Effect of Low Concentrations of Organophosphorus Additives on Tribological Performance of Polyalkylene Glycol-Based Oils for Tin Bronze on Tungsten Carbide Applications

Abstract

Low concentrations of organophosphorus additives in PAG oil for friction units are under consideration. Concentrations of 0.1–0.5 percent maintain the environmental friendliness of the oil, but can improve its tribological properties. The friction pair (tin bronze versus tungsten carbide) was made from a plunger and compressor seal. It was tested in the reciprocating sliding mode under a load of 600 N (contact pressure ≈ 13 P), a frequency of 8.33 Hz, an amplitude of 15 mm and a temperature of +70 °C. It was found that the phosphate concentration of 0.2 percent provides the most stable values of the friction coefficient (0.04). The mechanism of action of the modifier was determined by analyzing the results of SEM and profilometry. Organophosphates make both interacting surfaces smoother; and there is an optimal concentration of additives that provides optimal roughness.

1. Introduction

Choice of an optimal combination of additives to base oils to give them required set of performance properties is still one of the most pressing research topics, since friction units are constantly being modernized and their number is growing. The efficiency of oils is determined by the additives used to reduce friction and wear, which can be of different nature. Their diversity is due to the search for optimal solutions for specific friction units. Numerous studies are aimed to analyze the effect of new additives on the tribological characteristics of lubricants. Several areas of these studies can be identified.

One of them is the development of carbon materials as additives. In [1], fluid-like graphene oxide organic hybrid materials (FL-GO) are studied as an additive to polyethylene glycol oil (PEG). Tribological studies made on four-ball friction tester demonstrated a significant decrease in wear and friction due to the addition of FL-GO to PEG. In [2], mesoporous carbon nanoparticles were used as a lubricating additive to PEG, which showed an increase in antiwear properties and a decrease in friction at concentration of 0.7 wt.%. In [3], a new eco-additive to lubricating oil based on carbon quantum dot particle-doped nickel is proposed, which improves the lubricating properties of PEG oils.

Another important area is the use of some ionic liquids (IL), which are often used as additives to base oils and are currently being actively studied, despite their high cost and insufficiently studied environmental friendliness [4]. Most researchers preferred standard laboratory tribological tests, the results of which are explained using modern methods of friction surface analysis (scanning election microscope, SEM, with energy-dispersive X-ray detector or X-ray photoelectron spectroscopy (EDX, XPS), in order to study the composition and structure of the formed tribolayer. For example, in [5], a comprehensive study of ionic liquids based on phosphonium and organophosphorus compounds (Phosphonium-organophosphate ionic liquids) allows to conclude that ionic liquid has great potential as an antiwear lubricant additive. In [6], nanotribological tests were performed on atomic force microscope (AFM), based on the results of which the authors concluded that 1 mol % solution of IL dissolved in oil lubricates the silica surface as effectively as pure IL, which makes them suitable for large-scale industrial applications. Some researchers consider the behavior of IL in combination with other additives. In [7], the compatibility of various IL with organic friction modifier was studied, and both synergistic and antagonistic effects were noted depending on the composition of the IL. In [8], the interaction of IL with common friction modifiers (molybdenum dithiocarbamate) and dispersants (polyisobutene succinimide) was assessed; it was shown that, depending on the chemistry of the IL, stable ultra-low boundary friction or the opposite effect was observed. In [9], it was shown that phosphonium organophosphate IL in combination with zinc dialkyldithiophosphates (ZDDP) provides excellent tribological characteristics when using engine dynamometer testing. But at the same time, a negative impact of ZDDP additive on engine filters is also noted. In [10], the lubricating ability of engine oil was experimentally studied in the presence of combination of IL and ZDDP antiwear additive. The oil was tested on tribological stands and on an engine dynamometer. The results demonstrated improved wear protection and reduced friction compared to the use of ZDDP alone. In [11], the relationship between the mechanical properties of tribolayers formed using IL, ZDDP and their combinations and their antiwear characteristics was studied, emphasizing that higher elastic moduli of tribofilms, as a rule, lead to better friction and wear indicators.

It is worth noting that some researchers initially pay attention to the environmental aspects of the additives being developed [4,12,13]. In [12], the toxicity and biodegradability of various IL and traditional additives, such as ZDDP, were assessed. Some of the IL showed better combination of tribological and environmental properties than ZDDP. In [14], ILs that do not contain halogens and other harmful substances were studied, while a decrease in tribocorrosion, friction and wear was recorded.

It should be noted that ILs can be used as multifunctional additives not only to PEG, but also in poly-alpha-olefin oils (PAO). In [15], based on the results of tribological studies, it was shown that ILs improve the antiwear characteristics and load-bearing capacity of PAO due to the formation of a boundary lubricating film of FePO₄ and nitrogen. In addition, in [16], it was shown that when using experimental oils containing ILs, contact fatigue is reduced compared to commercial transmission oil SAE 75W-90 in rolling-sliding tests.

As an alternative to developing additives based on ionic liquids or using new carbon materials as an additive to traditional synthetic oils, it is possible to consider using environmentally friendly base oils, such as PAG oils. This raises the task of choosing optimal additive to impart the required properties to the lubricant composition. Low concentrations of additives in PAG oil, on the one hand, improves its tribological properties, and on the other hand, maintains the relatively high environmental safety of such oil. For example, in [17,18] it is proposed to use a non-traditional polyalkylene glycol (PAG) as an oil base. It is obtained that PAG oil itself is good for friction and fuel economy, but greater reduction in wear and the elimination of piston deposits require the development of additional additives.

Typically, in industrial oils, the content of additives is in the range of 1–10 wt.% [19,20]. Here we study the effect of low concentrations (less than 1 wt.%) of organophosphorus additives in PAG oil. This will allow to obtain oils that can be used in the production of food polyethylene. On the other hand, by influencing friction, additives can improve the performance properties of oils intended for use in high-pressure compressors both for lubricating the plunger-sealing ring friction unit and for preventing leaks of working gas from the compression chamber.

2. Materials and Methods

2.1. Lubricant Composition

Synthetic polyalkylene glycol (PAG) was used as a base oil. It was obtained by copolymerization of ethylene oxide and propylene oxide at weight ratio of 70–30, respectively, in the presence of an alkaline catalyst. The initiator of the anionic polymerization process was a dihydric alcohol (ethylene glycol). PAG oil is (Figure 1) a mixture of molecules of statistical copolymers of ethylene oxide and propylene oxide (randomly distributed monomer units); at the ends of the polymer molecules are hydroxyl (-OH) groups (since PAG oils were obtained by alkaline catalysis). PAG oil molecules consist of linear/branched chains (but not cross-linked structures). The base PAG oil was supplemented with an antiwear organophosphorus component (mixture of phosphoric acid esters) in various dosages (See

Table 1. Properties and formulation of test oils.

	Base Oil	1	2	3	4	5
Density (20 °C) [g/cm3]	1.087	1.087	1.087	1.087	1.087	1.088
Viscosity index	233	233	230	229	228	228
Kinematic viscosity (40 °C) [mm2/s]	260.00	259.39	259.01	258.65	258.38	257.91
Kinematic viscosity (100 °C) [mm2/s]	45.27	45.01	44.38	44.02	43.89	43.89
Additive content (wt.%)
- Phosphorus additive	-	0.1	0.2	0.3	0.4	0.5

). The required amount of additive was included to the base oil heated to 50 °C. After that, the mixture was stirred for 1 h at a temperature of 50–60 °C. The oil was homogeneous and transparent. The phosphorus-containing organic additives are quite polar compounds, so they dissolve well in polar polyalkylene glycol oils.

2.2. Friction Pair Materials

Since the purpose of the experiments was to simulate the operating conditions of a real high-pressure compressor 4M40M, the test samples were made by the method of electric spark cutting from natural elements of the friction unit of this compressor. For this hyper compressor solid tungsten carbide plungers are used (with diameter of 85 mm and length of 1115 mm). Sealing rings made of tin bronze (3.7 mass percent tin) are used as sliding guides. Figure 2a shows plunger cut in half with sealing rings put on it (the compressor uses 5 sealing rings located at an equidistant distance along the entire length of the plunger). Figure 2b shows a sample cut from a plunger measuring 25 × 20 mm and a tin bronze sample measuring 6.9 × 9 mm. Figure 2c shows the friction pair in the tribometer holders (outside the lubricant bath). The upper holder allows for self-alignment of the tin bronze sample relative to the plunger sample.

2.3. Equipment

In order to evaluate the tribological properties of compressor oils, a laboratory tribometer UMT (CETR, Td. Bruker, Campbell, CA, USA) was used with a scheme of reciprocating motion of the lower sample relative to the fixed upper one. A model sample made of sealing ring from tin bronze is used as the upper sample.

The configuration of the tribometer is shown in Figure 3. The lower sample 2 is in a stainless steel lubricating bath (printed on SLM 3D printer TruPrint 1000 (TRUMPF, Ditzingen, Germany)), which is fixed on the object table (position 3 in Figure 3a); the drive 4 ensures the reciprocating motion of the table (position 3 in Figure 3a). The upper sample is in the holder 1 (see Figure 2), which in turn is mounted on the elastic element 9. The normal load is transmitted to the sample through the elastic replaceable element by screw transmission 5, which is driven through a gearbox 6 by an electric motor 7. The elastic element 9 is a 2-component sensor designed to measure the normal force and friction force (Fz and Fx) acting on a sample.

The measuring system of the device is equipped with feedback of the electric motor 7 with the force sensor 9, which allows maintaining the specified normal load Fz at a constant level. The system of the device generates a mismatch signal if there is a change in the linear size of the interacting samples along the direction of the normal load Fz. One of the reasons for such a change is wear. The mismatch signal is sent to the drive of the screw transmission, which shifts the nut 5 along the guides 8 until the previous value of the normal load Fz is ensured. Thus, the loading system of the tribometer allows testing samples at a specified constant load.

The studies can be realized in the frequency range of 0.1–60 Hz and normal loads from 5 to 650 N in air or in liquid medium. The friction path of the sample is specified based on the task of the study. The measuring system provides continuous of the normal force, the approach of the wearing bodies, friction force, friction path and friction coefficient. The device can be equipped with thermal chamber that ensures heating of the lubricant in the bath from room temperature to +130 °C. Heating of samples and maintaining their temperature occurs automatically. In this study, the test temperature was +70 °C.

2.4. Methodology of Experiment

When developing the testing method for lubricants, on the one hand, the compressor operating modes (temperature, frequency, specific load) were taken into account, and on the other hand, the duration of a single experiment was chosen based on the following considerations. Firstly, at the beginning of the experiment, the running-in process goes on for an indefinite period of time, which is caused by a change in the contact area between the rubbing surfaces with tendency for equilibrium roughness to arise, as well as the appearance of tribofilms on the interacting surfaces. Secondly, upon reaching the steady-state friction mode, the friction coefficient must be measured at a certain (uniform for all experiments) time interval. Taking into account the above, we will describe the testing procedure.

Before testing on the tribometer, the test samples, the upper sample holder and the lubricant bath were washed in an ultrasonic bath for 10 min successively in two solvents, namely 95% ethyl alcohol and then in isopropyl alcohol. After washing, the samples were dried under a stream of air created by a 12 V fan at room temperature for 15 min.

The cleaned lower sample made of plunger is installed in the lubricant bath, which is mounted on the tribometer stage. Using disposable plastic pipette, 4 mL of the tested lubricant is added to the lubricant bath with the lower sample in it. The tribometer stage is located in a thermal chamber, which is closed with a lid having an inlet for inserting the upper holder with the sample into the chamber. The upper stationary sample in a self-aligning holder is placed into the thermal chamber and, using the tribometer automation, is brought to the lower sample at a distance of 0.2 mm, completely immersing it in the lubricant. After this, slow heating of the friction pair is switched on. The heating time with subsequent holding is 45 min (see Figure 4).

To study the duration of the running-in stage, a series of preliminary experiments were carried out. Based on their results, it was established that a two-step procedure is required to achieve stable contact before the main experiment. The duration of the steps was found empirically, and the essence of the running-in procedure is a gradual increase in contact pressures to a given level. The sequence of steps is shown in Figure 4.

In Figure 4, in Stage I, the normal load Fz = 10 N is set, which means that the samples are loaded to 10 N. Upon reaching this load, the upper holder is raised by 0.2 mm and then held for 45 min. The heater of the thermal chamber is switched on. This stage allows to heat the oil and the friction pair up to +70 °C. Stage II contains 4 steps. Step 1, lasting 5 s, allows the experiment to be started. That is, when the load reaches 100 N and it stabilizes after a short time (5 s), the friction process begins. This step allows to bring the sample into the contact with self-adjustment of the upper tin bronze sample relative to the lower moving sample. Step 2 and 3 simulate the running-in of a new friction pair on a compressor. Step 2 is accelerated running-in at idle speed of the compressor, and then step 3, i.e., running-in at partial load. The duration of step 3 is 10 min, since the main change in the friction coefficient occurs during this time (see example curve for oil 1, Figure 5). This conclusion is also supported by the acoustic emission (AE) signal, which shows that a significant change in noise is observed during the first 5–10 min. Based on the experiments, it has been obtained that running-in period of 10 min is sufficient to stabilize the contact before the long-term 60-min experiment.

In Step 4, the qualifying experiment is performed, the duration of which (60 min) was limited both due to the preliminary tests and according to the ASTM D6079-11 standard [21], which regulates the total testing time to be limited to 75 min. The running-in takes 15 min, and the qualifying test takes 60 min. It should be noted that the normal load was selected based on the maximum possible compression force for the tribometer equal to 600 N (contact pressure ≈ 13 MPa). The amplitude of the lower sample movement was 15 mm (A = 0.0075 m) and guaranteed the oil supply after each displacement of the lower sample from under the upper one (having a width of 6.9 mm). The oscillation frequency was 500 rpm (8.33 Hz).

Figure 5 shows, as an example, the time dependences of the friction coefficient COF, oil temperature T1 and the walls of the thermal chamber TR, as well as acoustic emission AE and normal load Fz for oil 1.

To calculate the average values of the friction coefficient, a section with a steady-state friction mode was used (from 2400 to 3600 s, which is the last 30% of the experiment). For a reliable assessment of the friction coefficient, the experimental technique assumed three repetitions of the experiments for each sample of lubricant. In this case, the range of the average value of the random value of the friction coefficient from experiment to experiment can be significantly smaller or larger than the range of random points of the friction coefficient directly in the experiment. This fact is taken into account in the statistical processing of the measured data.

The measurements of surface topography of the sample from tin bronze were made using 3D confocal profiler S Neox (Sensofar-Tech, Barcelona, Spain) with motorized X-Y stage and 20X objective. Stitching was used for obtain an image of a surface area of 6.9 × 9 mm. In order to obtain reliable results, three measurements were taken in three different samples. Before determining the surface roughness parameters, the shape component (waviness) was removed from the data by the method of plane subtraction at three points. A smart filter was used to eliminate artifacts and noise. The roughness parameters were estimated according to the ISO 25178-605:2025 standard [22], which involves determination of three-dimensional height parameters of topography using tools included in the profiler software.

The surface of the samples was examined using a Quanta-650 scanning electron microscope (SEM) (FEI, Hillsboro, OR, USA) in high vacuum (10⁻⁵–10⁻⁶ Torr) at accelerating voltages of 5 kV. Elemental composition analysis was performed using energy-dispersive X-ray spectroscopy (EDX) at the same parameters. The spectrum was collected from the entire image area.

Additionally, the microhardness of tin bronze samples was measured (plunger hardness HV = 1004) using Vickers micro-hardness tester HMV-2000 (Shimadzu, Kyoto, Japan) under a load of 0.010 kg at 10 s.

3. Results

Figure 6a,b show typical diagrams of the friction coefficient versus test time (all repetitions) for the base oil and for sample 2. For all tested oil samples, the running-in process is short, then the friction coefficient is either established or demonstrates a very slow decrease. The average established values of the friction coefficient and their reproducibility in the experiment can be estimated from Figure 6c. The most stable result was demonstrated by sample 2; all repetitions showed a value of about 0.04. Other samples, including the base oil, have a significant spread of values (both below and above 0.04).

The topography of friction surface of tin bronze can be estimated based on the data shown in Figure 7. Figure 7a–f show images obtained with optical 3D profilometer, where the darkest areas (dots or lines) are surface damage. The smallest number of these areas is recorded on the friction surfaces with lubricant samples 2 and 5. The results of the quantitative assessment of the roughness are shown in Figure 7g. The presence of the modifier leads to decrease in roughness. Analysis of Figure 7g shows that, starting with the second sample, Sa stabilizes (with standard deviation being almost zero), and Sz decreases with an increase in the amount of modifier, which indicates smoothing of the friction surface.

The tin bronze samples were examined using scanning electron microscopy with EDX analysis. The results for samples 2 and 5 are presented in

Table 2. Friction surface of bronze. Elemental composition.

Counter Body	Content of Element, wt.%
	C	O	P	W	Cu + Sn
initial	2.4	1.0	-	-	bal.
2	9.0	4.0	-	8.0	bal.
5	6.1	2.2	-	22.9	bal.

. These two types of oil were chosen to show the surfaces for the most positively demonstrated sample (2) and for the sample with the highest concentration of the modifier. On the friction surface, compared to the initial one, the amount of carbon and oxygen slightly increased. Tungsten appeared on the friction surfaces, the nature of which is unambiguous—these are the products of wear of the plunger material. It is interesting that there is almost three times more of it on the surface of sample 5 than on sample 2. Phosphorus, indicating the presence of modifier on the surface, was not detected.

The surface of the sample made from the plunger was also examined using SEM. In all cases, the data were obtained in the region between the dead center and the middle of the stroke length. Figure 8 shows the images of the initial surface and friction surfaces using oil samples 2 and 5. The elemental mapping is also presented. The reflectivity of the surfaces in Figure 8b,c clearly indicates smoothing of the surface during friction, which is more active for the oil sample with the largest amount of modifier. The friction surfaces are partially covered with layers. As the copper and tungsten maps show, this is the result of the transfer of the counterbody material. Visually and according to

Table 3. Friction surface of tungsten carbide. Elemental composition.

Sample	Content of Element, wt.%
	C	O	P	Cu	W
initial	11.2	4.3	0.02	-	bal.
2	8.9	2.5	0.1	35.2	bal.
5	8.0	3.3	0.2	52.2	bal.

with the results of the elemental analysis, it is clear that there are more of these layers when using oil 5. Interestingly, phosphorus is present on the surface free of layers. It occupies a larger area in the case of sample 2, but in percentage terms (see Table 3) it is higher when using sample 5.

The hardness of the tin bronze samples was measured after testing in the friction zone, which is clearly visible on the surface (see Figure 9a). In some cases, based on microhardness measurements, it is possible to draw conclusions about the damage of the surface layers of the material. In this case, the hardness does not depend on the oil number (the difference is within the standard deviation, see Figure 9b) and corresponds to the initial value.

4. Discussion

The group of PAG samples have friction coefficient close to each other (µ ≈ 0.04), but different repeatability. Analysis of the obtained data allows us to conclude that oil 2 has the lowest friction coefficient and the best repeatability. The steady-state friction mode for this sample is achieved much faster in comparison with other oils.

The modifier is part of the surface tribofilm. It is interesting that it was found at the surface of tungsten carbide, but not at the tin bronze surface. It is likely that the structure of the ceramic surface is more conducive to retaining the tribofilm when the contact is opened than the structure of the tin bronze surface. Another possible reason is that tin bronze wears faster than tungsten carbide. It usually takes time for a tribofilm to form. Confirmation of these hypotheses could be obtained by examining the surfaces when the contact is opened at different stages of the test. It is also productive to obtain additional information about tribofilms using Raman microscopy or XPS using the methods proposed in [23,24]. But even based on the range of studies that have been conducted, it can be stated that the amount of additive affects resulting roughness. Probably, the tribochemical reaction leading to smoothing of the unevenness is more active, the greater the load on a given single unevenness. This can explain the decrease in Sz on tin bronze surfaces with an increase in the amount of modifier. The surface of tungsten carbide samples is also smoothed; the effect is stronger the more additives are present in the oil. It can be assumed that there is a roughness that is optimal for this type of friction contact. It provides the necessary amount of tribofilm and lubricant in the gap and partially prevents the adhesion of counterbody particles to the surface. Judging by the experimental results, the surface obtained using oil 2 turned out to be closest to this optimum.

5. Conclusions

A tribological testing technique has been developed that allows determining the friction coefficient of compressor oils using the kinematic diagram of the plunger- seal contact. The technique provides dependences of the friction force on time for the contact of two movably mating surfaces of model samples made from a full-scale compressor friction unit. Series of tests of compressor oils were analyzed, the results showed that:

• The friction coefficient depends on the amount of modifier in the tested oil and differs both in the magnitude of its value and in their random scatter at steady-state friction.
• Oil 2 demonstrates the lowest friction level (µ ≈ 0.04) with a minimum scatter of random values of the friction coefficient both between experiments and in the experiment itself, i.e., it is stable.
• The mechanism of action of the modifier is smoothing the surfaces of interacting bodies; there is an optimal amount of modifier, which provides an optimal balance between smoothing and adhesion. In our study, this concentration is almost the lowest (0.2 percent), which is important for maintaining environmental friendliness.

From the point of view of fundamental tribology, it would be interesting to study in more detail the tribofilm with additives, its properties, rate of formation and interaction with wear particles.