Investigating the Viscosity of Motor Oil Under the Influence of Contaminants

1. The process of experimental design is just as iterative as mechanical design: my team and I came up with several versions of our experiment before we settled on a plan that was viable to carry out.

Close-up shot of the TA Instruments HR30 Rheometer used in this project

Synthetic lubricants like motor oil are vital to the function of everyday machines, from car engine blocks to heavy-duty machinery to hand-held power tools. Through wear and tear, these machines often introduce contaminants into the lubricants. For example, in the case of internal combustion engines, incomplete combustion introduces soot and surface wear introduces metal particulates. This is the motivation of this project: to quantify how contaminants affect the viscosity of motor oil.

To simulate particle contaminants, we used oven-dried silica particles. We created samples of 20 wt% and 50 wt% silica-motor oil mixtures. Next, we ran shear rate ramp tests on these samples (as well as control samples) using the TA Instruments HR30 rheometer with the sandblasted Peltier plate attachments. We ran 4 tests for each sample at room temperature and at 100 degrees Celsius each (the latter to simulate the temperatures typical of an internal combustion engine).

We began by characterizing the behavior of our control samples before studying the contaminated samples. Across our 5 runs, we found various initial responses that all converged to a steady state at higher shear rates. The oil behaves as a Newtonian fluid at higher shear rates because it's designed for high shear rate applications; shear thinning and/or shear thickening would be undesirable properties for a lubricant, and we see that once we reach application conditions the motor oil's viscosity becomes independent of the shear rate.

At lower shear rates, we see varying results: some runs show high viscosities while others have no measurements. The high viscosities can be explained by human error: underfilling the rheometer plates leads to an incorrectly measured shear force that is artificially inflated. Due to the relationship between shear force, viscosity, and shear rate, a large shear force will correspond to a larger viscosity at a given shear rate.

When the rheometer plates were appropriately filled, there were no viscosity measurements at lower shear rates. We hypothesize that this is because the motor oil inherently has such a low viscosity that it is initially unable to overcome the internal "zero" of the machine. Once we approach higher shear rates, the liquid starts resisting flow by a large enough magnitude that the rheometer can make measurements. For the rest of our analysis, we focus on the steady-state region of the viscosity curves because the properties at these shear rates are most interesting when we consider the application conditions of motor oil.

For the contaminated samples, we begin by analyzing the samples tested at room temperature. Here, we found that, on average, the 20 wt% samples showed a 17% increase in viscosity while the 50 wt% samples showed a 2.5% increase in viscosity relative to the control samples. Moreover, one of the 20 wt% samples is an outlier, providing much lower viscosity measurements than the rest. This outlier was caused by the upper rheometer plate that was still warm from a prior experiment, leading to an unexpected drop in the measured viscosity. Since only the lower plate is actively cooled and contains the thermocouple, we ran our experiment when the machine indicated that the plate was at room temperature without considering the top plate.

It was unexpected that the sample with a lower contamination content registered a higher viscosity. After further investigation, we found that this result was due to inhomogeneous mixing. When we tested our 50 wt% samples, some of the silica particles had settled to the bottom of the container. This artificially reduced the density of contaminants in our sample, leading to a lower viscosity than the 20 wt% samples.

Looking at the samples tested at 100 °C, we found that, on average, the 20 wt% samples showed a 6% increase in viscosity while the 50 wt% samples showed a 50% increase in viscosity relative to the control samples. We hypothesize that this large change in viscosity is not only due to the larger content of contaminants; silica has a thermal conductivity of ~1.3 W/mK and motor oil has a conductivity of ~0.1 W/mK. This allows the silica to absorb more heat, and where there is a greater density of particles more heat will go to the particles than the motor oil. Therefore, compared to the control sample, the 50 wt% sample does not get as warm and as a result doesn’t have as low of a viscosity. 

Viscosity vs. shear rate for the contaminated samples at 100 °C