Gasoline Lubricity

Abstract

It is concluded that the lubricity of gasoline is the least well understood of all three fuels due largely to the lack of a reliable test method for measuring the lubricity of such a very volatile and contamination-sensitive material. To overcome this limitation, the development of a simple and easy methodology based on the general standard ASTM G-133 have been produced. This method is first used to investigate the lubricity of commercial gasolines to obtain some baseline data for further study. A comparison of the overall lubricity level of diesel fuel and gasoline fuel indicates that additive-free gasolines have significantly poorer lubricity than highly-refined, Swedish Class I diesel fuel, while commercial, detergent-containing gasolines range from slightly better to significantly poorer than a Swedish Class I diesel fuel. Especially LRP (lead replacement) gasolines developed a tests on refinery streams used to blend gasoline also show quite varied wear behaviour. Gasoline lubricity can be significantly improved by adding small amount of diesel lubricity additives. The results indicate that the type of fuel is a significant factor for discriminating the lubrication properties of each type of gasoline fuel and that lubricity is affected by bulk and trace composition characteristics of the fuel.

1. Introduction

In the late 1980s and early 1990s, environmental concern about the toxic and harmful emissions from diesel and gasoline engines led to large reductions in the amounts of sulphur and the development of reformulated gasoline fuels.

The topic of gasoline lubricity has recently become more urgent with the possible introduction of direct-injection gasoline engines, which will necessitate high-pressure gasoline injection pumps, a development that is most likely to place considerably more emphasis on the lubricating ability of gasoline, accelerating wear especially in rotary distributor fuel pumps. According to pump manufacturers this loss of lubricity may be the difference between fuels from a controlled laboratory environment and a cost-conscious production environment [1].

The stringent specifications for sulphur content in gasoline (from the year 2005, Euro 4 emissions specifications have defined the limit of 50 ppm S for the countries of E.U., and from the year 2008 have been 10 ppm) and may take off some of the fuel lubricating capability. The lubricity of aviation kerosene and diesel fuel appears to arise from very small quantities of polar, quite high boiling point components. It is realised that the overcoming increase in the severity of refinement of gasoline fuels makes very difficult to analyse these components and chemically identify them, as they vary greatly depending upon the origin of the fuel.

Fuel quality in recent years became increasingly important, not only for its role in the actual performance of the vehicles, but also for its impact on the emissions. However, the fuel pump at the service stations is the point at which the actual specifications of the fuels should be ascertained. This paper presents results of a survey of gasoline samples obtained from service stations in Athens area.

In Greece, three main types of gasoline are sold in the service stations: new super or LRP gasoline with a Research Octane Number of 96 (96 RON) for the non-catalytic cars, unleaded gasoline with a Research Octane Number of 95 (95 RON) and super unleaded gasoline with a Research Octane Number of 98 (98 RON) for newer cars equipped with a catalyst. Some service stations also sell super unleaded with a Research Octane Number of 99 or 100 (99+ RON) but the market share of this product is very limited. Unleaded gasoline is the cheapest gasoline and it is marked with quinizarine, while new super and super unleaded gasoline have similar prices (and they are quinizarine free). This price differential is the main motive to mix the cheaper with the more expensive fuel. Most gasoline adulteration cases involve the illegal mixing of the cheaper unleaded into the LRP or super unleaded gasoline. Less common is the mixing of much cheaper heating fuel into the gasoline. In such cases, the sulphur content can be used as a physical marker, which characterises the fuel quality [1]. Also, the viscosity of gasolines is 10 times less than diesel fuels and it is an indicator of adulteration, too.

Gasoline lubricity is a complex phenomenon, involving many complicated and intercorrecting factors, such as the presence of water, oxygenates, diolefins, aromatics, the effect of viscosity and the synergistic effect of different wear mechanisms. The lubricity mechanism of gasoline is quite different from that of diesel fuels that leads to severe adhesive wear. With low-sulphur fuels, adhesive wear is seen instead of corrosive and mild oxidative wear, and deposits build up on top land [1].

The emissions from motor vehicles contribute about 90% of airborne lead in urban areas. So, it was committed to phase out leaded petrol to reduce ambient lead concentrations as much as possible. On the other hand, valve seat recession (VSR) occurs when there is insufficient lubrication between the exhaust valve and seat. The mechanism of valve seat wear is a mixture of two major mechanisms. Iron oxide from the combustion chamber surfaces adheres to the valve face and becomes embedded. These hard particles then embed into the valve seat and cause abrasive wear or valve recession leading to early engine failure. For this reason, there are a number of anti-wear additives on the market that protect car’s valve seats. Additives with active ingredients of either potassium, sodium, phosphorous or manganese have been shown to give protection to exhaust valve-seats. Although no additive is as effective as lead, it has been shown that correct dosing will provide adequate protection to exhaust valve-seats under normal driving conditions [1]. The new specifications in the Greek market determined as appropriate additive the potassium at the concentration level of 10–20 ppm (mg/kg). Because there is a small possibility that mixing of some anti-wear additives on the market could result in engine damage, the potassium additive was mixed from the refinery production [1].

2. Commercial gasoline lubricity evaluation

Examination of the gasoline lubricity has shown that the majority of the samples6were above the acceptance limit of diesel lubricity, the 460-μm limit (Figure 1). We cannot include the repeatability limit calculated according to Eq. (2) for diesel fuels because such an assumption is not scientific tested and experiments must be carried out for the determination of the repeatability and reproducibility limit of gasoline fuels. This means that research studies must determine the effect of temperature and humidity on gasoline lubricity for wears greater than 600 μm. Regarding the effect of the test apparatus’ modification, mentioned above, this limit must be restricted to lower values. This enhances even more the experimental observation of greater lubricity values for gasoline than that of a common diesel fuel.

$$r=139-\left(0.1688\times \mathit{WS}1.4\right),360\le \mathit{WS}1.4\le 600$$E1

Figure 1.

Lubricity values for the three gasoline types.

On the contrary, most of the samples of new super gasoline were near the limit of 460-μm indicating that the presence of the potassium additive had a main effect on the lubricating properties of fuels. Adulterated new super gasolines with unleaded gasoline have poorer lubricating properties, as shown in Figure 2. The effect of sulphur content in gasoline lubricity is depicted in Figure 3. It is obvious that unleaded and super unleaded gasolines have much higher lubricity values than LRP gasolines. Especially, below the level of 50 ppm are observed extremely high lubricity values.

Figure 2.

Gasoline lubricity values vs. water pressure (from the humidity and temperature of the experiment).

Figure 3.

LRP gasoline lubricity values and adulterated samples.

There was no linear or other type of correlation between the concentration of potassium and the lubricity, but it seems that there is a limit of demanded potassium that may maintain a significant reduction of MWSD1.4 value near the limit of 460 μm. The factors most likely to cause the observed differences in lubricity are the bulk fuel composition, the use of additives and the use of oxygenates.

3. Fuel comparison

The adulterated fuel samples were isolated and two statistical computations were carried out each time, one with these samples and the other without.

The spread of the values can be depicted using boxplots. In Figure 4 is shown the median, quartiles, and extreme values of lubricity for each type of gasoline fuel. In each box plot is displayed the 50% of samples’ population in the square area, the 75% of them within the upper and lower limit and the extreme values which are cases with values more than three box lengths from the upper or lower edge of the box. It is shown that LRP gasolines have a much better representative sample population indicating good lubricating properties compared with the other two types of gasoline. One unleaded gasoline has shown extreme good lubricity value, 279 μm, but it is mainly caused by the use of special anti-wear or other additives. For the samples, which were not identified as adulterated, a descriptive analysis has been made. The results are shown in Table 1.

Figure 4.

Box plot analysis—First statistical graphic approach to the data.

Gasoline lubricity	Total	Unleaded	LRP	Super unleaded
Mean	595.2	639.0	477.9	661.3
Standard deviation	140.8	123.6	54.6	149.5
Standard error	141.5	125.5	53.7	151.7
Mean square deviation	140.773	123.6	54.6	149.473
Variance	19,817.2	15,266.5	2986.4	22,342.0
Kurtosis	−1.162	0.813	−0.376	−1.561
Skewness	0.125	−1.143	−0.462	−0.288
Minimum	279.2	279.2	346.2	403.8
Maximum	846.0	792.6	567.2	846.0
Sum	54,761.6	21,724.9	13,860.4	19,176.3
Number	92	34	29	29
Confidence limit (95.0%)	28.766	41.531	19.890	54.401
Mean + C. L. (95%)	624.0	680.5	497.8	715.7
Mean − C. L. (95%)	566.5	597.4	458.1	606.8

Table 1.

Data of descriptive analysis for gasoline lubricity.

Because all the properties were not normally distributed for correlation analysis with Pearson correlation coefficient, were chosen the correlation coefficients of Spearman and Kendall’s tau-b to be computed. The effect of the properties on the gasoline lubricity is different for each type of gasoline fuel. The chemical structure and the related individual physical properties seemed to interconnect in their effect on lubricity in different degree for each type of fuel. The results indicate that the type of fuel is a critical factor for discriminating the lubrication properties of each type of gasoline fuel.

More specifically, the statistically significant coefficients showed that unleaded gasolines seem to have lower values of wear as sulphur and nitrogen content, saturates and viscosity increased. On the contrary, unleaded gasolines seem to have greater values as toluene, oxygen, MTBE and vapour pressure increased.

LRP gasolines seem to have lower values of wear as sulphur and nitrogen content, conductivity (no-adulterated samples), saturates and viscosity increased. On the contrary, LRP gasolines seem to have greater values as the benzene, aromatics and xylene increased.

Finally, super unleaded gasolines seem to have lower values of wear as sulphur content, nitrogen content and olefins increased. On the contrary, super unleaded gasolines seem to have greater values as toluene, xylene, water, benzene, aromatics and oxygen increased.

The results above were extracted after bivariate correlation analysis to measure how variables are correlated and the values of the correlation coefficients are shown in Table 2.

Effect on gasoline lubricity	Total	Unleaded	LRP	Super Unleaded
AROMATICS	0.355*	0.102	−0.020	0.414*
BENZENE	0.405**	0.223	0.148	0.762**
CONDUCTIVITY	−0.408**	−0.079	−0.309	−0.210
DENSITY	0.275**	−0.231	−0.183	0.346
MTBE	0.400**	0.465**	0.140	0.265
NITROGEN	−0.438**	−0.574**	−0.069	−0.427*
OLEFINS	−0.410**	0.109	0.004	−0.540**
OXYGEN	0.350**	0.475**	0.143	0.108
POTASSIUM	_	_	0.005	_
SATURATES	−0.263**	−0.434*	−0.056	0.014
SULPHUR	−0.510**	−0.505**	−0.255	−0.687**
TOLUENE	0.479**	0.643**	0.169	0.534**
VAPOUR PRESSURE	0.116	0.365*	0.151	0.021
VISCOSITY	−0.008	−0.395*	−0.441**	0.067
XYLENE	0.272**	0.057	0.050	0.068
WATER	0.467**	0.333	0.100	0.574**

Table 2.

Data of correlation analysis between lubricity and physicochemical properties.

^**

Correlation is significant at the 0.01 level (2-tailed).

^*

Correlation is significant at the 0.05 level (2-tailed).

This differentiation of the properties’ effect on lubricity reinforce the idea of the complicated wear mechanism that take place under the specific conditions of the experiments and the important role of the compositional characteristics of the fuel. Oxygen content and MTBE seems to maintain or even increase unacceptable wear diameters and a possible development of a uniform system for fuel quality monitoring, including the control of MTBE’ content, should accept a lower upper limit of acceptance for this oxygenate.

4. Viscosity and density effect

Due to no specification limit of viscosity in gasoline, was decided to test all the samples at the temperature of 15°C. During the statistical process, was espied a linear correlation between the viscosity and density (R² = 0.76). In Figure 5 is shown that correlation linearity for the total of gasolines and each fuel severally.

Figure 5.

Graphs indicating linear correlation between viscosity and density at 150°C.

This is an obvious interconnecting factor as concerns the effect of density or/and viscosity on gasoline lubricity and each gasoline type separately. Both these properties are greatly influenced from the composition of the fuel, chlorine, nitrogen, sulphur and MTBE content. Organic chlorine content is connected with dioxin emissions but is anti-wear role is unknown. It was detected using SEM on the wear surface.

That enhances the opinion that the compositional characteristics of the fuel do influence the gasoline lubricity in considerable degree.

5. Potassium content

With confidence we can say that the potassium additive for valve recession plays an important role in the boundary-forming characteristics of LRP gasolines. As long as a “minimum” is maintained, the lubricity of the fuel seems to be more acceptable than that of unleaded and super unleaded samples. It is not easily to determine this limit but as shown in Figure 6, we can expect good results even when the potassium concentrate is less than 4 ppm. The amount that is added to the fuel does not seem to affect the final result of WSD1.4 proportionally.

Figure 6.

Potassium content and gasoline lubricity.

Also, conductivity of LRP gasolines was much greater than that of unleaded and super unleaded gasolines. The main effect on that is due to the organic salt of potassium, but there is not good linear correlation between conductivity and potassium content (R² = 0.51). In Figure 7 could see the difference between the conductivity for each type of fuel.

Figure 7.

Conductivity values at 200°C for each type of gasoline fuel. Effect of the potassium concentration.

Due to its incompatibility to modern catalytic converters, we could not use potassium additives—alkyl, aryl or alkoxy potassium compounds or other—as additives for gasoline lubricity.

6. Model predicting the value of gasoline lubricity

The model described below is based on the observed values of 106 samples of automotive gasoline fuels that were collected during the years 2001, 2002 and 2003.

All the measured gasoline fuel properties were used for the development of regression statistical analysis. In particular, it was found through trial and error that we must not exclude any variable from the input data in order the methodology to produce the smallest error in the validation data and to obtain as much greater R-squared value as possible. Indeed, a 30 input–1 output network was set up using the above-mentioned variables as inputs and the lubricity as output. The predicted lubricity values by this linear regression analysis showed a standard error of 81.5 μm and the experimental values correspond to a correlation coefficient R² = 0.78. This model cannot be used as a predictor for gasoline lubricity. In Figure 8 it is obvious the difference between observed and predicted gasoline lubricity values.

Figure 8.

Observed versus predicted gasoline lubricity values after linear regression analysis.

The information obtained through univariate analysis—where the effect of each input variable on the output was examined separately—show that the values of each input variable and the direction in which it affects the fuel lubricity cannot indicate an acceptable accuracy for the degree of wear totally. More in depth analysis of the compositional constituents and their effect on lubricity will have to be done. Especially, properties such as the acid value and a GC analysis of the content of diaromatics, diolefins and other compositional characteristics would provide better results in the direction of a model predicting the gasoline lubricity.

7. Elastohydrodynamic (EHD) film formation

7.1 Introduction

It is reasonable to expect that elastohydrodynamic (EHD) film formation by gasoline may also play an important role in gasoline pump lubrication. However, up to date there has been no published work, on gasoline, or indeed, diesel fuel EHD lubrication. In this chapter, EHD traction and optical film thickness measurements of gasolines and MTBE in point contacts are described and compared with those of diesel fuel and hexadecane.

Section 7.3 describes an experimental study of EHD film formation properties of gasolines and diesel fuels. Research is directed at understanding how thin the films formed by gasoline fuels will be, and the transition from elastohydrodynamic to boundary lubrication. It was found that at high speed, the EHD behaviour of gasolines, MTBE, and diesel fuels obeys Hamrock-Dowson theory. At low speed, gasolines give very thin films and show significant boundary film formation.

To make further progress in understanding gasoline EHD behaviour, mini traction machine (MTM) tests were carried out to measure friction/traction in both fixed slide-roll ratio and variable slide-roll ratio conditions. Results are described in Section 7.4. Diesel fuels exhibited EHD behaviour similar to hexadecane and other lubricants, but gasolines gave EHD Stribeck curves and traction curves significantly different from those of lubricants.

In Section 7.5 the EHD behaviour of gasolines and diesel fuels is further examined by cross-plotting test results as log (traction coefficient) versus log (film thickness). The lubrication regimes of gasolines and diesel fuels involved in current study can then be discussed in terms of lambda ratio. Results suggest that the Stribeck curves of gasolines and MTBE obtained in this study are incomplete. Due to their extremely low viscosity and the rolling speed limitation of the current MTM test device, it was not possible to enter the full fluid film regime.

Section 7.6 summarises the main conclusions drawn from these results.

7.2 Test fuels

The properties of test fuels are listed in Table 3. These are a subset of the fuels studied in HFRR work previous HFRR research.

Test fuels	Properties
High Lub Diesel	Vis. = 3.75 mm^2/s at 25°C
Low Lub Diesel	Shell Class 1, Vis. = 2.82 mm^2/s at 25°C
Gasoline D	Commercial UK gasoline, 140 ppm S, 33% aromatics, 9% olefins; Vis. = 0.50 mm^2/s at 25°C.
Gasoline E	Finland City Gasoline, 27 ppm S, 26% aromatics, 11% olefins, 11% oxygenate, Vis. = 0.55 mm2/s at 25°C.
MTBE	purity 99.8%. Vis. = 0.46 mm^2/s at 25°C.

Table 3.

The composition and properties of gasolines, MTBE and diesel fuels used in this study.

7.3 The film formation of diesel and gasoline fuels

7.3.1 Ultra-thin film interferometry (UTFI)

The ultra-thin film interferometry technique provides a method of measuring the thickness of very thin lubricating films in rolling contact between a glass flat and a steel ball. In co-operation with traction measurements, film thickness measurements can provide valuable information concerning the rheological and friction behaviour of fuels at high pressures and high shear rates. A major limitation of conventional optical interferometry is that it cannot be used to measure films less than approximately one quarter the wavelength of the visible light used (approximately 75 nm), since this corresponds to the first destructive interference fringe of the shortest wavelength visible light. An ultra-thin EHD film thickness measurement technique was therefore developed by Johnson, Wayte and Spikes in the early 1990s to overcome this limitation [2].

This method used a transparent SiO₂ or Al₂O₃ “spacer layer” coating, typically 430 nm thick, applied on the top of a semi-reflecting film on a transparent glass or sapphire flat. This coating enables interference fringes to be obtained even in the absence of an oil film. A schematic representation of EHD test device used in this study is shown in Figure 9.

Figure 9.

Schematic representation of EHD film thickness test device.

A 19.05 mm in diameter steel ball is loaded against a rotating glass disk. The glass disc is driven by a motor and the disc drives the steel ball in nominally pure rolling. The rotating speed of the glass disc can be continually adjusted down to 0.0002 m/s to allow the measurement of very thin films formed by fuels.

The underside surface of the glass disc is coated with a very thin, sputtered layer of chromium. A transparent silica layer, of thickness greater than half of the wavelength of visible light, is deposited on the top of the semi-reflecting chromium layer. White light is shone through the glass disc into the contact between the glass disc and the steel ball. Some light is reflected from the semi-reflecting chromium layer and some light passes through the fuel film and is reflected off the steel ball. Since the intensity of the two reflected beams is similar, constructive or destructive interference produces an interference pattern based on the thickness of fuel film. A spectrometer is used to determine the wavelength of maximum constructive interference. A menu-driven computer program is employed for image grabbing and analysis [3].

As shown in Figure 9, the test fuel is enclosed in a chamber to reduce evaporation during the test. All tests in the current study were carried out at 25 ± 0.5°C. The load applied was 20 N, corresponding to a maximum Hertz pressure of 0.48 GPa in the contact and the composite roughness of the undeformed surfaces was 11 nm.

7.3.2 Hamrock-Dowson Equation

A number of equations have been developed for predicting EHD film thickness of lubricants. The most widely-used are the formulae proposed by Hamrock and Dowson in the 1970s (186), which can be employed with confidence for many material combinations including steel on steel up to maximum pressure of 3–4 GPa [4]. They were developed by regression-fitting numerical solutions of the EHD contact problem over a range of loads, speeds, geometries and materials. The Hamrock-Dowson equation for central film thickness hc is expressed by

$$\mathit{hc}/R_x=K\cdot U^{0.67}{\mathrm{G}}^{0.53}{\mathrm{W}}^{-0.067}$$E2

where

U = (uη₀/E′R′)—the non-dimensional speed parameter.

G = (αE′)—the non-dimensional material parameter.

W = (w/E′R′²)—the non-dimensional load parameter.

K = 2.69 (1–0.61e − 0.73k).

k = a/b—ellipticity parameter.

K ≈ 1.0339 (R′_x/R′_y)^0.636.

R_x, R_y reduced radii of curvature in the ‘x’ and ‘y’ directions, respectively.

η₀—the viscosity at atmospheric pressure of the lubricant (Pas).

E′–the reduced Young’s modulus (Pa), 2/E′ = (1 + ν₁²)/E₁ + (1 − ν₁²)/E₂.

u—the entraining surface speed (m/s) = (u₁ + u₂)/2.

R_x—the reduced radius of curvature in the entrainment direction (m), 1/R_x = 1/R_1x + 1/R_2x.

α—the pressure-viscosity coefficient (m²/N).

w—the contact load (N).

For a fixed lubricant, load and contact geometry, EHD film thickness is thus proportional to (U) 0.67, so that a log (film thickness) versus log (speed) plot for a given fluid at a given temperature should yield a straight line of gradient 0.67.

7.3.3 Film thickness test results

Figure 10 shows the variation of film thickness with rolling speed for five fluids, two diesels, two gasolines and MTBE. The following main features can be seen (Figure 11).

1. At a given speed, diesel fuels form significantly thicker films than those of gasolines.

2. For low lubricity diesel fuel (Shell Class 1), the linearity of log (film thickness) versus log (rolling speed) persists down to about 5 nm at a speed of 0.08 m/s, below which a deviation from linear behaviour is observed. The high lubricity diesel fuel obeys the EHD theory only down to about 17 nm. Below this, the film is thicker than predicted, probably due to boundary film formation. This can be explained by the fact that Shell Class 1 is much purer than high lubricity diesel. and impurities in high lubricity diesel fuel contribute to film thickness at low speed.

3. In the case of gasolines, the film formation behaviour of the two fluids are also quite different. Gasoline E, an oxygenated, highly refined gasoline (27 ppm sulphur) obeys EHD film theory down to a very low film thickness of 3 nm. For gasoline D, a commercial UK gasoline (140 ppm sulphur), the linearity only persists down to about 7 nm at a speed of about 0.3 m/s. Below this, a thicker than predicted film is formed.

4. MTBE has the lowest viscosity and form the thinnest film. It continues to obey EHD theory down to 1 nm.

Figure 10.

Film thickness of gasolines and diesel fuels.

Figure 11.

Film thickness of purified hexadecane compared with diesel fuels.

7.3.4 Discussion of film thickness results

7.3.4.1 Thick film behaviour

At high speeds, all fluids appeared to give behaviour consistent with EHD theory. Table 4 lists the film thicknesses of all fuels at 2 m/s together with their viscosities at 25°C.

Test fluids or fuels	Viscosity (mm^2/s)	Film thickness at 2 m/s (nm)
high lubricity diesel fuel	3.75	54
low lubricity diesel fuel	2.82	42
Gasoline D	0.50	15
Gasoline E	0.55	12
MTBE	0.46	6.4
Hexadecane	4.02	32*

Table 4.

The EHD film thickness of test fluids and fuels at 2 m/s and their viscosity at 25°C. * by extrapolation.

Table 4 indicates the film formation of hexadecane and Figure 12 compares the film thickness of hexadecane taken from reference [7] with the two diesel fuels. It can be seen that the latter give similar film thicknesses comparable to hexadecane. It is possible to use these film thickness results to obtain approximate value of pressure-viscosity coefficient, or “α-value” of gasolines and diesel fuels [8]. Assuming that EHD behaviour is occurring, from the Hamrock and Dowson equation.

$$h=k{\left(\mathit{U\eta }\right)}^{0.67}{\alpha }^{0.53}$$E3

Figure 12.

The α-values versus temperature for some hydrocarbons (completed from references [5, 6]).

By comparing two fluids at the same speed, load, solid geometry and materials.

$$h_1/h_2={\left({\eta }_1/{\eta }_2\right)}^{0.67}{\left({\alpha }_1/{\alpha }_2\right)}^{0.53}$$E4

The film thickness, viscosity, and α-value of mineral oil were used to determine the α-value of gasolines and diesel fuels in this experiment. The results are shown in Table 5.

Test fluids or fuels	Viscosity (mm^2/s) at 25°C	α value (1/GPa)
high lubricity diesel fuel	3.75	10.1
low lubricity diesel fuel	2.82	8.0
Gasoline D	0.50	7.1
Gasoline E	0.55	6.3
MTBE	0.46	2.5
Hexadecane	4.02	11.6

Table 5.

The effective pressure-viscosity coefficient of gasolines and diesel fuels.

The viscosity-pressure dependence of several hydrocarbons boiling in the gasoline-diesel fuel range has been studied previously. Bispo use a vibrating-wire viscometer to measure the viscosity of toluene and some C5–C10 normal alkanes in the pressure range of 0.1–300 MPa and temperature range 30–70°C. Ducoulombier et al. determined the viscosities of some alkyl benzenes and C10–C18 normal alkanes using a falling body viscometer. Table 5 shows α-values of hydrocarbons in the gasoline-diesel boiling range which are derived from these data using the Barus isothermal viscosity pressure equation:

$$\alpha =\left[ln\left(\eta /{\eta }_0\right)\right]/p$$E5

where p is the pressure.

η₀ is dynamic viscosity at p = 0 and at a constant temperature.

The effect of temperature on the α-values of these components is shown in Figure 12.

Figure 12 shows how alpha value varies with temperature. It can be seen that for gasoline model compounds (C5–C10), the curves are almost flat, i.e. α-values are independent on temperature in range of 30–70°C. In contrast, diesel model compounds (C12 ∼ C18) exhibit an obviously higher temperature gradient. Toluene gives the lowest α-value as compared with other hydrocarbons.

As shown in Tables 4 and 6, the values of pressure viscosity coefficient for gasolines (6.3–7.1 GPa⁻¹) are comparable to those for C7–C10 aromatics (6.2–8.4 GPa⁻¹), and C5–C10 normal alkanes (7.6–8.3 GPa⁻¹) if one bears in mind the difficulties associated with the EHD film determination of such volatile and extremely low viscosity liquids. MTBE gives a particularly low α-value of 2.5 GPa⁻¹. For diesel fuels similar phenomena are observed. The α-values of diesel fuels (8.0–10.1 GPa⁻¹) are lower but comparable to that for C10–C18 normal hydrocarbons (8.3–14.5 GPa⁻¹). For gasolines the α-values are about 1 GPa⁻¹ lower than the corresponding average α-values of their main hydrocarbon components (C5–C10 hydrocarbons) measured using viscometers. In case of diesel fuels, the α-values of C10–C18 hydrocarbons is 8.3–14.5 GPa⁻¹ and diesel fuel α-values are in the range 8–10.1 GPa⁻¹ and thus the difference is about 2 GPa⁻¹. In general, the α-values of gasoline and diesel fuel obtained using film thickness results are about 20% lower than corresponding average α-values of their main components, measured using conventional high pressure viscometry. This is consistent with previous findings (187) [9]. Baskerville and Moore have indicated that estimated α-values obtained using film thickness results were about 26% lower than those determined by conventional viscometry, irrespective of fluid type or temperature [9]. This was explained by the response of lubricants to the severe conditions of shear or the systematic error in determining film thickness (187) [9].

Compound	α-Values at 25°C (1/GPa)
Toluene	6.2
Butylbenzene*	8.4
Pentane	7.9
Hexane	7.6
Heptane	7.7
Octane	8.1
Decane	8.3
Dodecane	10.6
Tetradecane*	10.3
Hexadecane*	12.3
Octadecane*	14.5

Table 6.

The α-values of some gasoline components and diesel components, derived from references [5, 6].

Note: * obtained from reference [6], others from reference [5].

7.3.4.2 Discussion of boundary film behaviour

It is clear from the results in Figure 10 that the diesel fuels and one of the gasolines gave pronounced boundary film behaviour.

The boundary film formation properties of some hydrocarbons and synthetic fluids in rolling concentrated contacts have been investigated by Gao and Spikes [10]. It was found that a deviation from EHD theory can occur at between 1 and 10 nm for some base fluids. The enhancement of film thickness was ascribed to the formation of high viscosity layers, a few monolayers thick, on solid surfaces. This effect becomes significant when the film thickness generated is comparable to the thickness of the viscous layers, typically 1–10 nm. No measurable boundary film were found with hexadecane or other pure hydrocarbons.

The boundary film effect was further investigated by Gao et al. using binary mixtures of synthetic lubricants with different polarity and viscosity [11]. As shown in Figure 13, addition of 10% wt. of high viscosity ester into low viscosity polyalphaolefin produces a positive divergence from theoretical value, to approach the pure ester behaviour in very thin film region. When 10% wt. of low viscosity ester is added to the high viscosity polyalphaolefin, the trend is reversed and negative divergence is observed (Figure 14).

Figure 13.

The effect of addition of 10% wt. high viscosity ester in low viscosity polyalphaolefin [11].

Figure 14.

The effect of addition of 10% wt. low viscosity ester in high viscosity polyalphaolefin [11].

This phenomenon was interpreted in terms of the fractionation of base fluids due to Van der Waals forces very close to polar solid surfaces [12]. The thickness of EHD films formed in the very thin film region (<10 nm) is then controlled by the viscosity of the more polar component rather than the viscosity of the blend [12].

In the current research, a similar departure from EHD theoretical line was observed with most fuels in the 5–20 nm range: however only positive divergence was observed. Very pure fuels, such as gasoline E and MTBE, exhibited less deviation in this region (Figure 10). This phenomenon may be explained based on the boundary film formation of polar constituents in fuels. Two different effects are possible.

1. The adsorption of naturally-occurred polar impurities and gasoline oxidation products

   As described, non-additised modern gasolines generally contain about a few tens of ppm wt. of sulphur (thiophenes, benzothiophenes and sulphides), about 10 ppm of nitrogen (pyridines and isoquinolines), and trace amounts of oxygen-containing impurities (alkylphenols, alcohols and alkoxyalcohols). Although these polar impurities are only present in a very small quantity in gasolines, they are polar and, in general, would be more viscous than gasoline hydrocarbon constituents. For example, phenol has a viscosity of about 6 cP at 25°C, about 10 times that of gasoline. The oxidation of unstable components, mainly olefins and diolefins, in long-term storage will also increase the polar impurity content in gasoline. These polar impurities may play a part in thin film gasoline EHD lubrication, by surface adsorption.

2. The polymerisation of diolefin components in gasolines

   The two gasolines had been stored in a tightly-closed vessel 1 year before tests. Both gasolines probably contained about 0.5% wt. diolefin which has very strong polymerisation tendency (Table 4). It is possible that a polymer film was formed by a chain reaction mechanism. Polymers in solution have been shown to form enhanced EHD film formation at low speeds [13]. As a polymer, detergent can have a considerable influence on the viscosity of gasoline fuels and thus on the consequent EHD film-forming properties.

Gasoline E exhibit less divergence probably related to the presence of MTBE. The free radicals produced during oxidation of MTBE (as an ether MTBE is more easily oxidised to peroxide than gasoline hydrocarbons) will interfere with gasoline autoxidation and gum forming tendency.

It is reasonable to expect that elastohydrodynamic (EHD) film formation by gasoline may also play an important role.

Conflict of interest

The author declare no conflict of interest on the information delivered in this chapter.