**1. Introduction**

Electric vehicles (EVs) are emerging as the future of transportation as the market shifts from internal combustion engine vehicles (ICEVs) to electrification [1,2]. Although EVs are more energy efficient than ICEVs, energy losses in electric motors (EMs) are still considerable [3,4]. Mechanical losses in EMs mainly derive from friction in bearings [3,5,6]. Further, about 40–60% of early EM failures are said to be premature bearing faults [4,7], with most failures being due to improper lubrication [7]. Grease is used in 80–90% of rolling bearings [3,8], and, consequently, failed grease lubrication is the predominant cause of EM bearing failure [9–11]. Another source of premature bearing failure that adversely affects EM life is exposure to electrical environments, like those found in EVs [4]. Formulating greases for EV applications is a particular challenge because of key differences between the environment and operating conditions in EVs and those experienced by greases in traditional ICEVs, particularly, speed, temperature, and materials. These conditions are also experienced by greases in industrial EMs.

First, EMs in industrial applications and EVs are operated at high speeds. Grease lubrication is very dependent on speed and can exhibit inverse Stribeck behavior where friction is low at low speeds [12,13]. This deviation from the behavior of lubricating oils is most significant at low *λ* ratios, i.e., small film thickness to effective surface roughness ratios [12]. In addition, at low speeds or nominal boundary conditions, friction is determined by grease thickener alone and is lower than that predicted for base oil [12,14,15]. Grease film thickness is larger than calculated for a base oil at low speeds, and the same or smaller than calculated at higher speeds [5,16–21]. Therefore, the wide range of speeds expected for EMs introduces additional challenges when selecting or designing greases for EV or industrial applications.

High temperature is another challenge for grease lubrication in EMs. Although some increase in temperature can be beneficial because it helps grease bleed and, thus, resupplies lubricant to the bearing contact track [16,22], high temperatures can also generate harsh operating environments. High rotor speeds generate heat [3], and EMs can reach operating temperatures of 150 ◦C [23] or even 180 ◦C [24] for some applications. Therefore, greases used in EM bearings can experience thermal degradation in the form of oxidation and decreased lubricating capabilities as a result of EM operating environments [9–11]. More specifically, grease can suffer thermo-oxidation degradation as a result of high temperature during bearing operation [9], and, at temperatures >120 ◦C, oxidation ages grease which affects its lubricity and decreases grease life [8,25]. Note that aging will affect the rheological properties of the grease differently depending on its formulation [14]. The result of high temperature and high speed is grease degradation, which affects the chemical composition of the grease, as well as its physical properties that can adversely affect film formation, leading to ineffective lubrication [9].

Another challenge for grease lubrication of EMs is that the materials of bearings may differ from those in traditional motors. Specifically, many EM bearings have ceramic components that act as insulators to mitigate issues related to stray current. Stray current can damage both the bearing and the grease, as well as generate heat, which can cause localized melting of metal surfaces, cause pitting, break particles loose, and embrittle materials [26,27]. In addition, grease that is electrically conductive can amplify these effects and accelerate bearing damage [27]. Current discharge also causes grease degradation by thermal-oxidation and evaporation of the base oil and additives, which then makes grease rigid [26]. The adverse effects of stray current on bearing failure will be more prevalent as EVs become a larger portion of the transportation sector [4]. Therefore, it is necessary to better understand how the electric environment influences EM bearing/grease systems [4] and develop tribological knowledge of non-traditional bearing materials operating in EM environments.

The most common ceramic material used for such applications is silicon nitride. Silicon nitride is suitable for bearings due to its mechanical properties across wide temperature ranges, electrical insulation, thermal shock resistance, excellent fracture toughness, wear resistance, long life, and reliable low maintenance operation [28,29]. Silicon nitride serves as a bearing insulator [26] that disrupts stray current in EMs and, thus, minimizes grease thermal degradation and melting of material that leads to wear. Often, silicon nitride is used in hybrid bearings that consist of ceramic rolling elements and traditional steel raceways [26]. Hybrid bearings have been found to last longer than predicted based on the Lundberg-Palmgren theory [30], and grease life with hybrid bearings was found to be up to four times longer than with traditional all steel bearings [26]. However, there are issues associated with the use of ceramic bearing elements, particularly related to lubricant additives. For example, phosphorus-based additives were found to not react with silicon nitride as they would with steel so the tribofilms formed were not effective in improving silicon nitride bearing life [31–33]. Another potential issue is that hybrid bearings experience higher contact stress than all steel bearings under the same applied load [30], due to the higher hardness of ceramics, which corresponds to smaller contact areas.

The aforementioned challenges with grease lubrication in EMs can be partially addressed through design or selection of greases specifically for EM environments. Thickener type, base oil type, and viscosity all have a significant impact on film thickness and friction for grease lubricated rolling/sliding contacts [12–14,34]. Greases are continually changing, as formulators and designers seek to optimize lubrication in different operating environments, while remaining compatible with component materials. For example, recent studies have used nanotechnology to create novel additives for grease formulations that improve lubricity and grease life [3,35–37]. Another study focused on extending EM bearing life by reducing grease degradation and found this can be achieved with the use of an antioxidant and high-temperature composite grease formulation [9]. Continued research in the area of grease formulation is crucial because grease behavior is extremely application dependent [22].

Grease optimization often focuses on identifying the best combination of base oil and thickener for EM applications. Studies have evaluated the lubrication mechanisms associated with synthetic or mineral base oil with urea or lithium thickener. For instance, an ester-polyurea grease used for EMs with silicon nitride ceramic rolling elements was found to have excellent life, resist high operating temperatures, and withstand high speeds experienced in the motor [26]. Synthetic base oils resist higher temperatures, while generating low friction and improving service life [3,38]. Although lithium thickened greases are currently the most widely used [8], some studies sugges<sup>t</sup> urea thickened grease may generate lower friction and thicker films, as well as have a closer correlation to typical Stribeck behavior than lithium thickened greases [12]. A study comparing custom polyurea and lithium thickened greases on bearing steel was performed to characterize performance at 25, 70, and 120 ◦C and average surface roughness of 10, 100, and 200 nm [12]. Polyurea greases were shown to have the lowest friction at low speed, average surface roughness of 100 nm Ra, and temperatures of 70 and 120 ◦C. Further, it was reported that polyurea had thicker low-speed films than lithium greases [12].

Based on the current findings, it is evident that both base oil and thickener affect grease performance and that optimizing this performance for EM applications requires characterization at the conditions in which the motor will operate. Specifically, EM greases are subject to higher temperatures and may be required to function with different bearing materials than traditional applications. Here, we tested the tribological performance of four commercially available greases with formulations/additives designed for EM applications, with different combinations of mineral or synthetic base oil with lithium (complex) or urea thickeners. The study focused on lubrication of silicon nitride sliding on steel across a range of temperature and surface roughness conditions. The tribological performance of these greases and bearing materials was quantified in terms of friction and wear. Characterization included both ball-on-disk and 4-ball tests, as well as an analysis of the results in terms of lubrication regimes. Finally, the four greases were evaluated based on a ranking system that emphasized priorities for EM applications.

#### **2. Materials and Methods**

Four commercially available greases were studied, all of which were designed for EM applications, per manufacture specifications. All EM greases had an International Standard Organization Viscosity Grade (ISO VG) of 100 and a National Lubricating Grease Institute (NLGI) grade of 2, but with different combinations of thickener and base oil types. The specific greases studied were: synthetic-polyurea (SP), mineral-polyurea (MP), mineral-lithium (ML), and synthetic-lithium complex (SL). Note that, since the tested greases were commercially available, formulation details, such as additive composition and concentration, were not known. Table 1 provides the reported information for each grease.



Two types of experiments were conducted to characterize the tribological performance of the EM greases. First, a Rtec Instruments Multi-Function Tribometer equipped with a temperature chamber was used to perform unidirectional sliding ball-on-disk tests, illustrated in Figure 1a. In those tests, a silicon nitride ceramic bearing ball with a 9.525 mm diameter and an average surface roughness (Ra) of 20 nm were used. The ceramic balls

met grade 5 quality specifications. The flat disk had a 50.8 mm diameter and was made of hardened 52100 steel. An Allied High Tech Metprep 3 polisher was used to polish the disks with the use of a silicon carbide abrasive pad in water suspension for non-directional surface finish. The disks were polished to achieve a final average surface roughness of 10, 35, 60, 120, or 200 nm. Based on the ball and disk roughness, average composite roughness cases evaluated were 22, 40, 63, 122, and 201 nm. The surface roughness of the disks was measured using a Bruker DektakXT contact mode profilometer. Prior to testing, all testing surfaces were ultrasonically cleaned in heptane.

**Figure 1.** Greases were characterized using two test configurations: (**a**) Ball-on-disk and (**b**) 4-ball. The tests were performed with various combinations of steel (S) and silicon nitride ceramic (N) samples.

Ball-on-disk test parameters closely adhered to ASTM D5707-16 with some modifications to capture key conditions expected in EM environments, specifically, bearing material, temperature and surface roughness. The load was 10 N, corresponding to a maximum Hertz contact pressure of 1.2 GPa, and the sliding speed was 250 mm/s. Temperatures tested were 40, 100, and 150 ◦C. For each test, about 500 mm<sup>3</sup> (pea size amount) of grease was applied to lubricate the samples. A grease scoop was employed for tests at 40 ◦C to avoid starvation by continuously pushing the grease back onto the track [12,13,17,18]. All tests were run to 400 m total sliding distance, and each test condition was repeated three times.

Second, a Falex Multi-Specimen Test Machine was used to perform 4-ball testing, illustrated in Figure 1b. The 4-ball tests enabled different combinations of silicon nitride and hardened 52100 steel to be evaluated. Three different material configurations were tested:


The SS<sup>3</sup> case resembled a traditional all steel bearing assembly, and the NS<sup>3</sup> case resembled a hybrid bearing assembly. The SN<sup>3</sup> resembled an inverted hybrid bearing assembly, meaning that the material typically used for the races was used as the rolling element and vice versa. All-ceramic bearings are infrequently used for standard applications, so the NN<sup>3</sup> configuration was not tested here. Test parameters followed the ASTM-D2266 standard. The load was 392 ± 2 N, which corresponds to a maximum Hertz contact pressure of 4.6 GPa for the SS<sup>3</sup> configuration and 5.2 GPa for the NS3/SN3 configurations. The silicon nitride ceramic test balls met grade 5 quality specifications and had a surface roughness of 20 nm Ra. The steel balls met grade 10 specifications and had 25 nm Ra. The speed was 1200 ± 60 revolutions per minute for 60 minutes, and the temperature was held at 75 ◦C ± 2 ◦C. All four greases were tested using each material configuration; the SS<sup>3</sup> and NS<sup>3</sup> results are averages of three tests, and the SN<sup>3</sup> results are averages of two tests.

Error was calculated as the difference between the maximum and minimum wear for each configuration and grease.

Images captured of the worn ball surfaces were obtained using a Leica Optical Microscope (Model DM 2500M) for both test methods. For the ball-on-disk tests, wear volume was calculated per ASTM G-133-05. Specific wear rate was then calculated by dividing the volume by the load and total sliding distance. For the 4-ball test, wear scar diameters were measured per ASTM D2266-01. Those measurements were then used to calculate the wear area of the scar.
