- Email: [email protected]
The relationship between flake graphite orientation, smearing effect, and closing tendency under abrasive wear conditions
Author's Accepted Manuscript
The relationship between flake graphite orientation, smearing effect, and closing tendency under abrasive wear conditions Rohollah Ghasemi, Lennart Elmquist
www.elsevier.com/locate/wear
PII: DOI: Reference:
S0043-1648(14)00188-4 http://dx.doi.org/10.1016/j.wear.2014.05.015 WEA101032
To appear in:
Wear
Received date: 12 December 2013 Revised date: 22 May 2014 Accepted date: 26 May 2014 Cite this article as: Rohollah Ghasemi, Lennart Elmquist, The relationship between flake graphite orientation, smearing effect, and closing tendency under abrasive wear conditions, Wear, http://dx.doi.org/10.1016/j.wear.2014.05.015 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
The relationship between flake graphite orientation, smearing effect, and closing tendency under abrasive wear conditions Rohollah Ghasemi a,Ö, Lennart Elmquist b Department of Mechanical Engineering, Materials and Manufacturing, School of Engineering, Jönköping University, P.O. Box 1026, SE-551 11 Jönköping, Sweden E-mail address: [email protected]; [email protected] Ö
Corresponding author, Tel: +46 36101179; fax: +46 36166560
E-mail address: [email protected]
Abstract Plastic deformation of the matrix during the wear process results in closing the graphite flakes. In this study, the relationship between the deformation of the matrix and the closing tendency of flake graphite was investigated, both qualitatively and quantitatively. Two representative piston rings, which belonged to the same two-stroke marine engine but were operated for different periods of time, were studied. Initial microstructural observations indicated a uniform distribution of graphite flakes on unworn surfaces, whereas worn surfaces demonstrated a tendency towards a preferred orientation. Approximately 40% of the open flakes of the unworn surfaces were closed during sliding, which may result in the deterioration of the self-lubricating capability of cast iron. Moreover, flakes within the orientation range of 0 to 30° relative to the sliding direction showed a maximum closing tendency when subjected to sliding. The closing tendency gradually decreased as the angle increased, approaching a minimum between 30 and 70°. A slight increase in the closing tendency was observed for flakes with orientations between 70 and 90°. A similar trend was observed on both rings. Furthermore, SEM and EDS analysis indicated substantial deformation of the matrix in the area around the flakes. An insignificant corrosion attack was observed on both worn piston ring surfaces. Keywords: Grey cast iron, Piston ring-cylinder liner, Graphite flake orientation, Sliding wear, Abrasive wear, Graphite closing tendency
1. Introduction Over the years, grey cast iron (GCI) components have commonly been used in a variety of tribological applications, particularly in systems involving sliding, such as disc brakes, clutch and piston rings, and cylinder liner systems [1]. In tribological terms, the operating conditions for piston rings bring about a complex wear situation caused by high temperature, pressure, and mechanical friction, all of which contribute to multifaceted, high-stress conditions [2]. The oil lubrication system, clearance, initial surface roughness [3], and microstructure of the materials used [4] are other critical factors for high performance piston ring-cylinder liner materials. Low production costs, excellent tribological performance, superior machinability, outstanding thermal conductivity, and
high wear resistance, for both lubricated and dry sliding conditions, [5] have made GCI a great material for such complex applications [6]. From the tribological point of view, the excellent wear resistance of GCI stems from the selflubricating nature of the graphite particles which are distributed as flakes throughout the matrix. During sliding, carbon atoms are released from the graphite pocket, smeared onto the surfaces, and serve as solid lubricating agents by forming a thin graphite film between the sliding surfaces [7, 8]. Formation of tribofilm results in a decrease in specific wear rate by several orders of magnitude, as it limits direct contact between the sliding surfaces [7-9]. In addition, the graphite flake sites act as oil reservoirs, improving the supply of oil between sliding parts during dry starts or similar conditions involving oil starvation [9]. The advantageous contributions of open graphite flakes have been highlighted by Sugishita and Fujiyoshi [10]. For many years, the high wear rate reported in piston rings-cylinder liner systems seemed to be associated with the presence of a high sulphur content in the fuel. The presence of sulphur leads to a corrosive environment through the formation of sulphuric acid on the liner wall, which in turn accelerates the wear rate during sliding. Recently, however, it has been reported that the high wear rate and surface degradation of the piston rings are related to abrasion and scuffing, rather than corrosive wear [6]. In most investigated applications, the wear process proves to be significantly more complex than expected [11]. Scuffing occurs when a strong adhesive force develops between the piston rings and the cylinder liner under non-optimal lubrication conditions. As reported, scuffing happening is almost always accompanied by a substantial increase in friction [9], with a severe plastic deformation of the matrix [6]. The occurrence of scuffing is shown by micro-welding and/or severe wear scars, which represent plastic deformation, abrasive ploughing, and adhesive wear [12]. As has been proven so far, a direct relationship exists between the increased volume of graphite on the sliding surface and decrease in specific wear rate, discussed by Sugishita [13]. Abrasion plays a major role in sliding wear conditions by changing the matrix texture during sliding [12, 14]. Montgomery has studied the importance of matrix deformation under abrasive wear conditions, including both fully lubricated and dry sliding conditions [15, 16]. He reports that the formation of a graphite film significantly reduced the risk of scuffing under abrasive conditions by ploughing the metal matrix. Eyre et al. [14] indicated a dramatic increase in wear rate in piston rings and cylinder liner caused by abrasive wear. The hard abrasive particles could either be introduced into the system by the fuel in the form of catalyst fines (CAT fines), or be present as a wear-produced particles (debris) [6, 17]. A study performed by Riahi [17] showed that plastic deformation of the matrix accelerates debris formation. The abrasive particles are sufficiently hard to indent and scratch the sliding surfaces, which leads to severe plastic deformation of the metal matrix [12]. Apart from the microstructural standpoint the presence of the phases so as carbides and phosphides, graphite morphology and graphite flakes play important roles under abrasive condition [17-19]. The orientation of the graphite beneath the sliding surface affects the wear behaviour of the cast iron as it alters the deformation tendency in the matrix around the flake [20]. Micro-interactions, including micro-ploughing and micro-cutting; between abrasive particles and the matrix may also cause the graphite flakes to close [10], detrimentally deteriorates the self-lubricating performance of the cast iron. Therefore, keeping the graphite flakes open during sliding is tribologically beneficial, as their lubricating effect is retained.
The aim of this study is to investigate the relationship between flake graphite orientation and plastic deformation in a cast iron matrix during sliding conditions. An understanding of this relationship clarifies the deformation response of the flakes and the matrix when the piston ring is subjected to sliding and abrasive wear conditions. Both qualitative microscopic observations and quantitative measurements of the graphite distribution of worn samples are considered.
2. Materials and methods 2.1. Piston rings from a large two-stroke marine diesel engine The rings used in this study were 800 mm in diameter and designed for two-stroke marine engines, as shown in Fig. 1(a). The outer surface of the as-cast piston rings are essentially horizontally machined roughly 3 mm before applying the coating. This step is performed to remove the skin defects (undesired structure) from the surface. The surface roughness average (Ra) of the rings are checked to be within 3.2-6.3 m before coating. Fig. 1(b) illustrates the as-manufactured ring surface without coating. The piston rings were protected by two different types of thermally sprayed coatings, including cermet and aluminium-coating. The coatings and base material specifications for each of the piston rings are given in Table 1 [21]. Table 1. Specifications for the piston rings [21].
Piston ring
Base material
1st 2nd 3rd 4th
Compacted graphite iron Grey cast iron Grey cast iron Grey cast iron
Hard coating Cermet thickness (mm) 0.5 0.3
Running-in coating Aluminium-coat thickness (mm) 0.1 0.3 0.3 0.1
Cermet is a composite material which is used for hard coating, and consists of a ceramic part (chromium carbide) and a metallic part (molybdenum, nickel, and chrome). Aluminium-coating (a bronze-based coating containing alumina oxide insoluble, which was first introduced by MAN B&W Diesel A/S) was used as a semisoft running-in coating. The running-in coating is thermally deposited over the hard coating in order to shorten the running-in time, improve sliding characteristics, and increase scuffing resistance during the running-in period. The aluminiumcoating provides a smooth and profiled running-in surface, and is gradually worn off until it is entirely removed approximately after 1-2,000 hours. As the location of the second piston ring subjects to more severe wear conditions (e.g. higher temperature and pressure) than the third and fourth rings, it was selected for the present investigation. The chemical composition of the rings is presented in Table 2. The piston ring segments were provided by MAN Diesel and Turbo.
Table 2. Chemical composition of the investigated piston rings (wt%). Element C Si Mn Ni Mo Cu Ti Cr V S P Fe Composition 3.30 1.55 0.85 0.35 0.60 0.85 0.07 0.15 0.15 0.08 0.10 Balance
2.2. Sample preparation The present investigation was carried out on two different piston ring specimens extracted from the same ship engine, but were running for different time periods; approximately 16,000 and 20,000 hours. Both rings had a good run-in condition and had been running on heavy fuel oil. A schematic illustration of a typical ring sample used in this study is shown in Fig. 2. The black vertical dashed lines denote the scratches present on the outer surface. In order to examine the microstructure of the deformed matrix, the rings were sectioned perpendicular to the sliding direction, as indicated by the grey horizontal dashed line (see Fig. 2).
2.3. Microstructure characterizations and composition analysis The selected specimens were ground and polished following the standard sample preparation procedure for cast iron. The final polishing step was carried out using MD-Dac 3 μm diamond cloth with DiaPro suspension in order to obtain a scratch-free surface and a surface finish of roughly 0.05 μm. The polished samples were subsequently etched with different concentrations of Nital solution. This was done to reveal either the micro-constituents, or the carbide phases present in the cast iron matrix, depending on the purpose of the investigation. Further characterizations were carried out on worn and unworn samples using a light optical microscope (LOM) and a scanning electron microscope (SEM) in the secondary electron mode (SE), equipped with energy dispersive X-ray spectroscopy (EDS).
2.4. Experimental method 2.4.1.
The approach used to determine the orientation of the flake graphite
Fig. 3(a) schematically illustrates the approach employed to determine the relative degree of orientation of the flakes with respect to the sliding direction, denoted by . An optical microscope with a magnification of 50 u was used for these measurements. The measured angles shown in Fig. 3(b) correspond to flake orientation with respect to the x-axis. For ease of interpretation, the yaxis (the sliding direction) was selected as the reference axis, i.e. 0°. Thus, the orientation values were converted to a corresponding degree, from 0 to 90°. In each measurement, a surface area of approximately 100 mm2 was investigated. As the middle region of the piston ring has the most pronounced contact with the cylinder during sliding, this region was selected for investigation. The method that was used considered the flakes to be two-dimensional particles whereas, in reality, they are three-dimensional objects; this may limit the usefulness of the orientation measurements. Similarly, due to the natural shape of the graphite flakes, there is presently no accurate method for determining the angle value with absolute certainty. This may prove to be even more difficult for worn samples, where the worn surfaces are scratched. Consequently, the first attempts at determining the angle value of individual flakes were limited to narrowing them down to within a certain range of values using an optical microscope.
2.4.2.
Data handling
The graphite particles were characterized using an optical microscope, together with the Leica QWin image analysis system. Graphite particles smaller than 50 μm in length found in the examined area had a low aspect ratio (graphite length/graphite width) and were nearly rounded in shape, and thus could no longer be considered as flakes. Furthermore, it was discovered that some of these particles were not even graphite, but rather most probably other particles or simply dirt on the material surface. Therefore, for simplicity of comparison, these undesired particles were excluded from this investigation. The data obtained was grouped into nine different categories, each with an angular increment of 10°, and presented as frequency percentage. Normalizing, i.e. relative frequency percentage, was employed in order to compare the number of graphite flakes for worn surfaces with that of unworn ones (the bulk section at zero hours of operation); this was calculated by dividing the frequency of the graphite flakes on worn samples by the total number of flakes registered within each orientation-range class. The measurements of frequency percentage were taken as an average of three individual surfaces for each ring. Microsoft Excel was used to perform the statistical analysis. 2.4.3.
Automated vs. Manual measurement
Although it is possible to assess the orientation of the flake graphite automatically using an optical microscope equipped with image analysing software, this method is only applicable for unworn samples without scratches. In addition, this automated method detects and measures all black particles, including any scratches present on the worn surface, which is undesirable. Fig. 4 shows the cumulative percentage of graphite flake frequency with respect to graphite flake orientation, obtained through both automated and manual measurements performed on a subsurface section of a sample which was operated for 20,000 hours. Both types of measurement were conducted on a polished surface which was free of scratches, in order to validate the manual measurement method. In the automated measurements, the angle reported as the orientation of the graphite corresponds to the longest ferret measurement. The manual measurement displays a near-linear trend, indicating a uniform distribution of flakes throughout the matrix. The black dashed curve represents the result obtained through automated measurement and, although this curve does not show a perfectly linear shape (owing to the inclusion of fine particles), it comes very close to the manual measurement. Applying the length criterion moved the result of the automated measurements even closer to the manual one. Thus, it can be concluded that the manual measurement method proposed is a valid one for determining the orientation of the flake graphite.
3. Results and discussion 3.1. Microstructural observation Fig. 5(a) shows the microstructural features of the GCI, etched with Nital 5% solution. Graphite particles appeared as randomly oriented and distributed flakes in a pearlitic matrix. Etching with a
10% Nital solution for three minutes revealed hard phases such as carbides, shown as white areas in Fig. 5(b). These hard phases were distributed fairly evenly throughout the metal matrix. Image analysis was carried out using the Leica image analysis system to determine the volume fraction of ferrite, graphite, and carbide phases. An average of 20 readings (corresponding to 100 mm2) determined a very limited quantity of free ferrite (less than 1%), 8.7% graphite, and 1.7% hard phases.
3.2. Surface characterization of worn piston rings Fig. 6(a) and (b) represent optical micrographs of two worn ring surfaces selected from the same ship, but run for different periods of time; approximately 16,000 and 20,000 hours, respectively. The surface of the latter featured wider scratches and more elongated and coarser graphite flakes when compared to the 16,000 hours ring. These dissimilarities are quite logical. Apart from differences in base material conditions (graphite flake width and length), this is simply explained by the difference in the operating environments which were not identical. The SEM images, Fig. 6(c) and (d) further revealed this dissimilarity by the width of the scratches which were wider for 20,000 (2–6 m) than 16,000 hours (1–3.5 m). However, both surfaces demonstrated in the same way a severe plastic deformation of the metal matrix caused by abrasion and/or adhesive wear mechanisms, visible deep parallel grooves as and dark patches, respectively. The SEM images presented in Fig. 7(a) and (c) clearly highlight the occurrence of such severe deformation (sections marked A). The sliding surfaces were scratched as a result of interactions between the sharp and hard abrasive particles, and the metal matrix [22]. EDS analysis showed that these hard abrasive particles are either CAT fines or wear-produced particles (debris) indicated in Fig. 7(b) and Fig. 8, respectively. This is in agreement with literature [6, 17]. The abrasive particles identified as CAT fines are impurities rich in aluminium, silicon and oxygen, which remain in fuel oil after the refining process [23]. These particles are trapped between the ring and cylinder liner and act as third body elements, indenting and scratching the piston ring surface. Studies by Nadal et al. [6], Riahi et al. [17] and Jones et al. [24] showed that these particles severely intensify the abrasive wear process by continuous ploughing the sliding parts, so that in worst case may result in scuffing issue. The deformed matrix during the abrasive wear condition resulted in closing the graphite flakes. Compared to the unworn surface (see Fig. 3(b)), there exists a number of regions without graphite flakes which confirms the present explanation. These are the most obvious and common microstructural appearances of the worn GCI piston ring [25]. In addition to the deformed matrix as the main cause, the corrosion products might also give rise to the closure of flakes, as illustrated in Fig. 8(a) and (b). Small hard phases dispersed throughout the metal matrix (white areas) were also observed on worn surfaces. The scratches were prevented while passing the hard phases, as marked by the white arrow in Fig. 7(c). As has been discussed in literature, these hard phases typically consist of iron phosphide and titanium, chromium, and molybdenum carbides; such constituents greatly improve wear resistance, as they stand out from the metal matrix on a fully running-in surface [2], changing the topography [8, 26], and thereby minimizing the wear rate of the sliding surfaces [27, 28].
No significant evidence of corrosive attacks was observed in either of the examined worn surfaces, showing that this type of wear was very uncommon for these rings. The examination results for these piston ring surfaces are similar to those of Nadel and Eyre [6], in that they demonstrated very low wear rates. Frequently, the corrosive environment and the last stage of the removal of the graphite particles leave a large number of small pits and empty pockets on the cylinder liner wall [22]. Fig. 9 demonstrates such an empty pocket, where the graphite has been removed entirely. The role of this type of flake graphite on cast iron wear performance will be discussed in section 3.6
3.3. Investigation of worn surfaces and subsurface regions The as-received sample was marked with a diamond indenter. Fig. 10(a) and (b) show the worn surface before and after applying a slight polishing step, respectively. This was carried out only using an MD-Dac 3 μm diamond for three minutes on the worn surface, in order to obtain a lightly polished sample without removing too much material; although only a few microns (roughly 3-5 m) of material was removed by light polishing during this stage. A scratch-free subsurface was achieved after this polishing step. A qualitative comparison between the worn surface and the subsurface showed that the majority of the flakes that had already been covered and closed during sliding reappeared on the polished surface; this was also quantitatively confirmed (see Table 3). Measurements of the frequency and total fraction area of the graphite particles present on the worn sliding surfaces showed substantially lower values. The graphite fraction values presented in Table 3 are normalized values with reference to the unworn sample. The lowest frequency of open graphite flakes was found in samples worn for 16,000 and 20,000 hours, corresponding to a reduction of approximately 52% and 34%, respectively, in terms of the effective graphite fraction area. Table 3. Frequency and fraction area of the graphite flakes on investigated surfaces 100 mm² (flake size larger than 50 μm). Material (2nd piston ring) Reference surface (bulk section, 0 hours) Worn surface (16,000 hours) Worn surface (20,000 hours) Subsurface section (16,000 hours) Subsurface section (20,000 hours)
Number of flakes 770 470 460 -
Normalized value of graphite fraction area 1.00 0.48 0.66 0.82 0.97
The deformed matrix, impurities such as corrosion products and/or debris, or a combination of both, may be the cause of the closure of the graphite flakes during sliding. Since this has not been explicitly clarified so far, additional attempts were needed to identify the main causes of the closing of the flakes. Furthermore, it was observed that most of the flakes which reappeared exhibited a tendency towards a specific orientation (regions marked by white arrows), as illustrated in Fig. 10(a) and (b). Hence, further investigations were performed for areas which were large enough to facilitate a comparison of the relative frequency of the graphite flakes, both before and after polishing. It is
important to note that the present investigation only considered the number of flakes, regardless of their size and distribution, to be of interest, although this number may slightly change as a consequence of wear and matrix deformation.
3.4. Flake graphite orientation in the as-cast and as-received worn subsurface Solidification independence of the graphite flake orientations on investigated samples was verified by conducting the measurement on as-cast ring. To apply the proposed technique the as-cast sample was grounded to remove the skin defects and the grimy casting surface. Sample preparation procedure was then followed by smooth diamond polishing until it received a similar surface finish to that of the sample described in section 2.3. The result obtained was compared with 20,000 hours bulk section sample, as illustrated in Fig. 11. Both surfaces demonstrated nearidentical distribution and orientation of the graphite flakes, confirming the solidification-independent characteristics of the samples used in this investigation.
3.5. Correlation between flake graphite orientation and sliding direction Fig. 12 presents the frequency percentage distribution of the graphite flakes for the reference and worn surfaces. The mean and confidence interval (0.95%) of the measured orientation of flakes is reported. The black dashed line shows a constant frequency percentage value of 11.11% (100% divided by nine) of flakes (ideal case). For the bulk section surface, the error bars overlap the constant line, thus confirming the uniformity of the distribution and orientation of graphite flakes, which is in accord with literature [29]. For each category, the results presented below the reference surface “Bulk surface 0 hours”, are considered as the worn affected area. Adding up the relative frequency percentages and comparing them to the unworn sample yielded a significant reduction; approximately 39% and 42% of the samples worn for 16,000 and 20,000 hours, respectively. Surprisingly, a similar trend was observed for both rings within the whole range between 0 and 90°. The largest portion of this decrease was associated with the flakes which had a lower value. This suggests that most of the flakes with an orientation of between 0 and 30° had either disappeared from the surface, or being compressed and closed during the wear process. This is also confirmed by the reappearance of graphite flakes after polishing, which have been marked with white arrows in Fig. 10. These flakes were parallel, or near-parallel, to the sliding direction. In the study performed by Liu et al. [30] regarding the influence of flake graphite orientation on weight loss, and the coefficient of friction in aluminium alloy composites containing flake graphite, it was found that a sliding angle of around 45 degree provides the minimum frictional value and wear rate. It is believed that this phenomenon can be attributed to the interactions between the hard abrasive particles and the metal matrix, which result in the closing of the graphite flakes, as illustrated in Fig. 7(a). For both worn rings, the corresponding graphs approach roughly constant values for the intervals 30-40° and 60-70°, without showing a substantial decrease in relative frequency percentage, which means that graphite flakes with orientations within this range were not closed during sliding. This in turn means that these flakes showed highest closing resistance and stayed open longer, thus providing a better smearing effect than those that were closed very early during the running-in period. It is important to note that ‘‘resistance’’ here is a qualitative measure which refers to how
long the graphite flake maintains its self-lubricating behaviour without restricting the supply of graphite during sliding. Moreover, a slight decline was observed for worn samples within the 70-80° and 80-90° intervals, which could be explained by adhesive wear and debris generation, in addition to the abrasive wear mechanism which is known to intensify the closing process of flake graphite. As revealed by the SEM micrograph in Fig. 7(a), these effects were more pronounced for the flakes which were more or less perpendicular to the sliding direction.
3.6. Mechanisms controlling the closure of graphite flakes during sliding Although a number of different phenomena, such as adhesive wear, wear-produced debris, and corrosive wear, may precipitate in the closing of the graphite flakes during sliding, abrasion was found to be the predominant mechanism in the investigated samples, as shown in Fig. 13. It was observed that graphite flakes with different orientations with respect to the sliding direction interacted differently with abrasive particles. In micro-level abrasive action, the displacement of material (i.e. piling up formation) to the sides of the wear groove were a result of the surface being scratched by abrasive particles passing through the matrix. For graphite flakes with orientations parallel to the sliding direction, this lateral displacement may have resulted in the closure of entire flakes, while, for the perpendicular ones, the deformed matrix in front of the cutting particles partially collapsed the graphite flakes and covered them, which is an indication of micro-cutting. Previous experimental studies have shown that the size [31], shape, and attack angle of the abrasive particles [32] are important factors in determining the ratio between micro-ploughing and micro-cutting mechanisms. In order to better understand the micro-interaction between abrasive particles and graphite flakes (marked by white arrows), the worn samples were cut perpendicular to the sliding direction, as shown in Fig. 14 and Fig. 15. These show two possible occurrences, before and after etching, where the sheared and deformed metal matrix near the graphite flakes resulted in the closing of the flakes. Given the qualitative and quantitative results, it is obvious that the orientation of the flake plays an important role in tribofilm formation during sliding, as its orientation can either assist smearing, or stop the supply of graphite and thus modify the nature of the wear the component is subjected to [33]. See e.g. Fig. 15(a) and (b), where two flakes with different orientations experienced different types of deformation, in both cases resulting in the closure of the graphite pocket. As discussed by Ma et al. [34], the consequence of the closing of the graphite flake and ceasing of further smearing is a major decrease in self-lubricating properties, combined with a significant increase in the coefficient of friction. In a worst case scenario, it may lead to micro-welding and scuffing failure [30, 34]. 3.6.1.
The effect of flake graphite position on smearing resistance under sliding conditions
Fig. 16 illustrates schematically three positions in which the flake graphite can reside in the matrix. The sliding direction is perpendicular to the images (i.e. in and out). Investigations by Liu and Sarmadi [30, 35] showed that surface and subsurface deformations of the matrix material play an important role in controlling the graphite film formation by closing the flake and restricting the continuous supply of graphite to the sliding parts. Strains are related to the stiffness of the material,
and plastic deformation appears when the stresses on the surface exceed the yield strength of the metal. With regard to this, it is clear that the orientation of the graphite can induce various types of deformation in the matrix around the flake [20]. Fig. 16(a), for example, demonstrates an instance where the stiffness of the matrix to the left side of the flake is lower than that of the right side. This is associated with the presence of higher amounts of graphite (soft material) and lower amounts of iron in the matrix beneath the surface, so that when an evenly distributed stress is imposed, the plastic deformations are more likely to occur on the left rather than the right side of the graphite. If this proves to be the case, the material is displaced with some degree of deviation relative to the sliding direction, through the matrix. This will compress the graphite and either partially or entirely close it. This fully agrees with the materials studied, as shown in Fig. 7(a) and Fig. 15. Moreover, for a single pass of asperities, with the same reasoning, case (a) would be more prone to failure than case (b). On the other hand, the larger the cross-section of the flake graphite, the more difficult it becomes for the matrix deformation to close it, as discussed by Liu [30]. From this point of view, when the orientation of the flake graphite is parallel to that of the sliding direction (a sliding angle of zero degrees; see Fig. 16(b)), the relative change in dimension, the subsurface shear, is very high, creating a severe plastic deformation of metal below the sliding surface, which easily closes the flake. Hence, the narrow transverse dimensions of the flake graphite ease their closing. Further, Fig. 16(c) displays another possibility, where the orientation of the graphite flake is parallel to that of the sliding surface. A larger quantity of graphite in contact with the sliding surface is beneficial for the formation of tribofilm, and so this facilitates the smearing of the flake. However, it is very prone to being removed at once, leaving an empty pocket. Thus, throughout the wear process, the graphite regions are detached from the bulk section of the ring, as observed in Fig. 9. Such a position of the graphite is an advantage at the beginning of the sliding, because it acts as a large occupied surface and serves locally as a well-positioned supplies of lubricant (large amount of graphite). However, the regions left empty can also store wear debris particles and contaminations, which is, of course, not a desirable situation [20].
4. Conclusions The following conclusions can be reached, based on the results obtained: 1. Compared to the unworn samples, a substantial decrease was observed in both the frequency and area fraction of the graphite flakes on worn samples. 2. The graphite flakes with orientations between 30° and 70° relative to the sliding direction appeared to contribute more efficiently to the wear performance of the GCI in terms of supplying graphite to the sliding surfaces and modifying the extent of tribofilm formation. 3. A similar trend, with no significant difference in wear appearance, was observed in both of the examined piston rings, which were operated in the same engine but for 16,000 and 20,000 hours, respectively. 4. The two worn samples revealed more or less the same behaviour, showing that the flakes with orientations parallel or near-parallel to the sliding direction were closed earlier.
5. As the orientation of the graphite flakes deviated more from the sliding direction, there was a higher chance of them maintaining their intrinsic self-lubricating nature and continuously supplying graphite to the sliding surface.
Acknowledgements The research leading to these results has received funding from the European Union Seventh Framework Programme (FP7/2007-2011) under grant agreement no. 265861 (Helios).
References [1] T.S. Eyre, Friction and wear of cast irons, Metals Handbook: Friction and Wear of Cast Irons, Friction, Lubrication, and Wear Technology, ASM, Materials Park, Ohio, USA1992, pp. 695-701. [2] T.S. Eyre, Wear characteristics of metals, Tribology International, 9 (1976) 203-212. [3] R.R. Fessler, G.R. Fenske, Multiyear Program Plan: Reducing friction and wear in heavy vehicles, (1999). [4] J. Keller, V. Fridrici, P. Kapsa, S. Vidaller, J.F. Huard, Influence of chemical composition and microstructure of gray cast iron on wear of heavy duty diesel engines cylinder liners, Wear, 263 (2007) 1158-1164. [5] B. Koppers Company. American Hammered Piston Ring Division, Engineers' Piston Ring Handbook: A practical manual of piston ring types, specifications and applications for designers and operators, (1941). [6] J. Nadel, T.S. Eyre, Cylinder liner wear in low speed diesel engines, Tribology International, 11 (1978). [7] F.H. Stott, A.G. Macdonald, The influence of acid strength on the corrosive wear of grey cast irons in oil-sulfuric acid mixtures, Wear, 122 (1988) 343-361. [8] T.S. Sudarshan, S.B. Bhaduri, Wear in cylinder liners, Wear, 91 (1983) 269-279. [9] W.A. Glaeser, Materials for tribology, Elsevier (1992). [10] J. Sugishita, S. Fujiyoshi, The effect of cast iron graphites on friction and wear performance I: Graphite film formation on grey cast iron surfaces, Wear, 66 (1981) 209-221. [11] P. Gupta, On a multi-process wear model, Lubrication Engineering-Illinois, 57 (2001) 19-24. [12] P. Lacey, R. Stockwell, Development of a methodology to predict cylinder liner scuffing in the 6V92TA engine lubricant, Tribology Transactions, 42 (1999) 192-201. [13] J. Sugishita, S. Fujiyoshi, The effect of cast iron graphite on friction and wear performance III: The lubricating effect of graphite under rolling-sliding contacts, Wear, 77 (1982) 181-193. [14] T.S. Eyre, R.F. Iles, D.W. Gasson, Wear characteristics of flake and nodular graphite cast iron, Wear, 13 (1969) 229-245. [15] R. Montgomery, Run-in and glaze formation on gray cast iron surfaces, Wear, 14 (1969) 99105. [16] R. Montgomery, The mild wear mechanism of nodular iron, Wear, 13 (1969) 337-343. [17] A.R. Riahi, A.T. Alpas, Wear map for grey cast iron, Wear, 255 (2003) 401-409. [18] B.K. Prasad, Sliding wear response of a cast iron under varying test environments and traversal speed and pressure conditions, Wear, 260 (2006) 1333-1341. [19] A. Vadiraj, G. Balachandran, M. Kamaraj, B. Gopalakrishna, D. Venkateshwara Rao, Wear behavior of alloyed hypereutectic gray cast iron, Tribology International, 43 (2010) 647-653. [20] A. Polak, J. Grzybek, The mechanism of changes in the surface layer of grey cast iron automotive brake disc, Materials Research, 8 (2005) 475-479. [21] D. Woodyard, Pounder's marine diesel engines and gas turbines, Butterworth-Heinemann (2009).
[22] A. Adamkiewicz, J. Drzewieniecki, Operational evaluation of piston ring wear in large marine diesel engines, Journal of Polish CIMAC Selected full texts, 7 (2012) 9-17. [23] L. Jackson, T.D. Morton, Reed's general engineering knowledge for marine engineers, T. Reed (1978). [24] C.H. Jones, Marine fuels and international standards, Marine fuels: a symposium, ASTM International (1985). [25] T.S. Eyre, K.K. Dutta, F.A. Davis, Characterization and simulation of wear occurring in the cylinder bore of the internal combustion engine, Tribology International, 23 (1990) 11-16. [26] M.D. Rogers, Metallographic characterisation of transformation phases on scuffed cast-iron diesel engine components, Tribology, 2 (1969) 123-127. [27] W.J. Tomlinson, G. Dennison, Effect of phosphide and matrix microstructures on the dry sliding wear of grey cast iron, Tribology International, 22 (1989) 259-264. [28] G.C. Barber, K.C. Ludema, The break-in stage of cylinder-ring wear: a correlation between fired engines and a laboratory simulator, Wear, 118 (1987) 57-75. [29] D.M. Stefanescu, J.R. Davis, J. Destefani, Properties and Selection: Irons, Steels, and HighPerformance Alloys, ASM, Materials Park, Ohio, USA, Materials Park, Ohio, USA, 1990, pp. 100120. [30] Y. Liu, P. Rohatgi, S. Ray, Tribological characteristics of aluminum-50 vol pct graphite composite, MTA, 24 (1993) 151-159. [31] J. Larsen-Badse, Influence of grit size on the groove formation during sliding abrasion, Wear, 11 (1968) 213-222. [32] K.-H. Zum Gahr, Microstructure and wear of materials, Elsevier (1987). [33] B.K. Prasad, Sliding wear response of cast iron as influenced by microstructural features and test condition, Materials Science and Engineering: A, 456 (2007) 373-385. [34] Y. Ma, W. Zhao, S. Li, Y. Jin, Y. Wang, T.C. Simon, Lubricating durability of two GF category engine oils, Industrial Lubrication and Tribology, 57 (2005) 161-167. [35] H. Sarmadi, A.H. Kokabi, S.M. Seyed Reihani, Friction and wear performance of copper– graphite surface composites fabricated by friction stir processing (FSP), Wear, 304 (2013) 1-12.
Highlights x x x x x
Two ship engine piston rings running for different hours were investigated. The closing tendency of the graphite flakes during sliding were evaluated. Orientation seems to play an important role on graphite lubricating behavior. Matrix plastic deformation mainly controls the graphite flake closing mechanism. Graphite flakes parallel or close to the sliding direction are closed easier.
Figure 1a
Figure 1b
Figure 2
Figure 3a
Figure 3b
Figure 4
Figure 5a
Figure 5b
Figure 6a
Figure 6b
Figure 6c
Figure 6d
Figure 7a
Figure 7b
Figure 7c
Figure 7d
Figure 8a
Figure 8b
Figure 9
Figure 10a
Figure 10b
Figure 11
Figure 12
Figure 13
Figure 14a
Figure 14b
Figure 15a
Figure 15b
Figure 16
Fig. 1. (a) Schematic of a typical piston ring pack with four rings; (b) Typical surface appearance of the as-manufactured piston rings before applying coating. Fig. 2. Schematic illustration of a worn ring piece. Fig. 3. (a) Schematic of the method used for determining the orientation of flake graphite; the sliding direction is used as the reference axis; (b) An example of a bulk surface section showing flakes orientations in the unworn matrix. Fig. 4. Cumulative percentage of graphite flake frequency for a scratch-free subsurface, with and without the 50 μm criterion for graphite flake length. Fig. 5. Optical micrographs showing the graphite and the matrix structure of alloyed GCI: (a) Graphite flakes in a pearlitic microstructure; (b) Hard phases distributed in a cast iron matrix. Fig. 6. LOM and SEM images of two worn GCI ring surfaces after (a), (c) 16,000, and (b), (d) 20,000 hours of operation. Fig. 7. (a) and (c) show typical surface features of the major part of the worn piston rings. Large scratches are present throughout the cast iron matrix; (b) and (d) depict the chemical composition of foreign particles and hard phases quantified on worn ring surfaces, respectively. Fig. 8. SEM image (a) and EDS spectrum (b), indicating the presence of wear debris generated as the result of wear, which has been trapped in the flake graphite. Fig. 9. SEM image showing a worn surface with an empty graphite pocket, which has been created as a result of the total removal of graphite from it. Fig. 10. Microstructure feature of (a) as-received worn piston ring surface operated for 16,000 hours; (b) subsurface achieved after very smooth polishing for three minutes. Fig. 11. The cumulative percentage of flake graphite orientation in relation to sliding direction, measured on the subsurface of the as-cast and bulk section of the as-received (worn) sample. Fig. 12. Graph representing the relative frequency percentage distribution of graphite flakes throughout the worn surfaces, in relation to the sliding direction for ring number 2 after 16,000 and 20,000 hours of operation, as well as for the bulk section of an unworn surface. Fig. 13. Several proposed mechanisms behind the closure of graphite flakes. Fig. 14. A transverse cross-sectional view of the worn surface, showing the microscopic aspects and the deformation of the matrix occurring near the flake graphite. Fig. 15. Metallographic transverse cross-sectional view of the pearlitic microstructure of a GCI worn surface, showing the plastic deformation of the matrix occurring close to the flake graphite, etched with Nital 2% solution. Fig.16.Schematicillustrationrepresentingdifferentpositionsofgraphiteflakesinrelationtothesliding surface.
The relationship between flake graphite orientation, smearing effect, and closing tendency under abrasive wear conditions Rohollah Ghasemi, Lennart Elmquist
www.elsevier.com/locate/wear
PII: DOI: Reference:
S0043-1648(14)00188-4 http://dx.doi.org/10.1016/j.wear.2014.05.015 WEA101032
To appear in:
Wear
Received date: 12 December 2013 Revised date: 22 May 2014 Accepted date: 26 May 2014 Cite this article as: Rohollah Ghasemi, Lennart Elmquist, The relationship between flake graphite orientation, smearing effect, and closing tendency under abrasive wear conditions, Wear, http://dx.doi.org/10.1016/j.wear.2014.05.015 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
The relationship between flake graphite orientation, smearing effect, and closing tendency under abrasive wear conditions Rohollah Ghasemi a,Ö, Lennart Elmquist b Department of Mechanical Engineering, Materials and Manufacturing, School of Engineering, Jönköping University, P.O. Box 1026, SE-551 11 Jönköping, Sweden E-mail address: [email protected]; [email protected] Ö
Corresponding author, Tel: +46 36101179; fax: +46 36166560
E-mail address: [email protected]
Abstract Plastic deformation of the matrix during the wear process results in closing the graphite flakes. In this study, the relationship between the deformation of the matrix and the closing tendency of flake graphite was investigated, both qualitatively and quantitatively. Two representative piston rings, which belonged to the same two-stroke marine engine but were operated for different periods of time, were studied. Initial microstructural observations indicated a uniform distribution of graphite flakes on unworn surfaces, whereas worn surfaces demonstrated a tendency towards a preferred orientation. Approximately 40% of the open flakes of the unworn surfaces were closed during sliding, which may result in the deterioration of the self-lubricating capability of cast iron. Moreover, flakes within the orientation range of 0 to 30° relative to the sliding direction showed a maximum closing tendency when subjected to sliding. The closing tendency gradually decreased as the angle increased, approaching a minimum between 30 and 70°. A slight increase in the closing tendency was observed for flakes with orientations between 70 and 90°. A similar trend was observed on both rings. Furthermore, SEM and EDS analysis indicated substantial deformation of the matrix in the area around the flakes. An insignificant corrosion attack was observed on both worn piston ring surfaces. Keywords: Grey cast iron, Piston ring-cylinder liner, Graphite flake orientation, Sliding wear, Abrasive wear, Graphite closing tendency
1. Introduction Over the years, grey cast iron (GCI) components have commonly been used in a variety of tribological applications, particularly in systems involving sliding, such as disc brakes, clutch and piston rings, and cylinder liner systems [1]. In tribological terms, the operating conditions for piston rings bring about a complex wear situation caused by high temperature, pressure, and mechanical friction, all of which contribute to multifaceted, high-stress conditions [2]. The oil lubrication system, clearance, initial surface roughness [3], and microstructure of the materials used [4] are other critical factors for high performance piston ring-cylinder liner materials. Low production costs, excellent tribological performance, superior machinability, outstanding thermal conductivity, and
high wear resistance, for both lubricated and dry sliding conditions, [5] have made GCI a great material for such complex applications [6]. From the tribological point of view, the excellent wear resistance of GCI stems from the selflubricating nature of the graphite particles which are distributed as flakes throughout the matrix. During sliding, carbon atoms are released from the graphite pocket, smeared onto the surfaces, and serve as solid lubricating agents by forming a thin graphite film between the sliding surfaces [7, 8]. Formation of tribofilm results in a decrease in specific wear rate by several orders of magnitude, as it limits direct contact between the sliding surfaces [7-9]. In addition, the graphite flake sites act as oil reservoirs, improving the supply of oil between sliding parts during dry starts or similar conditions involving oil starvation [9]. The advantageous contributions of open graphite flakes have been highlighted by Sugishita and Fujiyoshi [10]. For many years, the high wear rate reported in piston rings-cylinder liner systems seemed to be associated with the presence of a high sulphur content in the fuel. The presence of sulphur leads to a corrosive environment through the formation of sulphuric acid on the liner wall, which in turn accelerates the wear rate during sliding. Recently, however, it has been reported that the high wear rate and surface degradation of the piston rings are related to abrasion and scuffing, rather than corrosive wear [6]. In most investigated applications, the wear process proves to be significantly more complex than expected [11]. Scuffing occurs when a strong adhesive force develops between the piston rings and the cylinder liner under non-optimal lubrication conditions. As reported, scuffing happening is almost always accompanied by a substantial increase in friction [9], with a severe plastic deformation of the matrix [6]. The occurrence of scuffing is shown by micro-welding and/or severe wear scars, which represent plastic deformation, abrasive ploughing, and adhesive wear [12]. As has been proven so far, a direct relationship exists between the increased volume of graphite on the sliding surface and decrease in specific wear rate, discussed by Sugishita [13]. Abrasion plays a major role in sliding wear conditions by changing the matrix texture during sliding [12, 14]. Montgomery has studied the importance of matrix deformation under abrasive wear conditions, including both fully lubricated and dry sliding conditions [15, 16]. He reports that the formation of a graphite film significantly reduced the risk of scuffing under abrasive conditions by ploughing the metal matrix. Eyre et al. [14] indicated a dramatic increase in wear rate in piston rings and cylinder liner caused by abrasive wear. The hard abrasive particles could either be introduced into the system by the fuel in the form of catalyst fines (CAT fines), or be present as a wear-produced particles (debris) [6, 17]. A study performed by Riahi [17] showed that plastic deformation of the matrix accelerates debris formation. The abrasive particles are sufficiently hard to indent and scratch the sliding surfaces, which leads to severe plastic deformation of the metal matrix [12]. Apart from the microstructural standpoint the presence of the phases so as carbides and phosphides, graphite morphology and graphite flakes play important roles under abrasive condition [17-19]. The orientation of the graphite beneath the sliding surface affects the wear behaviour of the cast iron as it alters the deformation tendency in the matrix around the flake [20]. Micro-interactions, including micro-ploughing and micro-cutting; between abrasive particles and the matrix may also cause the graphite flakes to close [10], detrimentally deteriorates the self-lubricating performance of the cast iron. Therefore, keeping the graphite flakes open during sliding is tribologically beneficial, as their lubricating effect is retained.
The aim of this study is to investigate the relationship between flake graphite orientation and plastic deformation in a cast iron matrix during sliding conditions. An understanding of this relationship clarifies the deformation response of the flakes and the matrix when the piston ring is subjected to sliding and abrasive wear conditions. Both qualitative microscopic observations and quantitative measurements of the graphite distribution of worn samples are considered.
2. Materials and methods 2.1. Piston rings from a large two-stroke marine diesel engine The rings used in this study were 800 mm in diameter and designed for two-stroke marine engines, as shown in Fig. 1(a). The outer surface of the as-cast piston rings are essentially horizontally machined roughly 3 mm before applying the coating. This step is performed to remove the skin defects (undesired structure) from the surface. The surface roughness average (Ra) of the rings are checked to be within 3.2-6.3 m before coating. Fig. 1(b) illustrates the as-manufactured ring surface without coating. The piston rings were protected by two different types of thermally sprayed coatings, including cermet and aluminium-coating. The coatings and base material specifications for each of the piston rings are given in Table 1 [21]. Table 1. Specifications for the piston rings [21].
Piston ring
Base material
1st 2nd 3rd 4th
Compacted graphite iron Grey cast iron Grey cast iron Grey cast iron
Hard coating Cermet thickness (mm) 0.5 0.3
Running-in coating Aluminium-coat thickness (mm) 0.1 0.3 0.3 0.1
Cermet is a composite material which is used for hard coating, and consists of a ceramic part (chromium carbide) and a metallic part (molybdenum, nickel, and chrome). Aluminium-coating (a bronze-based coating containing alumina oxide insoluble, which was first introduced by MAN B&W Diesel A/S) was used as a semisoft running-in coating. The running-in coating is thermally deposited over the hard coating in order to shorten the running-in time, improve sliding characteristics, and increase scuffing resistance during the running-in period. The aluminiumcoating provides a smooth and profiled running-in surface, and is gradually worn off until it is entirely removed approximately after 1-2,000 hours. As the location of the second piston ring subjects to more severe wear conditions (e.g. higher temperature and pressure) than the third and fourth rings, it was selected for the present investigation. The chemical composition of the rings is presented in Table 2. The piston ring segments were provided by MAN Diesel and Turbo.
Table 2. Chemical composition of the investigated piston rings (wt%). Element C Si Mn Ni Mo Cu Ti Cr V S P Fe Composition 3.30 1.55 0.85 0.35 0.60 0.85 0.07 0.15 0.15 0.08 0.10 Balance
2.2. Sample preparation The present investigation was carried out on two different piston ring specimens extracted from the same ship engine, but were running for different time periods; approximately 16,000 and 20,000 hours. Both rings had a good run-in condition and had been running on heavy fuel oil. A schematic illustration of a typical ring sample used in this study is shown in Fig. 2. The black vertical dashed lines denote the scratches present on the outer surface. In order to examine the microstructure of the deformed matrix, the rings were sectioned perpendicular to the sliding direction, as indicated by the grey horizontal dashed line (see Fig. 2).
2.3. Microstructure characterizations and composition analysis The selected specimens were ground and polished following the standard sample preparation procedure for cast iron. The final polishing step was carried out using MD-Dac 3 μm diamond cloth with DiaPro suspension in order to obtain a scratch-free surface and a surface finish of roughly 0.05 μm. The polished samples were subsequently etched with different concentrations of Nital solution. This was done to reveal either the micro-constituents, or the carbide phases present in the cast iron matrix, depending on the purpose of the investigation. Further characterizations were carried out on worn and unworn samples using a light optical microscope (LOM) and a scanning electron microscope (SEM) in the secondary electron mode (SE), equipped with energy dispersive X-ray spectroscopy (EDS).
2.4. Experimental method 2.4.1.
The approach used to determine the orientation of the flake graphite
Fig. 3(a) schematically illustrates the approach employed to determine the relative degree of orientation of the flakes with respect to the sliding direction, denoted by . An optical microscope with a magnification of 50 u was used for these measurements. The measured angles shown in Fig. 3(b) correspond to flake orientation with respect to the x-axis. For ease of interpretation, the yaxis (the sliding direction) was selected as the reference axis, i.e. 0°. Thus, the orientation values were converted to a corresponding degree, from 0 to 90°. In each measurement, a surface area of approximately 100 mm2 was investigated. As the middle region of the piston ring has the most pronounced contact with the cylinder during sliding, this region was selected for investigation. The method that was used considered the flakes to be two-dimensional particles whereas, in reality, they are three-dimensional objects; this may limit the usefulness of the orientation measurements. Similarly, due to the natural shape of the graphite flakes, there is presently no accurate method for determining the angle value with absolute certainty. This may prove to be even more difficult for worn samples, where the worn surfaces are scratched. Consequently, the first attempts at determining the angle value of individual flakes were limited to narrowing them down to within a certain range of values using an optical microscope.
2.4.2.
Data handling
The graphite particles were characterized using an optical microscope, together with the Leica QWin image analysis system. Graphite particles smaller than 50 μm in length found in the examined area had a low aspect ratio (graphite length/graphite width) and were nearly rounded in shape, and thus could no longer be considered as flakes. Furthermore, it was discovered that some of these particles were not even graphite, but rather most probably other particles or simply dirt on the material surface. Therefore, for simplicity of comparison, these undesired particles were excluded from this investigation. The data obtained was grouped into nine different categories, each with an angular increment of 10°, and presented as frequency percentage. Normalizing, i.e. relative frequency percentage, was employed in order to compare the number of graphite flakes for worn surfaces with that of unworn ones (the bulk section at zero hours of operation); this was calculated by dividing the frequency of the graphite flakes on worn samples by the total number of flakes registered within each orientation-range class. The measurements of frequency percentage were taken as an average of three individual surfaces for each ring. Microsoft Excel was used to perform the statistical analysis. 2.4.3.
Automated vs. Manual measurement
Although it is possible to assess the orientation of the flake graphite automatically using an optical microscope equipped with image analysing software, this method is only applicable for unworn samples without scratches. In addition, this automated method detects and measures all black particles, including any scratches present on the worn surface, which is undesirable. Fig. 4 shows the cumulative percentage of graphite flake frequency with respect to graphite flake orientation, obtained through both automated and manual measurements performed on a subsurface section of a sample which was operated for 20,000 hours. Both types of measurement were conducted on a polished surface which was free of scratches, in order to validate the manual measurement method. In the automated measurements, the angle reported as the orientation of the graphite corresponds to the longest ferret measurement. The manual measurement displays a near-linear trend, indicating a uniform distribution of flakes throughout the matrix. The black dashed curve represents the result obtained through automated measurement and, although this curve does not show a perfectly linear shape (owing to the inclusion of fine particles), it comes very close to the manual measurement. Applying the length criterion moved the result of the automated measurements even closer to the manual one. Thus, it can be concluded that the manual measurement method proposed is a valid one for determining the orientation of the flake graphite.
3. Results and discussion 3.1. Microstructural observation Fig. 5(a) shows the microstructural features of the GCI, etched with Nital 5% solution. Graphite particles appeared as randomly oriented and distributed flakes in a pearlitic matrix. Etching with a
10% Nital solution for three minutes revealed hard phases such as carbides, shown as white areas in Fig. 5(b). These hard phases were distributed fairly evenly throughout the metal matrix. Image analysis was carried out using the Leica image analysis system to determine the volume fraction of ferrite, graphite, and carbide phases. An average of 20 readings (corresponding to 100 mm2) determined a very limited quantity of free ferrite (less than 1%), 8.7% graphite, and 1.7% hard phases.
3.2. Surface characterization of worn piston rings Fig. 6(a) and (b) represent optical micrographs of two worn ring surfaces selected from the same ship, but run for different periods of time; approximately 16,000 and 20,000 hours, respectively. The surface of the latter featured wider scratches and more elongated and coarser graphite flakes when compared to the 16,000 hours ring. These dissimilarities are quite logical. Apart from differences in base material conditions (graphite flake width and length), this is simply explained by the difference in the operating environments which were not identical. The SEM images, Fig. 6(c) and (d) further revealed this dissimilarity by the width of the scratches which were wider for 20,000 (2–6 m) than 16,000 hours (1–3.5 m). However, both surfaces demonstrated in the same way a severe plastic deformation of the metal matrix caused by abrasion and/or adhesive wear mechanisms, visible deep parallel grooves as and dark patches, respectively. The SEM images presented in Fig. 7(a) and (c) clearly highlight the occurrence of such severe deformation (sections marked A). The sliding surfaces were scratched as a result of interactions between the sharp and hard abrasive particles, and the metal matrix [22]. EDS analysis showed that these hard abrasive particles are either CAT fines or wear-produced particles (debris) indicated in Fig. 7(b) and Fig. 8, respectively. This is in agreement with literature [6, 17]. The abrasive particles identified as CAT fines are impurities rich in aluminium, silicon and oxygen, which remain in fuel oil after the refining process [23]. These particles are trapped between the ring and cylinder liner and act as third body elements, indenting and scratching the piston ring surface. Studies by Nadal et al. [6], Riahi et al. [17] and Jones et al. [24] showed that these particles severely intensify the abrasive wear process by continuous ploughing the sliding parts, so that in worst case may result in scuffing issue. The deformed matrix during the abrasive wear condition resulted in closing the graphite flakes. Compared to the unworn surface (see Fig. 3(b)), there exists a number of regions without graphite flakes which confirms the present explanation. These are the most obvious and common microstructural appearances of the worn GCI piston ring [25]. In addition to the deformed matrix as the main cause, the corrosion products might also give rise to the closure of flakes, as illustrated in Fig. 8(a) and (b). Small hard phases dispersed throughout the metal matrix (white areas) were also observed on worn surfaces. The scratches were prevented while passing the hard phases, as marked by the white arrow in Fig. 7(c). As has been discussed in literature, these hard phases typically consist of iron phosphide and titanium, chromium, and molybdenum carbides; such constituents greatly improve wear resistance, as they stand out from the metal matrix on a fully running-in surface [2], changing the topography [8, 26], and thereby minimizing the wear rate of the sliding surfaces [27, 28].
No significant evidence of corrosive attacks was observed in either of the examined worn surfaces, showing that this type of wear was very uncommon for these rings. The examination results for these piston ring surfaces are similar to those of Nadel and Eyre [6], in that they demonstrated very low wear rates. Frequently, the corrosive environment and the last stage of the removal of the graphite particles leave a large number of small pits and empty pockets on the cylinder liner wall [22]. Fig. 9 demonstrates such an empty pocket, where the graphite has been removed entirely. The role of this type of flake graphite on cast iron wear performance will be discussed in section 3.6
3.3. Investigation of worn surfaces and subsurface regions The as-received sample was marked with a diamond indenter. Fig. 10(a) and (b) show the worn surface before and after applying a slight polishing step, respectively. This was carried out only using an MD-Dac 3 μm diamond for three minutes on the worn surface, in order to obtain a lightly polished sample without removing too much material; although only a few microns (roughly 3-5 m) of material was removed by light polishing during this stage. A scratch-free subsurface was achieved after this polishing step. A qualitative comparison between the worn surface and the subsurface showed that the majority of the flakes that had already been covered and closed during sliding reappeared on the polished surface; this was also quantitatively confirmed (see Table 3). Measurements of the frequency and total fraction area of the graphite particles present on the worn sliding surfaces showed substantially lower values. The graphite fraction values presented in Table 3 are normalized values with reference to the unworn sample. The lowest frequency of open graphite flakes was found in samples worn for 16,000 and 20,000 hours, corresponding to a reduction of approximately 52% and 34%, respectively, in terms of the effective graphite fraction area. Table 3. Frequency and fraction area of the graphite flakes on investigated surfaces 100 mm² (flake size larger than 50 μm). Material (2nd piston ring) Reference surface (bulk section, 0 hours) Worn surface (16,000 hours) Worn surface (20,000 hours) Subsurface section (16,000 hours) Subsurface section (20,000 hours)
Number of flakes 770 470 460 -
Normalized value of graphite fraction area 1.00 0.48 0.66 0.82 0.97
The deformed matrix, impurities such as corrosion products and/or debris, or a combination of both, may be the cause of the closure of the graphite flakes during sliding. Since this has not been explicitly clarified so far, additional attempts were needed to identify the main causes of the closing of the flakes. Furthermore, it was observed that most of the flakes which reappeared exhibited a tendency towards a specific orientation (regions marked by white arrows), as illustrated in Fig. 10(a) and (b). Hence, further investigations were performed for areas which were large enough to facilitate a comparison of the relative frequency of the graphite flakes, both before and after polishing. It is
important to note that the present investigation only considered the number of flakes, regardless of their size and distribution, to be of interest, although this number may slightly change as a consequence of wear and matrix deformation.
3.4. Flake graphite orientation in the as-cast and as-received worn subsurface Solidification independence of the graphite flake orientations on investigated samples was verified by conducting the measurement on as-cast ring. To apply the proposed technique the as-cast sample was grounded to remove the skin defects and the grimy casting surface. Sample preparation procedure was then followed by smooth diamond polishing until it received a similar surface finish to that of the sample described in section 2.3. The result obtained was compared with 20,000 hours bulk section sample, as illustrated in Fig. 11. Both surfaces demonstrated nearidentical distribution and orientation of the graphite flakes, confirming the solidification-independent characteristics of the samples used in this investigation.
3.5. Correlation between flake graphite orientation and sliding direction Fig. 12 presents the frequency percentage distribution of the graphite flakes for the reference and worn surfaces. The mean and confidence interval (0.95%) of the measured orientation of flakes is reported. The black dashed line shows a constant frequency percentage value of 11.11% (100% divided by nine) of flakes (ideal case). For the bulk section surface, the error bars overlap the constant line, thus confirming the uniformity of the distribution and orientation of graphite flakes, which is in accord with literature [29]. For each category, the results presented below the reference surface “Bulk surface 0 hours”, are considered as the worn affected area. Adding up the relative frequency percentages and comparing them to the unworn sample yielded a significant reduction; approximately 39% and 42% of the samples worn for 16,000 and 20,000 hours, respectively. Surprisingly, a similar trend was observed for both rings within the whole range between 0 and 90°. The largest portion of this decrease was associated with the flakes which had a lower value. This suggests that most of the flakes with an orientation of between 0 and 30° had either disappeared from the surface, or being compressed and closed during the wear process. This is also confirmed by the reappearance of graphite flakes after polishing, which have been marked with white arrows in Fig. 10. These flakes were parallel, or near-parallel, to the sliding direction. In the study performed by Liu et al. [30] regarding the influence of flake graphite orientation on weight loss, and the coefficient of friction in aluminium alloy composites containing flake graphite, it was found that a sliding angle of around 45 degree provides the minimum frictional value and wear rate. It is believed that this phenomenon can be attributed to the interactions between the hard abrasive particles and the metal matrix, which result in the closing of the graphite flakes, as illustrated in Fig. 7(a). For both worn rings, the corresponding graphs approach roughly constant values for the intervals 30-40° and 60-70°, without showing a substantial decrease in relative frequency percentage, which means that graphite flakes with orientations within this range were not closed during sliding. This in turn means that these flakes showed highest closing resistance and stayed open longer, thus providing a better smearing effect than those that were closed very early during the running-in period. It is important to note that ‘‘resistance’’ here is a qualitative measure which refers to how
long the graphite flake maintains its self-lubricating behaviour without restricting the supply of graphite during sliding. Moreover, a slight decline was observed for worn samples within the 70-80° and 80-90° intervals, which could be explained by adhesive wear and debris generation, in addition to the abrasive wear mechanism which is known to intensify the closing process of flake graphite. As revealed by the SEM micrograph in Fig. 7(a), these effects were more pronounced for the flakes which were more or less perpendicular to the sliding direction.
3.6. Mechanisms controlling the closure of graphite flakes during sliding Although a number of different phenomena, such as adhesive wear, wear-produced debris, and corrosive wear, may precipitate in the closing of the graphite flakes during sliding, abrasion was found to be the predominant mechanism in the investigated samples, as shown in Fig. 13. It was observed that graphite flakes with different orientations with respect to the sliding direction interacted differently with abrasive particles. In micro-level abrasive action, the displacement of material (i.e. piling up formation) to the sides of the wear groove were a result of the surface being scratched by abrasive particles passing through the matrix. For graphite flakes with orientations parallel to the sliding direction, this lateral displacement may have resulted in the closure of entire flakes, while, for the perpendicular ones, the deformed matrix in front of the cutting particles partially collapsed the graphite flakes and covered them, which is an indication of micro-cutting. Previous experimental studies have shown that the size [31], shape, and attack angle of the abrasive particles [32] are important factors in determining the ratio between micro-ploughing and micro-cutting mechanisms. In order to better understand the micro-interaction between abrasive particles and graphite flakes (marked by white arrows), the worn samples were cut perpendicular to the sliding direction, as shown in Fig. 14 and Fig. 15. These show two possible occurrences, before and after etching, where the sheared and deformed metal matrix near the graphite flakes resulted in the closing of the flakes. Given the qualitative and quantitative results, it is obvious that the orientation of the flake plays an important role in tribofilm formation during sliding, as its orientation can either assist smearing, or stop the supply of graphite and thus modify the nature of the wear the component is subjected to [33]. See e.g. Fig. 15(a) and (b), where two flakes with different orientations experienced different types of deformation, in both cases resulting in the closure of the graphite pocket. As discussed by Ma et al. [34], the consequence of the closing of the graphite flake and ceasing of further smearing is a major decrease in self-lubricating properties, combined with a significant increase in the coefficient of friction. In a worst case scenario, it may lead to micro-welding and scuffing failure [30, 34]. 3.6.1.
The effect of flake graphite position on smearing resistance under sliding conditions
Fig. 16 illustrates schematically three positions in which the flake graphite can reside in the matrix. The sliding direction is perpendicular to the images (i.e. in and out). Investigations by Liu and Sarmadi [30, 35] showed that surface and subsurface deformations of the matrix material play an important role in controlling the graphite film formation by closing the flake and restricting the continuous supply of graphite to the sliding parts. Strains are related to the stiffness of the material,
and plastic deformation appears when the stresses on the surface exceed the yield strength of the metal. With regard to this, it is clear that the orientation of the graphite can induce various types of deformation in the matrix around the flake [20]. Fig. 16(a), for example, demonstrates an instance where the stiffness of the matrix to the left side of the flake is lower than that of the right side. This is associated with the presence of higher amounts of graphite (soft material) and lower amounts of iron in the matrix beneath the surface, so that when an evenly distributed stress is imposed, the plastic deformations are more likely to occur on the left rather than the right side of the graphite. If this proves to be the case, the material is displaced with some degree of deviation relative to the sliding direction, through the matrix. This will compress the graphite and either partially or entirely close it. This fully agrees with the materials studied, as shown in Fig. 7(a) and Fig. 15. Moreover, for a single pass of asperities, with the same reasoning, case (a) would be more prone to failure than case (b). On the other hand, the larger the cross-section of the flake graphite, the more difficult it becomes for the matrix deformation to close it, as discussed by Liu [30]. From this point of view, when the orientation of the flake graphite is parallel to that of the sliding direction (a sliding angle of zero degrees; see Fig. 16(b)), the relative change in dimension, the subsurface shear, is very high, creating a severe plastic deformation of metal below the sliding surface, which easily closes the flake. Hence, the narrow transverse dimensions of the flake graphite ease their closing. Further, Fig. 16(c) displays another possibility, where the orientation of the graphite flake is parallel to that of the sliding surface. A larger quantity of graphite in contact with the sliding surface is beneficial for the formation of tribofilm, and so this facilitates the smearing of the flake. However, it is very prone to being removed at once, leaving an empty pocket. Thus, throughout the wear process, the graphite regions are detached from the bulk section of the ring, as observed in Fig. 9. Such a position of the graphite is an advantage at the beginning of the sliding, because it acts as a large occupied surface and serves locally as a well-positioned supplies of lubricant (large amount of graphite). However, the regions left empty can also store wear debris particles and contaminations, which is, of course, not a desirable situation [20].
4. Conclusions The following conclusions can be reached, based on the results obtained: 1. Compared to the unworn samples, a substantial decrease was observed in both the frequency and area fraction of the graphite flakes on worn samples. 2. The graphite flakes with orientations between 30° and 70° relative to the sliding direction appeared to contribute more efficiently to the wear performance of the GCI in terms of supplying graphite to the sliding surfaces and modifying the extent of tribofilm formation. 3. A similar trend, with no significant difference in wear appearance, was observed in both of the examined piston rings, which were operated in the same engine but for 16,000 and 20,000 hours, respectively. 4. The two worn samples revealed more or less the same behaviour, showing that the flakes with orientations parallel or near-parallel to the sliding direction were closed earlier.
5. As the orientation of the graphite flakes deviated more from the sliding direction, there was a higher chance of them maintaining their intrinsic self-lubricating nature and continuously supplying graphite to the sliding surface.
Acknowledgements The research leading to these results has received funding from the European Union Seventh Framework Programme (FP7/2007-2011) under grant agreement no. 265861 (Helios).
References [1] T.S. Eyre, Friction and wear of cast irons, Metals Handbook: Friction and Wear of Cast Irons, Friction, Lubrication, and Wear Technology, ASM, Materials Park, Ohio, USA1992, pp. 695-701. [2] T.S. Eyre, Wear characteristics of metals, Tribology International, 9 (1976) 203-212. [3] R.R. Fessler, G.R. Fenske, Multiyear Program Plan: Reducing friction and wear in heavy vehicles, (1999). [4] J. Keller, V. Fridrici, P. Kapsa, S. Vidaller, J.F. Huard, Influence of chemical composition and microstructure of gray cast iron on wear of heavy duty diesel engines cylinder liners, Wear, 263 (2007) 1158-1164. [5] B. Koppers Company. American Hammered Piston Ring Division, Engineers' Piston Ring Handbook: A practical manual of piston ring types, specifications and applications for designers and operators, (1941). [6] J. Nadel, T.S. Eyre, Cylinder liner wear in low speed diesel engines, Tribology International, 11 (1978). [7] F.H. Stott, A.G. Macdonald, The influence of acid strength on the corrosive wear of grey cast irons in oil-sulfuric acid mixtures, Wear, 122 (1988) 343-361. [8] T.S. Sudarshan, S.B. Bhaduri, Wear in cylinder liners, Wear, 91 (1983) 269-279. [9] W.A. Glaeser, Materials for tribology, Elsevier (1992). [10] J. Sugishita, S. Fujiyoshi, The effect of cast iron graphites on friction and wear performance I: Graphite film formation on grey cast iron surfaces, Wear, 66 (1981) 209-221. [11] P. Gupta, On a multi-process wear model, Lubrication Engineering-Illinois, 57 (2001) 19-24. [12] P. Lacey, R. Stockwell, Development of a methodology to predict cylinder liner scuffing in the 6V92TA engine lubricant, Tribology Transactions, 42 (1999) 192-201. [13] J. Sugishita, S. Fujiyoshi, The effect of cast iron graphite on friction and wear performance III: The lubricating effect of graphite under rolling-sliding contacts, Wear, 77 (1982) 181-193. [14] T.S. Eyre, R.F. Iles, D.W. Gasson, Wear characteristics of flake and nodular graphite cast iron, Wear, 13 (1969) 229-245. [15] R. Montgomery, Run-in and glaze formation on gray cast iron surfaces, Wear, 14 (1969) 99105. [16] R. Montgomery, The mild wear mechanism of nodular iron, Wear, 13 (1969) 337-343. [17] A.R. Riahi, A.T. Alpas, Wear map for grey cast iron, Wear, 255 (2003) 401-409. [18] B.K. Prasad, Sliding wear response of a cast iron under varying test environments and traversal speed and pressure conditions, Wear, 260 (2006) 1333-1341. [19] A. Vadiraj, G. Balachandran, M. Kamaraj, B. Gopalakrishna, D. Venkateshwara Rao, Wear behavior of alloyed hypereutectic gray cast iron, Tribology International, 43 (2010) 647-653. [20] A. Polak, J. Grzybek, The mechanism of changes in the surface layer of grey cast iron automotive brake disc, Materials Research, 8 (2005) 475-479. [21] D. Woodyard, Pounder's marine diesel engines and gas turbines, Butterworth-Heinemann (2009).
[22] A. Adamkiewicz, J. Drzewieniecki, Operational evaluation of piston ring wear in large marine diesel engines, Journal of Polish CIMAC Selected full texts, 7 (2012) 9-17. [23] L. Jackson, T.D. Morton, Reed's general engineering knowledge for marine engineers, T. Reed (1978). [24] C.H. Jones, Marine fuels and international standards, Marine fuels: a symposium, ASTM International (1985). [25] T.S. Eyre, K.K. Dutta, F.A. Davis, Characterization and simulation of wear occurring in the cylinder bore of the internal combustion engine, Tribology International, 23 (1990) 11-16. [26] M.D. Rogers, Metallographic characterisation of transformation phases on scuffed cast-iron diesel engine components, Tribology, 2 (1969) 123-127. [27] W.J. Tomlinson, G. Dennison, Effect of phosphide and matrix microstructures on the dry sliding wear of grey cast iron, Tribology International, 22 (1989) 259-264. [28] G.C. Barber, K.C. Ludema, The break-in stage of cylinder-ring wear: a correlation between fired engines and a laboratory simulator, Wear, 118 (1987) 57-75. [29] D.M. Stefanescu, J.R. Davis, J. Destefani, Properties and Selection: Irons, Steels, and HighPerformance Alloys, ASM, Materials Park, Ohio, USA, Materials Park, Ohio, USA, 1990, pp. 100120. [30] Y. Liu, P. Rohatgi, S. Ray, Tribological characteristics of aluminum-50 vol pct graphite composite, MTA, 24 (1993) 151-159. [31] J. Larsen-Badse, Influence of grit size on the groove formation during sliding abrasion, Wear, 11 (1968) 213-222. [32] K.-H. Zum Gahr, Microstructure and wear of materials, Elsevier (1987). [33] B.K. Prasad, Sliding wear response of cast iron as influenced by microstructural features and test condition, Materials Science and Engineering: A, 456 (2007) 373-385. [34] Y. Ma, W. Zhao, S. Li, Y. Jin, Y. Wang, T.C. Simon, Lubricating durability of two GF category engine oils, Industrial Lubrication and Tribology, 57 (2005) 161-167. [35] H. Sarmadi, A.H. Kokabi, S.M. Seyed Reihani, Friction and wear performance of copper– graphite surface composites fabricated by friction stir processing (FSP), Wear, 304 (2013) 1-12.
Highlights x x x x x
Two ship engine piston rings running for different hours were investigated. The closing tendency of the graphite flakes during sliding were evaluated. Orientation seems to play an important role on graphite lubricating behavior. Matrix plastic deformation mainly controls the graphite flake closing mechanism. Graphite flakes parallel or close to the sliding direction are closed easier.
Figure 1a
Figure 1b
Figure 2
Figure 3a
Figure 3b
Figure 4
Figure 5a
Figure 5b
Figure 6a
Figure 6b
Figure 6c
Figure 6d
Figure 7a
Figure 7b
Figure 7c
Figure 7d
Figure 8a
Figure 8b
Figure 9
Figure 10a
Figure 10b
Figure 11
Figure 12
Figure 13
Figure 14a
Figure 14b
Figure 15a
Figure 15b
Figure 16
Fig. 1. (a) Schematic of a typical piston ring pack with four rings; (b) Typical surface appearance of the as-manufactured piston rings before applying coating. Fig. 2. Schematic illustration of a worn ring piece. Fig. 3. (a) Schematic of the method used for determining the orientation of flake graphite; the sliding direction is used as the reference axis; (b) An example of a bulk surface section showing flakes orientations in the unworn matrix. Fig. 4. Cumulative percentage of graphite flake frequency for a scratch-free subsurface, with and without the 50 μm criterion for graphite flake length. Fig. 5. Optical micrographs showing the graphite and the matrix structure of alloyed GCI: (a) Graphite flakes in a pearlitic microstructure; (b) Hard phases distributed in a cast iron matrix. Fig. 6. LOM and SEM images of two worn GCI ring surfaces after (a), (c) 16,000, and (b), (d) 20,000 hours of operation. Fig. 7. (a) and (c) show typical surface features of the major part of the worn piston rings. Large scratches are present throughout the cast iron matrix; (b) and (d) depict the chemical composition of foreign particles and hard phases quantified on worn ring surfaces, respectively. Fig. 8. SEM image (a) and EDS spectrum (b), indicating the presence of wear debris generated as the result of wear, which has been trapped in the flake graphite. Fig. 9. SEM image showing a worn surface with an empty graphite pocket, which has been created as a result of the total removal of graphite from it. Fig. 10. Microstructure feature of (a) as-received worn piston ring surface operated for 16,000 hours; (b) subsurface achieved after very smooth polishing for three minutes. Fig. 11. The cumulative percentage of flake graphite orientation in relation to sliding direction, measured on the subsurface of the as-cast and bulk section of the as-received (worn) sample. Fig. 12. Graph representing the relative frequency percentage distribution of graphite flakes throughout the worn surfaces, in relation to the sliding direction for ring number 2 after 16,000 and 20,000 hours of operation, as well as for the bulk section of an unworn surface. Fig. 13. Several proposed mechanisms behind the closure of graphite flakes. Fig. 14. A transverse cross-sectional view of the worn surface, showing the microscopic aspects and the deformation of the matrix occurring near the flake graphite. Fig. 15. Metallographic transverse cross-sectional view of the pearlitic microstructure of a GCI worn surface, showing the plastic deformation of the matrix occurring close to the flake graphite, etched with Nital 2% solution. Fig.16.Schematicillustrationrepresentingdifferentpositionsofgraphiteflakesinrelationtothesliding surface.