Rheological and Electro-Pitting Performance of Electric Vehicle Motor Greases With Various Nanoparticle Greases

While the term Electric Vehicles (EVs) has historically implied battery-powered passenger vehicles, modern EVs take various shapes and sizes. From plug-in hybrid EVs to monorails powered by a continuous power supply, EVs have emerged as a popular and economically viable alternative to combustion technologies from different political and cultural factors. With an increased demand for these advanced and efficient technologies, the automotive industry has invested billions to expand EV manufacturing capabilities and improve EV performance and efficiency with new drivetrain technologies [1,2]. With these cutting-edge improvements, new tribological challenges emerge. Compared to combustion engines, electric drivetrains introduce higher speeds (15,000+ RPM) and, increased temperatures (150 °C) [3–5]. Combustion technologies have been improving for decades, and lubricants have followed suit. New lubricants are required to support the sustainability and performance expectations of EVs. Previous work has suggested that nearly half of all EV and HEV drivetrain failures have been bearing related [1,2]. Due to this and other failure mechanisms, EVs are shown to be less reliable than combustion-powered vehicles which have been improved upon for decades [3]. This may be due to electrically induced bearing damage.

Most gears and rolling element bearings are grease-lubricated due to favorable load support and resistance to leakage [6,7]. Grease is often used in bearings and other components where lubricant can be circulated or replenished, such as wheel bearings and electric motor bearings. The advantage is that it can stay in place due to its solid-like properties. While lubricating oils are comprised of a mineral or synthetic base oil and a carefully selected additive package, greases have an added thickening component called a thickener, which allows the grease to remain in place and form different lubricant channels depending on loading conditions [8,9]. Soap, polyurea, clay, and silica-based thickeners are popular thickener options, offering unique properties and capabilities [6]. Greases may degrade more rapidly than lubricating oils alone due to changes in loses in thickener consistency, over time. However, some works show that the oil begins oxidizing prior to the thickener [4]. This work also uses polyurea grease, known for having a longer life than most other greases [5,6]. Likewise, polyurea greases are usually considered the best grease for high-temperature and demanding applications. In electrified applications, mechanical components are susceptible to morphological damage caused by stray voltage and bearing currents discharging across lubricated contacts. Grease chemistry and molecular structure can significantly affect material conductivity and the ability to prevent or mitigate electrical-related damage [10–12]. Tailoring lubricant properties with different additives, such as conductive nanoparticles (NPs), presents an opportunity to mitigate bearing damage in EV motors without mechanically altering systems or introducing additional mechanical complexity.

In electrified systems, a capacitance charge can build when an insulating medium, like a dielectric lubricant, separates two conductive surface components. Discharged current inevitably passes through the path of least resistance, potentially through shaft bearings or couplings. The damage could also occur in gears that are lubricated with liquid oil rather than grease. Most research agrees that extremely thick and thin films effectively mitigate arcing, but each has drawbacks. Thus, the optimal lubricant conductivity should ensure the film is thick enough to prevent unwanted current leakage yet thin enough to avoid excessive charging and discharging [13]. The resulting morphological effects from this discharge, often called pitting, leave crater-like damage along the highest points of the surface [14]. Figure 1 depicts electrically induced pitting resulting from electrified rolling ball-on-disk tests. In addition to causing premature bearing failure, electrical damage will influence friction and the efficiency of rolling element bearings due to surface roughening. This roughening could reduce the maximum travel distance for one battery charge and emit noise that can be considered unpleasant to vehicle passengers and pedestrians. Understanding key mechanical and electrical conditions is essential in determining solutions for problematic currents.

Lubricants enriched with NPs exhibit favorable friction and wear performance qualities [15–34]. These nanolubricant solutions typically contain suspended material concentrations ranging from 0.04 to 5 wt%. NPs come in various shapes and sizes, with smaller sizes preferred for even distribution within the contact area [35]. Common NP geometries include fullerenes, sheet-like layers, nanotubes, or spheres [35–37].

The performance advantages of NPs stem from several proposed mechanisms, including rolling (nano-ball bearing effect) [10,17], mending [15], tribofilm formation [32,33], contact area reduction [38], polishing [27,34], and viscosity/thermal property enhancement. Unlike conventional friction and wear-reducing additives like ZDDP, which deposit a chemically bound, non-conductive tribofilm, the most appealing property of nanolubricants is their improved electrical conductivity without relying explicitly on film formation as the working mechanism. Bond et al. [39] reduced electrical pitting in rolling ball-on-disk tests with base grease enriched with silver (Ag) NPs. This work will expand upon these findings, offering results of similar greases tested under different electrical and mechanical conditions.

Four different bulk polyurea greases are tested in this work. These greases include mineral base grease (MGBase), synthetic base grease (SYNBase), fully formulated mineral grease (FFMG), and fully formulated synthetic grease (FFSYN). Each grease is National Lubricating Grease Institute (NLGI) Grade Two and has a base oil within the ISO 100 viscosity specification. The greases were fully homogenized by the original manufacturer. When incorporating the NPs, the greases were worked using a mechanical mixer to ensure an even distribution of the particles. The finished greases were analyzed under Tunneling Electron Microscopy to ensure proper particle distribution. The greases were not worked with a traditional mill, as this process could degrade the particles and change the original consistency of the grease. The base greases are a mineral or synthetic (PAO) base oil and a polyurea thickener. The fully formulated greases are a mineral or synthetic base oil, a polyurea thickener, and a commercial additive package without extreme pressure (EP) additives. Although not as common as lithium-thickened greases, polyurea-thickened greases are selected for this work due to their improved oxidation and thermal resistance without environmental or health concerns [40]. Polyurea greases are commonly used in electric motor applications for these reasons. Other thickener types may behave differently due to differences in molecular orientation and reactiveness, and therefore, the results of this work are restricted to polyurea.

When incorporating NPs into a bulk lubricant, surface modification of the NP is required. Surface modification is a critical step when creating stable suspensions. As liquid oil is excreted from the solid grease thickener during rolling, it is important that the nanoparticles remain in a stable suspension to maximize performance and longevity. Stable dispersions are created using a long-chain, high molecular-weight alkane and a stabilizer. This combination results in a colloid (solid particle suspended in a liquid). This study employed Ag as the primary NP due to its high electrical and thermal conductivity. Additionally, previous work shows that Ag reduced friction in the boundary and EHL lubrication regimes [41].

To synthesize the Ag colloid, 0.1 M oleoyl sarcosine (OS) is added to dodecane (DO) while being stirred and heated to 150 °C. Dodecane is selected as the bulk hydrocarbon for the reaction due to its high boiling point and the fact that it remains a stable liquid at room temperature [42]. The oleoyl sarcosine acts as a stabilizer, preventing NP agglomeration and colloid precipitation. Once heated, silver neodecanoate (AgOOR) powder is quickly mixed into the solution. The temperature of the solution is increased until the thermolysis reaction is complete at 180 °C. The resulting particles are spherical and have an average diameter of approximately 5 nm [42]; 2 mL of the Ag colloid solution is mixed with 10 g of each grease sample, resulting in NP-enriched greases with 0.19 wt% of Ag. The particles were mixed into the grease using a hand mixer with a specialized mixing head. Using a burst motion, the greases were mixed until the grease was smooth and the NP colloid was evenly dispersed. Additional details regarding the colloid synthesis can be found in the work published by Darvin et al. [42].

Due to the relatively high cost of Ag, these greases may only be suitable for some applications. Likewise, an alternative, cheaper option is also selected for testing. The chosen alternative is a commercially available colloid (EMG 900 Oil-Based Ferrofluid—Ferrotec (USA) Corporation) containing magnetic iron oxide (Fe₃O₄) stabilized with an oleic acid dispersant in a hydrotreated petroleum distillate. The Fe₃O₄ NPs had a nominal particle diameter of 10 nm and were suspended at a concentration of 17.7 vol%, as outlined by the manufacturer [43]. Approximately 0.1 mL of the ferrofluid was blended into 48.8 g of each of the four greases, resulting in similar concentrations to the Ag greases (0.19 wt%).

The approximate number of NPs present in the grease sample can be calculated after considering the relative particle size, concentration in grease, and the density of Ag (1.049 × 10⁴ kg/m³) and Fe₃O₄ (5.25 × 10³ kg/m³). Assuming a 1 g grease sample, approximately 6.96 × 10¹⁶ Ag and 7.02 × 10¹⁴ Fe₃O₄ particles are present. While the wt% was kept the same, the number of particles differed due to density and particle size differences. Abbreviations used for each grease/NP combination are outlined in Table 1.

The methodology and test rig used in the present work is a modified adaptation of the tests conducted by Bond et al. [39,44]. In these tests, 1 cm reciprocating rolling ball-on-disks were performed on flat samples with bolted electrical connections. One of the advantages of the reciprocating motion is the representation of start/stop conditions commonly experienced by real-world vehicles. These scenarios represent some of the most demanding tribological scenarios as the transition to boundary lubrication is possible. Like prior reciprocating tests, a Bruker UMT-3 Tribolab is the primary equipment for this work. The machine applies a constant load in the vertical direction using a power screw loading mechanism. A load cell mounted to a leaf spring suspension accounts for irregularities and ensures load continuity during the test cycle. A slider mechanism, also driven via a power screw, allows for horizontal movement along the sample. The lower drive of the machine consists of a belt-driven spindle that rotates at user-defined speeds.

The ball-on-disk tests incorporate SAE 52100 bearing steel rolling elements and flat samples. Cylindrical test samples are used, and specialized fixtures are needed to secure the test samples. A 3D-printed polylactic acid (PLA) compression ring secures the sample to the lower drive's universal holder since the test samples are relatively thin and round with no mounting holes. The tested samples were 6.35 mm (0.25 in.) thick and have an average roughness (Ra) of 1.284 µm. To support the rolling element, a specialized fixture incorporates two rollers to support the applied load while enabling free rotation. Each element is 9.525 mm (0.375 in.) in diameter, weighs approximately 3.5 g, and has a Rockwell Hardness between HRC60 and HRC67.

When designing the machine, Bruker did not intend to apply electrical loads to the sample pieces. The bolted connections employed in prior work enabled quick and reliable conduction of electricity, but the wires prevent longer moving patterns at the risk of tangling cables. Because of this, 3D-printed parts made of non-conductive filament secure a spring-loaded rolling electrical contact against the surface of the flat sample while preventing accidental conduction or discharge through the machine. This electrical contact is held against the top of the surface without making contact anywhere else, resembling a record player arm and needle. The rolling capabilities of the electrical contact prevent abrasive wear while ensuring solid contact as the sample rotates. The modified sample holder and electrical contact holder for the Bruker UMT-3 test rig are depicted in Fig. 2. Orange labels denote electrical components. The 3D-printed components were manufactured with a relatively high infill and strategic reinforcements to ensure the pieces were able to withstand the effects of loading or non-parallel test surfaces. Additionally, the inherent movement of the spring-loaded electrical contact provides a degree-of-freedom perpendicular to the surface of the plate in the event of severe loading. When setting up the test, care was given to ensure the spring-loaded component of the contact was adequately compressed without “bottoming out.” While supervising the tests, no unforeseen electrical signals, wear patterns, or permanent deformation of the 3D-printed components were noticed.

As seen in Fig. 2, a thin layer of grease was applied to the intended contact path with a spatula for each test. The grease was applied to the surface of the test disk and scraped even with the top lip of the compression ring. The initial grease film was approximately 1 mm. Once the tests are completed, the surfaces are cleaned with acetone and methanol, and tests are run on different positions on the surface of the samples.

As detailed in the previous section, 3D-printed components were designed to accommodate a spring-loaded, rolling electrical contact. This contact, which requires a soldered connection, facilitates rolling instead of sliding across the surface. The rounded, rolling tip of the electrical contact enables nearly continuous contact as the test piece rotates. Building on the electrical components introduced in Fig. 2, a four-wire resistance circuit is established between the rolling element holder and the electrical contact. The positive lead of the DC power supply is soldered to the electrical contact, while the negative lead is bolted to the rolling element holder. A digital multimeter, connected in parallel to the power supply, measures the voltage drop across the mechanical contact. The applied DC voltage is maintained at 31 V throughout the circuit, with a 275-W heating element placed in series with the positive lead to introduce an electrical load. This setup achieves a current of approximately 0.5 ADC. The idealized electrical circuit is illustrated in Fig. 3. Key electrical parameters were calculated using this power supply and the relevant mechanical loading conditions, and the results are presented in Table 2. The electrical current is chosen based on typical leakage values observed in electric vehicles. In this work, the bearing voltage ratio (BVR) was independent of the bearing material, including hybrid and ceramic composites or coatings, and is less than 10% for commercial motors. Since typical operating voltages for popular EVs range from 300 V to 400 V, a bearing voltage of ∼30 V is expected as an extreme condition.

As mentioned in previous sections of this work, the test set-up is a spherical ball rolling against a cylindrical rotating disk. In the initial work, the rotating test was only partially rotated (reciprocated) since the electrical leads were bolted to the disk. This resulted in a short arc-shaped wear track on the disk whose length can be specified. Subsequently, the rig was modified so that the disk could fully rotate, which resulted in a circular wear track. In a previous work, tests were conducted with full rotation in one direction and also partial rotations that paused and alternated directions [14]. Testing with MGBase grease (mineral base oil + polyurea thickener) showed that the 2-cm reciprocating tracks had the most damage out of the tested moving patterns (full rotation, 1 rev/cycle, 10 rev/cycle, and 1 cm/2 cm/3 cm/4 cm reciprocating) [14]. In non-reciprocating (full rotation) tests, run at velocities ranging from 0.01 m/s to 1 m/s, the generated EHL film may be too thick to facilitate discharge in ball-on-disk testing. Another consideration was that continuous motion did not allow adequate time for charge accumulation. Instead of charging and discharging at concentrated locations, the electric current is potentially dissipated across a relatively large space as the disk rotates. Additionally, results varied significantly between the four tested reciprocating track lengths (between 1 and 4 cm long, and with maximum velocities between 12 and 25 mm/s).

Based on the experimental results with MGBase greases at different lengths, the two synthesized NP greases were tested at 2-cm track lengths and at a maximum speed of 17.6 mm/s. For 2-cm track lengths, the damage accumulated mostly at the ends of the track. The authors have recently conducted a numerical model that suggests the generated film thickness at the ends of the 2-cm tracks decreases enough to facilitate electrical discharges [45]. However, it could also be that the short pause during directional changes allows time for a charge to build at a single location and discharge. The model is not yet capable of including the effect of the nanoparticle additives.

Rather than visually inspecting SEM images, a matlab code was developed to quantify pitting areas. The code combines manual and automatic processes to remove scratches and grooves using the Image Processing Toolbox in matlab. The SEM image with the most severe damage at 2-k magnification is selected for each test, typically from the track ends. The scale bar and data label are removed, and the image is converted to binary using an optimal threshold. Pixels are grouped by size to identify and remove machining grooves and stray pixels. Manual touchups are performed after automated and semi-automated steps. The damaged area is a percentage of the total image area, as illustrated in Fig. 4. For MGBase tests, the average damage was 5.13% [14]. These rolling ball-on-disk results will serve as a baseline for the remainder of this work.

Most rolling element bearings are dominated primarily by rolling motion. However, certain geometries, such as cylindrical bearings, could be more susceptible to slipping behavior, especially in the case of thrust needle or roller bearings. In the event of slipping, electrical pitting damage is still expected to accumulate, but the pits observed in the current analysis might change with the rest of the surface due to abrasive wear associated with sliding contacts.

Based on the results from testing MGBase grease without NPs, 2 cm tracks resulted in the most severe pitting damage. Likewise, all eight NP-enriched greases are tested at 2 cm reciprocating track lengths. With the same test methodology, adding NPs reduced the amount of electrical pitting damage. The damage for each tested NP is depicted in Fig. 5. The bar represents the average damage recorded for each triplicate test, and the error bar represents the standard deviation between each test. Figure 5 shows that Ag NPs had a more positive effect on grease performance, but Fe₃O₄ alternatives still protected against electrical pitting. This trend may be attributed to the two-order-of-magnitude difference in the number of NPs between the Ag and Fe₃O₄ greases. Future tests will use the same number of individual NPs rather than equivalent weight concentrations. Performance differences may also be attributed to Ag's superior thermal and electrical conductivity. Similarly, additives in the FFSYN and FFMG greases also positively reduced electrical pitting. The relatively large standard deviation exhibited by the Syn Base grease with the Fe₃O₄ sample was surprising. However, the damage on certain plates appeared more severe than others. Because the discharge depends on the surface roughness of the two contacting surfaces, even the slightest changes in pretest surface geometry can affect the extent and location of the observed damage. Figures 6 and 7 show examples of SEM scans of surfaces with little and severe damage, respectively.

The damage values were compared to specific test parameters to understand why the NPs aided in damage reduction. The key parameter was the measured voltage across the contact. During the beginning, middle, and end of testing, the multimeter was used to measure and record 5000 data points at a frequency of 50 kHz. The voltage values for all three data sets were averaged and plotted in Fig. 8. Across all testing with NP greases, the average voltage was 0.559 V with a standard deviation of 0.238 V. While no clear correlations were drawn between the damage and voltage, damage and voltage were not linearly related. The difference in recorded voltage was likely due to variances with the bolted or rolling connections rather than with the grease itself. The nature of the spring-loaded electrical contact introduces inherent rolling or sliding. Additionally, the parallel connections for the multimeter used bolted contacts. Previous test configurations used bolted contacts to reduce transient effects. However, even disconnecting and connecting the bolted contacts can result in changes in overall electrical resistance from slight differences in the contact force, contact wear, or surface contamination.

Several mechanisms may be responsible for the performance gains experienced by NP-enriched greases. First, suspended NPs may flow throughout the contact, creating a conductive pathway for electron flow. The movement of the NPs may be affected by the magnetic and electrical fields generated by the flow of electrons. Alternatively, a highly viscous thickener layer may be deposited along the surface as the contact rolls over the grease. In that case, the NPs may become embedded in the layer, offering a solid conduction channel that improves the overall conductivity of the contact. This theory will be discussed later in this work. Additionally, the NPs may also act as a heat sink during the arcing process. Rather than all the interface power being directed toward the cathode surface, the NPs may dissipate thermal energy to the bulk lubricant, effectively lowering the energy available to melt the surface. Nanoparticles can also act as electron scavengers, becoming negatively charged by electrons in the ionization zone. This charging ultimately modifies the lubricant's electrodynamics. Similarly, the presence of additives in the FFSYN and FFMG greases also showed a positive effect on reducing electrical pitting, as reported by Bond et al. [39].

The authors believe the bulk resistivity of the grease remains largely unchanged due to the relatively small number of particles. At the maximum concentration considered in the study, there are 6.59 × 10¹⁶ iron-oxide or silver particles in each gram of grease. This seems like a large number, but the volume percentage is 3%. By the rule of mixtures, and taking the resistivity of Fe₃O₄ to be 0.3 mohm·m, the resistivity of the grease might change from 10¹² ohm·m to approximately 3 × 10¹⁰ ohm·m [46]. This is still a relatively high resistance or low conductivity material.

Considering the chemical differences between the colloids, the impact of grease consistency and rheology on pitting occurrence is particularly intriguing. Likewise, the remainder of this work will further explore the rheological effects of adding NPs to grease. Regardless of the specific mechanism(s) at work, the results show that NP-enriched greases provide enhanced protection from electrical pitting damage in 2-cm tests compared to neat bulk greases.

Traditionally, grease consistency and rheology have been assessed using the cone penetration test (ASTM D217) [8,47]. Standardized NLGI consistency grades are assigned based on the cone's penetration depth. While this test remains industry-standard, modern equipment like digital rheometers offers comprehensive grease characterization with smaller samples and precise temperature control. Recent studies have compared traditional penetrometer measurements with rheometer oscillatory amplitude sweeps. These sweeps can assess viscoelastic material properties without sample fracturing or requiring high sensitivity to plate roughness [8].

Amplitude sweeps measure storage and loss modulus at a desired angular frequency with varying strain amplitude. The storage modulus reflects stored deformation (elastic) energy, resembling solid-like behavior. Conversely, the loss modulus indicates the portion of deformation energy lost to internal friction during flow [48]. From amplitude sweep curves, crucial points like the linear viscoelastic region (LVER) limit and crossover (flow) point can be identified. The LVER limit marks the approximate end of Newtonian behavior and the onset of strain dependency. The flow point, at the intersection of the storage and loss modulus, denotes the crossover stress and the prevalence of viscous (liquid-like) behavior.

To analyze the impact of NPs on grease rheology, a rheometer was used to conduct amplitude sweeps under conditions detailed in Table 3, similar to those described by Gurt et al. [49]. A 25-mm diameter parallel sandblasted measuring device (Ra = 6.684 µm) is used at a 1-mm gap. A pea-sized amount of each grease was placed between the measuring surfaces before the sample was trimmed at 0.025 mm above the measuring gap. When the grease sample is loaded and the gap between the plates is set to the trimming position, excess grease squeezes beyond the plates' edges. Trimming involves using a spatula to remove this excess grease. This process reduces edge effects and maintains uniform sample geometry to ensure reproducibility and accuracy of testing. Once the sample is trimmed, the measuring height is set (see Fig. 9). The trim height is larger than the measuring height to avoid disturbing the sample in the measuring zone.

The test program for the rheometer calculates the flow point and LVER limit for each test. Likewise, these key points for each amplitude sweep are identified with a unique marker in Figs. 10–13 and represent the average values amongst triplicate tests. The flow points and LVER limits for each grease are plotted against each other to compare the effect of adding NP colloids to the grease. Results for the Ag greases showed significant shifts from the neat grease. Because of the relatively large amount of liquid hydrocarbon in the Ag colloid, the dodecane and stabilizer without NPs were also mixed into the grease at the same concentration. These tests aid in determining if the rheological changes are attributed to the Ag particles or the dodecane and stabilizer solution.

When looking at the LVER limits for each grease, shown in Figs. 10 and 11, adding the Ag colloid or the dodecane/stabilizer solution greatly reduced the storage modulus. The extra liquid from the Ag nanoparticle colloid solutions may be responsible, as it decreases the thickener concentration and weakens the structural integrity of the grease. Lowering the storage modulus may be advantageous for pumpability but can limit load-carrying capacity and shear stability. While significant in all cases, this effect was more noticeable with the base greases. Additionally, it is worth noting that adding larger Fe₃O₄ particles reduced the shear strain needed to deform the structure of the fully formulated grease. Adversely, adding Fe₃O₄ particles to the base grease had the opposite effect. This effect was less noticeable with the smaller Ag particles. Adding the NPs to the base grease increased the shear strain needed for grease deformation.

The flow points are also compared. Bulk greases without NP additives often reached their flow point at higher strain rates than those with NPs. This observation suggests that the colloid additives accelerated the transition to liquid-like behavior. In many cases, adding Fe₃O₄ had less effect on the flow point than the Ag. One possible reason for this trend could be the increased concentration of the Fe₃O₄ colloid. Likewise, the lower NP concentration Ag colloid greases were compared to those containing the dodecane and dodecane/stabilizer. Ultimately, the extra liquid in the Ag NP colloid played a more significant role in shifting the flow point than the presence of NPs. This trend is more prominent in the fully formulated greases. The crossover stress, or the storage modulus at the flow point, varied more for base greases. This shift could underscore the absence of essential interactions between additive packages and base oils, significantly affecting grease behavior and consistency. Flow points are compared in Figs. 12 and 13.

The crossover stress at the flow point is also significant. A comparison study conducted by Gurt et al. suggests that the crossover stress can be approximately correlated to cone penetration tests with a curve fit equation [8]. However, this methodology should be used with caution. With this correlation, equivalent cone penetration values can be estimated from amplitude sweep data, as shown in Fig. 14. This is done to provide a single quantity to directly compare between the greases. Compared to the neat grease, adding the Ag colloid had a more profound effect on the equivalent cone penetration value than the Fe₃O₄. Comparable results were evident with the dodecane and stabilizer solution, indicating that the driving factor for this shift was the increased volume of liquid hydrocarbon added via the Ag colloid. The more concentrated Fe₃O₄ solution introduced less liquid to the grease, resulting in less penetration variation. All greases were initially NLGI Grade Two, but none are within specification based on the calculated values. This discrepancy may be attributed to errors in the empirical model and grease softening with age.

With the approximate cone penetration values calculated, the results are compared to the damage from Fig. 5. This relationship is illustrated in Fig. 15, with the control group representing the values for 2 cm reciprocating tests with MGBase. The plotted results indicate that greases exhibiting the lowest cone penetration values showed the most severe damage. Minitab Statistical Software is used to determine if any statistically significant relationships exist among the data collected. Instead of triplicate averages, individual data points are utilized to increase the sample size. A linear regression was created with this data to determine an appropriate p-value. The p-value, or probability value, is used in hypothesis testing to support rejecting the null hypothesis in favor of the alternative hypothesis. The calculated p-value of 0.002 is well within the statistically significant range (p < 0.05). Additionally, 32.90% of the variation in damage can be explained by the regression model (R² = 0.3290). This value is considerably low, suggesting that while the model may be statistically significant, a larger sample size is needed to obtain a more robust and precise linear fit. While the model may not be perfect, the negative correlation in Fig. 16 suggests that damage tends to decrease with increasing cone penetration. Figure 16 also shows that the data fall within the range of this model's upper and lower 95% confidence intervals. While there appears to be a relationship between increased cone penetration and decreased damage, other mechanisms may influence this trend. Likewise, future statistical analysis is needed for expanded data sets representing more diverse lubricant samples. Again, it is important to remember that these are not true cone penetration tests and should be used with caution.

The section “Rheological Testing and Results” of this work explored the performance of Ag and Fe₃O₄ NPs at an equal weight percentage (0.19 wt%). While this is an appropriate comparison for some applications, MGBase + Fe₃O₄ grease samples were blended at higher NP concentrations to achieve the same number of NPs as the MGBase + Ag colloid. This high concentration MGBase + Fe₃O₄ grease was blended by combining 5.6 mL of Fe₃O₄ colloid into 25 g of MGBase grease. The resulting weight percentage was approximately 15.18 wt%, and the number of NPs was comparable to that of the 25-gram MGBase + Ag sample. When tested at 2-cm reciprocating tracks, the high-concentration Fe₃O₄ grease outperformed the lower-concentration Fe₃O₄ grease and was comparable to the Ag grease. The performance differences between the low and high concentrations of MGBase + Fe₃O₄ suggest that the number of particles, not just the weight percentage and particle size, plays a deciding role in grease performance under electrified conditions. Figure 17 compares the behavior of MGBase grease for each of the tested NPs and concentrations.

The rheological behavior of the high-concentration Fe₃O₄ grease was compared to that of the previously tested low-concentration greases. The resulting amplitude sweep curve is depicted in Fig. 18. From this curve, the LVER and flow points were determined and plotted in Fig. 19. As evident in Fig. 19, the high concentration of relatively large Fe₃O₄ NPs offered initial elastic support. With that, the 5.6 mL of added liquid in the 15.18 wt% Fe₃O₄ colloid significantly reduced the shear strain needed to induce flow. This grease sample contained the highest volume of NP colloid, followed by the Ag grease, low concentration Fe₃O₄, and MG Base grease. The descending order of added liquid coincides with increased flow point shear strain values. This trend reaffirms that adding additional liquid hydrocarbons to the grease solution may prematurely induce viscous behavior.

In addition to testing a higher concentration of Fe₃O₄-enriched grease, a commercially available, but proprietary, conductive grease (Fuchs NyoGel^® 758G, Nye Lubricants, Fairhaven, MA) was obtained and tested under similar conditions. This grease is synthetic ester base oil thickened with a lithium soap and its conductivity is enhanced with carbon black particles. This carbon black–enhanced lithium (CBEL) grease was specially developed for applications requiring increased electrical conductivity. According to the manufacturer, the CBEL grease has a bulk resistivity of 623 ohm/cm at 25 °C and a kinematic viscosity of 6.6 cSt at 100 °C. Unlike the previously tested grease, this sample is an NLGI Grade 3 grease. When tested under the same conditions previously used, little definitive pitting damage was evident under SEM. A representative SEM image is depicted in Fig. 20.

Amplitude sweeps were also conducted for the carbon black enhanced lithium (CBEL) grease samples, with results shown in Fig. 21. As expected, the CBEL grease has a much higher storage modulus at the LVER limit than the other tested greases Additionally, the flow point occurred at a significantly low shear strain value. However, these values are not comparable to previous results, as the CBEL grease is a different NLGI grade. Additionally, the lithium thickener is expected to behave and perform differently than the polyurea-thickened greases previously tested. For these reasons, the CBEL grease was not tested further. The preliminary results show that the tested electrically conductive greases are a promising solution for mitigating electrical pitting damage with tailored additive packages.

When studying grease rheology, it' is essential to consider the interaction between the thickener and base oil. Traditional grease EHL calculations assume the base oil viscosity applies to the bulk grease. Despite accounting for the largest volumetric percentage of the grease, using base oil properties does not account for the solid-like behavior of the thickener. The generated film is likely a complex combination of compressed thickener and base oil coverage via capillary forces. Scarlett suggests that degraded thickener makes its way into the EHL film and results in the deposition of a highly viscous thickener layer along the surface of the sample [50]. Others support the belief that sheared thickener particles enter the contact and affect film formation [40,51–53]. Zhang et al. modeled this approach to determine if a thickener layer contributes to film thickness, suggesting that thickener concentration and structure variations affect the thickener deposition rate and base oil bleed [54]. A deposited layer can affect the film thickness and the electrical conductivity of the tribological interface, potentially affecting the formation of electrically induced damage.

Additionally, the test rig used in this work neglects a continuous grease supply, possibly resulting in base oil starvation due to the squeeze film effect [55]. After running the 2-cm reciprocating tests, striated grease patterns intersecting the wear track at 45 deg were noticed. Microscope images of these formations are shown in Fig. 22. These patterns indicate the lack of grease flow back into the contact area once expelled [56]. This work underscores the theory that additional mechanisms beyond simple EHL film generation may be at play, and a combination of a thickener layer and some base oil replenishment could sustain a steady-state film smaller than expected for fully flooded conditions [56]. It is also possible that the elevated temperatures of an EV drivetrain environment, and the nanoparticles themselves, could accelerate the degradation of the grease.

With the transportation industry's ambitious electrification plans, EV manufacturers are maximizing performance from compact drivetrain packages. Advanced motor and battery systems enhance control and efficiency but increase energy leakage across mechanical components. Understanding lubricant behavior in electrified applications helps tailor performance and prolong equipment life. This research examined the impact of conductive nanoparticles (Ag and Fe₃O₄) in polyurea-thickened greases with and without additives during rotating ball-on-disk tests. matlab analysis revealed the most significant surface damage was present with 2-cm tracks. This damage could be attributed to film thickness and charge dissipation.

Blending two different NP colloids into base and fully formulated greases and testing at a 2-cm track length showed that NP-enriched greases outperformed neat grease, with Ag particles showing the highest damage reduction. Another batch of Fe₃O₄ greases was blended at a higher concentration so that the number of NPs matched the number of NPs in the Ag-enriched NP grease. The high-concentration Fe₃O₄ greases offered comparable performance gains to the Ag greases, suggesting that the number of NPs in the grease may play a more influential role than particle size or electrical/thermal properties of the bulk NP material.

The performance gains of NP-enriched greases are promising for EVs. These gains may stem from improved electrical and thermal conductivity within the lubricant film or NP embedding in the thickener layer. NPs may also trap electrons or reduce the distance between conductive elements, lowering the voltage needed to bridge the lubricated contact and potentially reducing destructive ablation. Rheometry showed that added fluid in the grease mixture decreased the flow point, indicating quicker viscous behavior dominance than neat grease. It is important to remember that particles were not added alone but were part of a larger colloid solution. This colloid affected grease composition and consistency. Despite the tradeoff between increased tribological performance and resulting grease rheology, NP-enriched greases demonstrated potential for reducing pitting in electrified applications. Note that the results of this work are confined to polyurea-thickened greases, but other greases may be explored in the future, such as the CBEL.

However, there are limitations to the usage of the proposed nanoparticle additives. This work does not consider elevated temperatures that might be observed in electric vehicle operations. It is possible that the nanoparticle additives would lose their effectiveness over a long period of operation. However, the particles have been observed to have a long shelf stability. In addition, the nanoparticles could accelerate the degradation of the thickener. Over a longer duration test, the nanoparticles could lose their effectiveness.

Future research aims to replicate complex loading scenarios in multi-element bearings and address the study's limitations, including extreme temperatures, vibrations, and environmental factors in real EV applications. The effect of changing these operational parameters was not the primary focus of this work. However, in theory, the temperature and velocity will both influence the film thickness. If the film is very thick or very thin, less damage may occur. Therefore, extremes in temperature would influence the viscosity and that combined with velocity would influence the film thickness. Varying the electrical signal will also be explored in future work and it is believed that this would influence the type and extent of damage seen. The study provides critical insights into electrical damage mechanisms, emphasizing the importance of understanding operational conditions affecting electrical damage. This understanding will lead to improved design and maintenance practices, enhancing system efficiency and longevity.

The authors would like to recognize Dr. Paul Slade for his valuable consultation on arc propagation, Shell Oils for providing the tested polyurea greases, Fuchs Lubricants Co. for providing the tested carbon black enhanced lithium grease, Bruker for donating the rolling element holder, AntonPaar for allowing us to use their rheometer, and Chad Chichester from DuPont for his support as the NLGI project liaison. Dr. Samuel Bond deserves recognition for synthesizing all nanoparticle solutions and providing chemistry support when needed. The first author was also awarded the Bob Jackson Award from the Independent Lubricant Manufacturers Association (ILMA).

• The National Lubricating Grease Institute (NLGI) Academic Outreach Research Grant.

There are no conflicts of interest.

The datasets generated and supporting the findings of this article are obtainable from the corresponding author upon reasonable request.

Ag =

silver

DO =

dodecane

DO/OS =

dodecane + oleoyl sarcosine

EHL =

elastohydrodynamic lubrication

EV =

electric vehicle

Fe₃O₄ =

iron-oxide

FFMG =

fully formulated mineral grease

FFSYN =

fully formulated synthetic grease

MGBase =

mineral base grease

NP =

nanoparticle

OS =

oleoyl sarcosine

PAO =

polyalphaolefin

PLA =

polylactic acid (3D print filament)

SEM =

scanning electron microscope

SYNBase =

synthetic base grease