Improved Operating Behavior of Self-Lubricating Rolling-Sliding Contacts under High Load with Oil-Impregnated Porous Sinter Material

Abstract

Resource and energy efficiency are of high importance in gearbox applications. To reduce friction and wear, an external lubricant supply like dip or injection lubrication is used to lubricate tribosystems in machine elements. This leads to the need for large lubricant volumes and elaborate sealing requirements. One potential method of minimizing the amount of lubricant and simplifying sealing in gearboxes is the self-lubrication of tribosystems using oil-impregnation of porous materials. Although well established in low-loaded journal bearings, self-lubrication of rolling-sliding contacts in gears is poorly understood. This study presents the self-lubrication method using oil-impregnated porous sinter material variants. For this, the tribosystem of gear contacts is transferred to model contacts, which are analyzed for friction and temperature behavior using a twin-disk tribometer. High-resolution surface images are used to record the surface changes. The test results show a significant increase in self-lubrication functionality of tribosystems by oil-impregnated porous sinter material and a tribo-performance comparable to injection-lubricated tribosystems of a sinter material with additionally solid lubricant added to the sinter material powder before sintering. Furthermore, the analyses highlight a significant influence of the surface finish, and in particular the surface porosity, on the overall tribosystem behavior through significantly improved friction and wear behavior transferable to gear applications.

1. Introduction

Gearboxes are widely used in mechanical engineering due to their versatility in terms of performance class, transmission ratio, and axis positions. They offer a simple yet reliable and efficient design, making them suitable for various applications, including industry, automotive engineering, aerospace, or wind turbines. A liquid-lubricated gearbox consists of machine elements, including gears, bearings, and seals, as well as lubricants and the housing. The use of liquid lubricants reduces power losses during operation and increases lifetime by separating the surfaces in tribological contacts by a hydrodynamic lubricant film, enabling low friction and good heat transfer. Integrating the required lubricant volume for contact lubrication into the pores of a sintered component can significantly simplify or even eliminate the lubrication and sealing system, thereby reducing no-load gearbox power losses, weight, and the risk of oil leakage [1].

Commonly used gearbox lubrication methods are dip lubrication, such as in grease or oil sumps, and injection lubrication, including interval or continuous, circulating, or loss lubrication. The selection of a lubrication method for a gearbox depends, among various factors, on the type of gear, the power loss, the prevailing speed, the type of lubricant, the lubricant supply to tribological contacts, and the lifetime requirements. The lubrication methods presented are used in applications from low to high loads and speeds in continuous and intermittent operation [2,3]. The back-coupling of hydrodynamics with elastic deformation of rolling-sliding elements results in elasto-hydrodynamically lubricated (EHL) contacts [4]. In self-lubricating EHL contacts with oil-impregnated sintered rolling-sliding components, similar to the human knee joint [5], compression of the pore structure causes oil to extrude out, lubricating the contact.

Self-lubrication is already used in low-load applications, such as sintered journal bearings or roller-bearing cages, whose porous structures serve as oil reservoirs [6]. For example, the use of sintered journal bearings is standardized by ISO 2795/DIN 1850 III [7]. The formation of an oil film in such self-lubricating contacts is affected by the flow of oil from the pore reservoir in the component and its exposed surface and vice versa. The flow is determined by a number of factors, including centrifugal force as a function of rotational speed, oil viscosity [8], surface roughness [9], diffusion due to concentration gradients [10], capillary flow, and the different thermal expansion of porous solid material and oil. Additionally, the compression or deformation behavior of the material [11,12] and the operation-induced pressure fluctuations [13] must be considered. In their studies of oil-impregnated sintered bearings, Roberts [8] and Fote et al. [9] identified surface porosity and oil viscosity as significant factors affecting journal-bearing lubrication. For example, surface treatments such as grinding can cause covering of the pores necessary for oil flow, thereby reducing the porosity-dependent permeability. Also, the polarity and viscosity of the oil can affect the surface tension, thereby influencing the flow of oil into and out of the porous surface. The studies of centrifugal force by Marchetti et al. [11,12] also show that no oil is extruded below a pore-structure-dependent threshold. For sintered journal bearings, the lubricant flow in the porous material is described by the effect of self-circulation: Oil is intruded into the porous material at the contact area under high pressure, while oil can extrude from the material in unloaded areas. Also, concentration differences result in compensating flows within the porous structure [14,15]. In porous journal bearings, oil extrusion is particularly affected by the surface porosity. Dizdar [16] and Lipp [17] define the transition zone from open to closed porosity at a density of the porous matrix of ρ = 7.1–7.2 g/cm³. In open porosity, the surface pores are mostly directly connected to the inner porous body, allowing the resulting channels to transport oil into and out of the porous material [18]. Oil-carrying pores of various sizes are randomly distributed on the surface, resulting in locally inhomogeneous hydrodynamic oil film formation and load distribution in the tribological contact. Li and Olofsson’s [19] investigations of externally lubricated sliding contacts with sintered bodies demonstrate the correlation between the measured coefficient of friction and wear and pore size. According to Balasoiu et al. [20], increased oil permeability leads to increased oil extrusion but decreased surface load-carrying capacity.

The use of powder metallurgical sinter materials offers the possibility of economically producing near-net-shape parts with high durability. The thermomechanical properties are mainly determined by the density and pore structure of the sintered component [21], which are adjusted by the powder raw material and process conditions. In direct comparison with solid steel, sinter materials often exhibit poorer mechanical properties such as lower tensile strength, fatigue strength, ductility, and stiffness with increasing porosity [22]. However, Ebner et al. [23,24] transferred in their pioneering works the known self-lubricating tribosystem of oil-impregnated sinter materials to highly-loaded EHL rolling-sliding contacts, prevailing in gears and roller bearings. When considering highly loaded contacts, the lubricant flow driving causes are analogous to those reported for journal bearings. These include centrifugal force, elastic deformation of the contact partners, and thermal expansion of the oil due to frictional heat production during operation [25]. Transferring the knowledge of the self-lubrication mechanism, the general functionality of the system with porous sinter materials is confirmed using an optical tribometer and a twin-disk tribometer, as well as accompanying tribosimulation. The best functionality of self-lubrication is demonstrated with case-hardened sinter material density of ρ_sinter = 7 g/cm³ [26], whereas the operating behavior is characterized in stable, metastable, and unstable conditions by a stability coefficient Ξ determined by the in-/decreasing or stationary trends of the measured friction and temperature [25]. Further, Ebner et al. correlated the operating behavior with the global and local surface porosity of the test specimens and the impregnated oil. In most cases, the tribosystem runs in severe mixed lubrication regimes, determined with contact resistance measurements. The investigations reveal cases of starved lubrication, where the amount of oil is not sufficient to form a lubricant film thickness as high as in the case of external lubrication [27]. Further, investigations using optical interferometry by Omasta et al. [28] show the local formation of a lubricant film on porous surfaces, which has a thickness of up to 40% compared to external lubrication. An ISO viscosity grade (VG) of 100 with friction modifier additives such as plastic deformation additives has proven to be the most suitable for self-lubricating tribosystems with severe mixed lubrication [29]. Nevertheless, the EHL tribosystems tend toward failure within specific load and speed limits due to reaching their thermal limits, in particular because of the avoidance of the heat convection due to external oil volumes.

Zhang et al. [30], Rabaso et al. [31], and others ([32,33]) investigated the friction and wear-reducing effects of molybdenum (Mo) and tungsten (W) nanoparticles as oil additives in sliding contacts in a vacuum atmosphere. In addition, Zhu et al. [34] reviewed the friction and wear resistance benefits. Vazirisereshk et al. [35] showed in their review limitations of coatings and composite materials, including MoS₂ and WS₂, and identified current applications such as journal bearings or sealings especially suited for operations in aerospace, with extremely low or high temperature or radioactive surrounding, and also highlighted the oxidation tendency in humid oxygenic environments. Among others (e.g., Kovalchenko et al. [36]), Dhanasekaran et al. [37] mixed solid lubricant MoS₂ powder into the sinter material powder mixture (here: Fe-C-Cu) during the sintering process for a dry-lubricating sintered structure, achieving with an addition of <5% of MoS₂ in the reference material decreased friction in pin-on-disk tests compared to non-compound dry-running material variants, by showing similar to better mechanical properties in terms of strength and wear resistance.

The literature on self-lubricating oil-impregnated sinter materials indicates that it is necessary to properly design the porosity and its distribution as well as the oil properties. For self-lubricating oil-impregnated machine elements with low-load applications like in journal bearings, the compromise between a supportively higher porosity of the tribosystem partners for oil extrusion for lubricant film formation within the tribocontact and a lower porosity for oil intrusion out of the tribocontact back into the porous structure in order to achieve the required stable operating conditions within a broad transmittable power is already found. To increase the power density of self-lubricating oil-impregnated machine elements for high-load applications like cylindrical gears, suitable material pairings with sufficient mechanical properties as well as frictional and thermal behavior are determined in this study by, e.g., varying the amount of oil-supplying oil-impregnated EHL tribosystem partners. Further, as self-lubricating tribosystems primarily operate in severe mixed lubrication regimes with comparable smaller lubricant film thicknesses, the oil intrusion inhibiting and the relative film thickness increasing effect of the surface needs to be controlled by adjusting, next to the surface roughness, the surface porosity due to surface finishing, which is evaluated in this study as well. Because of the thermal limitations due to the limited heat removal from the tribological contact, further tribological performance increasing effects such as the solid-friction reducing effect of solid lubricants have to be implemented in the self-lubricating EHL tribosystem, which are investigated first within this study. To finally increase the power density of the self-lubricating EHL tribosystem with oil-impregnated sinter materials, material–surface variants are systematically investigated in this study regarding their influence on the frictional performance at various operating conditions. Also, the stability of the operating behavior as well as of the material and surface condition alteration with further operation progress are analyzed to find suitable specifications of the EHL tribosystem with reduced frictional power loss and proper durability to enhance the performance of self-lubricating materials at higher loads and speeds. To summarize, this study investigates, in an experimental approach, the influence of material pairing and surface finish on frictional and thermal behavior.

2. Materials and Methods

The section below describes the materials and methods to generate the results of this study. First, the material and surface variants of the test specimens are introduced. Second, its testing on the twin-disk tribometer is described. Third, the measurement procedure and the analysis of the results are outlined.

2.1. Specimen

The materials for the self-lubricating EHL tribosystems investigated in this study are 16MnCr5, a conventional case-hardening steel (hereafter referred to steel), and variants according to DIN-30910 [38] open-pore sintered metal, Sint-D31 (hereafter referred to as sinter). The mechanical material properties are given in

Table 1. Mechanical properties of the considered materials.

Material	Young’s Modulus E in GPa	Poisson’s Ratio ν	Surface Hardness	Density ρ in g/cm3
steel	210	0.30	745 HV5	7.8
sinterref/sl	130	0.27	755 HV5	7.0
sinterpni	140	0.27	785 HV5	7.0

. The sintered test specimens were produced in three material variants according to DIN EN ISO 5755 [39] and consist of conventional case-hardened P-FL-05M1 sinter material (sinter_ref), solid lubricant-alloyed case-hardened P-FL-05M1 sinter material (sinter_sl) with an addition of 2% of molybdenum sulfide MoS₂ and tungsten sulfide WS₂ solid lubricant additives in the powder mixture, which does not affect the mechanical properties regarding, e.g., Young’s Modulus, and plasma-nitrided P-FL-05Cr3M sinter material (sinter_pni). Nearly homogeneous pore distributions and isotropic material properties can be achieved through standard-compliant production in the axial press sintering process. The sintered metal test specimens have a density of ρ = 7 g/cm³, which Ebner [25] classified as the most functional for the self-lubricating EHL tribosystems, resulting in an open porosity (Φ = 10%). During the manufacturing process, each material underwent a quality assurance measurement of density to ensure repeatability and consistency in global porosity. The porosity requirement is derived from a trade-off between material properties (particularly strength and hardness) and a maximum pore volume for oil storage.

The specimen geometry is shown in Figure 1. The materials used are machined from a cylindrical blank to the target geometry for test purposes. The rolling-sliding contact is formed by the contact of two specially machined cylindrical disks of width l_eff = 5 mm and radius R = 40 mm (see Section 2.2 and Section 2.3). Based on reference surfaces in a longitudinally ground state with a parallel relative orientation to the running direction, various surface finishes (see surface images made with a Leica M205A from Leica Microsystems GmbH (Wetzlar, Germany) encoded stereo microscope in

Table 2. Surfaces of the considered specimen.

Longitudinally Ground (lgr)	Axially Ground (agr)	Superfinished (suf)	Mechanically Polished (mpo)	Plasma Nitride (pni)

) are tested to investigate the influence of the surface on the self-lubricating EHL tribosystem. To consider the transfer to gear contacts in particular, axially ground disks are considered, as they represent the characteristic surface structure of a gear contact with conventionally ground tooth flanks with a surface structure perpendicular to the rolling-sliding direction. The arithmetic mean roughness Ra of the surface variants used is ≤0.25 µm for longitudinally and axially ground and non-finished at sinter_pni, ≤0.10 µm for superfinished, and ≤0.01 µm for polished. Surface roughness measurements were performed using the profile method according to DIN EN ISO 13565 1-3 [40], measured with a length of L_t = 4.0 mm transversely to the current surface structure orientation and a cut-off wavelength of λ_c = 0.8 mm.

For self-lubrication tests, the sintered specimens were impregnated with oil using a pressure impregnation device (see [25]). The theoretical maximum impregnation quantity of a component is m_max ≈ 2 g and can be determined according to [25] from the porosity volume V_por and the density of the designated oil ρ_oil = 0.849 g/cm³. V_por is set at approximately 10% of the specimen volume V_spe = 24.2 cm³ through a validated manufacturing process (see Equations (1) and (2)). The impregnability of sintered components depends mainly on the porosity distribution inside the material and on the surface, the radius of the pore channels (equivalent to capillaries), the oil properties like viscosity, and the environmental conditions, e.g., process pressure during impregnation and temperature [41,42]. Since the pore network is randomly distributed and, according to Lados et al. [43], there may be isolated or hard-to-reach pores, the measured impregnation quantity m_oil is lower than m_max. The impregnation amount m_oil of all specimens is consistently lower than the theoretically possible impregnation amount m_max, ranging from m_oil = 1.501 g to = 1.863 g. The average oil impregnation amount across all tests is $\overline{\mathrm{m}}$_oil = 1.671 g, corresponding to an impregnation degree of χ_imp ≈ 75%. Prior to weighing, the oil-impregnated specimens are cleaned with a compressed air pistol and dry tissues for the surfaces, as well as cotton swabs for the small bore of the Pt100 and for sharp edges.

V_por = 0.1·V_spe

(1)

m_max = V_por·ρ_oil

(2)

The test oil PAO100+PD used is a synthetic oil based on polyalphaolefin. It is a fully formulated oil of the ISO VG 100 class with a plastic deformation (PD) additive. The PD additives belong to the class of friction modifiers, which provide solid friction-reducing properties [44] through the tribo-induced reaction of additives with metallic surfaces. The oil data can be found in

Table 3. Data of the considered oil PAO100+PD [45].

ISO VG	Density ρoil at 15 °C in kg/m3	Kin. Viscosity η40°C at 40 °C in mm2/s	Kin. Viscosity η100°C at 100 °C in mm2/s
100	849	105.0	15.7

.

2.2. Test Rig

The experimental tests in this study were conducted using a twin-disk tribometer, which was also used by Ebner et al. (e.g., [23,25]) to investigate self-lubricating rolling-sliding contacts. Figure 2 shows a schematic of the mechanical configuration of the twin-disk tribometer. The description and formulations presented here are based on the publications of Ebner et al. [23] and Lohner et al. [44].

A normal force F_N is applied to the disk contact by a pneumatic cylinder mounted at the end of a pivot arm. The lower disk is mounted in a skid that is attached to the frame by thin steel sheets. The skid is laterally supported by a load cell, which measures the frictional force F_R in the disk contact as a reaction force of the operating conditions in the contact. The maximum load that can be applied on the disk contact is F_N = 16 kN, resulting in individual pressure depending on the disk geometry and material. The maximum sum velocity is v_Σ = 12 m/s. In this study, the upper disk (v2) rotates at a speed that is either the same or faster than the lower disk (v1), both of which rotate with a speed between 2 m/s and 8 m/s, depending on the test method (see Section 2.3). Based on the considered normal loads, the calculated Hertzian pressure p_H in the disk contact is between 600 N/mm² and 1.043 N/mm². Within this study, the upper disk (v₂) rotates at a faster speed than the lower disk (v₁). The measurands are the normal force (F_N), the frictional force (F_R), the surface velocities of the lower and upper disk (v₁ and v₂) derived from the measured shaft speeds, and the bulk temperature (ϑ_M) of the upper test disk. ϑ_M is measured 5 mm below the disk surface using a Pt100 resistance temperature sensor, whose signal is transmitted from the rotating shafts via a mercury rotary signal transmitter and whose signal is transformed to a voltage signal in a Wheatstone Bridge after the transmitter. The location of the Pt100 sensor is selected as a compromise in order to obtain the temperature of the surface region as closely as possible, but also to ensure that there is sufficient distance between the sensor and the surface to prevent any interference with the contact stiffness and material strength when the disk is loaded. For evaluation, the bulk temperature measurement can be utilized to determine the level of dissipated energy and, thus, the frictional behavior of the tribosystem. Before each test, the contact pattern is carefully evaluated to ensure an evenly distributed load during the line contact of the disks, and any misalignment is mechanically corrected. To analyze the frictional behavior, the coefficient of friction (CoF) µ is determined from the measured friction force F_R and normal force F_N at the disk contact. According to Vojacek [46], the considered twin-disk tribometer has an accuracy for measuring µ of Δμ = 0.0025.

v_∑ = v₁ + v₂

(3)

s = ((v₁ − v₂)/v₁)·100% with v₁ < v₂

(4)

µ = F_R/F_N

(5)

For a comparative test with external lubrication, an injection-lubrication unit is used. Thereby, the PAO100+PD oil is injected directly into the contact inlet with a volume flow rate of $\dot{V}$ = 1.5 L/min at ambient oil temperature (ϑ_oil = 25 °C) without any temperature regulation. For all measurements, the disk pair of a longitudinally ground, oil-impregnated sinter_ref disk with a longitudinally ground steek disk was established as the reference system. The experimental procedure outlined in Section 2.3 also includes other material–surface combinations to evaluate the influence of lubrication methods, material pairing, solid lubricant addition, and surface condition.

2.3. Test Method

The test method involves two steps under line contact conditions at the twin-disk tribometer (see Section 2.2), first starting with measuring friction curves of self-lubricating oil-impregnated sinter material and surface combinations compared to conventional injection lubrication, followed by long-term operation tests to evaluate the operating stability for stationary operating systems.

2.3.1. Friction Curves

For all tests, two new disk pairs with identical specifications are used. Prior to each test under self-lubrication, an initial oil volume of V_init = 0.02 mL is applied to the dry tissue-cleaned contact surfaces to prevent initial damage during running-in. The measurements with injection lubrication are carried out with an initial oil temperature at ϑ_oil = 25 °C. All measurements are started at disk bulk temperatures ϑ_M ≈ ϑ_ambient with a running-in of t = 60 s without an imposed pressure p_H < 500 N/mm² at moderate speed v_Σ = 1 m/s and s = 2%. While testing, termination criteria are used to detect possible failures of a disk pair. These criteria are a bulk temperature ϑ_M greater than the critical bulk temperature ϑ_crit > 140 °C or a CoF µ greater than the critical CoF µ_crit > 0.12. Light microscopy images of the specimen surfaces are captured before and after the tests to evaluate the lubricating effect and the corresponding surface alteration. The disk pair of a longitudinally ground sinter_ref disk impregnated with the PAO100 with a longitudinally ground steel disk was established as the reference tribosystem for all measurements.

The friction curves are measured at operating points with a constant pressure of p_H = 600 N/mm² and different sum velocities v_∑ = {2; 4; 8} m/s. Each measurement at a sum velocity is made at discrete, increasing slip ratio s = {0; 2; 5; 10; 20; 30; 40; 50}%. The measurements with each disk pair are repeated once, with the 2^nd measurement taken without re-impregnation of the sinter disk with oil. The repeatability of the two measurements is also analyzed to investigate the difference in the tribological behavior of a variant. Therefore, a repetition test is performed with a 2^nd new disk pair, and both initial and subsequent measurements are made. Each operating point on the friction curve is maintained until quasi-stationary operating conditions are reached. An operating point is considered quasi-stationary when the bulk temperature difference Δϑ_M remains within Δϑ_M < 1 K for a period of t = 60 s. If a stationary operating condition is achieved, the average of the measurement data from the last t = 30 s is used. To evaluate the tribological behavior, µ as well as ϑ_M is plotted for each sum velocity v_Σ at each prevailing slip ratio s. The measurements are started with the lowest sum velocity v_Σ, starting with pure rolling at s = 0% at disk bulk temperatures ϑ_M close to ϑ_ambient. Then, as the system reaches a stationary operating point, the slip ratio s is gradually increased. The friction curve for the next higher sum velocity v_Σ is run after cooling the system back to ϑ_ambient. The procedure is continued until all operating points have been investigated or until a termination criterion is reached. To achieve the objectives of this study, the experiments are performed sequentially with the following structure:

• Demonstration of the functionality of self-lubricating rolling-sliding contacts across a wide range of operating points and comparison with injection-lubricated rolling-sliding contacts.
• Determination of the tribologically optimal number of oil-impregnated elements in a self-lubricating rolling-sliding contact.
• Investigation of the influence of surface condition and finish on self-lubricating rolling-sliding contacts.

2.3.2. Long-Term Tests

In the long-term operation tests, constant operating points are maintained. Hereafter referred to as the reference operating conditions, the operating point at a pressure of p_H = 1043 N/mm², a sum velocity of v_∑ = 4 m/s, and slip ratio of s = 20% is used for evaluation of the test results. The maximum duration of the tests is t = 48 h, which corresponds to a maximum load cycle number N₂ ≈ 1.2 m on the faster rotating disk, provided that the automatic shutdown mechanism does not terminate the test prematurely due to reaching the termination criteria ϑ_crit or µ_crit. The repeatability of the long-term test results is indicated by the maximum difference of µ or N₂ of the 1^st and a repetition test performed with a 2^nd new disk pair. To further investigate the study’s objectives, the long-term operating tests are conducted as follows:

• Demonstration of long-term durability of self-lubricating contacts and comparison with injection-lubricated contacts.
• Investigation of the influence of surface finish on the frictional and stress behavior in self-lubricating contacts.
• Investigation of the improvement of the frictional behavior in self-lubricating contacts through solid lubricant additives in the sinter material.
• Determination of the performance of self-lubricating contacts under higher load and speed.
• Evaluation of the transferability to gear applications with typical tooth flank surfaces.

The measurements of each variant are examined immediately after the running-in. The test run is conducted intermittently in order to enhance the detection of surface alterations of the upper disk after a defined number of load cycles N₂, using the Infinite Focus Sensor R25 from Alicona Imaging GmbH (Raaba, Austria), an optical measuring device based on the principle of focus variation for high-resolution surface imaging. The number of load cycles between two surface measurements is shorter at the beginning of the tests to document the running-in-like process of the first load cycles and increases after each measurement. An overview of the measurement intervals, expressed in terms of load cycle numbers and the corresponding times in hours, is presented in

Table 4. Intermittent measurement intervals during long-term operation tests.

	Measurement Intervals
Load cycle number N2	>2.5 k	>10 k	>30 k	<100 k	>250 k	>750 k	>1200 k
Time t in h	0.08	0.33	1	3	8	24	48

. During the intermittent measurement intervention, the system is interrupted for t < 5 min to prevent excessive cooling of the bulk temperature ϑ_M. The surface measurement of the upper disk captures individual areas in the shape of a square with an edge length of a = 1 mm and a 20-fold magnification to analyze the local surface condition. To obtain a global qualitative impression of surface alteration across the entire disk width from the composite local images, six overlapping surface images are taken along the direction of the cleaned area from one disk edge to the other. In this study, the surface image in the middle of the upper disk width is evaluated. Moreover, an expert qualitatively evaluated a surface area in the middle of the upper disk width using the optical appearance of pores in 2D images and the matching geometrical hole-like shape of pores in 3D topography measurements.

Each test run is evaluated by its levels and differences of friction µ and bulk temperature ϑ_M as well as its operating behavior coefficient Ξ according to Ebner [25]. The operating behavior coefficient Ξ is calculated by the division of the load cycles during the longest test sequence, with a maximum difference of the minimum and maximum of µ being smaller than Δµ < 0.005, and the load cycles for the whole lifetime of the disk pair. An operating behavior is defined as unstable when the load cycle of the upper disk for the whole lifetime is N₂ < 0.25 M, as metastable when N₂ > 0.25 M and Ξ < 0.40, and stable when N₂ > 0.25 M and Ξ > 0.40.

3. Results

In the following section, first, the friction curves and, second, the long-term operation tests with applicability promising disk pairs are shown.

3.1. Friction Curve Tests

At first, the functionality of self-lubricating rolling-sliding contacts with oil-impregnated sinter material across a wide range of operating points is shown and compared to an injection-lubricated pure steel contact. Further on, the tribologically optimal number of oil-impregnated sinter disks in a self-lubricating rolling-sliding contact and the influence of surface condition are evaluated.

3.1.1. Demonstration of Self-Lubricating Rolling-Sliding Contacts and Comparison with Injection-Lubricated Rolling-Sliding Contacts

Figure 3a shows the measured friction curves of the reference disk pair with a longitudinally ground, oil-impregnated sinter_ref disk and a longitudinally ground steel disk under self-lubrication. Additionally, the maximum values µ_max of each curve are shown in a bar diagram. The reference disk pair can be evaluated at all operating points as it did not exceed the termination criteria µ_crit or ϑ_crit during the tests. The measurements with the 1^st and the 2^nd disk pair both show the same trend: µ decreases with increasing v_Σ. The friction curves of the 1^st disk pair increase steadily within the first three measured slip ratios of s ≤ 5% and maintain a plateau at v_Σ = 2 m/s over the remaining slip ratios at approximately µ ≈ 0.064. At v_Σ = 4 m/s and 8 m/s, the CoF curves decrease again after reaching their respective local maxima at s = 10%, before increasing again at the highest slip ratio at s = 50% to the global maximum µ_max = 0.051 and 0.047, respectively. The friction curves of the tests with the 2^nd disk pair are at a continuously lower level compared to the CoFs of the 1^st disk pair, with plateau-like curves at v_Σ = 4 m/s and 8 m/s, reaching µ_max = 0.042 and 0.047 at the highest slip ratio, respectively. Figure 3b shows the measured bulk temperature curves and the maximum values ϑ_M,max, which correspond to the increasing frictional power depending on the sum velocity, so that the levels of the bulk temperature are at v_Σ = 2 m/s at the lowest, at 4 m/s at the middle, and at 8 m/s at the highest level. Within the respective sum velocities, the bulk temperature also increases with increasing slip ratio. In the bulk temperature curves, the temperatures of the 1^st disk pair are higher than the temperatures of the 2^nd disk pair. Specifically, the bulk temperature maxima of the 1^st disk pair are ϑ_M,max = 56 °C, 76 °C, and 109 °C, while with the 2^nd disk pair they are 53 °C, 64 °C, and 109 °C. Figure 4a shows the measured friction curves of the 1^st and 2^nd test with the 1^st disk pair, which was re-run in the 2^nd test after the 1^st test without re-impregnation. Additionally, the maximum difference of µ between the 1^st and the 2^nd test is shown for each sum velocity. As for the 1^st test, every operating point of the 2^nd test can be evaluated, because no termination criteria were reached. The friction curve of the 2^nd test at v_Σ = 2 m/s increases with increasing slip ratio up to s ≤ 10% to a higher level than in the 1^st test and, at s = 10%, reaches a local maximum for µ. However, with increasing slip ratio up to s ≤ 20%, µ decreases before increasing again to µ_max = 0.083 at s = 40%. At the highest slip ratio s = 50%, µ decreases again. However, the friction curves of the 2^nd test at sum velocities of v_Σ = 4 m/s and 8 m/s follow a similar trend as the 1^st test, with smaller increases in the curves up to the highest slip ratio at s = 50%. The global maxima of the curves at v_Σ = 4 m/s and 8 m/s are µ_max = 0.051 and 0.046, respectively. Additionally, Figure 4b shows the bulk temperature curves ϑ_M and the maximum values ϑ_M,max of the 1^st and 2^nd test with the 1^st disk pair. The curves of both tests are very similar at v_Σ = 8 m/s. At a velocity of 4 m/s, minimal temperature differences between the 1^st and 2^nd test are observed over most of the range of slip ratio variation, with slightly lower temperatures of the 2^nd test at the highest level of slip ratio at s = 50%. In comparison, at v_Σ = 2 m/s, higher bulk temperatures are measured at the 2^nd test over the entire range of slip ratio variation. The maximum bulk temperatures of the 2^nd test recorded for v_Σ = 2 m/s, 4 m/s, and 8 m/s are ϑ_M = 60 °C, 67 °C, and 104 °C, respectively.

Figure 5a shows the measured friction curves of the 1^st and 2^nd test with the 1^st disk pair with two longitudinally ground steel disks under injection lubrication. The friction curves initially increase with larger gradients over the slip ratios of s = 2–5%, then exhibit a smaller increase. The maximum coefficients of friction are µ_max = 0.057, 0.038, and 0.035 in the 1^st test run and µ_max = 0.058, 0.034, and 0.033 in the 2^nd test run. Figure 5b shows the bulk temperature curves of both tests, the 1^st and 2^nd test with longitudinally ground steel disks under injection lubrication, correlating to the curves shown in Figure 5a. The bulk temperature differences between the sum velocities remain largely constant over the slip ratios, with almost constant temperature differences. The 2^nd test at v_Σ = 2 m/s resulted in slightly higher bulk temperatures, approximately Δϑ_M = +2 K higher than those of the 1^st test. The maximum bulk temperatures at v_Σ = 2 m/s, 4 m/s, and 8 m/s are ϑ_M,max = 38 °C, 43 °C, and 52 °C in the 1^st test and 40 °C, 43 °C, and 52 °C in the 2^nd test.

3.1.2. Determination of the Tribologically Optimal Number of Oil-Impregnated Sinter Elements in Self-Lubricating Rolling-Sliding Contacts

Figure 6a shows the measured friction curves of the 1^st tests of two polished steel disk pairs, each lubricated only once with V_init. Because of an expected harsh mixed lubrication regime in once-lubricated EHL tribosystems due to the low amount of prevailing oil, the steel disks were mechanically polished to reduce the influence of the surface roughness on the relative lubricant film thickness and thus improve the expected friction to be better performing compared to once-lubricated EHL tribosystems with longitudinally ground surfaces. Due to the termination criterion of µ > 0.12 reached at v_Σ = 2 m/s with s = 20% or 30%, both tests can only be evaluated for a few operating points. Figure 6b shows the bulk temperature curves of both tests correlating to Figure 6a, which both show increasing ϑ_M with increasing slip ratios. Repeating the test with the 1^st or the 2^nd disk pair is not possible because of surface damage.

Figure 7a shows the measured friction curves and maxima µ_max of the 1^st tests of two disk pairs, each with two longitudinally ground, oil-impregnated sinter_ref disks. The trends observed with the 1^st disk pair are clearly related to the sum velocity. The data shows that µ gradually decreases with increasing v_Σ, with all curves starting with a steep gradient of µ within the initial slip ratios, then decreasing after reaching a global maximum µ_max. For instance, at v_Σ = 2 m/s, the global maximum CoF is µ_max = 0.076 at s = 10%. Similarly, at v_Σ = 4 m/s, the global maximum of CoF is µ_max = 0.060 at s = 20%, and at v_Σ = 8 m/s, it is µ_max = 0.052 at s = 5%. The bulk temperatures shown together with ϑ_M,max in Figure 7b appear to increase as the sum velocity increases, with relatively constant increases for each sum velocity. At v_Σ = 2 m/s, the maximum bulk temperature is ϑ_M,max = 51 °C, at 4 m/s it is 63 °C, and at 8 m/s it is 79 °C. The 1^st test with a 2^nd disk pair of the same specification shows trends that follow less clearly the previous patterns of sum velocity and slip ratio variation (see Figure 7a,b). In particular, at v_Σ = 2 m/s, the coefficient of friction is initially higher at s ≤ 5% and reaches a maximum of µ_max = 0.072 at s = 5%, before decreasing at s ≤ 40% and remaining approximately constant up to s = 50%. At v_Σ = 4 m/s, the coefficient of friction initially follows the trend of the 1^st disk pair before a peak of µ_max = 0.079 at s = 20%, which is higher than the maximum at v_Σ = 2 m/s, and then declining to a level comparable to the 1^st test. The friction curve at v_Σ = 8 m/s lies, as typically expected, at the lowest level of the sum velocities within the tests of one disk pair. The curve reaches a local maximum at s = 5%, before decreasing again to s = 20%, then increasing significantly and reaching the global maximum at s = 50% with µ_max = 0.044. The correlation between the bulk temperatures and the coefficient of friction curves is evident. The curve at v_Σ = 2 m/s is the lowest with ϑ_M,max = 51 °C, the curve at v_Σ = 4 m/s is higher with ϑ_M,max = 78 °C and shows a significant increase at the peak of the CoF at s = 20%, and the curve at v_Σ = 8 m/s is comparatively the highest, increasing significantly from the slip ratio s = 20% until it reaches a maximum with ϑ_M,max = 122 °C, significantly higher than that of the 1^st test.

Figure 8a shows the friction curves and maxima µ_max measured with the 1^st and 2^nd test of the 1^st disk pair of two longitudinally ground, oil-impregnated sinter_ref disks. The 2^nd test of the 1^st disk pair can be evaluated at all operating points but shows significantly higher µ soon after initially lower values at all three sum velocities investigated, reaching near the termination criterion at v_Σ = 2 m/s with µ_max ≈ 0.12 and increasing to similar values at v_Σ = 4 m/s and 8 m/s, with µ_max = 0.063 and 0.062, respectively. Figure 8b shows the bulk temperature curves and maxima ϑ_M,max measured during the 2^nd test of the disk pair shown in Figure 8a. The bulk temperatures initially start at slightly higher levels before increasing along with the significant increases in the coefficients of friction, all reaching the bulk temperature maxima at s = 50% at v_Σ = 2 m/s, 4 m/s, and 8 m/s, at ϑ_M,max = 66 °C, 90 °C, and 136 °C, respectively.

3.1.3. Investigation of the Influence of Surface Finish on Self-Lubricating Rolling-Sliding Contacts

The investigation of the surface influence on the tribological behavior of self-lubricating, oil-impregnated sintered rolling-sliding contacts is carried out systematically. Initially, a mechanically polished steel disk surface is paired with a longitudinally ground, oil-impregnated sinter_ref disk. Subsequently, a mechanically polished steel disk is paired with different surface variants of the oil-impregnated sinter_ref disk. Therefore, the surface of the sinter_ref disk is investigated in different conditions for the reference material, including superfinished and mechanically polished surfaces. Furthermore, the influence of a plasma-nitrided sinter_pni disk with high hardness is investigated.

Figure 9a,b summarize the evaluated maximum µ_max and ϑ_M,max from measured friction and bulk temperature curves of the two test runs with the 1^st disk pair (1^st and 2^nd test), as well as the 1^st test with a 2^nd disk pair. Additionally, Figure 10a,b show the maximum difference of the friction curves Δµ_max and of the bulk temperature curves Δϑ_M,max of the 1^st and 2^nd test with the 1^st disk pair. The results obtained with the reference disk pair with a longitudinally ground steel disk and a longitudinally ground, oil-impregnated sinter_ref disk are also included for comparison.

The surface variation of the steel disk from a longitudinally ground to a polished condition shows improved tribological behavior compared to the reference measurements with both disk surfaces in a longitudinally ground condition, indicated by low levels of µ_max and ϑ_M,max in the 1^st test runs of the 1^st and 2^nd disk pair and especially with the 2^nd disk pair, with µ_max = {0.033, 0.025, 0.020} and ϑ_M,max = {46, 66, 71} °C. There was good repeatability of the 1^st test with the 1^st disk pair through the 1^st test with the 2^nd disk pair, with very similar ϑ_M,max and smaller µ_max at all v_Σ (see Figure 9a,b), and a low maximum difference of Δµ_max = 0.003 … 0.010 and Δϑ_M,max ≤ 2 K between the 1^st and 2^nd test with the 1^st disk pair (see Figure 10a,b).

Figure 11a and Figure 12a show the measured friction curves with disk pairs with mechanically polished steel and superfinished sinter_ref disk surfaces; the curves with a superfinished sinter_ref disk surface show, on one hand, a similar tribological beneficial trend to the curves with the mechanically polished steel disk surface, and on the other hand, the best tribological behavior compared to the other surface variants. Measurements taken with superfinished contact surfaces show µ_max ≤ 0.042 for all tests with both disk pairs down to µ_max ≤ 0.022 at v_Σ = 8 m/s with the 2^nd disk pair as well as a very good repeatability between the 1^st tests with the 1^st and 2^nd disk pair, with slightly lower levels of µ_max with the 2^nd disk pair (see Figure 11a) and the least maximum difference Δµ_max = 0.002 between the 1^st and 2^nd test with the same 1^st disk pair (see Figure 12a). Figure 11b and Figure 12b additionally show the measured bulk temperature curves of the disk pairs with mechanically polished steel and superfinished sinter_ref disk surfaces. Overall, the levels of the maximum bulk temperature ϑ_M,max are the lowest compared to the tests with the other surface variants, with the lowest maximum bulk temperatures at the 1^st test with the 2^nd disk pair with ϑ_M,max = {43, 53, 69} °C. Comparing the 1^st and 2^nd test with the 1^st disk pair, a low maximum difference of Δϑ_M,max ≤ 3 K is shown.

In contrast, the tests with disk pairs with mechanically polished steel and sinter_ref disk surfaces show, at the 1^st tests of the 1^st and 2^nd disk pair, initially at a sum velocity of v_Σ = 2 m/s, high to very high CoF, with µ_max = 0.086 and = 0.12 (see Figure 9a). During both test runs (1^st and 2^nd test) with the 1^st disk pair, the friction curve levels decrease with increasing sum velocity, while the 2^nd test shows a lower level of µ_max at v_Σ = 2 m/s and comparable levels of µ_max at v_Σ = 4 m/s and = 8 m/s. During the 1^st test with the 2^nd disk pair, the friction curve level first slightly decreases with an increasing sum velocity at v_Σ = 4 m/s before leading to termination at the highest sum velocity at v_Σ = 8 m/s. The bulk temperatures of the two tests with the 1^st disk pair shown in Figure 9b are comparable to the bulk temperature levels with the disk pairs of a polished steel disk and ground sinter disk and show moderate to low maximum difference with Δϑ_M,max ≤ 6 K. The bulk temperatures of the 1^st test with the 2^nd disk pair follow the trend of µ and, compared to the tests with the 1^st disk pair, also start at a high level at v_Σ = 2 m/s with ϑ_M,max = 101 °C, and decrease slightly to ϑ_M,max = 83 °C at v_Σ = 4 m/s before thermally increasing at v_Σ = 8 m/s. With the comparison of the tests of the two disk pairs with identical specifications, which can show both a good frictional behavior with a low to moderate difference of the maximum Δµ_max ≤ 0.025 and Δϑ_M,max ≤ 6 K between the 1^st and 2^nd test with one disk pair (see Figure 10a,b), and a non-functional behavior with another disk pair, a significantly worse repeatability of the tribological behavior of disk pairs with polished sinter disk surfaces is shown.

The tests with the disk pairs with a polished steel disk and a plasma-nitrided sinter_pni disk with high hardness cannot be evaluated due to a failure within the initial sum velocity of v_Σ = 2 m/s in the 1^st tests of the 1^st and 2^nd disk pair. The failure is indicated by an acceleration detector, which causes the tribometer to stop immediately. An analysis of the sinter_pni disk surfaces revealed delamination of the compound layer from the underlying diffusion layer.

3.1.4. Summary of the Friction Curve Tests

Table 5. Maximum friction µ_max, maximum bulk temperature ϑ_M,max, and total load cycles N₂ of the upper disk of all friction curve tests with both the 1^st and the 2^nd disk pair as well as maximum difference of the friction Δµ_max and maximum difference of the bulk temperature Δϑ_M,max between the 1^st and the 2^nd test of the 1^st disk pair.

Lubrication Method Upper/Lower Disk pH = 600 N/mm2 s = 0…50%		µmax at vΣ			ϑM,max in °C at vΣ			N2 in M
		2 m/s	4 m/s	8 m/s	2 m/s	4 m/s	8 m/s	Total
once-lubrication steel mpo/steel mpo	1.1 4	>0.120 1	x 2	x 2	>140 1	x 2	x 2	x 2
	1.2 5	x 2	x 2	x 2	x	x 2	x 2	x 2
	Δµmax	x 2	x 2	x 2	x	x 2	x 2
	2.1 6	>0.120 1	x 2	x 2	>140 1	x 2	x 2	x 2
injection lubrication steel lgr/steel lgr	1.1 4	0.057	0.038	0.035	38	43	52	0.03
	1.2 5	0.055	0.034	0.033	40	43	52	0.03
	Δµmax	0.002	0.001	0.002	1	0	0
	2.1 6	- 3	- 3	- 3	- 3	- 3	- 3	- 3
self-lubrication sinterref lgr/steel lgr	1.1 4	0.064	0.051	0.047	56	76	109	0.15
	1.2 5	0.083	0.051	0.046	60	67	104	0.17
	Δµmax	0.019	0.008	0.005	9	9	5
	2.1 6	0.057	0.042	0.047	53	64	109	0.14
self-lubrication sinterref lgr/sinterref lgr	1.1 4	0.076	0.060	0.052	51	63	79	0.13
	1.2 5	0.120	0.063	0.062	66	90	90	0.20
	Δµmax	0.055	0.023	0.033	17	17	15
	2.16	0.072	0.079	0.044	51	88	122	0.16
self-lubrication sinterref lgr/steel mpo	1.1 4	0.057	0.047	0.025	46	66	72	0.13
	1.2 5	0.049	0.046	0.027	45	66	70	0.13
	Δµmax	0.010	0.003	0.004	1	0	2
	2.1 6	0.033	0.025	0.020	46	66	71	0.13
self-lubrication sinterref suf/steel mpo	1.1 4	0.042	0.033	0.026	43	53	69	0.12
	1.2 5	0.040	0.034	0.026	44	53	69	0.12
	Δµmax	0.002	0.002	0.002	3	1	3
	2.1 6	0.034	0.026	0.022	43	56	79	0.13
self-lubrication sinterref mpo/steel mpo	1.1 4	0.086	0.041	0.031	42	56	76	0.14
	1.2 5	0.061	0.041	0.034	46	58	77	0.13
	Δµmax	0.025	0.007	0.004	6	2	2
	2.1 6	0.120	0.110	>0.120 1	101	83	>140 1	0.08 1
self-lubrication sinterpni/steel mpo	1.1 4	>0.120 1	x 2	x 2	>140 1	x 2	x 2	x 2
	1.2 5	x 2	x 2	x 2	x 2	x 2	x 2	x 2
	Δµmax	x 2	x 2	x 2	x 2	x 2	x 2
	2.1 6	>0.120 1	x 2	x 2	>140 1	x 2	x 2	x 2

¹: termination at operating point; ²: no value because of termination; ³: no test conducted; ⁴: 1^st disk pair, 1^st test; ⁵: 1^st disk pair, 2^nd test; ⁶: 2^nd disk pair, 1^st test.

summarizes the evaluated data of the friction curve tests (see Section 3.1.1, Section 3.1.2 and Section 3.1.3), including the total test duration N₂ of each friction curve test of the two disk pairs.

The results of the friction curve tests were used to reduce the number of variants to the most promising disk pairs for long-term testing. As a result, the disk pair variants terminating after one test under self-lubrication with a mechanically polished sinter_ref disk as well as with a sinter_pni disk are not considered for the long-term tests. The disk pair of two sinter_ref disks with its increased material preparation effort for oil impregnation by simultaneously increased friction and bulk temperature levels at all tests and increased differences of friction and bulk temperature between the 1^st and 2^nd test with the 1^st disk pair are also not considered for the long-term tests.

3.2. Long-Term Tests

This section presents the results of the long-term tests with the selected material pairing and surface condition variants from Section 3.1. The operating behavior of the reference disk pair with a longitudinally ground steel and a longitudinally ground sinter_ref disk is compared to that of an injection-lubricated disk pair with two longitudinally ground steel disks. Furthermore, the influence of surface finish on the operating behavior during the long-term tests is examined, followed by an investigation of the influence of the solid lubricant for friction reduction in both once-lubricated and self-lubricating disk pairs. Also, the performance at higher loads is evaluated, and the functionality of gear-application-like surfaces on the disks is examined.

3.2.1. Demonstration of Long-Term Durability of Self-Lubricating Contacts and Comparison to Injection Lubrication

Figure 13 shows the measured friction curves and the calculated stability coefficients of the two long-term tests with a self-lubricating EHL tribosystem with a polished steel disk and an oil-impregnated ground sinter_ref disk as well as an EHL tribosystem with longitudinally ground steel disk pairs under injection lubrication without temperature regulation. Both the 1^st and 2^nd test of the injection-lubricated disk pair show for the whole test run of 48 h (i.e., N_2|1st and N_2|2nd > 1.20 M) very good stability according to Ebner [25], with Ξ = 0.95 and = 0.76. Initially, each of the curves starts with a coefficient of friction level of µ = 0.045 and exhibits the level for the main time of the test duration. In comparison, both tests with the self-lubricating EHL tribosystem start with a lower CoF of µ = 0.037. Then, the test with the 1^st disk pair without intermediate interruptions for surface measurements increases steadily with a small gradient, passing the level of CoF of the injection-lubricated tribosystem after t ≈ 25 h and reaching µ_max = 0.052. The determined stability coefficient is Ξ = 0.51 and indicates stable operating behavior. However, the test with the 2^nd disk pair slightly decreases at the first section of the tests and is maintained at a comparably low level of µ = 0.035 for at least t ≈ 24 h, when the level slowly starts increasing to µ_max = 0.045. The operating behavior of the 2^nd disk pair can be classified as stable, as the determined stability coefficient is Ξ = 0.71. In addition to the comparable lower µ and the stable operating behavior, the EHL tribosystem under self-lubrication shows less repeatability because of the lower level of µ of the 2^nd test, with Δµ_max = 0.020, whereas both tests with the injection-lubricated EHL tribosystem show a difference of Δµ_max = 0.005 from the beginning of the tests and a neglectable difference for the rest of the measurement time.

As there are intermediate interruptions for surface measurements, the following EHL tribosystem behavior after continuing the test has to be evaluated to verify the friction and bulk temperature trends deriving from the disk pair specifications and not from cool-down periods. At the intermittently interrupted test with the self-lubricating disk pair shown in Figure 13, one can see the very short peaks of CoF directly after re-starting the test and, after a few load cycles, the return of the friction curves to the previous level the moment before the interruption of the twin-disk tribometer. With the curves returning to the previous state before the interruption, the influence of the test interruptions on the trend of the tribosystem operating behavior can be assumed as negligible.

3.2.2. Investigation of the Influence of Surface Finish on the Frictional and Stress Behavior in Self-Lubricating Contacts

Figure 14a shows the measured friction curves and the calculated stability coefficients of the long-term tests with self-lubricating disk pairs with the reference EHL tribosystem of a longitudinally ground steel disk and a longitudinally ground oil-impregnated sinter_ref disk and surface variants of the disk pair with disk pairs of a mechanically polished steel disk and an oil-impregnated longitudinally ground as well as a mechanically polished sinter_ref disk. When varying the surface finish of the pairing disks in the self-lubricating EHL tribosystem, the trends of the measured friction during the long-term tests are comparable to the trends obtained at the friction curve tests, with no termination until the end of the 48 h test runs at N_2|1st and N_2|2nd > 1.20 M (see Figure 9 in Section 3.1.3). In nearly all tests, the disk pairs with at least one smoothed surface show lower levels of µ compared to the reference system with longitudinally ground surfaces, except in the 1^st test with the disk pair of a mechanically polished steel disk and a longitudinally ground sinter_ref disk. When taking the pair of the polished steel disk and the superfinished sinter_ref disk evaluated as the best during the friction curve tests into account, the comparably very low level of µ ≤ 0.040 and the very stable operating behavior of both disk pairs with Ξ = 1.00 and = 0.69 as well as the good repeatability with Δµ_max < 0.008 verify the trend of the best-performing self-lubricating EHL tribosystem. Further, the bulk temperatures ϑ_M measured during the long-term tests are shown in Figure 14b, in which every curve of ϑ_M correlates to the trend of µ evaluated, leading to the lowest ϑ_M measured with the disk pair of a mechanically polished steel disk and a superfinished sinter_ref disk. The short intermediate interruptions for surface measurements lead at all tests to a difference of the bulk temperature of Δϑ_M,int ≤ 37.5 K while cooling down. In every test run, the offset Δϑ_M,int is compensated at a few load cycles and has little significant influence on the further levels as well as trends of the bulk temperature curves and friction curves measured before the interruption.

3.2.3. Investigation of the Improvement of the Frictional Behavior in Self-Lubricating Contacts through Solid Lubricant Additives in the Sinter Material

Additionally to the measurements made with the sinter_ref material, tests with an oil-impregnated and a non-impregnated solid lubricant alloyed sinter_sl material are made. Figure 15 shows the measured friction curves and the stability coefficients during long-term tests with disk pairs of a mechanically polished steel disk and variants of the longitudinally ground sinter disk in terms of a non-oil-impregnated sinter_ref or sinter_sl disk for investigations under once-lubrication as well as an oil-impregnated sinter_ref or sinter_sl disk for investigations under self-lubrication.

Both tests with the once-lubricated steel-sinter_ref disk pairs show unstable operating behavior with Ξ = 0.07, leading to termination after just a few load cycles of the upper disk. The operating behavior obtained with the once-lubricated EHL tribosystems with a non-oil-impregnated sinter_sl disk is characterized by an initial friction curve level of µ ≈ 0.044, which is slightly higher than the friction level of the EHL tribosystem with a disk pair of a mechanically polished steel disk and a longitudinally ground, oil-impregnated sinter_ref disk under self-lubrication. During the initial sequence of <7 h (i.e., N_2|1st and N_2|2nd < 0.15 M) stable operating behavior is obtained. After that, µ increases in the test with the 1^st disk pair with a steep gradient and demonstrates unstable operating behavior, promptly reaching the termination criterion after t ≈ 14 h at N_2|1st ≈ 0.35 M. The test with the 2^nd disk pair reaches unstable operating behavior, and thus its termination limit is after a longer period of increasing µ, after t ≈ 19 h at N_2|2nd < 0.50 M. When additionally impregnating the sinter_sl disk with PAO100+PD, the solid lubricant additive consistently results in low levels of µ over the whole measurement time. Additionally, it exhibits repeatably larger load cycle ranges without any increase in µ, thus demonstrating the most stable operating behavior with Ξ = 1.00. The difference between the tests with the two disk pairs of Δµ_max < 0.002 is the lowest when compared to all other tests.

3.2.4. Determination of the Performance of Self-Lubricating Contacts under Higher Load and Speed

Based on the long-term tests under reference operating conditions, maximum load tests are carried out at both higher v_∑ and higher p_H. The variation in kinematics is tested with the promising self-lubricating EHL tribosystems with a polished steel disk and, as the best-performing sinter material variant evaluated, a longitudinally ground sinter_sl disk. Figure 16 shows the friction curves of the long-term tests with disk pairs of a mechanically polished steel disk and a longitudinally ground sinter_sl disk under self-lubrication at v_∑ and p_H, each at a higher level compared to the reference operating conditions shown in Section 3.2.1, Section 3.2.2 and Section 3.2.3.

When the total speed is increased to v_∑ = 8 m/s (equivalent to a maximum of N_2|1st and N_2|2nd > 2.4 M for a full 48 h test run), the tests with the 1^st and 2^nd steel–sinter_sl disk pair reach their termination limits at different times, after t ≈ 43 h at N_2|1st ≈ 2.17 M and after t ≈ 17 h at N_2|2nd ≈ 0.89 M due to exceeding ϑ_M,max = 140 °C. Due to the higher v_Σ and, therefore, the sliding velocity, the measured CoF of µ ≈ 0.043 is comparably higher than at the lower sum velocity of v_Σ = 4 m/s. Whereas the 1^st disk pair shows stable operating behavior with Ξ = 1.00, the operating behavior of the earlier terminating 2^nd disk pair is evaluated as metastable with Ξ = 0.18. The difference between the tests with the two disk pairs is Δµ_max < 0.04. Nevertheless, the test with the 1^st disk pair shows a 244% higher lifetime than the test with the 2^nd disk pair.

At tests at higher pressure p_H = 1200 N/mm², the two tests with a steel–sinter_sl disk pair initially differ from each other more (Δµ_max = 0.011) than after further test progression, which leads to a measured µ in both tests at the same level (Δµ_max ≈ 0) of µ = 0.038. The overall trend shows initially metastable operating behavior with µ = 0.040, which changes to a more stable operating behavior with a decreasing µ after increasing load cycles, N. Although the frictional power is higher, both tests reach the measurement time limitation at N_2|1st and N_2|2nd > 1.2 M without termination.

3.2.5. Evaluation of the Functionality of Self-Lubrication with Tooth Flank Similar Surfaces

Figure 17 shows the friction curves measured during the long-term tests under self-lubrication with disk pairs of an axially ground steel disk and a sinter_ref as well as a sinter_sl material variant of an axially ground, oil-impregnated sinter disk. The axially ground surface structure orientation represents the typical surface structure of cylindrical gear tooth flank surfaces.

In both variants of the EHL tribosystem with axially ground sinter_ref and sinter_sl material, the measured friction curves of all tests are consistently higher than the measured friction curves of the tests with the reference EHL tribosystem with longitudinally ground surface structures. The two tests with the sinter_ref disk both terminate due to exceeding the upper limits of ϑ_M,max after different times with sequences of short-term stable, metastable, and unstable operating behavior, approximately after t ≈ 35 h at N_2|1st ≈ 0.88 M and after t ≈ 10 h at N_2|2nd ≈ 0.25 M of the upper disk. The test with the 1^st disk pair shows stabilization of operating behavior after an initial increase in µ. This is demonstrated by the formation of plateaus in the measured friction curves. However, there is a brief decrease in µ after t > 9 h at N_2|1st ≈ 0.20 M, followed by an increase of µ with stronger metastable operating behavior from t ≥ 10 h at N_2|1st ≈ 0.22 M until the test run terminates, which leads to an overall stability coefficient of Ξ = 0.30, characterizing a metastable operating behavior, although the system is terminated. The test with the 2^nd disk pair shows a continuous increase in µ and, with this, an overall stability coefficient of Ξ = 0.15, characterizing an unstable operating behavior. In both tests with the axially ground disk pair with a sinter_sl disk, the solid lubricant additive of sinter_sl proves advantageous. In both test runs, the measured levels of µ are lower compared to the tests with the axially ground disk pairs with a sinter_ref disk, resulting in both the 1^st and the 2^nd test lasting until the end of the designated test time of t = 48 h at N_2|1st and N_2|2nd > 1.2 M. Generally, a temporarily decreasing level of µ can be observed in both tests, which leads to the stability coefficients Ξ = 0.27 and = 0.56. During the 1^st test only, the operating behavior changed from a stable to a metastable status, with rapidly increasing and then decreasing trends of µ starting from t ≈ 34 h at N_2|1st ≈ 0.83 M, with the decreasing trend continuing and the experiment not having to be terminated.

3.2.6. Summary of the Long-Term Tests

Table 6. Mean friction $\overline{\mathsf{\mu }}$_1st and $\overline{\mathsf{\mu }}$_2nd, mean bulk temperature $\overline{\vartheta }$_M|1st and $\overline{\vartheta }$_M|2nd, relative load cycles of the upper disk N_2|1st and N_2|1st, and stability coefficients Ξ_1st and Ξ_2nd of all long-term tests and maximum difference of friction Δµ_max between the long-term tests with the 1^st and 2^nd disk pairs.

Lubrication Method Upper/Lower Disk pH = 1043 N/mm2 vΣ = 4 m/s s = 20%	$\overline{\mathbf{\mu }}$1st $\overline{\mathbf{\mu }}$2nd	Δµmax	$\overline{\mathbf{\vartheta }}$M|1st $\overline{\mathbf{\vartheta }}$M|2nd in °C	N2|1st/N2|48h N2|2nd/N2|48h in %	Ξ1st Ξ2nd
once-lubrication sinterref lgr/steel mpo	0.094 0.052	0.059	92 52	<1 <1	0.07 0.07
once-lubrication sintersl lgr/steel mpo	0.051 0.047	0.057	93 86	≈29 ≈40	0.46 0.48
injection lubrication steel lgr/steel lgr	0.047 0.047	0.005	56 56	100 100	0.76 0.95
self-lubrication sinterref lgr/steel lgr	0.051 0.043	0.013	85 78	100 100	0.45 0.38
self-lubrication sinterref lgr/steel mpo	0.045 0.035	0.020	82 70	100 100	0.51 0.71
self-lubrication sinterref suf/steel mpo	0.031 0.036	0.008	62 70	100 100	1.00 0.69
self-lubrication sintersl lgr/steel mpo	0.032 0.033	0.002	65 68	100 100	1.00 1.00
self-lubrication sinterref agr/steel agr	0.073 0.057	0.041	110 92	≈73 ≈21	0.30 0.15
self-lubrication sintersl agr/steel agr	0.053 0.047	0.010	91 87	100 100	0.27 0.56
self-lubrication 1 sintersl lgr/steel mpo	0.045 0.039	0.011	86 87	100 100	0.46 0.43
self-lubrication 2 sintersl lgr/steel mpo	0.042 0.042	0.016	117 116	≈90 ≈37	0.48 0.18

¹: operating condition variation at p_H = 1200 N/mm²; ²: operating condition variation at v_Σ = 8 m/s.

shows the evaluated data of the long-term tests conducted at the reference operating conditions (see Section 3.2.1, Section 3.2.2, Section 3.2.3 and Section 3.2.5) as well as at operating conditions at increased load and increased sum velocity (see Section 3.2.4). The lifetime reached by each disk pair is shown in percentage of the reached load cycles N_2|1st and N_2|2nd compared to the aimed load cycle N_2|48h after t = 48 h. For the tests with v_Σ = 4 m/s, the load cycles of a test without termination are N_2|48h > 1.2 M, and for the tests with v_Σ = 8 m/s, N_2|48h > 2.4 M.

None of the disk pairs of the two disk pair variants tested under once-lubrication reach the aimed lifetime, whereas the disk pair with a sinter_ref disk shows a negligible lifetime with unstable operating conditions, and the disk pair with a sinter_sl disk shows a lifetime of minimum 29% to maximum 40% of the aimed lifetime with a stability coefficient of Ξ_1st and Ξ_2nd > 0.40, leading to a stable operating condition according to Ebner [25]. Further, seven of the eleven disk pair variants reach the aimed lifetime of N_2|48h > 1.2 M and > 2.4 M after t = 48 h, even the disk pair tested at an increased applied load. All the disk pairs reaching the aimed lifetime operate in stable operating conditions, except the metastable operating 2^nd disk pair with the both longitudinally ground sinter_ref and steel disks and the 1^st disk pair with the both axially ground sinter_sl and steel disks. The lowest friction results from the disk pairs with a mechanically polished steel disk and a superfinished sinter_ref disk or a longitudinally ground sinter_sl disk, with $\overline{\mathsf{\mu }}$_1st = 0.031 and $\overline{\mathsf{\mu }}$_2nd = 0.036, and $\overline{\mathsf{\mu }}$_1st = 0.032 and $\overline{\mathsf{\mu }}$_2nd = 0.033, respectively.

4. Discussion

In the following section, the influences of the material, the surface finish, and operating conditions on the self-lubricating EHL tribosystem are discussed.

4.1. Influence of the Material on the Operating Behavior

Tests were conducted on different material pairings, i.e., steel–steel, steel–sinter, and sinter–sinter material disk pairs. To evaluate the operating behavior of the EHL tribosystem according to the considered material combination, the measured friction and bulk temperature curves of the friction curve tests (see Section 3.1) and the long-term tests (see Section 3.2) as well as surface measurements are analyzed. For this, Figure 18a–d shows the surface alteration of the upper disk of the 1^st disk pair after the 2^nd friction curve test run compared to the new surfaces. Additionally, the absolute and relative difference of the arithmetic mean surface roughness ΔRa before and after the test is determined.

4.1.1. EHL Tribosystem without Oil-Impregnated Sinter Disks under Once-Lubrication and with One Oil-Impregnated sinter_ref Disk under Self-Lubrication

The tests with once-lubricated disk pairs with two steel disks or with one steel and one sinter_ref disk without oil impregnation but with the initially added oil quantity V_init could not be fully completed in both friction curve (see Figure 6a,b) and long-term tests (see Figure 15), due to continuously increasing µ and ϑ_M reaching the termination limits after a comparatively short number of load cycles (see Table 5 and Table 6). With the failure of the once-lubricated tribosystem without an oil-impregnated partner, the functionality of self-lubrication due to extruding oil of the pores supplementing the initial lubricant quantity V_init could be verified as expected and could be separated from a possible presumption that the measured frictional behavior with at least one oil-impregnated specimen is achieved just by the PAO100 oil with its PD additives. Further, the surfaces of the EHL tribosystems under self-lubrication via oil-impregnated sinter disks show less severe surface alteration compared to the once-lubricated EHL tribosystems (see Figure 18a,b), which show heavy alteration characterized by abrasively and/or adhesively altered regions of the initially very smooth polished steel disk surface after the comparatively short test runs. The heavy alteration is also verified by a significant relative increase of surface roughness of ΔRa = 366% compared to the moderate increase of surface roughness of ΔRa = 67% of the self-lubricating disk pair with one longitudinally ground, oil-impregnated sinter_ref disk. With this, an improvement of the heavy mixed lubrication regime by thermal expansion-, elastic deformation-, and centrifugal force-driven extruding oil out of the material structure into the tribocontact (see [25]) can be assumed, which can increase the lubricant film thickness to a sufficient film thickness (see [27]) to further separate the mating surfaces, which consequently results in less surface alteration.

To further evaluate the operating behavior due to surface alteration analysis during the long-term tests, surface measurements were taken at intermittent intervals (see Section 2.3). Figure 19 shows the surface images sequentially taken during the long-term test with the 1^st disk pair with the longitudinally ground steel and sinter_ref disk pair under self-lubrication. Initially, nearly no pores are optically visible in the 2D images of the 1 mm² square area in the center of the disk surface. However, immediately after the first load cycles, there is an optically recognizable increasing change in the surface appearance, with increasing visible pores in terms of number and size. By the end of the test, the surface appearance is characterized by regions smoothed due to wear but also longitudinally oriented marks scattered with numerous pores as qualitatively determined. Areas around the pores appear less stressed, indicating local oil extrusion and possible locally improved lubrication regimes, which was also seen in the lubricant film investigation with an optical ball-on-disk tribometer by Ebner [47]. To verify the subjective identification of pores based on the 2D representation of the surface images using simultaneously measured 3D data, the depth information from the 3D topography measurement of the surface can be used. Round shaped areas deeper than the reference level in the form of holes in the material surface represent potential oil extruding pores connected to the inner oil-storing pore structure. Figure 20 shows the topography measurements of the sinter_ref disk of the 1^st disk pair with a longitudinally ground steel and a longitudinally ground sinter_ref disk before the test, after N₂ = 2.5 k, and at the end of the test (here: N₂ > 1.2 M). The sequential measurements show that initially, even with the grinding marks, pores can be detected through round holes that are deeper than the marks. Thus, initially there are a few randomly distributed pores on the unworn surface. After the first load cycles, the number of identifiable pores increases slightly. In addition, an influenced area of oil transported to the surface can be seen around the pores, where the initial grinding structure is still visible, indicating locally relatively low solid contact ratios and ongoing abrasion. By the end of the test, the number of pores has increased slightly again, with the diameter of the pores appearing larger. The surface seems to be smoothed due to alteration, but the initially uniform grinding direction remains over the whole test.

4.1.2. EHL Tribosystem with One Oil-Impregnated sinter_pni Disk under Self-Lubrication

The friction curve tests with sinter_pni were automatically and abruptly terminated by the test rig acceleration detector after a few load cycles (see Section 3.1.3). Further, Figure 18d shows the surface images of the sinter_pni disk before and after the friction curve test. Before the test, the images of the sinter_pni disk surfaces of the new disks indicate a low to no presence of surface pores that enable the oil to flow across the surface, thus enabling the self-lubrication principle, leading to predicted failure of the test. The failure presumably resulted from the surface pores covered by the compound layer. Figure 21a,b shows the sectional images of a new longitudinally ground sinter_ref and a sinter_pni disk and verifies the presumption of the covered near-surface pores by the compound layer, which prevents oil flow to the surface, resulting in insufficient support for the lubricant film and reduced friction-reducing effects as with, e.g., sinter_ref disk pairs. As a consequence, the compound layer delaminated during the friction curve tests (see image of sinter_pni disk after friction curve test in Figure 18d), which led to heavy surface damage, causing significant vibration.

4.1.3. EHL Tribosystem with Two Oil-Impregnated sinter_ref Disks under Self-Lubrication

The test with self-lubricating disk pairs of two longitudinally ground, oil-impregnated sinter_ref material resulted in no improvement of tribological performance due to mainly higher levels of µ_max and ϑ_M,max in all tests with both disk pairs as well as higher differences of Δµ_max and Δϑ_M,max between the 1^st and the 2^nd tests with the 1^st disk pair compared to the reference EHL tribosystem with just one oil-impregnated sinter_ref disk paired with a steel disk (see Figure 7a,b and Figure 8a,b, Table 5 and Table 6). Surface alteration of both disk pair variants, the sinter_ref–sinter_ref disk pair (see Figure 18c) and the reference steel–sinter_ref disk pair (see Figure 18b), can be seen in a subjectively determined local increase of pore numbers, pore size, and regions with tribofilms, as well as an objectively measured relative increase of ΔRa = 40% and ΔRa = 67%, respectively. A disk pair comprising two oil-impregnated sinter disks exhibits less benefits in tribological behavior, i.e., primarily repeatable low levels of µ and ϑ_M, than a disk pair with one steel disk and one sinter disk. Consequently, the latter disk pair is considered the most effective. This can be due to a one-sided flow of the oil extruding from the surface for EHL lubricant film formation as well as intrusion back from the tribocontact under high hydrodynamic pressure into the porous material. Presumably because of the back-intrusion in permeable bodies, a smaller lubricant film thickness-reducing effect dominates the tribocontact compared to the tribocontact within an EHL tribosystem with two porous structures and thus a two-sided oil intrusion with a higher total permeability of the tribocontact, which led in Ebner’s [25] calculations of self-lubricating EHL tribosystem configurations with increasing permeability to a decreasing lubricant film thickness.

4.1.4. EHL Tribosystem with One Non-Impregnated sinter_sl Disk under Once-Lubrication or One Oil-Impregnated sinter_sl Disk under Self-Lubrication

Sinter_sl material with MoS₂ and WS₂ solid lubricant additives showed in once-lubricated long-term tests at the reference operating conditions compared to the unstable running once-lubricated steel–steel tribosystem a longer lifetime up to N₂ = 0.35 M at moderate operating conditions with at least half of the lifetime running with stable operating behavior before failing, with decent increasing friction and bulk temperatures (see Figure 15). With this, the expected solid friction reducing influence of the solid lubricant in the sinter material is proven. When impregnating the sinter_sl material additionally with oil, the effect of solid friction reduction via a solid lubricant as well as liquid lubrication via extruded oil out of the pores are combined, which repeatably leads to the lowest levels of µ and ϑ_M measured, with the lowest difference between the tests with two identical disk pair specifications, and continuously stable, quasi-stationary operating behavior (see Table 6).

4.2. Influence of the Surface on the Operating Behavior

Test were also conducted on different surface finish pairings, i.e., longitudinally and axially ground, superfinished, and mechanically polished. To evaluate the influence of surface finishes on the frictional behavior, stability of the operating behavior, and lifetime behavior of self-lubricating rolling-sliding contacts with oil-impregnated sintered material, again the measured friction and bulk temperature curves of the friction curve tests (see Section 3.1) and the long-term tests (see Section 3.2) as well as surface measurements are analyzed. For conventionally lubricated EHL tribosystems, the direct link between surface roughness and frictional behavior is commonly known. For self-lubricating EHL tribosystems with oil-impregnated sintered materials, the study suggests that smoother surfaces may result in lower friction, but surface finishing can affect the possible deviations of surface porosity and thus the prevailing oil flow within the tribocontact. This, in turn, affects the global and local lubrication and friction regime. For this, Figure 22a–c shows the surface alteration of the upper disk of the 1^st disk pair after the 2^nd friction curve test run compared to the new surfaces, and the absolute and relative difference of the arithmetic mean surface roughness ΔRa before and after the test is determined.

4.2.1. EHL Tribosystem with Ground Steel Disk and Sinter Disk Surfaces under Self-Lubrication

Tests with ground EHL tribosystem partners were conducted with longitudinally and axially ground disk surfaces.

The tests with the reference tribosystem with longitudinally ground steel and sinter_ref disks show in the friction curve tests a typical tribological behavior of injection-lubricated rolling-sliding contacts, with a transition from a linear increase of the coefficient of friction at a low slip ratio to a thermal regime with decreasing coefficient of friction at a higher slip ratio, as well as lower values of µ and higher ϑ_M with increasing v_∑ (see Figure 3a,b and Figure 4a,b). All operating points can be tested in both test runs with the 1^st disk pair as well as with 2^nd disk pair. In long-term tests under reference conditions, the reference tribosystem (see Figure 14) shows slightly higher µ and ϑ_M compared to injection-lubricated disk pairs (see Figure 13), which can be caused by less cooling via convection by the surrounding external oil as well as lower oil amounts for sufficient lubricant film formation. However, the reference tribosystem reaches the aimed end of the test at N₂ > 1.2 M without termination. The tests of the self-lubricating EHL tribosystem with an axially ground steel and sinter_sl disk with both surfaces similar to gear tooth flank surfaces could be completed without termination (see Figure 17) as well as with a proper tribological performance with decreasing levels of friction with test progress and a mostly stable to metastable operating behavior. The EHL tribosystem with a steel disk and a sinter_ref disk shows severe tribological performance with a metastable to unstable operating behavior, leading to friction and bulk temperature increase with test progress and, thus, termination at N2 < 0.88 M and < 0.25 M.

4.2.2. EHL Tribosystem with a Smoothed Steel Disk Surface Paired with a Longitudinally Ground Sinter Disk Surface under Self-Lubrication

The variation in the steel disk surface, from ground to polished finish, expectedly results in a significant improvement in the tribological behavior of a self-lubricating contact with a ground oil-impregnated sinter material. This improvement can be seen in the better repeatability of the 1^st friction curve tests of the two disk pairs, as well as the lower difference between the 1^st and 2^nd test with the 1^st disk pair and the overall slightly lower levels of µ and ϑ_M compared to the reference tribosystem (see Figure 9a,b and Figure 10a,b and Table 5). Compared to the operating behavior of the injection-lubricated EHL tribosystem in the long-term tests (see Figure 13), the levels of µ and ϑ_M with the self-lubricating disk pair with a polished steel disk and a longitudinally ground sinter_ref disk are lower for part of the test duration up to the whole test duration. This might be mainly caused by the improved lubrication regime, due to lower arithmetic mean surface roughness of the contact partners. Because there is no heat convection by an external oil, the friction-induced temperature of the tribosystem increases during operation, which decreases the oil viscosity and thereby increases the influence of the surface roughness of both mating partners on the lubricant film thickness. This consequently leads to an increase of solid friction and thus an increase of total friction measured. This can also be seen in the documentation in Figure 22a of the surface alteration, in which the surface of the ground sinter_ref disk shows nearly no alteration in terms of tribofilm formation, pore number, or size increase as well as ΔRa (ΔRa = 3%).

4.2.3. EHL Tribosystem with a Smoothed Steel Disk Surface Paired with a Smoothed sinter_ref Disk Surface under Self-Lubrication

The friction curve test results with an additional polished steel smoothed surface of the sinter disk also show the trend of an improved tribological behavior of a reduction in surface roughness by superfinishing or polishing, resulting in lower measured levels of µ and ϑ_M compared to tests with conventionally ground surfaces (see Figure 9a,b, Figure 10a,b, Figure 11a,b and Figure 12a,b and Table 5). In the following sections, the influence of the finish variants of mechanically polishing and superfinishing of the sinter_ref disk surface on the operating behavior is evaluated in detail.

Superfinishing of the sinter_ref Disk Surface

In addition to the results that the tests with disk pairs with a superfinished sinter_ref disk surface show better repeatability of the 1^st tests of two disk pairs and lower difference of the measured µ and ϑ_M between the test of the 1^st and the 2^nd disk pair compared to the reference system as well as the lowest levels of µ and the most stable operating behavior during the long-term tests compared to the other surface variants in Figure 14a,b (see also Table 6), Figure 22b shows the low surface roughness increases with ΔRa = 18%, with an absolute increase of ΔRa = 0.006 µm, which is negligible. Further, Figure 23 shows the 2D image documentation of the surface alteration during the long-term tests at reference conditions with the mechanically polished steel and superfinished sinter_ref disk pair. Similar to the tribosystem with two longitudinally ground disks, at the initial state, only a few surface pores are visibly recognizable. Additionally, there is a less uniform direction of the surface structure obtained. After a few load cycles, the surface pores are numerous and increase in size, whereas the initial surface structure shows for a longer time less global alteration. Furthermore, the surface area surrounding the pores shows significant deviations locally in the pore-free surface regions. The initial state of the altered porosity and surface modification after the first load cycles persists, with a similar appearance throughout the whole test. Figure 24 shows, additionally, the topography measurements of the sinter_ref disk of the 1^st disk pair with a mechanically polished steel disk and a longitudinally ground sinter_ref disk before the test, after N₂ = 2.5 k, and at the end of the test. The measurements confirm the observation, as evidenced by the identifiable pores, verified by depth information, which initially increase slightly and then remain constant over time. The less severe alteration of the disk surfaces can be interpreted as another result of the comparably best tribological behavior evaluated within the surface variants, which again verifies Ebner’s findings [25].

With the surface images and topography measurements of the best-performing disk pair surface combinations, it can be stated that in tests that show good tribological behavior and do not lead to premature termination due to metastable or unstable operating behavior, the porosity becomes more uncovered initially and, thus, is more exposed, which can be interpreted as a running-in process of oil-impregnated sinter materials with surface finish. Subsequently, once there is sufficient oil exchange between the inner material structure and the tribocontact through the opened surface pores, there are no significant changes in surface porosity.

Mechanical Polishing of the sinter_ref Disk Surface

In contrast, the friction curve test with the comparably smoothest sinter_ref disk, a mechanically polished surface, shows greater difference between the 1^st and 2^nd tests with the 1^st disk pair and a worse repeatability of the tests with two disk pairs compared to an EHL tribosystem with a mechanically polished steel disk and a longitudinally ground or superfinished sinter_ref disk. It is postulated that the manual polishing process results in inconsistencies in abrasive material removal and pore closure. While this facilitates the requisite smoothing of the surfaces and thus a comparatively high increase of the minimum lubricant film thickness, it also gives rise to unexpected and divergent measurement results due to local surface porosity emphasis. Further, the additional effort of mechanically smoothing the sinter disk surface is less useful when evaluating the surface alteration shown in Figure 22c, which shows a significant optical change of the surface with a measured increase of ΔRa = 380%. With this, the surface finish of mechanically polishing the sinter disks was excluded for the subsequently run long-term tests.

4.3. Influence of the Operating Conditions on the Operating Behavior

Tests with disk pairs of a mechanically polished steel disk and a longitudinally ground sinter_ref disk were also conducted with a comparably higher load or sum velocity. To evaluate the operating behavior of the EHL tribosystem according to the prevailing operating condition, the measured friction and bulk temperature curves of the friction curve tests (see Section 3.1) and the long-term tests (see Section 3.2) as well as surface measurements are analyzed.

4.3.1. EHL Tribosystem under Comparably Higher Load

The increase in transmitted power during the long-term tests at higher loads initially exhibits areas with metastable operating behavior, which transitions into stable operating behavior after further load cycles in both tests of the two disk pairs (see Figure 16). Initially, there is a measurable difference of µ between the tests of the two disk pairs, which gradually converges to nearly identical stable operating behavior at the same level. Thus, the influence of load increase in long-term tests results in a slightly higher level of both μ and ϑ_M compared to the tests at the reference load (see Table 6), but a functional tribosystem is still maintained throughout the entire intended test. The increased measured friction is primarily attributed to the higher specific friction in the tribocontact at higher pressure.

4.3.2. EHL Tribosystem under Comparably Higher Sum Velocity

In contrast to the increased load, the impact of an increased speed of v_Σ = 8 m/s exerts a more pronounced influence on the operating behavior of the self-lubricating EHL tribosystem. During the tests shown in Figure 16, the severe operating condition results in varying lifetimes, and in both tests with the two disk pairs, the test runs fail due to the emergence of unstable operating behavior at disparate times. The primary cause of this phenomenon is not only the higher specific friction power but also the increased dissipated energy generated over time, which results in increased system heating and a concomitant decrease of oil viscosity. This decrease of viscosity also reduces the lubricant film thickness, potentially increasing solid contacts. Further, Figure 25 and Figure 26 show, through 2D images and 3D topography measurements, the surface alteration of the sinter_sl disk of the 1^st long-term tests at the increased sum velocity of v_Σ = 8 m/s with the disk pairs of a polished steel and a longitudinally ground sinter_sl disk pair that terminated before the designated test time. Visually, there is only a slight alteration in the appearance of the ground surface within the first load cycles up to about N₂ ≈ 100k. However, at N₂ = 250k, corresponding to an abrupt decrease in µ, the porosity increases significantly in both number and size. At all subsequent time intervals when the surface condition was recorded, the increasing porosity correlates with the increasing measured friction until the tribosystem fails, resulting in highly pronounced surface damage characterized by numerous visible and measurable large surface pores as well as, presumably, severe wear.

With the surface images and topography measurements of the disk pairs that terminated during the long-term tests, it can be stated that in tribosystems that experience more severe tribological stress, due to factors such as the initial limited presence of surface pores or operating conditions that exceed the performance limit of the self-lubricating EHL tribosystem, porosity increases significantly. This, in turn, enables more oil to intrude from the tribocontact back into the inner material structure, reducing the lubricant film thickness. As a result, solid contacts increase, leading to increased friction and wear as well as temperature increase, which reduces oil viscosity, further promoting the formation of thinner lubricant film thickness and enhancing oil intrusion. Under these conditions, the EHL tribosystem will continue to degrade, leading also to thermal and chemical alteration of the oil with further load cycles, until the EHL tribosystem reaches its limits and fails.

The consistently good measurements of the tribological behavior of the self-lubricating EHL tribosystem at moderate to high reference operating conditions, together with the stable or metastable operating behavior at higher pressures and/or speeds, suggests the suitability of self-lubricated tribosystems for use in various operating modes, including, e.g., intermittent high-load operation followed by less loaded or unloaded cool-down periods.

5. Conclusions

This study investigated the functionality and influence of specific material and surface specifications on the operating behavior of oil-impregnated, self-lubricating sintered rolling-sliding contacts under moderate to high operating conditions. A twin-disk tribometer was used to evaluate the tribological behavior of various material disk pairs and surface finishes. Friction curve tests were performed to measure the coefficient of friction and bulk temperature, while long-term tests were performed to evaluate the lifetime behavior of promising material disk pairs and surface finishes, including friction and bulk temperature measurements. The main findings are as follows:

• A once-lubricated EHL tribosystem without external oil lubrication performs better when a solid lubricant is added to the sintering powder prior to sintering.
• A self-lubricating EHL tribosystem with longitudinally ground surfaces and one oil-impregnated sinter disk exhibits adequate friction behavior and, thus, mostly stable operating behavior during the tests. Two oil-impregnated sinter disks show worse tribological performance.
• Smoothing the surface of the steel disk in the self-lubricating EHL tribosystem improves tribological performance, while finishing both contact partner surfaces by mechanical polishing of the steel disk and superfinishing of the sinter disk provides the best repeatable tribological performance.
• For axially ground surfaces similar to typical gear tooth flanks, the best tribological performance of a self-lubricating EHL tribosystem can be achieved with a solid lubricant addition to the sintered material prior to sintering.
• The surface alteration during the tests with EHL tribosystems with good tribological performance is characterized by a consistent running-in-like increase in pore size and quantity within the initial load cycles and an only slight additional change during the further lifetime.
• Nevertheless, great differences in operating behavior and surface alteration are observed for disk pairs manufactured with identical requirements for mechanical properties and surface conditions.

Systematic studies of material and surface porosity, the resulting permeability, and the influence of manufacturing methods such as heat treatment, surface finishing, etc., on the surface porosity as well as on the sinter material density in near-surface regions are further needed to understand these differences in operating behavior and surface alteration.

The tests of this study were conducted under stationary operating conditions. The results demonstrated that, for a sufficient number of load cycles, self-lubricating EHL tribosystems can operate within specific load ranges, with a highly stable operating behavior and low levels of µ and ϑ_M. A stationary operating application with medium-high loads, such as geared motors for logistic conveyor belts, could be a viable option. Further investigations at adapted model contact tests are required to ascertain the suitability of the self-lubrication technology for promising applications under intermittent operating conditions with peaks of higher load or speeds and operating sequences under low load to cool down, e.g., in machine tools such as hand screwdrivers. In addition, the so-far limited characterization of surface porosity reveals the need for a quantitative method to determine pore size distribution and quantity, allowing for a more objective characterization of surface porosity, as the surface is considered one of the main components of future mathematical models for the design or prediction calculations of self-lubricating EHL tribosystems with oil-impregnated sintered materials.