Influence of PMA on the anti-scuffing properties of AW/EP additives

Scuffing is becoming a quite common failure mode in gears and bearings. It has been shown that AW/EP additives are effective in preventing scuffing, but only if they are able to form a thick tribofilm before encountering severe scuffing-type conditions. This study has employed a contra-rotating, step-sliding speed scuffing test to explore the impact of PMAs on the ability of ZDDP and a commercial SP additive-containing package to prevent scuffing when subjected to immediately severe conditions. It is found that some PMAs can greatly enhance the anti-scuffing performance of these AW/EP additives. They do this by forming thick, adsorbed boundary films that can withstand high speed sliding conditions and protect the rubbing surfaces long enough for tribofilms to form.


Introduction
Scuffing was first identified as a problem when hypoid gears were introduced into the rear axle gearboxes of automobiles in the late 1920s [1].Such gears operate at high sliding speeds and high tooth pressures.This combination of conditions led to catastrophic collapse of the protective lubricant films between gear teeth and to consequent severe surface adhesive damage along the direction of sliding.This was termed scuffing or scoring.An extensive testing effort determined that this type of failure could be prevented by adding organosulphur-and organochlorine-based compounds to the lubricant and these became known as extreme pressure (EP) additives [2,3].Subsequent research showed that, as the surface temperature of gear teeth rises sharply due to very high friction during incipient scuffing, these adsorbed additives can react very rapidly to form mixed ferrous sulphide/oxide and ferrous chloride-based tribofilms [4,5].The relatively low friction of these tribofilms, combined with a local decrease in contact pressure due to their sacrificial wear, then enables the gears to recover steady-state operating conditions.
Since the 1930s, in addition to its occurrence in gears, scuffing has been an intermittent problem in many other lubricated machine components, including piston/cylinder interfaces, sliding cam/followers and some types of rolling bearings.To improve their efficiency, over recent years the power densities of these components in many applications have increased and they are being lubricated by progressively lower viscosity oils to reduce churning and hydrodynamic friction losses.Both these trends mean that scuffing is becoming a more prevalent issue.
Scuffing requires the breakdown of the separating fluid film (hydrodynamic or elastohydrodynamic) followed by that of any tribofilm present on the rubbing surfaces, so ability of the lubricant to form strong tribofilms is a key factor in preventing scuffing, especially when using low viscosity lubricants.As outlined above, most sulphur-based EP additives are believed to form tribofilms predominantly in response to incipient scuffing.With the more reactive S-based EP additives these films are very thin and formed and removed extremely rapidly, and this can lead to high wear rates.A few S-based additives such as some carbamates and heterocyclic compounds may form thicker films, but research is very limited [6].Other types of additive that can be used to control scuffing are based on compounds containing both S and P.These include the treated reaction products of P 2 S 5 with olefins (SP additives) and more well-defined compounds such as dithiophosphates and thiophosphate esters.These additives may be less effective at preventing scuffing than sulphur-based EP ones, but they also have a very important anti-wear ability and are thus termed AW/EP additives.They limit wear by forming relatively thick (20-200 nm) tribofilms based on metal phosphate/polyphosphate on top of an underlying sulphide film.It has been suggested that some S-free, P-based AW additives such as the amine phosphates and phosphonates may also provide scuffing resistance [5,7].
In a recent paper by the authors it was shown that the additive zinc dialkyldithiophosphate (ZDDP), that contains both S and P, in addition to being very effective in controlling wear, could also prevent scuffing of a heavily-loaded rolling-sliding contact as present in many machine components [8].However, it could only do this if it was able to form a protective tribofilm under relatively mild conditions before the contact was subjected to more severe conditions.Thus, when a step-sliding speed test was started at a low sliding speed of 0.1 m/s the contact did not scuff even when the sliding speed was raised to 6 m/s.However, when a test was started at a sliding speed of 1.0 m/s, scuffing occurred immediately.Tribofilm thickness measurements indicated that starting a test at 0.1 m/s allowed a thick, protective tribofilm to form before conditions became severe enough to challenge the mixed boundary/EHD lubricating film.In practice, scuffing may be initiated by a sudden change in conditions such as an increase in load or temperature, or passage of debris through a contact [9].This may occur before an AW/EP additive can form a suitable tribofilm.The question that arises is how to protect rubbing surfaces well enough to enable ZDDP or other additives containing both S and P to form tribofilms under severe contact conditions, especially in low viscosity lubricants where only thin hydrodynamic or elastohydrodynamic films are present.
One possible way to protect against scuffing is to increase the oil film thickness in the boundary and mixed lubrication regime by forming adsorbed films using polar-functionalised polymer molecules.For example, it has been well studied that functionalised olefin polymers (OCPs) and some polymethacrylates (PMAs) can adsorb on metallic surfaces to form thick boundary films [10][11][12][13][14]. Smeeth et al. [10] suggested that this behaviour was due to the presence of a viscous boundary film of about 20 nm thickness, formed by an effective increase in polymer concentration close to the surface due to polymer adsorption.At low speeds, the contact inlet is filled with this adsorbed polymer layer, which, being much more viscous than the bulk solution, leads to greater fluid entrainment and thus a thicker than expected film.In subsequent work, Gunsel et al. [13] showed that these thick, viscous boundary films reduced friction and mild wear in sliding and sliding-rolling contacts.Such thick, adsorbed films may reduce contact severity enough to enable AW/EP additives to form protective tribofilms under conditions when they would normally be unable to form such a film and consequently scuff.It is, of course, important, that these viscous boundary films do not themselves screen the rubbing surfaces so thoroughly as to prevent AW/EP additives from reacting to form tribofilms, and also that the polymers adsorb on ZDDP tribofilms as well as on steel surfaces.However, there has been very limited previous work on the tribological behaviour of combinations of PMAs with other additives, and apparently none on scuffing.
Based on the above, the aim of the current study was to determine the extent to which PMAs can work in concert with ZDDPs and other additives containing both S and P to provide protection against scuffing, and to explore the origins of this response.The results are presented in three sections.First the impact of a range of PMAs on scuffing of ZDDP solutions is measured using step-sliding speed scuffing tests starting at severe sliding conditions.ZDDP tribofilm thicknesses is monitored during these tests and it is found that some but not all the PMAs studied enable ZDDP to form tribofilms and prevent scuffing in tests starting at high sliding speed.The second section of results describes optical elastohydrodynamic (EHD) and boundary film thickness measurements on ZDDP solutions with and without added PMAs.These indicates that those PMAs that can extend the anti-scuffing performance of ZDDP are all themselves able to form thick boundary films in rolling-sliding conditions.Finally, while ZDDPs are widely used in engine and hydraulic oils they are only rarely used in gear oils.Therefore, to assess relevance to gear oil design, the ability of PMAs to extend the anti-scuffing performance of a conventional, fully-formulated axle gearbox oil containing a SP additive is measured.The results obtained provide practical insight into the formulation of lubricants with anti-scuffing performance and some understanding of the relevant mechanisms by which lubricant additives mitigate scuffing.

Scuffing test rig
All scuffing tests were carried out using a step-sliding speed, rollingsliding method based on that introduced by Ingram et al. [15] and further developed by Peng et al. [16] and Bayat et al. [17].Scuffing is well known to depend critically on both applied load and sliding speed, and most scuffing tests are based on increasing the applied load stepwise at constant sliding speed until scuffing occurs.The disadvantage of this approach is that when the load is increased the contact area increases, so that fresh, unrubbed surface is introduced into the contact and this can precipitate scuffing.This can be avoided by carrying out tests in which sliding speed is increased in steps at constant load.Such an approach requires the ability to apply a large range of sliding speeds at a fixed entrainment speed and Ingram et al. showed that this could be obtained by using a rolling-sliding test rig in which the two surfaces could be rotated in opposite directions; i.e. operating in contra-rotation [15].
In this study a mini-traction machine (MTM) was employed to generate scuffing.This is a ball on disc tribometer, as shown schematically in Fig. 1 and was supplied by PCS Instruments, UK.A ball is loaded against the flat surface of a steel disc immersed in lubricant and ball and the disc are driven by separate electric motors at user-specified rolling/ sliding conditions that can include contra-rotation.A spacer-layer imaging attachment (SLIM) is used to monitor tribofilm thickness during tests as shown schematically in Fig. 1.To capture SLIM images, the test is paused periodically and the ball is raised and loaded against a glass flat coated with a semi-reflective chromium layer and a transparent silica spacer layer.Light reflected from the chromium layer and the ball surface undergo optical interference upon recombination, generating an interference image of the contact between the ball and the coated glass surface that is captured by an RGB camera [18].The thickness of any solid tribofilm present on the ball is calculated from this interferometric image using calibrated RGB values.Table 1 lists the properties of MTM specimens used in this study.The steel tests balls and discs were supplied by PCS Instruments and had roughness less than 10 nm Rq.

Scuffing test protocol
The scuffing test sequence employed in this study is based on that developed by Peng et al. [16].After a running-in stage at very low entrainment speed, the load and entrainment speed are held constant while the sliding speed is increased in steps until scuffing occurs.The entrainment speed is defined by U = (u disc + u ball )/2 while the sliding speed U s = (u disc -u ball ), where u ball and u disc are the velocities of the ball and disc surface with respect to the contact.The slide-roll ratio, SRR is the ratio of the sliding speed to the entrainment speed.To obtain sliding speeds greater than twice the entrainment speed (the case in pure sliding when one surface is stationary), the MTM is operated in contra-rotation mode, with the ball and disc rotating in opposite directions with respect to the contact so that u ball and u disc have opposite signs.This enables very Fig. 1.Schematic image of MTM-SLIM.
M. Ueda et al. high sliding speeds to be reached even at low entrainment speed.Between each speed-step stage in the test sequence there is a rest stage where the ball and disc are rotated at zero applied load to enable the surfaces to cool.At the start of each rest stage a SLIM interference image is obtained to determine the presence and thickness of any tribofilm on the ball surface.Table 2 shows the conditions used in each stage and Fig. 2 shows a schematic image of the various stages in each scuffing test.
Two types of test were carried out.These differed only in that the test stage; in test type1 started at 0.1 m/s while in test type2 it started at 2.0 m/s.In all tests the scuffing test sequence was carried out at 45 N, corresponding to 1.05 GPa maximum Hertzian pressure and all tests were conducted out at 100 ºC.All tests were carried out at least twice to verify repeatability.Further detail and rationale of each of the three stages, running-in stage, test stage and rest stage are described in [19].It should be noted that tests on GTL4 +ZDDP solution with an initial sliding speed of 1.0 m/s and above gave scuffing immediately after the test started, while those starting below 1.0 m/s did not scuff even up to a sliding speed of 5 m/s.By comparison, as shown in Fig. 6, some GTL3 +PMAs without ZDDP gave scuffing at above 2.0 m/s sliding speed.Based on this, an initial sliding speed of 2.0 m/s initial sliding speed was chosen to explore the impact of PMAs on scuffing of ZDDP and SP additive blends.

Test lubricants
The PMAs studied in this work are listed in Table 3. Various types of PMAs with a wide range of molecular weights (Mw) are employed, including conventional type, dispersant type, comb type [20] and film-forming, functionalized type [21].For comparison, a non-functionalised olefin copolymer (OCP) without polar groups in its structure, was also studied.
The oil formulations used in this study are listed in Table 4.To ensure thin film and thus boundary and mixed lubrication conditions, two gasto-liquid (GTL) base oils with quite low kinematic viscosities (KV) were employed; (i) GTL3 of viscosity 9.7 mm 2 /s at 40 ºC and 2.7 mm 2 /s at 100 ºC and (ii) GTL4 with viscosity 18.5 mm 2 /s at 40 ºC and 4.1 mm 2 /s at 100 ºC.GTL4 was used for polymer-free blends while GTL3 was employed in polymer-containing blends.By selecting the appropriate polymer dosage this enabled all oils to be formulated to the same viscosity at the 100 ºC test temperature.
Two AW/EP additives were studied, (i) a commercial ZDDP and (ii) a commercially used axle oil package that contains a SP type additive.In addition to the SP AW/EP additive, this axle oil package also contained succinimide dispersants -some borated, and antifoam agent.The ZDDP and SP package were blended at the concentrations shown in Table 4, which are typical of those used in engine oils and axle gear oils respectively.
Various PMAs and OCP were blended with ZDDP and the SP package in GTL3, and to understand the effect of PMAs themselves on scuffing PMAs were also studied alone in GTL3 without antiwear additives.The friction modifier glycerol monooleate (GMO), which is believed to physically and/or chemically adsorb on surfaces, was also studied, both alone in solution in GTL4 and blended with ZDDP, to compare its impact on ZDDP with that of the polymers.
All blends were formulated to have a KV of 4.1 mm 2 /s at the test temperature of 100 ºC by controlling the GTL type and the dosage of the polymers, as shown in Table 4.The corresponding dynamic viscosity at the test temperature of 100 ºC was 3.2 cP.At the entrainment speed of 0.2 m/s used in the main test sequence, the calculated minimum oil film thickness was ca.6 nm.This corresponds to an initial, theoretical lambda ratio (ratio of EHD film thickness to composite surface roughness) of ca.0.5, and thus to mixed lubrication conditions.In the runningin stage where the entrainment speed was 0.003 m/s, the calculated minimum oil film thickness was ca.0.3 nm, corresponding to lambda = ca.0.03 and thus to predominantly boundary lubrication conditions.

EHD film thickness
The EHD film-forming properties of the test oils were measured at 100 • C using optical interferometry (EHD2, PCS Instruments, UK).Oil film thickness was measured using two tribopairs; a steel ball/sapphire disc and a steel ball/glass disc, as shown in Table 5.
In this type of film thickness measurement, a steel ball/glass disc tribopair is most often employed to measure EHD film thickness, but this gives a relatively low contact pressure since, because of the low elastic modulus of glass, the reduced elastic modulus of the steel/glass combination is only 58 Ga as compared to that of steel/steel of 115 GPa.To ensure equivalent contact pressure to the MTM scuffing tests, in this study the coated glass disc was replaced by a much higher elastic modulus coated sapphire disc.The steel/sapphire tribopair had a reduced elastic modulus of 150 GPa, which is higher than that of steel/ steel, so the AISI 52,100 steel ball was loaded against a coated sapphire disc at 24 N to give a maximum contact pressure 1.1 GPa, equivalent to the contact pressure in the 45 N MTM scuffing tests [22].In this high contact pressure contact, film thickness was measured over a range of entrainment speeds in nominally pure rolling.
To study the effect of sliding on oil film formation, additional tests were carried out in rolling/sliding contact conditions.When sliding and thus shear stress is introduced to a contact, the coatings on the disc are easily removed and avoid this problem a lower load at 10 N, corresponding to 0.4 GPa contact pressure, was applied on a steel ball/glass disc tribopair.Under such contact conditions, sliding was applied by increasing SRR from 0 % to 450 %.Note that tests were conducted over an entrainment speed range of 0.01-1 m/s because of the restricted maximum speed of the ball drive motor in the EHD2 rig.

Influence of initial sliding speed
In a previous study by the authors, the influence of initial sliding speed stage was studied by starting the main step-speed test sequence at different sliding speeds [19].Results showed that for additive-free PAO, the sliding speed at scuffing was largely independent of the initial sliding speed used.However, for ZDDP solution, the scuffing speed depended strongly on the sliding speed at which the test sequence started.If the test sequence started at 1.0 m/s sliding speed, scuffing occurred at much lower speed than if it started at 0.1 m/s.This is believed to result from the fact that when test stages start at a low sliding speed, ZDDP is able to develop a protective tribofilm in the early speed stages and this then provides protection against scuffing up to higher sliding speeds.
In the current study, since the main interest was in whether PMAs can enable ZDDP tribofilms to form in more severe sliding conditions than ZDDP alone can, the effect on scuffing of an initial sliding speed of 2.0 m/s in the main test sequence was studied, in addition to the more normal 0.1 m/s initial sliding speed.The rationale for using 2.0 m/s was described in Section 2.2.Fig. 3 shows the variation of friction during the step-speed sequence for GTL4+ZDDP and for GTL4.It should be noted that the running-in and the rest stages are not shown in this and all subsequent figures.Also, it is noted that the momentary blips in the friction traces between speed steps in Fig. 3 (and other figures in this paper) are artefacts of the measurement method and occur immediately after motion is paused to capture SLIM images during the rest stages.Additive-free GTL4 gave a sudden friction increase characterising scuffing at 0.8 m/s sliding speed in the test started at 0.1 m/s sliding speed, but in accord with previous work, GTL4 +ZDDP did not show scuffing up to the maximum sliding speed of 5.0 m/s, though there were considerable fluctuations of friction during speed steps.By contrast, when the initial sliding speed was 2.0 m/s, GTL4+ZDDP gave scuffing immediately after sliding began in the test stage.
SLIM images shown in Fig. 8 indicate that when the step-speed tests started from 0.1 m/s sliding speed, GTL4 + ZDDP formed approximately 20 nm of tribofilm by the time 2.0 m/s sliding speed was reached.This tribofilm could then protect against scuffing at sliding     speeds when scuffing would occur in the absence of effective tribofilm (e.g. in GTL4 alone at around 0.8 m/s sliding speed).When tests were started from 2.0 m/s sliding, only 2-3 nm of tribofilm was present after the running-in and this was too thin to prevent immediate scuffing.
Since the primary aim of the current study was to understand the effect of PMAs on scuffing in severe conditions when there is not enough protective antiwear tribofilm on the surfaces to prevent scuffing, subsequent scuffing tests started at 2.0 m/s sliding speed, the condition at which PMA-free ZDDP solution scuffed immediately.

The effect of PMAs on scuffing of ZDDP
Fig. 4 shows how friction coefficient varied in the test sequences for twelve ZDDP-containing lubricants listed in Table 4. GTL4+ZDDP and GTL3 +ZDDP with PMA2, D-PMA3, D-PMA5, film-PMA8, OCP and GMO all gave scuffing immediately after the tests started at 2.0 m/s sliding speed.By contrast, addition of PMA1, D-PMA4, comb-PMA6 and Dcomb-PMA7 to GTL3+ZDDP prevented scuffing up to the maximum sliding speed tested of 5.0 m/s.It is noteworthy that the significant variations of friction coefficient during speed steps that could be seen in the GTL4+ZDDP test started from 0.1 m/s sliding speed in Fig. 3, are much less when PMA1 is added and are not seen at all when D-PMA4, comb-PMA6 and D-comb-PMA7 are present.This is further discussed in Section 3.4.
Fig. 5 shows SLIM images at the end of tests (after the 5.0 m/s sliding speed step) with blends containing PMA1, D-PMA4, comb-PMA6, and Dcomb-PMA7.The figure also shows the evolution of tribofilm thickness during tests for these four blends.The values shown at 0 s are tribofilm thicknesses after the running-in stage.All four PMA blends formed tribofilms of ca. 10 nm thickness after 150 s rubbing, corresponding to the 2.4 m/s sliding speed step.Tribofilms formed from the comb-PMA6 and D-comb-PMA7 blends then grew continuously up to the end of the tests, to reach approximately 80 nm and 90 nm respectively.The oil with D-PMA4 formed a thicker tribofilm, to reach 100 nm after 570 s rubbing, followed by a reduction to 40 nm at the next step.This tribofilm removal may result from the high shear stress caused by high sliding.This tribofilm removal corresponds with a slight drop of friction coefficient from 0.09 to 0.08 between 570 s and 600 s rubbing.This may result from ZDDP tribofilm removal as Dawczyk et al. [23] showed that ZDDP forms a rough and solid tribofilm surface which reduces fluid entrainment to increase lambda ratio, resulting in increase of friction.After this tribofilm removal, ZDDP tribofilm growth resumed to reach 110 nm at the end of the test.Tribofilm formed from the oil with PMA1 grew to ca. 30 nm up to 300 s, corresponding to 2.9 m/s sliding speed, and then stayed around 20 nm to the end of the tests with fluctuations.This tribofilm fluctuation may result from micro-scuffing, as shown by the slight sudden increase of friction shown in Fig. 4 that can be seen especially after 300 s rubbing.

The effect of PMAs alone on scuffing
The scuffing performance of ZDDP-free PMA solutions was also studied for tests starting at 2.0 m/s sliding speed.Fig. 6 shows the variation of friction during the main step-speed sequence after runningin.PMA2, D-PMA3, D-PMA5 film-PMA8, OCP and GMO solutions gave scuffing immediately after the tests started.By contrast, PMA1, D-PMA4, comb-PMA6 and D-comb-PMA7 slightly delayed scuffing onset above 2.0 m/s.Note that since ZDDP was not present to form a solid-like tribofilm, observable tribofilms from these four oils were not seen in SLIM images.|It is noteworthy that these four PMAs are the ones that enabled ZDDP to form a tribofilm and prevent scuffing up to 5.0 m/s sliding   speed.These PMAs appear to form adsorbed films that provide enough protection to enable ZDDP to form a tribofilm at high sliding speeds.

The effect of PMAs on ZDDP micro-scuffing
As shown in Fig. 4, fluctuations in friction indicative of microscuffing were prevented in the ZDDP oils with D-PMA4, comb-PMA6 and D-comb-PMA7, compared to the friction behaviour of GTL4+ZDDP without PMA seen in Fig. 3.However, since GTL4+ZDDP gave scuffing immediately after the test started at 2.0 m/s sliding speed, a clear comparison between GTL4+ZDDP and GTL3+ZDDP with PMA blends on micro-scuffing could not be made.To allow such a comparison, tests starting at 0.1 m/s sliding speed were carried out for some ZDDP/PMA blends.Fig. 7 shows how friction coefficient varied in tests starting at 0.1 m/ s for six lubricants containing ZDDP listed in Table 4, GTL4+ZDDP, GTL3 +ZDDP with PMA1, D-PMA4, comb-PMA6 and D-comb-PMA7, all of which gave an explicit effect on scuffing protection, and GTL3+ZDDP with GMO.The last of these was for comparison with the PMAs.GTL4+ZDDP gave significant variations of friction coefficient throughout the test due to micro-scuffing.By comparison, the oils with D-PMA4, comb-PMA6 and D-comb-PMA7 gave a stable friction coefficient without variation throughout the tests, while PMA1 gave slight variations of friction especially after 960 s, corresponding to 3.1 m/s sliding speed.In the case of the oil with GMO, micro-scuffing was not seen, but full scuffing occurred at 3.1 m/s.Since ZDDP solution can survive without scuffing up to 5 m/s in tests starting at 0. 1 m/s sliding speed, this shows that GMO degrades the scuffing performance of ZDDP though it reduces the level of friction.
Fig. 8 shows the evolution of tribofilm thickness of lubricants during the test stages at 0.1 m/s sliding speed, and also SLIM images at the end of test.For the GTL4+ZDDP and the GLT3+ZDDP with PMA blends, the SLIM images are after the 5.0 m/s sliding speed stage, while that for GMO is after the 2.9 m/s sliding stage, i.e. just before the stage at which scuffing occurred.GTL4+ZDDP tribofilm grew to reach ca.30 nm after 420 s rubbing, corresponding to 1.4 m/s sliding speed, and then stayed at approximately 20 nm to the end of the test.As shown in Fig. 5, tribofilms formed from the oils with D-PMA4, comb-PMA6 and D-comb-PMA7 grew continuously up to the end of the tests, to reach approximately 170 nm, 120 nm and 90 nm, respectively.Tribofilm formed from with PMA1-containing blend grew to ca. 60 nm up to 1020 s rubbing, corresponding to 3.3 m/s sliding speed, and then reduced to remain around 20-30 nm to the end of the tests.GMO deterred tribofilm growth, then stayed at low level of around 5 nm until scuffing at 3.1 m/s sliding speed.It is noteworthy that the initial growth rates of tribofilms in PMAs and GMO up to 420 s rubbing were slower than that of GTL4+ZDDP.This may result from competitive adsorption between ZDDP and the PMAs and GMO.However, the final tribofilm thicknesses with D-PMA4, comb-PMA6 and D-comb-PMA7 blends were significantly thicker than that formed by ZDDP alone.This probably resulted from there being no tribofilm removal caused by micro-scuffing.
These results suggest that some PMAs prevent micro-scuffing and help maintain protective ZDDP tribofilms on surfaces even at high sliding speed.This should help to protect surfaces from scuffing, and also mild wear.

Results: oil film thickness of PMAS
As described above, some PMAs improved the anti-scuffing behaviour of ZDDP by enabling antiwear tribofilms to form in more severe sliding conditions that was the case in the absence of PMA.It is well known that some PMAs can adsorb on steel surfaces to thicken oil film thickness in boundary and mixed lubrication conditions [10][11][12][13][14] and this is a probable origin of at least some of the observed effect on scuffing.To explore this further, the effect of PMAs in ZDDP solution on fluid film thickness was studied using optical interferometry.The behaviours of OCP and GMO blends were also studied for comparison.Two sets of tests were carried out, (i) in nominally pure rolling using a sapphire disc on steel ball contact to give high contact pressure (max.1.1 GPa) and (ii) in mixed rolling sliding using a relatively low contact pressure (max 0.4 GPa) glass disc on steel ball.Fig. 9 shows measured EHD oil film thickness of ZDDP blends at 100 ºC in high pressure, pure rolling conditions.Since all oils have the  same viscosity at 100 ºC, they gave very similar EHD film thicknesses of 80-85 nm at a high entrainment speed of 3 m/s.However, boundary film thickness at low entrainment speed, for example 10 mm/s, varied between lubricants.GTL4 and GTL4 +ZDDP gave a low oil film thickness of ca. 3 nm at 10 mm/s.This is slightly above the predicted central EHD film thickness of ca 1.5 nm.Similar to these two lubricants, the GTL3 oils with OCP gave a low oil film thickness of 3.5 nm at 10 mm/s while GMO produced a quite thick boundary film to reach 8 nm at 10 mm/s.GMO has a polar group to adsorb on surfaces and form boundary films [24].By comparison, all the PMAs blends gave boundary oil films greater than 10 nm at 10 mm/s.This behaviour has been ascribed to the presence of a viscous boundary film of about 10-20 nm thickness formed by an increase in polymer concentration close to the surface due to adsorption [10].
It was not possible to apply a significant amount of sliding at this high contact pressure damaging the optical coatings in thin film conditions.Therefore, film thickness tests were also carried out using a lower pressure, steel ball on glass disc contact.Fig. 10 shows measured oil film thicknesses at 0 % SRR, 150 % SRR and 450 % SRR.The latter corresponds to contra-rotation conditions.Results in pure rolling are broadly similar to the high contact pressure measurements.At 150 % SRR, GMO, PMA2 and PMA8 no longer formed a thick boundary film and behaved very similarly to GTL4, while at 450 % SRR PMA3 was no longer able to form a boundary film.The remaining PMAs, D-comb PMA-7, D-PMA5, comb-PMA6, PMA1 and D-PMA4 all continued to form a boundary film of between 10 and 15 nm thickness at 450 % SRR.It is noteworthy that four of these PMAs (all except D-PMA5) were the ones able to extend the scuffing performance of ZDDP.Test at 250 % and 350 % SRR gave similar response to that at 150 % SRR, while it was not possible to test at 550 % SRR without damaging the glass disc coatings.The results show that the boundary film forming properties of PMAs are influenced by the degree of sliding, and thus the shear stress experienced by the film in the contact.

Results: PMAS in an axle oil package
The tests using ZDDP-containing oils described above show that PMAs can enhance the anti-scuffing performance of ZDDPs and allow ZDDP tribofilms to develop under more severe contact conditions than possible with ZDDP alone.However, gear oils very rarely contain ZDDPs and instead are generally based on metal-free SP AW/EP additives.They also usually contain several other surface-active additives such as dispersants that might affect tribofilm formation.To determine whether the above effects of PMAs on scuffing improvement are applicable to fully-formulated gear lubricants, the behaviour of PMAs in a commercially employed axle oil additive was studied.Four PMAs, PMA1, D-PMA4, comb-PMA6 and D-comb-PMA7, all of which prevented scuffing in a ZDDP solution in tests starting at 2.0 m/s sliding speed, were added to the package solution in GTL3.
Fig. 11 shows how friction coefficient varied during tests.The oil containing just the axle oil additive package gave scuffing immediately after the start of the test at 2.0 m/s sliding speed, the same behaviour as seen with the blend containing just ZDDP.By contrast, the oils including each of the four PMAs prevented from scuffing up to 5.0 m/s sliding speed, similar to the response seen with ZDDP+PMAs.Friction coefficient was stable at a low level, between 0.06 and 0.1, without fluctuations indicative of micro-scuffing.It should be emphasised that all five oils were blended to have the same viscosity at the test temperature.
Fig. 12 shows the evolution of tribofilm thickness of the four PMA blends during the test stages from 2.0 m/s sliding speed, together with SLIM images after the 5.0 m/s sliding stages.Tribofilms formed from the oils with PMAs reached approximately 10-20 nm over a few stages after the tests started, and then became stable until the end of the tests without a significant drop of thickness.It has been well reported that most SP type additives form much thinner tribofilms than those formed by ZDDPs [25].However, even such relatively thin tribofilms were enough to protect surfaces from scuffing.These results show that some  PMAs help to protect surfaces from scuffing not only in a simple-ZDDP solution but also with a gear oil additive package.
Fig. 13 shows EHD oil film thickness measurements of the oils including an axle oil package in a high-pressure steel ball on sapphire disc contact.Because axle oil package contains adsorbing molecules such as dispersant, the blend of GTL4+axle oil package gave a slightly thicker boundary film, of 6 nm at 10 mm/s than the base oil.The addition of PMA1, D-PMA4, comb-PMA6 and D-comb-PMA7 all resulted in thicker boundary films, of 10-15 nm thickness at 10 mm/s.The results show that these PMAs can provide thick boundary films in both ZDDP solutions and a fully-formulated axle oil.

Discussion
In this study, ZDDP solution and an axle oil package solution both gave scuffing in step-speed tests that started at relatively severe conditions, i.e. that began with the immediate imposition of a sliding speed of 2 m/s.However, it was found that when some types of PMA were added to the blends, both the ZDDP solution and the AW/EP additive containing axle oil were able to prevent scuffing from occurring up to extreme sliding conditions of at least 5 m/s.These PMAs thus markedly extended the anti-scuffing performance of both ZDDP and the SP additive present in the axle oil package.
Scuffing tests on ZDDP-free PMA solutions showed that, while most of these scuffed immediately at the initial sliding speed of 2.0 m/s, the four PMAs that were found to augment the scuffing performance of ZDDP were also able to survive a 2.0 m/s sliding condition without scuffing; i.e. they possessed some resistance to scuffing of their own.This, combined with SLIM measurements of tribofilm thickness during tests, suggests that these four PMAs provided sufficient boundary film protection for the AW/EP additives tested to have time (or perhaps sliding distance) to form thick tribofilms, despite initially severe conditions.
It is well known that some polymer additives, including functionalised PMAs, can form thick boundary films of ca.20 nm thickness [14] and Gunsel et al. [13] have shown that these boundary films can reduce mild wear in sliding-rolling contacts.In the current study, optical interferometric EHD/boundary film thickness measurements showed that all the eleven PMAs blends tested formed adsorbed boundary films of over 10 nm in nominally pure rolling conditions.The question that thus arises is why only four of them enhance the anti-scuffing performance of ZDDP?Müller et al. showed that although a functionalised PMA formed a thick boundary film at 0 % SRR and lowered friction in an MTM at 50 % SRR, its ability to reduce friction in pure sliding HFRR conditions was minimal [26].The starting sliding speed of 2.0 m/s in the current study corresponds to an SRR of 1000 %.It is thus possible that although all eleven PMAs were able to form thick boundary films in pure rolling contacts, only the films formed by four of them could survive the very large shear stresses present in the contact inlet at high sliding.This was explored by carrying out some optical EHD film thickness measurements in high sliding conditions and it was indeed found that most PMAs lost the ability to form thick boundary films at 450 % SRR.In the context of our scuffing tests this is a relatively mild condition of 0.9 m/s sliding speed.Five of the PMAs retained a boundary film at 450 % SRR, four being those that supported ZDDP anti-scuffing response (D-PMA4, comb-PMA6, D-comb-PMA7 and PMA1) and one other, D-PMA-5.Possibly, the boundary film formed by D-PMA5 was weaker than the other four and was removed at higher sliding speed, but unfortunately tests at 550 % SRR removed the optical coatings from glass disc so tests above 450 % SRR were not possible.
As stated in the introduction to this paper, for a polymer additive to be effective in supporting ZDDP's anti-scuffing ability it must (i) be able to form a boundary film not just on steel but also on a ZDDP tribofilm surface and (ii) not inhibit to a significant extent the ability of ZDDP to react with rubbing surface to form tribofilms.The ability of the four beneficial PMAs to prevent the micro-scuffing seen with PMA-free ZDDP solution in Fig. 7 indicates that the PMAs are continuing to protect the ZDDP tribofilms once these are formed and thus may be absorbing on the latter.Fig. 8 shows that the PMAs may be having a slight effect of slowing ZDDP tribofilm growth in tests starting at a sliding speed of 0.1 m/s, but this is minor and, as the sliding speed is increased, appears to be more than compensated by the ability of the PMAs to protect the ZDDP film from partial removal.As previous studies have shown, the chemical structure and morphology of ZDDP tribofilms evolve during rubbing, resulting in the films becoming stronger [27][28][29].It is possible  that adsorbed PMAs may influence this evolution chemically, but more probably they simply reduce the contact condition severity, enabling the tribofilms to have time to form and evolve.
The superior scuffing behaviour of combinations of some PMAs with ZDDP and with SP package can be regarded as a synergistic response and there are several different mechanisms by which such a response can arise.Most previous work on the synergistic or antagonistic behaviour of combinations of other additives with ZDDP have identified direct interactions between the ZDDP and other additives in solution, i.e. ligand exchange and complexing, or at surfaces, i.e. additive molecules may react together directly at metal surfaces [30].By comparison, the synergistic combination of ZDDP and PMA appears to be a complementary effect in which an adsorbed PMA film ameliorates the contact conditions enough to enable a much stronger ZDDP tribofilm to form [30].
There has been very limited previous work on the tribological behaviour of combinations of PMAs with other additives, and apparently none on scuffing.Müller et al. [31] measured the friction properties of blends of two dispersant-functionalised PMAs with a range of other lubricant additives in the MTM.They found that both PMAs continued to be effective in reducing friction, and thus presumably forming boundary films, when blended with most other additives, including ZDDPs and S-based EP additives.The PMAs were only ineffective when combined with additives having basic aminic groups such as one friction modifier and a succinimide dispersant.The functional groups of the two PMAs contained basic nitrogen atoms and it was conjectured that other additives with similar basic nitrogen might compete with polymer adsorption on the steel surfaces.
The differing response of the various PMAs is of considerable interest, though unfortunately since all were commercial samples little structural detail was available.In order to form boundary films, Müller et al. [14] suggested that there are three general rules for designing PMA copolymers (i) functionalized groups in a structure, (ii) clustered functionalized groups within a polymer molecule rather than a statistical distribution, and (iii) medium to high molecular weight rather than low molecular weight.The authors also found PMAs in accord with such rules reduced boundary friction.It should, of course, be noted that in their study, oil film formation and friction were measured at 0 % and 50 % SRR, respectively, while in the current study discrimination between effective and ineffective PMAs was only seen at much higher SRRs.The various PMAs tested did not vary systematically enough, nor was sufficient chemical structure information available, to test Muller et al.'s criteria above.Certainly, at least two of the effective PMAs are dispersant-functionalised.It should be noted that the thick tribofilm formed by ZDDP is based on phosphate/polyphosphate and is more likely to be acidic than basic in nature and thus quite likely to adsorb basic functionalities.However, so far as the authors are aware, PMA1 is a relatively low Mw and conventional non-functionalised polymer so it is perhaps surprising that it is so effective at supporting ZDDP anti-scuffing response.Evidently further study is needed to understand the relation between chemistry of PMA and durability of boundary films under sliding.In this study, in order to have the same viscosity at 100 • C, the concentrations of PMA employed varied depending on the PMA Mw.Although no correlation was observed between PMA concentration and scuffing prevention, it would be desirable to explore the influence of concentration on PMA behaviour in future work.
Based on the results in this study, the relevant mechanism by which scuffing is prevented in PMA blends is suggested schematically in Fig. 14.At the start of tests at 2 m/s, those PMAs whose boundary films are able to survive this high sliding speed, i.e.PMA1, D-PMA4, comb-PMA6 and D-comb-PMA7, provide a thick boundary film at the initial phase of test where otherwise a protective tribofilm is absent.This helps separate the surfaces and alleviates the contact conditions, resulting in the prevention of scuffing and allowing the ZDDP or SP package to start generating a tribofilm.
The role of PMA at higher sliding speed is unclear but since microscuffing was prevented by the effective PMAs in ZDDP solution during the tests shown in Fig, 7, some adsorbed PMA is still likely to be present on the tribofilm even up to the highest sliding speed.Regardless of such PMA adsorption, as sliding distance increases, the ZDDP or SP tribofilm becomes thick enough to protect the sliding surfaces until the end of the tests.

Conclusions
The impact of PMAs in a ZDDP solution and a SP additive-containing axle oil package solution on scuffing has been studied using a test method based on contra-rotation and a step-sliding speed sequence.The results show that some PMAs markedly improve the anti-scuffing performance of both AW/EP additives.They do this by adsorbing on steel and tribofilm surfaces to form thick boundary lubricating films.These are then able to protect the surfaces long enough for the EP/AW additives to react with the rubbing surfaces to form protective tribofilms.It is found that all the PMAs tested are able to adsorb to form thick boundary films on surfaces but only some of these are able to survive high speed sliding conditions; these latter PMAs are the ones that can augment the anti-scuffing properties of ZDDP and the SP additive in the axle oil package.The insights presented here should help with the design of low viscosity lubricants that are effective in controlling scuffing by optimizing both the boundary oil film and the tribofilm.

Fig. 2 .
Fig. 2. The sequence of the scuffing tests with step-varying sliding speed.Each test point consists of a 30 s test stage followed by a rest stage of 30 s to capture a SLIM image.The 600 s running-in stage is denoted by a +.

Fig. 4 .
Fig. 4. The effect of PMAs, OCP and GMO on friction coefficient in ZDDP solutions.The initial sliding speed was 2.0 m/s.

Fig. 5 .
Fig. 5.The development of tribofilm thickness formed by various lubricants as quantified by SLIM images captured during the rest stages.SLIM images of the four lubricants after the end of the tests are also shown.

Fig. 6 .
Fig. 6.The effect of PMAs on friction coefficient in GTL3.The initial sliding speed was 2.0 m/s.

Fig. 7 .
Fig. 7.The effect of PMAs and GMO in ZDDP solutions on friction coefficient.The initial sliding speed was 0.1 m/s.

Fig. 8 .
Fig. 8.The development of tribofilm thickness by various lubricants as quantified by SLIM images captured during the rest stages.SLIM images of the six lubricants after the end of the tests are also shown.SLIM image of GTL4+ZDDP+GMO is after the 2.9 m/s sliding speed step.The initial sliding speed was 0.1 m/s.

Fig. 9 .
Fig. 9. EHD oil film thickness of PMAs, OCP and GMO in ZDDP solutions using sapphire disc-steel ball at 24 N with 0 % SRR.

Fig. 11 .
Fig. 11.The effect of PMAs on friction coefficient in axle oil package solutions.The initial sliding speed was 2.0 m/s.

Fig. 12 .
Fig. 12.The development of tribofilm thickness formed by various lubricants as quantified by SLIM images captured during the rest stages.SLIM images of the four lubricants after the end of the tests are also shown.The initial sliding speed was 2.0 m/s.

Fig. 13 .
Fig. 13.EHD oil film thickness of PMAs in axle oil package solutions using sapphire disc-steel ball at 24 N with 0 % SRR.

Fig. 14 .
Fig. 14.Schematic illustration of the primary mechanism by which PMA blends in a ZDDP or axle oil package solution prevent scuffing.M.Ueda et al.

Table 1
Properties of MTM specimens.

Table 2
Scuffing test conditions using MTM.

Table 3
Molecular weight of viscosity modifiers used in this study.

Table 4
Test oil formulations.

Table 5
EHD film thickness test conditions using EHD2 rig.