller than the variance of repeated measurements.

2.3. Surface/Wear Analysis

The analysis of wear scars was performed with a Keyence VHX-1000 digital optical microscope (Keyence International, Mechlin, Belgium) and VH-Z100R objective. The average of the width and length of the wear scars was used to calculate the average wear-scar diameter (AWSD), which is a representative parameter for the worn surface. In accordance with ISO 19291 [10], the measurement of the average ball wear-scar diameter offers sufficiently precise information regarding the wear characteristics of the contact surfaces.

2.4. Oil Analysis

The oils were analyzed using an Anton Paar SVM 3001 viscometer (Anton Paar GmbH, Graz, Austria) and a Bruker INVENIO-S Fourier-transform infrared spectrometer (Bruker Corporation, Billerica, MA, USA). The viscometer was used to investigate the viscosity-modifying properties of ethanol on lubricating oil. The kinematic viscosity of the oils was measured based on ASTM D7042-21 [11] at five temperatures at which the tribometer tests were run (52 °C, 60 °C, 80 °C, 100 °C, and 108 °C).

Fourier-transform infrared spectroscopy was used to compare the reference oil with the used oil contaminated with E20 fuel, as this method is not suitable for determining absolute values. The oxidation and anti-wear additive content (zinc dialkyldithiophosphate—ZDDP) of the oils were investigated according to ASTM E2412-10R [12], which determined the range of compounds/bonds as a function of wavelength. The spectrum analysis was necessary to support the assumption that 10 h of running did not cause any significant change in the used oil except for oil dilution.

3. Results and Discussion

3.1. Spectrum Analysis

The spectrum of the reference and E20-contaminated used oil showed minor differences in the range of oxidation (1800 to 1660 cm⁻¹) and anti-wear additive (1020 to 930 cm⁻¹). Figure 2a depicts the region of the FTIR spectrum that corresponds to oxidation. As this methodology is not suitable for determining absolute values, further studies were required to establish whether the oxidation and degradation of the used oil had a significant effect on the friction and wear properties. To better describe the extent of oxidation, the kinematic viscosity of the oils was also measured, which showed a slight decrease, as illustrated in Figure 2b. Correlating the two results suggests very minor oxidation, which is to be expected with low-mileage engine oils. The slight decrease is possibly the results= of fuel dilution, the amount of which could be determined either through Gas Chromatography–Mass Spectroscopy or through viscosity blending methods; however, these were outside of the scope and reach of the study.

At all temperatures tested, the used oil contaminated with E20 fuel showed a lower viscosity than the reference. The difference in viscosity seems to be reduced around 80 °C, due to the evaporation of ethanol at 78.9 °C. This suggests that while, at low temperatures, the viscosity-modifying property of ethanol is significant, at higher temperatures, the effect on the boundary layer becomes more prominent [5]. As shown by Costa et al. [5] and Lenauer et al. [8], the reduction in the anti-wear additive content is also a possible consequence of ethanol on the boundary layer. However, this study could not demonstrate the phenomenon due to lack of appropriate methodology, as FTIR is not suitable for the detection of reaction products. On the other hand, FTIR spectroscopy indicated a slight decrease in ZDDP content in the used oil (Figure 3), thus indicating a minor amount of depletion. Based on the results of the oil analysis, it can be assumed that the difference between the coefficient of friction and the average wear coefficient of the reference oil and the oil used is due to the presence of E20 fuel.

3.2. Average Wear-Scar Diameter (AWSD) Analysis

The analysis of the average wear-scar diameter showed that the ethanol-contaminated used oil caused higher wear scars due to the effect of ethanol on the ZDDP–tribofilm [5]. Since three measurements were carried out at the center point (80 °C, 200 N), the deviation of the wear-scar diameters was also investigated. Despite the fact that the ethanol increased the average wear-scar diameter, it also reduced the deviation between measurements, which was 5.01 µm (0.97%) compared to 17.1 µm (3.5%) for the reference oil, as shown in Figure 3.

3.3. Coefficient of Friction (CoF) Analysis

The coefficients of friction were used to develop the models based on the method described in Section 2.1. It can be seen in Figure 4 that the presence of E20 fuel moved the minimum coefficient of friction towards higher loads. Costa et al. also confirms that the presence of ethanol in a hydrodynamic lubrication condition reduces the value of the coefficient of friction [5]. In addition, as shown in the analysis of standard deviation of the average wear-scar diameters, the presence of E20 fuel increased the stability of the lubricating-oil properties. On the other hand, for the oil contaminated with E20, the model predicted the minimum friction coefficient at 95 °C and 300 N, where it was estimated to be 0.1008. In order to validate the model, three more measurements were run under these conditions.

The average coefficient of friction of the three measurements was 0.1118 (Figure 5), which deviated from the estimated value by 9.8%. This means that the deviation of the model built for ethanol contaminated oil is comparable to the previously reported ~1% standard deviation between measurements.

In this relatively slow system operating at a low stroke length and extreme contact pressure (~3.6 GPa), 11 measurements were used to build an oil model that provided a good approximation of the expected point of the lowest coefficient of friction and hence the corresponding temperature and load variables.

4. Conclusions

Within the scope of this study, the friction and wear parameters of a reference oil and a used oil contaminated with E20 fuel were investigated. During the oil analysis, it was assumed that the difference between the properties of the two oils was mostly affected by the presence of E20 fuel. The increased wear-scar diameters were found to be caused by a degradation in the anti-wear additive content, for which the ethanol component of the fuel was responsible, in addition to reducing the variance of the wear scars. The investigation of the models developed using the Design of Experiments method showed that ethanol moved the minimum value of the coefficient of friction towards higher loads under hydrodynamic lubrication conditions.

However, it is important to consider that the reduction in coefficient of friction values for E20 fuel might not be exclusively attributed to the influence of ethanol. Increased wear can lead to an increase in the effective contact area and an accompanying reduction in surface contact pressure, which can consequently impact the variation in the coefficient of friction. Ball-on-disc arrangements are widely used as a cost- and time-efficient method for lubricant testing but are notoriously affected by a continuously changing contact area during individual experiments. Although assessing the effect of this phenomenon on the outcomes was not the objective of this study, a further investigation will be conducted to explore the extent of the problem, as well as potential solutions.

On the other hand, the Design of Experiments-based model developed for the oil contaminated with E20 fuel was not only suitable for screening but also for predictive purposes. The ethanol has stabilized the tribosystem and has also determined a local extreme value, which, during validation, showed a deviation of 9.8% from the actual measured value.